Twisting, bending, stretching: Strategies for making ferromagnetic [Mn III3] triangles

Article abstract of DOI:10.1039/b911820a

The synthesis and characterisation of a large family of trimetallic [Mn^III₃] Single-Molecule Magnets is presented. The complexes reported can be divided into three categories with general formulae (type 1) [Mn^III₃

Full text of DOI:10.1039/b911820a

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Twisting, bending, stretching: strategies for making ferromagnetic
[MnIII ] triangles†
3
Ross Inglis,a Stephanie M. Taylor,a Leigh F. Jones,a Giannis S. Papaefstathiou,b Spyros P. Perlepes,c
Saiti Datta,d Stephen Hill,d Wolfgang Wernsdorfere and Euan K. Brechin*a
Received 17th June 2009, Accepted 14th September 2009
First published as an Advance Article on the web 2nd October 2009
DOI: 10.1039/b911820a
The synthesis and characterisation of a large family of trimetallic [MnIII3] Single-Molecule Magnets is
presented. The complexes reported can be divided into three categories with general formulae (type 1)
-
[MnIII3O(R-sao)3(X)(sol)3-4] (where R = H, Me, tBu; X = O2CR (R = H, Me, Ph etc); sol = py and/or
-
H2O), (type 2) [MnIII3O(R-sao)3(X)(sol)3-5] (where R = Me, Et, Ph, tBu; X = O2CR (R = H, Me, Ph
etc); sol = MeOH, EtOH and/or H2O), and (type 3) [MnIII3O(R-sao)3(sol)3(XO4)] (where R = H, Et,
Ph, naphth; sol = py, MeOH, b-pic, Et-py, tBu-py; X = Cl, Re). We show that deliberate structural
distortions of the molecule can be used to tune the observed magnetic properties. In the crystals the
ferromagnetic triangles are involved in extensive inter-molecular H-bonding which is clearly manifested
in the magnetic behaviour, producing exchange-biased SMMs. These interactions can be removed by
ligand replacement to give “simpler” SMMs.
(where H2bamen is 1,2-bis(biacetylmonoximeimino)ethane was
Introduction
only reported in 2002,22 with a second example, [Mn3O(mpko)3-
The synthesis and study of polymetallic clusters containing
(O2CR)3](ClO4) (mpkoH = methyl 2-pyridyl ketone oxime) ap-
paramagnetic transition metal ions has provided scientists with
pearing in 2005.23 Both molecules can be considered simple
molecules displaying fascinating new physics.1-19 The emergence
structural analogues of the well known basic carboxylates of
of Molecular Nanomagnets in proposed applications as diverse as
general formula [MnIII3O(O2CR)6L3]+ (R = Me, Et, Ph; L = H2O,
information storage, molecular spintronics, quantum computation
py, MeCN etc) in which the bridging carboxylates (Mn-O-C-O-
and magnetic refrigeration has seen synthetic chemists, physicists,
Mn) have been replaced with bridging oximes (Mn-N-O-Mn) - all
theoreticians and materials scientists working in tandem to create,
six in the former and only the “lower” three in the latter (Fig. 1).
understand and design molecules with specific properties. For
The publication of these two molecules was fascinating and
example, combining the organic chemistry of rotaxanes with the
prompted us to ask the following questions: why do these
inorganic chemistry of heterometallic wheels recently resulted in
oxime-bridged triangles display dominant ferromagnetic exchange
the assembly a beautiful family of inorganic-organic molecular
between the metal centres and thus S = 6 ground states,
shuttles comprising inorganic rings assembled around organic
while the carboxylate-only [MnIII3O] triangles display dominant
threads.20 The amalgamation of two previously unconnected areas
antiferromagnetic exchange? And why did our own (and at
of chemistry is a timely illustration that with some imagination
the time unpublished) oxime-based triangles of general formula
chemists can create molecules with untold potential.
[MnIII3O(sao)3(O2CR)L4] (where saoH2 is salicylaldoxime and
Oxide-centred [MnIII3] triangles have interested inorganic co-
L = H2O, py; Fig. 1)24 display dominant antiferromagnetic
ordination chemists for many years not only for the study of
exchange and S = 2 ground states?
their intrinsic magnetism but because they represent the basic
An initial inspection of the molecular structures (Fig. 1)
building block from which a plethora of beautiful polymetallic
of the three molecules provided a possible clue to the puz-
clusters with fascinating physical properties are constructed.21
zle and implanted an idea. Both [Mn3O(bamen)](ClO4) and
The first ferromagnetic [MnIII3O]7+ triangle, [Mn3O(bamen)](ClO4)
[Mn3O(mpko)3(O2CR)3](ClO4) (Fig. 1) have structures that are
essentially analogous to the well known [MnIII3O(O2CR)6L3]+
aSchool of Chemistry, The University of Edinburgh, West Mains Road,
family25 where the bridging oximes simply replace the carboxylates
Edinburgh, UK EH9 3JJ. E-mail: ebrechin@staffmail.ed.ac.uk; Fax: +44
11–275–4598; Tel: +44 131 650 7545
and thus occupy a plane at approximately 60◦ to the [Mn3O]
plane, while the oxime bridges in the [MnIII3O(sao)3(O2CR)L4]
family are in exactly the same plane as the [Mn3O] unit (Fig. 1).
Clearly these three family types, despite being closely related, are
different because their formulae and structural architectures are
not the same, nevertheless we wondered if forcing or “twisting” the
planar equatorial Mn-N-O-Mn unit out of the [MnIII3O] plane in
the latter (R-saoH2) family would have any significant effect and
if so whether this would be large enough to switch the pairwise
exchange from antiferromagnetic to ferromagnetic.
bLaboratory of Inorganic Chemistry, Department of Chemistry, National and
Kapodistrian University of Athens, Panepistimiopolis, 157 71, Zogragfou,
Greece
cDepartment of Chemistry, University of Patras, 26504, Patras, Greece
dNational High Magnetic Field Laboratory and Florida State University,
Tallahassee, FL32310, USA
eInstitut Ne´el, CNRS & Universite´ J. Fourier, BP 166, 38042, Grenoble,
France
† Electronic supplementary information (ESI) available: Fig. SI1–SI3.
CCDC reference numbers 736436–736445. For ESI and crystallographic
data in CIF or other electronic format see DOI: 10.1039/b911820a
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9157
©

View Online
the [Mn6] SMMs; (b) by employing small “pincer” type tripodal
-
-
ligands (ClO4 , ReO4 etc) to sit on the “upper” triangular face;
and (c) by employing large sterically bulky ligands to occupy the
“lower” triangular face.27
Results and discussion
The twenty-one complexes (Table 1) can be divided into three
structural types. Complexes 3-6 and 8-10 can be represented by
the general formula (type 1) [MnIII3O(R-sao)3(X)(sol)3-4] (where
t
-
R = H, Me, Bu; X = O2CR (R = H, Me, Ph etc); sol = py
and/or H2O). Complexes 7, 11-12, 14-15 and 20 can be described
by (type 2) [MnIII3O(R-sao)3(X)(sol)3-5] (where R = Me, Et, Ph, tBu;
-
X = O2CR (R = H, Me, Ph etc); sol = MeOH, EtOH and/or
H2O), and complexes 1-2, 13, 16-19 and 21 can be represented
by the formula (type 3) [MnIII3O(R-sao)3(sol)3(XO4)] (where R =
t
H, Et, Ph, naphth; sol = py, MeOH, b-pic, Et-py, Bu-py; X =
Cl, Re) (Fig. 2). Interatomic distances and angles relevant to the
discussion herein are shown in Table 1. Each complex consists of
the same core comprising a [MnIII3O]7+ triangular unit (Fig. 2) with
three R-sao2- ligands bridging between adjacent MnIII centres in
1
1
1
a h :h :h :m-fashion. For type 1 complexes the three axial sites at
the base of the molecule are occupied by pyridine molecules. The
h :h :m-bridging carboxylate ligand connects two MnIII centres
(Mn1 and Mn2) on the upper face of the molecule. The remaining
site on the upper face is occupied by a pyridine or H2O molecule
in 3, 6 and 9, but remains unoccupied in 4, 5, 8 and 10 (Mn3 is
five-coordinate) owing to the presence of a bulkier carboxylate
group. Each MnIII centre adopts a distorted octahedral geometry
and displays Jahn–Teller elongation, with Mn-O2CR bond lengths
1
1
˚
in the range 2.13-2.27 A and Mn-N bond lengths in the range
Fig.
1
Top-bottom: the molecular structures of [Mn3O(bamen)]+,
˚
2.22-2.58 A; i.e. the JT axes are perpendicular to the [Mn3]
[Mn3O(mpko)3(O2CR)3]+ and [MnIII3O(sao)3(O2CR)L4] (R = Me; L = py,
H2O). Colour code: Mn = purple, O = red, N = blue, C = gold.
plane. In 8, a six-coordinate Na+ ion connects two symmetry-
equivalent {MnIII3O(Me-sao)3(O2C15H9)(py)3} moieties together
forming a [{Mn3}-Na-{Mn3}] dimer, with charge balanced by
We speculated that a change from salicylaldoxime to its
alkyl/aryl-substituted derivatives (R-saoH2, Scheme 1) would
provide the steric perturbation required to do this, and indeed this
has already proven successful for a family of hexametallic single-
-
two symmetry-equivalent ClO4 counter ions and one pyridinium
cation, which is H-bonded to the pyridine solvent molecule. Type
2 molecules have the same core structure as type 1, the only
difference being the terminally bonded pyridine molecules are
replaced by alcohol and/or H2O molecules. Complexes 7 and 11
have the carboxylate bonded terminally with the vacant site on the
other Mn centre taken up by another solvent molecule. In 11 the
Mn centre attached to the carboxylate is five-coordinate. Complex
14 has two carboxylate ligands in its structure, with one bridging
Mn1 and Mn2 on the upper face and the other terminally bound
to Mn3 on the lower face. The charge is balanced by a protonated
triethanolamine ligand which is connected to Mn3.
molecule magnets (SMMs) of formula [Mn6O2(R-sao)6(X)2(L)4-6
]
(R = H, Me, Et, Ph; X = O2CR¢, halide; L = EtOH, MeOH, H2O)
with spin ground states ranging from S = 4 to S = 12.26 Herein
we return to the [MnIII3O(sao)3(O2CR)L4] molecules and report a
family of over twenty [MnIII3] triangles whose ground state spin
values appear to be controlled by the puckering or twisting of their
central cores. Initial studies by us,27 and more recently by others
on related systems28 have suggested this to be the case.
Type 3 molecules have the upper face of the molecule capped by
1
1
1
-
a h :h :h :m3 coordinated XO4 (X = Cl, Re) anion with the three
axial sites on the lower face occupied by solvent molecules.
Type 2 molecules (and complex 13) have extensive inter-
molecular interactions propagated by the terminal alcohol/H2O
molecules and carboxylate ligands in all three dimensions
that severely affect their magnetic behaviour (vide infra). For
example, the molecules in 12 are arranged with their upper faces
face-to-face. The water molecule which is attached on the upper
face of the Mn3 triangle is hydrogen bonded to a carboxylate
Scheme 1 The R-saoH2 family of pro-ligands. saoH2, R = H; Me-saoH2,
R = Me; Et-saoH2, R = Et; Ph-saoH2, R = Ph.
The puckering of a planar [Mn3O(oxime)3] triangle can be
achieved in three ways: (a) through the use of derivatized
salicylaldoxime (R-saoH2) ligands that occupy the equatorial
([Mn3]) plane - in the same manner as that already achieved for
◦
˚
O-atom [∠O2-H22 ◊ ◊ ◊ O24 160.0 , O ◊ ◊ ◊ O 2.913(2) A and H ◊ ◊ ◊ O
˚
2.130 A] that belongs to the same cluster and to a phenolate
9158 | Dalton Trans., 2009, 9157–9168
This journal is The Royal Society of Chemistry 2009
©

View Online
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9159
©

View Online
cluster [∠O25-H251 ◊ ◊ ◊ O11 (1 - x, -y, 1 - z) 165.0◦, O ◊ ◊ ◊ O
˚
˚
2.783(2) A and H ◊ ◊ ◊ O 1.980 A] to create a zig-zag chain that
runs parallel to the a axis (Fig. SI1†). All Mn3 mean planes
within a chain are parallel, with the inter-plane distances being
˚
˚
4.011 A within a dimer and 3.471 A between dimers, respectively.
Clusters of 15 assemble through the coordinated water molecules
of the upper face of the Mn3 triangle and a phenolate O-atom to
form a hyd◦rogen-bonded dimer [∠O47-H472 ◊ ◊ ◊ O26 (1 - x, -y,
˚
˚
2 - z) 156.0 , O ◊ ◊ ◊ O 2.812(4) A and H ◊ ◊ ◊ O 2.060 A]. The dimers
assemble via the coordinated water and ethanol molecules of the
Mn3 triangle base with the lattice ethanol and water molecules to
form a two-dimensional hydrogen-bonded layer that runs parallel
to the ab plane (Fig. SI2†). All Mn3 mean planes within a layer
are again parallel and form an angle of approximately 42◦ with
respect to the mean plane of the layer and the ab plane. The
˚
inter-plane Mn3 ◊ ◊ ◊ Mn3 distance within a dimer is 3.929 A.
The presence of the capping ClO4- anion on the upper triangular
face of 13 and the absence of solvate molecules forces the Mn3
clusters to self-assemble through the three MeOH molecules
which are attached on the base of the triangle. Each Mn3 is
hydrogen-bonded to three neighbours through six complementary
hydrogen bonds (one unique) that involve the terminal MeOH
molecules and the phenolate O atoms of◦the Et-sao2- ligands
˚
[∠O15-H1 ◊ ◊ ◊ O8 (1 - x, 1 - y, 1 - z) 168 , O ◊ ◊ ◊ O 2.734(1) A
˚
and H ◊ ◊ ◊ O 1.95 A]. In this arrangement an undulated two-
dimensional hydrogen bonded layer forms that conforms to a (6,3)
net and lies parallel to the ab plane (Fig. 3). The Mn3 triangles
are arranged with their upper faces above and below the plane of
the hydrogen-bonded framework with their mean planes parallel
to each other and to the mean plane of the framework. The
layers stack in an off-set fashion with the Mn3 triangles of one
layer lying above and below the hexagonal cavities of the two
neighbouring frameworks. The Mn3 clusters in 20 have assembled
with a Ph-saoH2 molecule of crystallisation. The MeOH molecule
at the upper face of the Mn3 triangle forms an intra-molecular
hyd◦rogen bond with one carboxylate O-atom [∠O18-H18 ◊ ◊ ◊ O24
˚
˚
162 , O ◊ ◊ ◊ O 2.876(3) A and H ◊ ◊ ◊ O 2.09 A]. Two of the three
MeOH molecules at the base of the triangle are hydrogen-bonded
◦
˚
with themselves [∠O17-H17 ◊ ◊ ◊ O15 167 , O ◊ ◊ ◊ O 2.833(3) A and
◦
˚
˚
H ◊ ◊ ◊ O 2.02 A and ∠O15-H15 ◊ ◊ ◊ O16 171 , O ◊ ◊ ◊ O 2.695(2) A
and H ◊ ◊ ◊ O 1.90 A] with the third being attached to the Ph-saoH2
molecule [∠O16-H16 ◊ ◊ ◊ N99 163 , O ◊ ◊ ◊ O 2.751(3) A and H ◊ ◊ ◊ O
1.95 A]. The salicyl OH group of the Ph-saoH2 molecule forms an
˚
◦
˚
˚
intra-molecular hydrogen bond with the oximic O-atom [∠O19-
◦
˚
˚
H19 ◊ ◊ ◊ O109 168 , O ◊ ◊ ◊ O 2.577(3) A and H ◊ ◊ ◊ O 1.69 A] while
the latter attaches to a phenolate O-atom of the triangle [∠O109-
◦
˚
˚
H109 ◊ ◊ ◊ O103 163 , O ◊ ◊ ◊ O 2.643(2) A and H ◊ ◊ ◊ O 1.82 A]. The
Mn3·Ph-saoH2 units interact with five neighbouring assemblies
through ten (five unique) C-H ◊ ◊ ◊ p interactions to create a three-
dimensional framework with 46.64-bnn topology (Fig. 3).29
Fig. 2 The molecular structures of 9 (top, type 1), 12 (middle, type 2) and
16 (bottom, type 3) representing the three different structural types in the
[Mn3] family. Colour code: Mn = purple, O = red, N = blue, C = gold,
Cl = green.
Magnetic studies
O-atom o◦f a neighbouring cluster [∠O2-H21 ◊ ◊ ◊ O13 (1 - x, -y,
Dc magnetic susceptibility studies were carried out on powdered
crystalline samples of 1-21 in the 5-300 K temperature range in
a field of 0.1 T. The magnetic susceptibility data obtained for
each were simulated using the program MAGPACK30 employing
the Hamiltonians in equations (1)-(3) (Scheme 2) to provide the
isotropic parameters summarised in Table 2. No attempt was made
˚
˚
-z) 174.0 , O ◊ ◊ ◊ O 2.791(2) A and H ◊ ◊ ◊ O 1.980 A] creating
a hydrogen-bonded dimer (Fig. SI1†). The dimers assemble
with the bases of the Mn3 triangles face-to-face through two
complementary hydrogen bonds that involve a methanolic OH
group of one cluster and a phenolate O- atom of a neighboring
9160 | Dalton Trans., 2009, 9157–9168
This journal is
The Royal Society of Chemistry 2009
©

View Online
ˆ
ˆ ˆ
ˆ ˆ
ˆ ˆ
H = -2J1 (S1S2 + S2S3) -2J2 (S1S3)
(2)
(3)
ˆ
ˆ ˆ
ˆ ˆ
ˆ ˆ
H = -2J1 (S1S2) -2J2 (S2S3) -2J3 (S1S3)
The cMT vs. T data are plotted in Fig. 4 along with their associ-
ated simulations (solid lines). Room-temperature cMT values for
1-11 range from 7.25 to 8.69 cm3 K mol-1. This is lower than the
expected spin-only (g = 2) value for three non-interacting MnIII
Fig. 3 (top) The hydrogen bonded (6,3) layer in 13. Most hydrogen and
carbon atoms have been omitted for clarity. Symmetry code: a 1 - x,
1 - y, 1 - z. (bottom) The 46.64-bnn hydrogen bonded network found in
20. Purple spheres represent the Mn3·Ph-saoH2 assemblies while the gold
lines represent the C-H ◊ ◊ ◊ p interactions.
Scheme 2 Schematic detailing the 1, 2 and 3-J models employed to
simulate the experimental data for 1-21.
to simulate the data for 5 and 18 as the crystal structures contain
two independent molecules with different geometries (Table 1),
and for 3 and 6 as the crystal structures are highly disordered. We
also stress that what follows should be regarded as a qualitative
interpretation. It is clear that these molecules (and their [Mn6]
forefathers) are complicated molecules in which excited states (a
direct result of the weak magnetic coupling) and inter-molecular
interactions play an important role. The analysis is thus confined
within our simplistic model, but allows us to identify and define
clear trends in behaviour which is invaluable for future synthetic
molecular design. More detailed studies on individual family
members will be presented in future papers.
Fig. 4 Plots of cMT vs. T for complexes 1-21. The solid lines represent
simulations of the experimental data. For parameters see Table 2.
ˆ
ˆ ˆ
ˆ ˆ
ˆ ˆ
H = -2J (S1S2 + S2S3 + S1S3)
(1)
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9161
©

View Online
Table 2 Magnetostructural parameters for complexes 1-21; Mn-N-O-Mn torsion angles (a) vs. J and S
Crystal
Complex System
Space a/◦ Mn1-2/
Group Mn2-3/Mn1-3
JT tilt/◦ Mn1/
Mn2/Mn3
J/cm-1a
J1/J2/J3
Sb 1st
(cm-1)b gc
D/cm-1d t0/se
U
eff /Ke q/Ka
exc. st.
(1)
(2)
(3)
(4)
(5)
Cubic
Pa-3
P-3
P-3
P-1
P-1
4.11
13.11
8.96
5.29
1.43
8.45
-3.10
0
0
2
2
1(6.2)
1(6.04)
n.a.
1.94 n.a.
1.98 n.a.
1.96 -2.39
1.98 -2.33
n.a. n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
-36.00
-43.73
-28.75
-21.44
n.a.
Trigonal
Trigonal
Triclinic
Triclinic
-3.02
n.a.
-2.95, +0.16
n.a.
27.88, 35.82, 6.64 11.19, 2.53
31.99, 11.06, 3.04 2.64, 11.74
24.38, 5.33, 11.06 6.62, 16.24
1(13.08)
n.a. n.a.
(6)
(7)
Trigonal
Monoclinic P21
P-3
7.32
26.78, 30.70,
38.11
16.35, 11.74,
26.58
10.16
4.51, 7.35, 1.30
n.a.
-1.20,
2
2
n.a.
3(6.46)
n.a. n.a.
1.98 n.a.
n.a.
n.a.
n.a.
n.a.
-44.52
0.545
-0.96, +0.56
-1.20,
(8)
Triclinic P-1
4.68, 15.75
2
2
2
2
6
1(2.89)
1(3.36)
1(16.93)
1(7.40)
5(22.08)
2.02 -3.61
2.01 n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
-2.96
-12.46
-20.51
-1.32
27.82
-1.94, -0.40
-1.56, -0.36
(9)
Monoclinic P21/n 4.15, 4.45, 23.44 10.06, 14.72,
5.73
Monoclinic C2/c
(10)
(11)
(12)
14.38, 35.36, 1.31 12.80, 17.95,
6.92
13.07, 20.23,
32.09
44.60, 38.17,
39.07
42.12
44.60, 35.76,
37.18,
46.66, 38.56,
40.35
-3.85,
-3.05, +0.4
-1.65, +0.10
1.98 -3.77
2.02 n.a.
Triclinic
Triclinic
P-1
P-1
2.34, 6.89
11.37, 7.95, 5.84 +1.84
2.02 -0.59
1.27 ¥ 10-9 43.69
(13)
(14)
Trigonal
Triclinic
R-3
P-1
5.93
+2.80
6
6
5(33.60)
5(16.80)
2.06 -0.77
1.98 -0.92
1.98 ¥ 10-9 57.02
8.40 ¥ 10-9 25.73
30.77
19.82
3.97, 16.29, 4.19 +1.40
(15)
Triclinic
P-1
6.71, 1.73, 9.00
+1.52
6
5(18.24)
1.98 -0.82
7.40 ¥ 10-9 42.53
16.98
(16)
(17)
(18)
Trigonal
Trigonal
Trigonal
P-3 c1 44.96
R-3
R 3 c
2.78
10.20
10.35
2.11
+3.40
+4.10
n.a.
6
6
5(40.80)
5(49.20)
1.99 -0.52
1.98 -0.48
n.a. n.a.
2.97 ¥ 10-8 42.74
1.46 ¥ 10-8 47.97
n.a.
37.47
38.19
n.a.
46.78
45.40
40.42
41.65, 40.25,
43.53
32.98, 34.41,
41.44
n.a. n.a.
n.a.
(19)
(20)
(21)
Triclinic
Triclinic
P-1
P-1
3.68, 1.98, 8.91
+4.02
6
5(48.24)
2.00 -0.75
1.98 -0.51
1.98 -0.37
5.55 ¥ 10-9 48.31
40.78
20.44
15.51
3.27, 12.53, 5.25 +0.85, +1.44 6
4.49, 7.58, 12.07 +1.20
5(13.80)
5 (14.40)
n.a.
n.a.
n.a.
n.a.
Monoclinic P21/n 46.22, 39.31,
40.78
6
a Calculated from dc susceptibility studies. b Calculated from both dc susceptibility and magnetization measurements. The latter were collected in the field
and temperature ranges 0–7 T and 2–7 K. In each case the data were fit by a matrix-diagonalization method to a model that assumes only the ground state
2
2
ˆ
ˆ
ˆ
is populated, includes axial zero-field splitting (DSz ), and carries out a full powder average. The corresponding Hamiltonian is H = D(Sz - S(S + 1)/3) +
mBgHS where D is the axial anisotropy, mB is the Bohr magneton, Sz is the easy-axis spin operator, and H is the applied field (see ref. 15). c Calculated from
dc susceptibility measurements. d Calculated from magnetization measurements. e Calculated from dc susceptibility data and/or single-crystal relaxation
measurements performed on a micro-SQUID; n.a. = not available.
ˆ
ˆ
centres of 9 cm3 K mol-1, suggesting the presence of dominant
antiferromagnetic exchange between the MnIII centres. The cMT
values then decrease gradually until approximately 100 K where
they decrease more rapidly to values between 0.93 and 3.92 cm3 K
mol-1 at 5 K. For all complexes the data was simulated using the
simplest model possible. Thus, the data for complexes 1-2 was
simulated using the 1-J model of equation (1) and Scheme 2,
giving the parameters S = 0, g = 1.94, J = -3.10 cm-1 and S =
0, g = 1.98, J = -3.02 cm-1, with the first excited state (S = 1)
6.20 cm-1 and 6.04 cm-1 above the ground state, respectively. The
data obtained for complexes 4, 9 and 11 were simulated using the
2-J model of equation (2), with J1 mediated through oxide, oxime
and carboxylate, and with J2 = J3 mediated through oxide and
oxime only. Despite many attempts using the 2-J model, the data
for complexes 7, 8 and 10 could only be simulated using the 3-J
model of equation (3) and Scheme 2 in which J1 π J2 π J3. These
results are all summarised in Table 2.
increasing more rapidly at lower temperatures, reaching maximum
values of between ~13 and ~19 cm3 K mol-1. A sharp drop in cMT
then occurs for each at ~20 K with the low temperature maxima
being significantly smaller than the expected value of 21 cm3 K
mol-1 expected for S = 6. The drop in cMT is attributed to the
strong inter-molecular interactions discussed above (and/or zero-
field splitting effects). The (“high temperature”) data for 12-15
and 19 was simulated (Fig. 4) using the 1-J model of equation
(1) and afforded the parameters S = 6, with J values ranging
from 1.40 to 4.02 cm-1 (Table 2). This model was unsuccessful
for 20 and required the 2-J model of equation (2), giving S = 6,
g = 1.98, J1 = 0.85 and J2 = 1.44 cm-1. The room temperature
cMT values for 16, 17 and 21 range from ~9.3 to ~10.7 cm3 K
mol-1 and increase gradually as temperature is decreased, reaching
maximum values of between ~18 and ~21 cm3 K mol-1 at the lowest
temperature measured. In each case the data could be simulated
with the simple 1-J model affording S = 6, g = 1.98, J = 3.40 cm-1;
S = 6, g =1.98, J = 4.10 cm-1 and S = 6, g = 1.98, J = 1.20 cm-1,
respectively. It is clear from Fig. 4 that inter-molecular interactions
are playing a very important role in the observed behaviour for
all the ferromagnetically coupled complexes. Thus we add a note
For 12-15, 19 and 20 the room temperature cMT values range
from 9.06 to 10.67 cm3 K mol-1, indicating the presence of
ferromagnetic exchange between the MnIII centres. In each case the
cMT values increase gradually as temperature is decreased before
9162 | Dalton Trans., 2009, 9157–9168
This journal is
The Royal Society of Chemistry 2009
©

View Online
of caution to the absolute validity/accuracy of the simulation
parameters, though they are well within the range observed for
all previously reported (and analogous) salicyaldoxime-bridged
[Mn(III)3]and [Mn(III)6] clusters.26,27
Variable field and temperature dc magnetisation data were
collected in the 0.5–7 T and 2–7 K field and temperature ranges. In
each case we attempted to fit the data with an axial ZFS plus Zee-
man Hamiltonian (4) in the whole field and temperature range,31
weak exchange and thus the likely population of excited states in
the temperature range studied, and the simplistic model employed,
do not allow for an accurate determination of D. Indeed, magneti-
sation and HF-EPR studies of analogous molecules in which there
are no inter-molecular interactions suggest |D| values ≥ 0.8 cm-1
which would be consistent only with 14 and 15.28 Further proof for
this assumption emerges in the analysis of the ac data and single
crystal hysteresis loop measurements discussed below.
Ac magnetic susceptibility measurements were carried out on
crystalline samples of 3-21 in the 2–10 K temperature range
in a 3.5 G ac field oscillating at frequencies ranging from 50
to 1000 Hz. Fully visible out-of-phase (cM¢¢) signals indicative
of SMM behaviour (Fig. 6 shows those obtained for com-
plex 17) were observed for all S = 6 family members except
for [Mn3O(Ph-sao)3(O2C-anthra)(MeOH)4](Ph-saoH2) (20) and
[Mn3O(Ph-sao)3(b-pic)3](ClO4) (21), in which only the tails of
the signals were observed. The ac data obtained were combined
with single-crystal dc relaxation measurements performed on a
m-SQUID32 (vide infra) and fitted to the Arrhenius equation
t = toexp(Ueff/kT), where to is the pre-exponential factor, t
is the relaxation time, Ueff is the barrier to the relaxation of
the magnetisation and k is the Boltzmann constant, to give the
effective barrier to magnetisation reorientation (Ueff) for each
[Mn3] complex. These data are summarised in Fig. 6 and Table 2
and span barrier heights of between ~25–57 K. These are amongst
the largest effective barriers observed for any low nuclearity
SMMs, but they are also larger than the theoretical upper limit
[U = S2|D|] calculated using the D values obtained from the
powder dc magnetisation measurements (~30 K (12); ~40 K (13);
~48 K (14); ~42 K (15); ~27 K (16); ~25 K (17); ~39 K (19)).
Indeed, only 14 has an effective barrier lower than the theoretical
upper limit, with 15 having an experimental value close to its
theoretical value. This is to be expected for exchange-coupled
SMMs since the spin reversal is hindered by the relatively weak
inter-molecular coupling. For stronger inter-molecular exchange
coupling one would move away from the biased-SMM regime into
a 3D system where the collective modes (domain wall propagation
etc) would reduce the effective barriers. It also points to a possible
underestimation in the zfs parameters obtained from the powder
dc measurements, which is confirmed in the single crystal low
temperature magnetisation studies (below).
2
ˆ
ˆ
ˆ
H = D(Sz - S(S + 1)/3) + mBgHS
(4)
where D is the axial zero field splitting parameter, mB is the Bohr
magneton, S is the easy-axis spin operator, and H is the applied
ˆ
field. The results are summarised in Table 2 with representative
plots given in Fig. 5. Complexes 3-11 possess a spin ground
state S = 2 with D values ranging from -2.33 to -3.77 cm-1,
while complexes 12-20 have the maximum ground state S = 6
with smaller D values ranging from -0.37 to -0.92 cm-1. Despite
the ferromagnetically coupled complexes being rather similar in
structure, the calculated D values span a wide range, and therefore
must be treated with caution. Though one would expect some
variation due to the varying Jahn–Teller tilts, it is clear that the
strong inter-molecular interactions, disorder (in 15, 16, 17) which
causes a distribution in molecular environments, the presence of
Single crystal hysteresis loop and relaxation measurements
Hysteresis loop and relaxation measurements were carried out
on single crystals of all the ferromagnetically coupled complexes
using a micro-SQUID assembly, with the field applied along the
easy axis of magnetisation.32 In each case temperature and sweep
rate dependent hysteresis loops were observed, confirming SMM
behaviour for all complexes. Representative examples are shown
for complexes 13 and 16 in Fig. 7. For all type 2 molecules, in
which significant inter-molecular interactions are observed in the
crystal lattice, the loops display step-like features separated by
plateaus. After saturating the magnetisation, the first resonance is
seen in negative fields, indicative of the presence of small and
antiferromagnetic inter-molecular interactions as was first ob-
served in the complex [Mn4O3Cl4(O2CEt)3(py)3] which crystallises
as a supramolecular H-bonded dimer of cubanes ([Mn4]2).33 The
hysteresis loops show that the collective spins of each [Mn3]
Fig. 5 Plots of reduced magnetisation (M/NmB) versus H/T for 10 (top)
and 15 (bottom) in the noted field ranges and the 2–7 K temperature range.
The solid lines correspond to the fit of the data as documented in Table 2.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9163
©

View Online
/
//
Fig. 6 Ac in-phase c MT vs. T (top) and out-of-phase c M vs. T (middle)
plots obtained for complex 17 in an oscillating field of 3.5 Oe and
frequencies of 50–1000 Hz. Plots of ln(1/t) vs. 1/T obtained from the
ac magnetic susceptibility data for a cross section of family members
(bottom).
Fig. 7 Magnetisation versus field hysteresis loops for single crystals of 13
(top) and 16 (middle) at the indicated temperatures and field sweep rates.
M is normalised to its saturation value. Arrhenius plot (bottom) using ac
and dc data for 16. The dashed line is the fit of the thermally activated
region.
molecule are coupled antiferromagnetically to its neighbouring
molecules, acting as a bias that shifts the quantum tunneling
resonances with respect to the isolated SMM. The majority of the
small steps observed in all the loops are therefore due to molecules
having one (or several) “reversed” (“spin up - spin down”)
neighbouring molecules - though some may be attributed to multi-
body quantum effects.34 The complexity of the three dimensional
H-bonding networks seen in the crystal structures makes it
essentially impossible (or at least extremely difficult) to determine
all of the active exchange paths and to identify all of the steps.
9164 | Dalton Trans., 2009, 9157–9168
This journal is
The Royal Society of Chemistry 2009
©

View Online
Type 3 molecules, however, in which the lower face of the [Mn3]
triangle is occupied by derivatized pyridine analogues show no
significant H-bonding in the crystal and this is reflected in the
appearance of much simpler looking hysteresis loops - particularly
at the lowest temperatures measured (below 1.5 K). Those for
16 are shown in the middle and lower panels of Fig. 7. Above
~1.5 K the loops appear more complicated with the emergence
of fine structure which can be attributed to the contribution of
excited states, crystal defects, nuclear spins, dipolar interactions
or most likely the disorder associated with one of the b-picoline
ligands (see CIF file for full details†). Data obtained at very
low temperatures however are remarkably simple showing only
resonances originating from the ground state. The data obtained
at 40 mK are shown in the lower-middle panel of Fig. 7 and show
hysteresis loops with a step at zero field. A step indicates a surge
in the magnetisation relaxation rate due to quantum tunnelling of
the magnetisation through the barrier, occurring at a field position
where there is an avoided level crossing. At H = -1 T all the
molecules are in the MS = +6 state. When the field is swept in a
positive direction there is resonance between the +6 and -6 MS
levels at H = 0, and some of the molecules tunnel. As the field
sweep rate is decreased a second step emerges at ca. 0.80 T. The
separation between the steps is related to D by the equation DH =
|D|/gmB. Measurement of the step positions for complex 16 thus
afford an average field separation of ~0.8 T and thus a |D|/g
value of ~0.38 cm-1. Assuming g = 2.00, this corresponds to a
|D| value of approximately 0.76 cm-1, somewhat larger than that
obtained from the dc magnetisation measurements (-0.52 cm-1),
and consistent with earlier comments regarding the validity of
the D values obtained from the fit of the magnetisation versus field
Analyzer (MVNA) served as a superheterodyne spectrometer.
Both single- (complexes 3 and 13) and double-axis (complexes
12 and 13) field-orientation dependent studies were performed in
order to insure magnetic field alignment either perpendicular or
approximately parallel to the magnetic easy axes of the crystals.
Details of the experimental technique can be found elsewhere.35,36
The most comprehensive study involved the high-symmetry
(R-3) complex 13, for which measurements were performed on
separate crystals in two different spectrometers with single- (15 T)
and double-axis (7 T) rotation capabilities. The quality of the
hard-plane spectra obtained in the former were superior, but we
first present high-frequency data obtained in the 7 T system.
Angle-dependent studies were performed to obtain approximate
alignment of the field with the sample’s easy-axis. Once aligned,
measurements were performed as a function of frequency and
temperature so as to provide data sets which maximally constrain
the axial ZFS parameters. Fig. 8 displays representative tempera-
ture dependent spectra obtained at 273 GHz (inset), together with
plots of EPR peak positions deduced from measurements at four
additional high frequencies (main panel, open symbols).
data, and with previous reports of analogous “uncoupled” [MnIII
]
3
triangles; i.e. the magnitude of D for all four complexes appears
to be somewhat underestimated. New clusters with these inter-
molecular interactions (and disorder) removed will be required to
get an accurate representation of the spin Hamiltonian parameters
and of the relaxation dynamics. 16 represents a rare example of
a rather simple SMM in which the influence of inter-molecular
interactions and excited states are negligible at low temperatures.
This is potentially an exciting discovery since it gives us access
to a molecule containing only three metal ions that displays
beautiful low temperature SMM behaviour and tunnel effects,
and thus a model system with which to go beyond the giant
spin approximation to potentially yield much fruitful physical
information. Full details of the low temperature physics will be
published in future papers.
Fig. 8 Frequency dependence of EPR peak positions (black squares)
obtained with the field applied approximately parallel to the easy-axis of
13; the temperature was 20 K for these measurements. Superimposed on
the data is the best simulation based on equation (5); the obtained ZFS
parameters are given in figure. The inset displays representative spectra
obtained at 273 GHz and at three different temperatures. The various fine
structures have been labeled according to the spin projection along the
easy axis.
Before discussing the simulations in Fig. 8 (solid lines in main
panel), we comment on the spectra. The peaks are rather broad,
indicating disorder in this crystals. It is also apparent that ad-
ditional fine-structure splittings appear at the lowest temperature,
which is a sign of intermolecular exchange interactions.37 The data
points in the main panel of Fig. 8 were extracted from the 20 K
spectra so as to avoid complications due to such intermolecular
interactions. The solid curves were simulated using the following
spin Hamiltonian, containing only axial ZFS parameters.
High frequency electron paramagnetic resonance studies
High-frequency EPR measurements were performed at various
discrete frequencies in the range from 52 to 344 GHz on single-
crystals of complex 3 (S = 2), 12 (S ª 5–6, vide infra) and 13
(S = 6) to verify their ground state spin values and zero field
splitting (ZFS) parameters. Prior to measurement, crystals were
quickly transferred from their mother liquor and coated in silicone
grease in order to avoid solvent loss. The samples were also
initially cooled under atmospheric helium gas, with a total transfer
time from the mother liquor to the cryostat of ~5 minutes. An
oversized cylindrical resonator was employed in order to provide
enhanced sensitivity, and a Millimeter-wave Vector Network
2
0
4
2
ˆ
ˆ
ˆ
H = DSz + B4 [35Sz - {30S(S + 1)}Sz ] + mBB·g·S
(5)
The operators and fundamental constants in equation (5) are
defined elsewhere,37 and the obtained ZFS parameters are given in
the figure. The most important observation is the robust agreement
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9165
©

View Online
with a spin S = 6 Hamiltonian. Attempts to fit to a spin state with
S < 6 were unsuccessful. Thus, these EPR studies confirm the
findings of the preceding magnetic studies. Nevertheless, evidence
for considerable mixing between the ground and excited spin states
is given by the significant 4th order ZFS parameter.38 Interestingly,
0
the values of D and B4 are quite comparable to those reported
recently for a family of very similar Mn3 complexes.28 In particular,
there is a good correlation between B40 and the exchange constant,
J, among these complexes, e.g. B4 = -7.2 ¥ 10-5 K and J = +4.0 K
0
for complex 13, while B40 ranges from -3.8 ¥ 10-5 K to -11 ¥ 10-5
K
and J from +2.3 to 4.7 K for the FM complexes in ref. 28. This
correlation is significant, because it provides further proof that
the fourth order terms are a direct consequence of spin mixing
between low-lying multiplets, i.e. J and B4 are directly related.38
0
91 GHz temperature dependent spectra obtained for 13 with
the field in the hard plane are displayed in Fig. 9. The spectra are
superior to those presented in Fig. 8, likely due to a higher quality
sample. The signal-to-noise is also vastly improved by virtue
of the lower frequency employed. Nevertheless, the appearance
of multiple peaks associated with each cluster of fine-structure
transitions suggests the presence of multiple species in the crystal,
very similar to recently reported results for a family of Ni4
complexes.39 These spectra and their temperature dependence are
fully consistent with a spin S = 6 state possessing significant
axial anisotropy. Indeed, simulations assuming S = 6 confirm the
axial ZFS parameters obtained from the easy-axis measurements.
However, the best simulation necessitates inclusion of transverse
ZFS terms (see Fig. SI3†). The nature of these terms is not
well determined, but this nevertheless suggests the presence of
transverse anisotropy. Attempts to fit to an S < 6 model were
unsuccessful.
Fig. 10 Frequency dependence of EPR peak positions (black squares)
obtained with the field applied approximately parallel to the easy-axis
of 12; the temperature was 10 K for these measurements. Superimposed
on the data are simulations based on equation (5) with (a) S = 5 and
(b) S = 6; the obtained ZFS parameters are given in the figure.
one would not expect good fits to equation (5), as confirmed in
Fig. 10. Furthermore, the anisotropy is significantly depressed
relative to complex 12: the simulations in Fig. 10 suggest a barrier
of only ~15 K (both for S = 5 and 6), in contrast to a value of 35 K
for 13. Note that both of these values are significantly lower than
those deduced on the basis of magnetic studies (Table 2). This
is not surprising given the marginal applicability of the models
used to estimate these numbers (e.g. equation (4)). What is more, it
has recently been shown that ac susceptibility measurements often
overestimate the barrier, particularly for SMMs with low-lying
excited spin multiplets.41
Finally, Fig. 11 shows hard-plane temperature dependent data
obtained at 52 GHz for complex 3. The overall appearance of the
spectrum is consistent with that of a molecule with a relatively low
spin quantum number (there are far fewer peaks as compared to
Fig. 9) and a very significant axial anisotropy. However, it is not
possible to make meaningful peak assignments or to fit data to
Fig. 9 Hard plane temperature dependence spectra obtained at 91 GHz
for complex 13; the fine structure peaks have been labeled according to a
scheme described elsewhere.40
Plots of the 10 K EPR peak positions obtained from multi-
frequency measurements on complex 12 with the field approx-
imately aligned with its easy axis are displayed in Fig. 10.
Simulations of these data are far less satisfactory than those for
13. Indeed, simulations assuming S = 5 (Fig. 10(a)) are marginally
better than those for S = 6 (Fig. 10(b)); the corresponding ZFS
parameters are given in the figure. It is notable from Table 2 that
complex 12 sits right at the borderline between the F and AF cases.
It is thus very likely that the S = 6 state for 12 is not well developed,
i.e. it is strongly mixed with low-lying S < 6 states. In this situation,
Fig. 11 Hard plane temperature dependence spectra of complex 3
obtained at 52 GHz.
9166 | Dalton Trans., 2009, 9157–9168
This journal is
The Royal Society of Chemistry 2009
©

View Online
equation (5). Nevertheless, one can make a crude estimate of the
anisotropy on the basis of the position of the highest field (ground
state) transition relative to the expected isotropic position (1.9 T
for g = 2); this deviation DB provides a rough measure of the
product of D and S or, rather, |12 D(2S - 1)| ª DB ¥ 28 GHz/T
ª 336 GHz or 16 K.37 If one assumes a spin S = 2 state, then
this implies an extraordinarily large molecular D value of nearly
-11 K (-7.5 cm-1). From this, one can estimate an equally large
single-ion D1 value of approximately -5 cm-1 for the MnIII ions
(this assumes identical values for all three ions and parallel Jahn–
Teller axes).42 Note that a similar estimate for 13 gives a value
for D1 of approximately -2.6 cm-1. While large, these values are
not unprecedented.43 Nevertheless, the value for 3 should not be
taken too seriously given the many assumptions that went into
estimating both the molecular and single-ion D values.
that this also appears to be true here, with the more puckered
triangles displaying larger (more positive) exchange constants
and the more planar triangles displaying more negative exchange
constants. There does not appear to be any obvious correlation
between J and the out-of-plane shift of the central oxide, the
Mn-(m3-O)-Mn angle, the MnIII-O2- distances, or the Mn-N-O-
Mn distances. For example, the presence of the carboxylate versus
perchlorate/perrhenate seems to have little influence and there are
antiferromagnetic triangles in which the central oxide is displaced
more from the [Mn3] plane than that seen in the ferromagnetic
triangles. For the [Mn6] family a Mn-N-O-Mn torsion angle of
around 31◦ proved to be the “switching point”, consistent with that
observed here. Of course this analysis is purely qualitative in nature,
since in each case we simulated the experimental susceptibility
data using the most simple model possible. Indeed even with 21
members to dissect, definitive answers are difficult to come by. It is
clear that the majority of family members are rather complicated
molecules and we will need to synthesise and analyse more of the
“isolated” siblings to gain more understanding.
Discussion and conclusions
We have demonstrated three simple strategies for making ferro-
magnetic salicylaldoxime-based [MnIII3] triangles, all of which
are based on the deliberate puckering of the magnetic core of
the molecule: a) when the parent oxime ligand (saoH2) occupies
the same plane as the [Mn3O] moiety, the Mn-N-O-Mn unit is
easily made non-planar by derivatising the oximic carbon, i.e.
by making R-saoH2 ligands. When R = Et or Ph the puckering
appears to be at its greatest. This effect is then re-inforced by
Experimental procedures
All manipulations were performed under aerobic conditions, using
materials as received. CAUTION! Although no problems were
encountered in this work, care should be taken when using the
potentially explosive perchlorate anion.
-
-
(b) employing small “pincer” type tripodal ligands (ClO4 , ReO4
etc) to sit on the “upper” triangular face and (c) employing large
sterically bulky ligands to occupy the “lower” triangular face.
The combined effect is the production of highly distorted “bowl-
shaped” molecules. When alcohol solvent molecules occupy the
lower face, the molecules are involved in extensive H-bonding
in the crystal lattice and the effect is clearly manifested in the
magnetic behaviour and the triangles behave as exchange-biased
SMMs. These inter-molecular interactions are removed and the
molecules isolated from each other by replacing the alcohols with
derivatized pyridine molecules such as b-picoline or ethyl-pyridine
etc. The result is much “cleaner” and “simpler” magnetic data
as reflected in both the bulk powder dc (and ac) data and the
single-crystal hysteresis loop measurements. The emergence of
families of such [Mn3] SMMs in which the influence of inter-
molecular interactions and/or excited states can be nullified is an
exciting discovery since it gives access to very simple molecules
(3 spins) that display beautiful SMM behaviour and tunnelling
effects, and model systems with which to go beyond the giant spin
approximation to yield much fruitful physical information.44
Understanding the relationship between the structure and
magnetic behaviour in these [Mn3] triangles, however, is a difficult
task (even with twenty members) since one must consider all
contributions to the exchange. This means we must consider the
combination of different bridging and non-bridging ligand types
(oxime, oxide, carboxylate or perchlorate, alcohol or pyridine);
their relative positions and the bond lengths and angles associated
with each. In earlier work on analogous [Mn6] complexes we
discovered that the dominant factor governing the exchange
between nearest neighbours appeared to be the twisting of the
Mn-O-N-Mn moiety with respect to the [Mn3] plane as induced
by the distortion imposed on the molecule by bulkier oximes
(R-sao2-). An examination of the data of Tables 1 and 2 suggests
The syntheses, structures and magnetic properties of complexes
1, 3, 5-6, 8-9, 12-13, 15, 19 and 20 have already been communicated
or reported. Compounds 2, 4, 7, 10-11, 14, 16-18 and 21 are
reported here for the first time.
General synthetic strategies:
For complexes 2, 16-18 and 21: MnII(ClO4)2·6H2O (1 mmol)
and the (derivatized) salicyaldoxime ligand R-saoH2 (R = H,
naphth, Et, Ph) (1 mmol) were dissolved in a mixture of pyridine
(or derivatized pyridine) (5 ml) and MeOH (20 ml) with NEt4OH
(1 mmol). For 21, NaReO4 (1 mmol) was also added to the reaction
mixture. The solutions were left stirring for ~1 h and then filtered.
Et2O was diffused into one half of the solution and the remainder
was left to slowly evaporate. Suitable crystals grew after 3–5 days
from both solutions.
For complex 4: A pyridine/EtOH solution (25 ml) of
MnII(O2CC10H7)2·H2O (1 mmol), saoH2 (1 mmol) and NEt4OH
(1 mmol) was stirred for ~1 h. After filtering single crystals grew
upon slow evaporation during 5 days.
For complex 7: MnII(O2CCH3)2·4H2O (1 mmol), Me-saoH2
(1 mmol) and NMe4ClO4 (1 mmol) were dissolved in MeOH
(25 ml) and NEt4OH (1 mmol) added. After stirring for ~1 h the
solution was filtered and allowed to evaporate. Crystals suitable
for X-ray diffraction grew after 5 days.
For complex 10: MnII(ClO4)2·6H2O (1 mmol), tBu-saoH2
(1 mmol) and HO2CPh(OMe)3 were dissolved in a mixture of
pyridine (5 ml) and EtOH (20 ml). NEt4OH (1 mmol) was added
and the mixture stirred for 2 hours before being filtered and
allowed to slowly evaporate. Crystals suitable for X-ray diffraction
grew after 5 days.
For complex 11: MnII(ClO4)2·6H2O (1 mmol), tBu-saoH2
(1 mmol) and NaO2CPh(CH3)2 were dissolved in MeOH (20 ml).
CH3ONa (1 mmol) was added and the mixture stirred for 2 hours
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 9157–9168 | 9167
©

View Online
before being filtered and allowed to slowly evaporate. Crystals
suitable for X-ray diffraction grew after 5 days.
M. Affronte and R. E. P. Winpenny, Nat. Nanotechnol., 2009, 4,
173.
20 C.-F. Lee, D. A. Leigh, R. G. Pritchard, D. Schultz, S. J. Teat, G.A
Timco and R. E. P. Winpenny, Nature, 2009, 458, 314.
21 G. Christou, Polyhedron, 2005, 24, 2065.
22 S. G. Sreerama and S. Pal, Inorg. Chem., 2002, 41, 4843.
23 T. C. Stamatatos, D. Foguet-Albiol, C. C. Stoumpos, C. P.
Raptopoulou, A. Terzis, W. Wernsdorfer, S. P. Perlepes and G. Christou,
J. Am. Chem. Soc., 2005, 127, 15380.
For complex 14: MnII(O2CPh)2·2H2O (1 mmol), Et-saoH2
(1 mmol) and teaH3 (1 mmol) were dissolved in EtOH (25 ml).
˜
After stirring for 1 h the solution was filtered and allowed to
slowly evaporate. Crystals suitable for X-ray diffraction grew after
5 days.
Elemental Anal. calcd (found) for dried 2 solvent free: C 50.09
(50.37), H 4.20 (4.08), N 8.34 (8.15). 4: C 58.17 (57.70), H
3.94 (3.93), N 9.13 (9.01). 7: C 46.90 (46.73), H 5.35 (4.71), N
5.61 (5.48). 10: C 58.48 (58.41), H 5.48 (4.88), N 7.33 (6.99).
11(-3MeOH): C 55.20 (55.75), H 5.60 (5.19), N 4.49 (4.73). 14:
C 49.84 (49.31), H 5.72 (5.46), N 4.74 (4.56). 16: C 51.52 (51.62),
H 4.61 (4.49), N 8.01 (8.09). 17: C 52.76 (52.94), H 5.40 (5.28), N
7.38 (7.60). 18: C 55.18 (55.41), H 5.66 (5.68), N 7.15 (6.92). 21: C
57.73 (57.75), H 4.05 (3.25), N 7.04 (6.48).
Variable temperature, solid-state direct current (dc) and alter-
nating current (ac) magnetic susceptibility data down to 1.8 K
were collected on a Quantum Design MPMS-XL SQUID magne-
tometer equipped with a 7 T dc magnet. Diamagnetic corrections
were applied to the observed paramagnetic susceptibilities using
Pascal’s constants. Magnetic studies below 1.8 K were carried out
on single crystals using a micro-SQUID apparatus operating down
to 40 mK.
24 (a) C. J. Milios, PhD Thesis, University of Patras, Greece, 2004; (b) C. J.
Milios, A. G. Whittaker and E. K. Brechin, Polyhedron, 2007, 26, 1927.
25 R. D. Cannon and R. P. White, Prog. Inorg. Chem., 1988, 36,
195.
26 (a) R. Inglis, L. F. Jones, C. J. Milios, S. Datta, A. Collins, S. Parsons,
W. Wernsdorfer, S. Hill, S. P. Perlepes, S. Piligkos and E. K. Brechin,
Dalton Trans., 2009, 3403; (b) C. J. Milios, S. Piligkos and E. K. Brechin,
Dalton Trans., 2008, 1809; (c) C. J. Milios, R. Inglis, A. Vinslava, R.
Bagai, W. Wernsdorfer, S. Parsons, S. P. Perlepes, G. Christou and E. K.
Brechin, J. Am. Chem. Soc., 2007, 129, 12505; (d) C. J. Milios, R. Inglis,
R. Bagai, W. Wernsdorfer, A. Collins, S. Moggach, S. Parsons, S. P.
Perlepes, G. Christou and E. K. Brechin, Chem. Commun., 2007, 3476;
(e) C. J. Milios, A. Vinslava, W. Wernsdorfer, A. Prescimone, P. A.
Wood, S. Parsons, S. P. Perlepes, G. Christou and E. K. Brechin, J. Am.
Chem. Soc., 2007, 129, 6547; (f) C. J. Milios, A. Vinslava, S. Moggach, S.
Parsons, W. Wernsdorfer, G. Christou, S. P. Perlepes and E. K. Brechin,
J. Am. Chem. Soc., 2007, 129, 2754; (g) C. J. Milios, A. Vinslava, P. A.
Wood, S. Parsons, W. Wernsdorfer, G. Christou, S. P. Perlepes and E. K.
Brechin, J. Am. Chem. Soc., 2007, 129, 8.
27 (a) R. Inglis, L. F. Jones, G. Karotsis, A. Collins, S. Parsons, S. P.
Perlepes, W. Wernsdorfer and E. K. Brechin, Chem. Commun., 2008,
5924; (b) R. Inglis, L. F. Jones, K. Mason, S. P. Perlepes, W. Wernsdorfer
and E. K. Brechin, Chem.–Eur. J., 2008, 14, 9117; (c) C. J. Milios, R.
Inglis, L. F. Jones, A. Prescimone, S. Parsons, W. Wernsdorfer and E. K.
Brechin, Dalton Trans., 2009, 2812.
28 P. L. Feng, C. Koo, J. J. Henderson, P. Manning, M. Nakano, E. del
Barco, S. Hill and D. N. Hendrickson, Inorg. Chem., 2009, 48, 3480.
29 (a) M. O’Keeffe, M. A. Peskov, S. J. Ramsden and O. M. Yaghi,
Acc. Chem. Res., 2008, 41, 1782; (b) V. A. Blatov, IUCr CompComm
Newsletter, 2006, 7, 4 http://www.topos.ssu.samara.ru.
30 J. J. Borras-Alemnar, J. M. Clemente-Juan, E. Coronado and B. S.
Tsukerblat, J. Comp. Chem., 2001, 22, 985.
31 S. Piligkos, MAGMOFIT, The University of Copenhagen.
32 (a) W. Wernsdorfer, Adv. Chem. Phys., 2001, 118, 99; (b) W. Wernsdorfer,
Supercond. Sci. Technol., 2009, 22, 064013.
Diffraction data were collected at 150 K on a Bruker Smart Apex
CCDC diffractometer, equipped with an Oxford Cryosystems LT
device, using Mo radiation. See CIF files† for full details.45
Notes and references
1 D. Gatteschi and R. Sessoli, Angew. Chem., Int. Ed., 2003, 42, 268.
2 R. Sessoli, D. Gatteschi, A. Caneschi and M. A. Novak, Nature, 1993,
365, 141.
3 W. Wernsdorfer and R. Sessoli, Science, 1999, 284, 133.
4 W. Wernsdorfer, N. Aliaga-Alcalde, D. N. Hendrickson and G. Chris-
tou, Nature, 2002, 416, 406.
33 R. Tiron, W. Wernsdorfer, N. Aliaga-Alcalde and G. Christou, Phys.
Rev. B: Condens. Matter Mater. Phys., 2003, 68, 140407.
34 (a) W. Wernsdorfer, S. Bhaduri, R. Tiron, D. N. Hendrickson and G.
Christou, Phys. Rev. Lett., 2002, 89, 197201; (b) W. Wernsdorfer, S.
Bhaduri, A. Vinslava and G. Christou, Phys. Rev. B: Condens. Matter
Mater. Phys., 2005, 72, 214429.
5 S. Hill, R. S. Edwards, N. Aliaga-Alcalde and G. Christou, Science,
2003, 302, 1015.
6 M. N. Leuenberger and D. Loss, Nature, 2001, 410, 789.
7 A. Caneschi, D. Gatteschi, N. Lalioti, C. Sangregorio, R. Sessoli, G.
Venturi, A. Vindigni, A. Rettori, M. G. Pini and M. A. Novak, Angew.
Chem., Int. Ed., 2001, 40, 1760.
8 S. Carretta, P. Santini, G. Amoretti, M. Affronte, A. Candini, A. Ghirri,
I. S. Tidmarsh, R. H. Laye, R. Shaw and E. J. L. McInnes, Phys. Rev.
Lett., 2006, 97, 207201.
9 F. Troiani, A. Ghirri, M. Affronte, S. Carretta, P. Santini, G. Amoretti,
S. Piligkos, G. Timco and R. E. P. Winpenny, Phys. Rev. Lett., 2005,
94, 207208.
10 J. Lehmann, A. Gaita-Arino, E. Coronado and D. Loss, Nat. Nan-
otechnol., 2007, 2, 312.
11 E. Coronado, J. R. Gala´n-Mascaro´s, C. J. Go´mez-Garc´ıa and V.
Laukhin, Nature, 2000, 408, 447.
12 M. Evangelisti, A. Candini, A. Ghirri, M. Affronte, E. K. Brechin and
E. J. L. McInnes, Appl. Phys. Lett., 2005, 87, 072504.
13 L. Bogani and W. Wernsdorfer, Nat. Mater., 2008, 7, 179.
14 D. Gatteschi, R. Sessoli and J. Villain, Molecular Nanomagnets, Oxford
University Press, Oxford, 2006.
35 M. Mola, S. Hill, P. Goy and M. Gross, Rev. Sci. Instrum., 2000, 71,
186.
36 S. Takahashi and S. Hill, Rev. Sci. Instrum., 2005, 76, 023114.
37 E. del Barco, A. D. Kent, S. Hill, J. M. North, N. S. Dalal, E. M.
Rumberger, D. N. Hendrickson, N. Chakov and G. Christou, J. Low
Temp. Phys., 2005, 140, 119.
38 A. Wilson, J. Lawrence, E-C. Yang, M. Nakano, D. N. Hendrickson
and S. Hill, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74,
140403.
39 J. Lawrence, E.-C. Yang, R. Edwards, M. M. Olmstead, C. Ramsey,
N. S. Dalal, P. K. Gantzel, S. Hill and D. N. Hendrickson, Inorg.
Chem., 2008, 47, 1965.
40 S. Takahashi, R. S. Edwards, J. M. North, S. Hill and N. S. Dalal, Phys.
Rev. B: Condens. Matter Mater. Phys., 2004, 70, 094429.
41 C. Lampropoulos, S. Hill and G. Christou, ChemPhysChem, 2009, 10,
2397–2400.
15 G. Arom´ı and E. K. Brechin, Struct. Bonding, 2006, 122, 1, and
42 S. Datta, E. Bolin, C. J. Milios, E. K. Brechin and S. Hill, Polyhedron,
references therein.
2009, 28, 1788.
16 R. Bircher, G. Chaboussant, C. Dobe, H. U. Gu¨del, S. T. Ochsenbein,
A. Sieber and O. Waldmann, Adv. Funct. Mater., 2006, 16, 209.
17 M. Evangelisti, F. Luis, L. J. de Jongh and M. Affronte, J. Mater. Chem.,
2006, 16, 2534.
18 M. Affronte, S. Carretta, G. A. Timco and R. E. P. Winpenny, Chem.
Commun., 2007, 1789.
43 J. Krzystek, A. Ozarowski and J. Telser, Coord. Chem. Rev., 2006, 250,
2308.
44 Detailed studies of the single crystal measurements will appear in future
papers.
45 D. J. Watkin, C. K. Prout, J. R. Carruthers, P. W. Betteridge and
R. I. Cooper, CRYSTALS, Issue 12, 2003, Chemical Crystallography
Laboratory, University of Oxford, Oxford, UK.
19 G. A. Timco, S. Carretta, F. Troiani, F. Tuna, R. J. Pritchard, C. A.
Muryn, E. J. L. McInnes, A. Ghirri, A. Candini, P. Santini, G. Amoretti,
9168 | Dalton Trans., 2009, 9157–9168
This journal is
The Royal Society of Chemistry 2009
©