Spin structure of surface-supported single-molecule magnets from isomorphous replacement and X-ray magnetic circular Dichroism

Article abstract of DOI:10.1021/ic102184n

Surface-supported arrays of Fe₄-type Single-Molecule Magnets retain a memory effect and are of current interest in the frame of molecule-based information storage and spintronics. To reveal the spin structure of [Fe₄(L)₂(dpm)₆] (1) on Au, an isomorphous compound [Fe₃Cr(L)₂(dpm)₆] was synthesized and structurally and magnetically characterized (H₃L is tripodal ligand 11-(acetylthio)-2,2-bis(hydroxymethyl)undecan-1-ol and Hdpm is dipivaloylmethane). The new complex contains a central Cr³⁺ ion and has a S = 6 ground state as opposed to S = 5 in 1. Low-temperature X-ray Magnetic Circular Dichroism studies at Fe- and Cr-L_2,3 edges revealed that the antiparallel alignment between Fe and Cr spins is preserved on surfaces. Moreover, the different Fe-L_2,3 spectral features found in the homo- and heterometallic species disclose the opposing contribution of the central Fe³⁺ ion in the former compound, proving that its ferrimagnetic spin structure is retained on surfaces.

Full text of DOI:10.1021/ic102184n

ARTICLE
pubs.acs.org/IC
Spin Structure of Surface-Supported Single-Molecule Magnets from
Isomorphous Replacement and X-ray Magnetic Circular Dichroism
Matteo Mannini,† Erik Tancini,‡ Lorenzo Sorace,† Philippe Sainctavit,§ Marie-Anne Arrio,§ Yu Qian,§
Edwige Otero,|| Daniele Chiappe,^ Ludovica Margheriti,† Julio C. Cezar,# Roberta Sessoli,† and
Andrea Cornia*,‡
†Department of Chemistry ‘Ugo Schiff’ and INSTM RU, University of Florence, 50019 Sesto Fiorentino, Italy
‡Department of Chemistry and INSTM RU, University of Modena and Reggio Emilia, 41100 Modena, Italy
§Institut de Minꢀeralogie et de Physique des Milieux Condensꢀes, CNRS UMR7590, Universitꢀe Pierre et Marie Curie, 75252 Paris, France
Synchrotron Soleil, Saint Aubin, 91192 Gif sur Yvette, France
^Department of Physics, University of Genova and CNISM, 16146 Genova, Italy
#European Synchrotron Radiation Facility (ESRF), 38043 Grenoble, France
S Supporting Information
b
ABSTRACT: Surface-supported arrays of Fe4-type Single-Molecule Mag-
nets retain a memory effect and are of current interest in the frame of
molecule-based information storage and spintronics. To reveal the spin
structure of [Fe4(L)2(dpm)6] (1) on Au, an isomorphous compound
[Fe3Cr(L)2(dpm)6] was synthesized and structurally and mag-
netically characterized (H3L is tripodal ligand 11-(acetylthio)-2,2-bis-
(hydroxymethyl)undecan-1-ol and Hdpm is dipivaloylmethane). The new
complex contains a central Cr3þ ion and has a S = 6 ground state as opposed
to S = 5 in 1. Low-temperature X-ray Magnetic Circular Dichroism studies at
Fe- and Cr-L2,3 edges revealed that the antiparallel alignment between Fe
and Cr spins is preserved on surfaces. Moreover, the different Fe-L2,3
spectral features found in the homo- and heterometallic species disclose the
opposing contribution of the central Fe3þ ion in the former compound, proving that its ferrimagnetic spin structure is retained on
surfaces.
’ INTRODUCTION
exhibit a “memory effect”, which is more pronounced in 2 because
of its tendency to afford partially oriented layers.5b These magnetic
investigations on molecules at surfaces, covering both static and
dynamic properties, have largely relied on X-ray Magnetic Circular
Dichroism (XMCD). XMCD is a spectroscopic technique derived
from X-ray Absorption (XAS) and provides direct information on
the magnetic polarization at specific elements.6 Furthermore, when
operated in the Total Electron Yield detection mode it becomes
surface-sensitive and allows to probe the magnetism of monolayers
or submonolayers,4a,5,7 in a way not easily accessible to tradi-
tional magnetometry or magnetic resonance techniques.7c,8
Complexes of the Fe4 series feature a central Fe3þ ion (Fec)
with spin si = 5/2 antiferromagnetically coupled to three peripheral
ones(Fep) toafford a ground state with total spin S=5, asprovedby
magnetic measurements and electron paramagnetic resonance
(EPR) spectroscopy.5,9 This peculiar spin arrangement, in which
uncompensation of antiparallel spins yields a large S value, is
referred to as a ferrimagnetic spin structure because of its
The prospect of exploiting molecular spins to store and
process information (molecular spintronics)1 is driving funda-
mental investigations on the electronic structure and magnetism
of open-shell molecular frameworks or molecular arrays at
surfaces.2 Single-Molecule Magnets (SMMs) in particular have
been proposed as miniaturized memory units to be electrically
written and read-out using an STM tip.2,3 The electronic
structure and functionality of these fragile metal complexes
may however easily undergo profound alterations upon deposi-
tion, as found for the Mn12 family.4 Special care must therefore be
applied to rule out unwanted structural reorganization, fragmen-
tation, or complete molecular disruption, which may take place
during the grafting process.
We have recently shown that iron(III)-based SMMs of the Fe4
family, [Fe4(L)2(dpm)6] (1) and [Fe4(L0)2(dpm)6] (2), rein-
forced and functionalized using tripodal ligands H3L =
11-(acetylthio)-2,2-bis(hydroxymethyl)undecan-1-ol5a and H3L0 =
7-(acetylthio)-2,2-bis(hydroxymethyl)heptan-1-ol,5b retain their
structure and electronic properties on a gold surface (Hdpm =
dipivaloylmethane). In fact, surface-supported complexes 1 and 2
Received: October 29, 2010
Published: February 25, 2011
r
2011 American Chemical Society
2911
dx.doi.org/10.1021/ic102184n Inorg. Chem. 2011, 50, 2911–2917
|

Inorganic Chemistry
ARTICLE
resemblance with the magnetic configuration of bulk ferrimag-
nets. The large magnetic moment in the ground state results in a
XMCD signal at low temperature, which has the same intensity
and fine structure in bulk samples and in monolayers.5,10 The
consistency of the observed XMCD response with a ferrimag-
netic spin structure has found theoretical support from Ligand
Field Multiplet (LFM) calculations,10 but has never been experi-
mentally verified.
We herein show that the XMCD response of 1 at Fe-L2,3 edges
is specific of its ground state spin configuration. To this aim,
we have prepared a related compound, [Fe3Cr(L)2(dpm)6] (3),
in which the central Fe3þ ion has been replaced by a Cr3þ ion
(si = 3/2). According to previous work, such an isomorphous
substitution can be carried out without metal scrambling, though
residual Fe4 species are found in the crystal lattice.11 The
resulting differences in the dichroic signal, detected both on
bulk phase and on monolayer samples, disclose the field-oppos-
ing contribution of the central ion and demonstrate that 1 retains
its ferrimagnetic spin structure on Au(111) surfaces.
Fe3.16Cr0.84C96H168O20S2: C, 59.85; H, 8.79; S, 3.33; Cr, 2.27; Fe, 9.16.
Found (average over two samples): C, 59.88; H, 8.80; S, 3.66; Cr, 2.33;
Fe, 9.38 (Fe:Cr atom ratio 3.75:1). Alternative room-temperature routes
were also tested, but they invariably afforded a lower Cr content.
X-ray Crystallography. The structure determination on 3 was
carried out at 120(2) K using a 4-circle Bruker-Nonius X8APEX
diffractometer, equipped with Mo-KR radiation and a Kryo-Flex nitro-
gen flow cryostat. The structure was solved by direct methods using the
2
SIR92 program.13a Full-matrix least-squares refinement on Fo was
performed using the SHELXL-97 program13b implemented in the
WINGX suite.13c All non-hydrogen atoms were refined anisotropically,
with the exception of the terminal portion of the alkyl chain, which
was found disordered over three positions with 0.61, 0.21, and 0.18
occupancies and refined with restraints over geometrical and displace-
ment parameters. In particular the C-S-C(O)CH3 moiety was refined
as a rigid group with the same geometry found in 1,2-ethanedithiol
diacetate at 183 K.13d Hydrogen atoms were treated as riding contribu-
tors with isotropic displacement parameters. These were fixed to UH
=
pUiso(C) where p = 1.2 for methine and methylene hydrogens, and p =
1.5 for methyl H atoms.
Magnetic Measurements. Direct Current (DC) magnetic data
for a polycrystalline sample of 3 (8.40 mg), pressed in a pellet, were
recorded on a Quantum Design MPMS SQUID magnetometer. Mag-
netic susceptibilities were measured in applied fields of 1 kOe from 1.9 to
35 K and 10 kOe from 35 to 300 K. Magnetization was also measured at
1.9, 2.5, and 4.5 K in fields up to 50 kOe. Data reduction was carried out
using a molecular weight of 1926 and a diamagnetic contribution
(estimated from Pascal’s constants) of -1148 ꢀ 10-6 emu/mol.
Alternating Current (AC) susceptibility in zero static field was measured
in the frequency range 100-25000 Hz on a microcrystalline powder
sample of 3 using an Oxford Instruments MAGLAB 2000 platform
equipped with a home-built inductive coil probe.
’ EXPERIMENTAL SECTION
Synthesis. All operations were carried out with strict exclusion of
moisture, unless otherwise stated. All chemicals were reagent grade and
used as received, except for solvents, which were dried and distilled
under nitrogen before use. Methanol was treated with Mg(OMe)2
(prepared from Mg turnings, I2, and MeOH) and distilled; diethylether
and tetrahydrofuran (THF) were distilled from their blue Na/benzo-
phenone solutions. 1,2-Dimethoxyethane (DME) and acetonitrile were
treated with CaH2 and distilled, acetonitrile being stored over 3A
molecular sieves. The dimer [Fe2(OMe)2(dpm)4] (4) was prepared
according to the procedure of Rossman et al.,12a modified as described in
a previous publication.9b The compound CrCl3(THF)3 was prepared
from CrCl3 by prolonged Soxhlet extraction using anhydrous THF and
zinc powder, as reported by Herwig and Zeiss.12b H3L = 11-(acetylthio)-
2,2-bis(hydroxymethyl)undecan-1-ol was synthesized as previously
described.9c Sodium methoxide solution (3.08 M in methanol) was
prepared by careful addition of sodium metal to anhydrous methanol under
nitrogen. The Fe and Cr contents were measured by complexometric
titration with EDTA (ca. 5 ꢀ 10-3 M), using back-titration with Pb(NO3)2
(ca. 5 ꢀ 10-3 M) and the technique of kinetic masking.12c
W-band (ca. 94 GHz) EPR spectra were recorded on a polycrystalline
powder sample of 3, embedded in wax, using a Bruker Elexsys E600 CW
spectrometer. Simulation of EPR spectra and spin-Hamiltonian calculations
were carried out using a dedicated software, as described elsewhere.9c,13e
Monolayer Preparation. All the procedures were carried out in a
portable glovebox inflated with nitrogen gas to avoid contamination of
the surface or damaging of the molecular deposits. Polycrystalline Au,
evaporated on a mica substrate (Agilent Inc.), was flame annealed with a
hydrogen flame to achieve Au(111) reconstruction and cleaned with
fresh ethanol (g99.8% purity Aldrich) and dichloromethane (g99.5%
purity, Aldrich). The substrate was plunged in a 2 mM dichloromethane
solution of 3 for 20 h so as to afford monolayer coverage. Overlayers of
physisorbed molecules not directly grafted to the gold surface were
removed by repeated washings with pure dichloromethane. A similar
procedure already described elsewhere5a was used for the preparation of
a monolayer of 1. We expect that molecules form an homogeneous
monolayer deposit, in analogy to our previous reports on 1,5a,14 with no
preferential orientation on the surface. The substrate was mounted on
the sample holder of the XMCD setup using a screwed copper mask.
Bulk reference samples were thick films obtained by drop casting of
2 mM solutions of 1 or 3 in dichloromethane on a gold support.
XMCD Experiments. The XAS and XMCD experiments on bulk
samples were carried out at the ID08 beamline of the European
Synchrotron Radiation Facility (ESRF), Grenoble (France),15a and at
the UE46-PGM beamline of BESSY II synchrotron, Helmholtz-
Zentrum Berlin f€ur Materialen und Energie (Germany).15b Monolayers
were measured at the SIM-X11MA beamline of Swiss Light Source
(SLS), Paul Scherrer Institut (PSI), Villigen (Switzerland).15c Drop cast
and monolayer samples were inserted in a sample holder attached to the
coldfinger of a 4He cryostat working down to 7 K. Spectra were recorded
at Fe- and Cr-L2,3 edges (2p-3d transitions). All the measurements were
carried out at very low X-rays photon flux, so as to avoid radiation
[Fe3Cr(L)2(dpm)6] (3). Compound 4 (0.5323 g, 0.5870 mmol) was
suspended in a MeOH:Et2O mixture (1:2 v/v, 50 mL) and stirred for
10 min. Solid CrCl3(THF)3 (0.1466 g, 0.3913 mmol) was then added in
one portion, and the mixture stirred until complete dissolution of the
solids to give a clear, dark-red solution (ca. 15 min). NaOMe (3.08 M in
MeOH, 0.40 mL, 1.23 mmol) was introduced dropwise with vigorous
stirring, and the mixture stirred for additional 30 min (a dark-red
precipitate appeared). A MeOH:Et2O mixture (1:4 v/v, 165 mL) was
then added, and the reaction mixture was stirred for 1 h and left
undisturbed overnight. The precipitated NaCl was filtered off using
a G4 frit and washed extensively with Et2O until colorless washings
(ca. 50 mL). The liquid phase was evaporated at reduced pressure with
gentle heating (30-40 ꢀC) to give a brown powder (0.579 g). A portion
of the solid (0.105 g, 0.0685 mmol Cr) and H3L (0.0600 mg,
0.196 mmol, 2.7 equiv) were refluxed in DME (2 mL) for 4 h to give
a dark brown solution. Slow evaporation of the solvent afforded
a microcrystalline solid which was collected and washed extensively
with methanol until colorless washings. The solid was suspended in
the minimum amount of acetonitrile:DME (1:1) at 85 ꢀC. Slow
cooling of the clear solution afforded the desired product as red-black
prisms that were collected by filtration, washed with acetonitrile
and dried in vacuum (0.073 mg, 55% yield). Anal. Calcd. (%) for
2912
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917

Inorganic Chemistry
Table 1. Selected Interatomic Distances (Å) and Interbond Angles (deg) in 3, with Estimated Standard Deviation in Parenthesesa
ARTICLE
Cr1 Fe2
3.0100(10)
5.1970(9)
1.961(2)
1.977(2)
1.980(2)
2.001(2)
90.15(9)
80.70(12)
166.63(14)
79.89(12)
86.28(9)
91.71(9)
79.60(9)
89.56(9)
97.48(9)
170.49(9)
92.79(11)
99.71(10)
Cr1 Fe3
3.0119(6)
120.37(2)
1.957(2)
Fe2 Fe3
5.2250(8)
119.25 (3)
1.956(2)
3 3 3
3 3 3
3 3 3
Fe3 Fe30
Fe2 Cr1 Fe3
Fe3 Cr1 Fe30
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
Cr1-O1
Cr1-O2
Cr1-O3
Fe2-O1
Fe3-O20
Fe2-O4
Fe3-O3
Fe3-O8
2.000(2)
Fe2-O5
Fe3-O6
Fe3-O9
1.9943(19)
2.002(2)
1.979(2)
Fe3-O7
2.003(3)
2.003(2)
O2-Cr1-O1
O10-Cr1-O1
O20-Cr1-O2
O10-Fe2-O1
O5-Fe2-O4
O50-Fe2-O1
O3-Fe3-O20
O3-Fe3-O6
O3-Fe3-O9
O6-Fe3-O9
O7-Fe3-O8
Cr1-O1-Fe2
O3-Cr1-O1
O20-Cr1-O1
O30-Cr1-O2
O4-Fe2-O1
O50-Fe2-O5
O40-Fe2-O4
O3-Fe3-O7
O20-Fe3-O6
O20-Fe3-O9
O3-Fe3-O8
O6-Fe3-O8
Cr1-O2-Fe30
90.62(9)
O3-Cr1-O2
O30-Cr1-O1
O30-Cr1-O3
O5-Fe2-O1
O40-Fe2-O1
O50-Fe2-O4
O20-Fe3-O7
O7-Fe3-O6
O7-Fe3-O9
O20-Fe3-O8
O9-Fe3-O8
Cr1-O3-Fe3
90.65(9)
100.07(9)
80.72(9)
166.16(9)
99.80(13)
96.92(9)
93.50(9)
168.75(14)
93.23(14)
169.96(11)
97.94(10)
89.66(10)
95.93(10)
87.49(10)
99.80(10)
172.82(9)
86.00(9)
92.19(10)
85.84(10)
88.16(10)
172.89(10)
85.43(10)
99.88(10)
a Primed atoms are related to unprimed ones by a 2-fold rotation along Cr1 Fe2.
3 3 3
Figure 2. Experimental and simulated EPR spectra of 3 at 94.24 GHz.
Asterisks and crosses mark the parallel type bands assigned to Cr-
centered Fe3Cr (S = 6) and Fe4 species (S = 5), respectively.
Figure 1. Core structure (a) and complete molecular structure (b) of 3
determined by X-ray diffraction at 120(2) K, omitting H atoms and
disorder on alkyl chains. Primed atoms are related to unprimed ones by a
2-fold rotation around Cr1 Fe2.
3 3 3
a Et2O:MeOH mixture, removing the precipitated NaCl, eva-
porating the solvent completely, and refluxing the solid residue
with excess H3L in DME (eq 1). A final recrystallization from
acetonitrile:DME afforded 3 as red-black prisms in 55% yield.
damaging effects,10 and in the Total Electron Yield (TEY) detection
mode to achieve the required surface sensitivity.15d XMCD spectra were
evaluated as (σ - - σ þ), where σ - and σ þ are the measured cross
sections for right and left circularly polarized light, respectively (that is,
for photon helicity vector antiparallel and parallel, respectively, to the
applied magnetic field). To minimize systematic errors, each XMCD
spectrum was obtained from eight measurements taken by switching the
polarization and the field.15e In the experiments, we also checked
the absence of any preferential orientation of molecules on the surface
ð3=2Þ½Fe2ðOMeÞ2ðdpmÞ4ꢁ þ CrCl3ðTHFÞ3 þ 3NaOMe
þ 2H3L f ½Fe3CrðLÞ2ðdpmÞ6ꢁ þ 3NaCl þ 3THF þ 6MeOH
ð1Þ
Chemical analysis indicates that the Cr:Fe ratio (1:3.75) in the
solid is lower than used in the synthesis (1:3). As described
below, the location of the chromium(III) ions was clearly
revealed by structural data, EPR spectra and magnetic
measurements.
(no X-ray Natural Linear Dichroism) so that (σ þ σ -)/2 was
þ
considered a good approximation to the isotropic spectrum in the
electric dipole approximation. All reported XMCD contributions are
normalized, that is, they are obtained by dividing the measured XMCD
signal by the edge-jump of the isotropic absorption spectra at the energy
of its maximum amplitude. Therefore, the XMCD spectra are propor-
tional to the XMCD response per metal atom.
Single-crystal X-ray diffraction was used to determine the
crystal and molecular structure of 3 (see Figure 1 and Supporting
Information, Figure S1).16 Selected structural parameters for 3
are gathered in Table 1. Although 1 and 3 are isomorphous,
differences stand out clearly when examining the coordination
geometry of the metal ions. While the average M-O distances
for the peripheral ions in 1 and 3 are identical within experimental
’ RESULTS AND DISCUSSION
Compound [Fe3Cr(L)2(dpm)6] (3) was synthesized by re-
acting [Fe2(OMe)2(dpm)4] with CrCl3(THF)3 and NaOMe in
2913
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917

Inorganic Chemistry
ARTICLE
Figure 4. XAS and XMCD spectra obtained at 30 kOe and 7 K for
monolayer and bulk samples of 3, along with Ligand Field Multiplet
Figure 3. Raw and corrected DC magnetic data for 3, with best fit
calculated curves (solid lines). The main panel displays χmT vs T data,
while the inset shows Mm vs H/T data recorded at three temperatures.
For the sake of clarity, raw magnetization data are omitted.
-
simulations. XAS is presented in the isotropic form (σ þ σ þ)/2;
XMCD is evaluated as (σ - - σ þ) and is normalized with respect to the
isotropic signal (see Experimental Section). A vertical offset has been
applied to the spectra for a better presentation.
error (1.993(11) vs 1.991(13) Å), the central ion in 3 forms
shorter bonds with the bridging O atoms of the tripod (1.958(3)
vs 1.983(5) Å). In addition, the pitch of the helical structure,
defined as the average dihedral angle between the M1(O)2Fe2
and M1(O)2Fe3 moieties (M = Fe, Cr) and the metal ion plane,
is distinctly smaller in 3 (62.2ꢀ vs 67.3ꢀ).9,11 These differences
suggest that the central ion in 3 is smaller than in 1 by about
0.02 Å, as expected for a Cr-centered species.17
DC magnetic data of 3, recorded in the form of low-field χm vs
T and isothermal Mm vs H curves, are in accordance with the
proposed composition and spin ground state. Low-field χmT vs T
data, displayed in Figure 3, are typical for antiferromagnetic
systems featuring a high-spin ground state.11 After correction for
the presence of 16% Fe4, based on the known susceptibility
values for 1,9a the χmT value at low temperature is very close
to the Curie constant for a S = 6 state with g = 2.00 (21 emu
K mol-1). Data at T g 5 K were accurately fitted using an iso-
tropic Heisenberg plus Zeeman spin-Hamiltonian with nearest-
neighbor (JFe-Cr) and next-nearest-neighbor (JFe-Fe) coupling
constants (eq 3).
The EPR spectra of a polycrystalline sample of 3were recorded at
ν = 94.24 GHz in the temperature (T) range 5-20 K (Figure 2).
Below the central field H0 = hν/(2 μB) = 33.66 kOe the spectrum at
5 K exhibits a pattern of almost equally spaced signals (marked with
asterisks in Figure 2). Their intensity variation with field (H) and
T is typical for parallel-type bands in a high-spin system with
easy-axis anisotropy (D < 0).18 From the position of the five
visible lines, it can be inferred that six such transitions appear
below H0. This indicates that the ground state has total spin S = 6,
which is the value expected for a Cr-centered Fe3Cr species with
antiferromagnetic Fe-Cr interactions. The line-to-line separa-
tion also provides |D| about 0.16-0.17 cm-1, while the broad-
ening of perpendicular bands at H > H0 is suggestive of a very
weak rhombic anisotropy.18 A further series of weaker signals,
marked by crosses in Figure 2 and spanning a wider field range at
H < H0, is easily assigned to Fe4 species on the basis of the line-to-
line separation and of the observed temperature and field
dependence.9 If the compound is taken as a solid solution of
Cr-centered Fe3Cr and Fe4, the observed Fe:Cr ratio implies the
presence of 84% Fe3Cr and 16% Fe4. Notably, very similar results
were obtained using a different tripodal ligand.11 The given
composition allowed an accurate reproduction of the spectra
using the spin-Hamiltonian9c,13e
^
0
0
0
H ¼ JFe-Cr^s1 ð^s2 þ^s3 þ^s3 Þ þ JFe-Feð^s2 ^s3 þ^s3 ^s3 þ^s3 3^s2Þ
3
3
3
^
^
þ μBgS H
ð3Þ
3
(1 is the central Cr3þ ion, with spin s1 = 3/2; 2, 3, and 30 label
the peripheral Fe3þ ions, each with spin si = 5/2). The best fit
parameters were JFe-Cr = 16.34(10) cm-1, JFe-Fe = 0.57(3) cm-1
,
and g = 1.9868(11)). Isothermal Mm vs H data, also reported in
Figure 3, were corrected as above and fitted using an axial zero-
field splitting Hamiltonian (eq 4)9c
2
^
^
^
^
Hzfs ¼ μBgS H þ D½Sz - SðS þ 1Þ=3ꢁ
ð4Þ
3
with S = 6, D = -0.154(3) cm-1, and g = 1.964(3).
On the basis of the D value determined by EPR, we obtain the
expected anisotropy barrier as U/kB = |D/kB|S2 = 8.5(1) K,
hence smaller than in 1 (ca. 16 K)9a but still large enough to
induce the slow relaxation of the magnetization typical of SMMs.
The frequency dependence of the real and imaginary compo-
nents of AC susceptibility was measured in zero static field, in the
frequency range ν = 100-25000 Hz and at temperatures down
to 1.8 K (Supporting Information, Figure S2). The tempera-
ture dependence of the relaxation time follows Arrhenius
law τ = τ0 exp(Ueff /kBT) with τ0 = 1.5(4) ꢀ 10-7 s and
Ueff /kB = 8.1(5) K and is thus indicative of a thermally activated
relaxation process (Supporting Information, Figure S3). The
value of the effective barrier, Ueff /kB, is in good agreement with
the total splitting of the S= 6 multiplet in zero field (U/kB =8.5(1) K).
2
2
2
^
^
^
^
^
^
HEPR ¼ μBS g H þ D½Sz - SðS þ 1Þ=3ꢁ þ EðSx - Sy Þ
3 3
0
0
^
þ B4O4
ð2Þ
^
0
^ ^
where S is the total spin operator (with components Sx, Sy, and
^
^
Sz), O4 is the fourth-order axial anisotropy operator, and the
other symbols have their usual meaning. The best-fit parameter
set was (for Fe3Cr) S=6, D=-0.165(2) cm-1, E= 0.005(1) cm-1
,
B40 = 0.0(1) 10-5 cm-1, gx,y = 1.990(2), gz = 2.000(2); (for Fe4)
3
S = 5, D = -0.405(2) cm-1, E = 0.018(4) cm-1, B04 = 1.5(2)
3
10-5 cm-1, gx,y = 1.995(2), gz = 2.005(2).
2914
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917

Inorganic Chemistry
ARTICLE
ascribed to the suppression of the opposing contribution of the
central Fe3þ so as to provide direct proof of the ferrimagnetic
ground state spin structure in 1. To this aim, we have calculated
the spin component in the applied field direction (z) at each
metal site (sz(i)), using the solid-state J-values with g = 2.000, H =
30 kOe, T = 7 K and disregarding magnetic anisotropy for
simplicity. Notice that the first-excited exchange multiplets in 1
and 3 lie 37.5 and 20.2 cm-1, respectively, above the ground
state, so that their thermal population is negligible in the
temperature range of interest. A rigid-spin behavior is then
followed and the sz(i) values are simply proportional to the
molecular Sz as given by the appropriate Brillouin function.20 The
numeric values of sz(i) calculated at iron sites of 1 and 3 are
presented in the inset of Figure 5 and in Supporting Information,
Table S1. The average spin component per iron site is then given by
[3sz(Fep) þ sz(Fec)]/4 = -0.934p in 1 and as sz(Fep) = -1.905p
in 3. Assuming the XMCD signal to be the same for Fep and Fec and
proportional to the local magnetic moment, the normalized XMCD
signal is expected to undergo an approximately 2-fold increase on
passing from 1 to 3, as experimentally observed.
A deeper comparison between the XMCD signals of 1 and 3 at
the Fe-L3 edge reveals another relevant difference. At an energy
of about 707.9 eV (marked by an asterisk in Figure 5) the
dichroic signal is negative in 3 but close to zero in 1. The former
behavior is typical for octahedral high-spin FeO6 complexes.19
Since the coordination geometry of Fep ions in the two com-
pounds is virtually identical, the feature observed in 1 at 707.9 eV
may reflect differences in the energy dependence of the XMCD
contribution for Fep and Fec, because of their inequivalent
coordination spheres.
Figure 5. XMCD contribution at the Fe-L2,3 edges (evaluated as a
fraction of the isotropic signal) measured on monolayers of 1 and 3 at 30
kOe and 7 K. A schematic view of the metal core in 1 and 3 is also
provided; numbers give the calculated spin component sz(i)/p along the
applied field H at each metal site; black arrows depict the corresponding
magnetic moment, -gμBsz(i); the asterisk marks the feature at 707.9 eV
dicussed in the text.
A slower relaxation process is also visible which we attribute to
the minority Fe4 component.
All magnetic data coherently indicate that compound 3 has
the same ferrimagnetic spin structure of 1. However, it is an
heterometallic system, and its central Cr3þ ion was exploited as
an additional internal spectroscopic probe for combined XAS/
XMCD studies at Fe-L2,3 and Cr-L2,3 edges. Monolayer deposits
of 3 on Au(111) were assembled from dichloromethane solution,
as described for 1 and 2.5,14 Their XAS and XMCD spectra at
30 kOe and 7 K are similar to those recorded on bulk references,
obtained by drop casting (Figure 4 and Supporting Information,
Figure S4). In both cases, the XMCD signal reaches its maximum
amplitude at 709.1 eV (Fe-L3 edge) and 577.5 eV (Cr-L3 edge)
and displays the same fine structure, indicating a similar electro-
nic structure for the monolayer and for the bulk phase (small
intensity differences may arise from the relevant baseline signal of
the gold surface in the monolayer deposit, which influences
the normalization procedure). Moreover, whereas the features
observed at Fe-L2,3 edges are characteristic for field-polarized
high-spin Fe3þ ions,19 the signal at Cr-L2,3 edges is of opposite
sign than for an isolated Cr3þ ion. This suggests that the
magnetic polarization of Cr opposes the applied field and that
Fe and Cr spins are antiparallel in both bulk and monolayer
samples. To check the consistency between the measured XAS/
XMCD spectra and the electronic structure of 3, as determined
by EPR and magnetic investigations, we have carried out Ligand
Field Multiplet (LFM) calculations (see Supporting Information
for details). The simulated spectra, reported in Figure 4 and
Supporting Information, Figure S5, compare well with the
observed ones in both intensity and fine structure and lend
quantitative support to the above analysis.
’ CONCLUSIONS
XMCD is becoming of widespread use to reveal the electronic
structure of magnetic molecules at surfaces. Many studies
focused on simple metal-ion complexes, such as metal-
porphirinato7a,b and metal bis(phthalocyaninato) species,7c
whose crystal field parameters and magnetic properties were
found to be robust upon adsorption on a metal substrate. With
their complex and fragile architectures, polynuclear SMMs pose
new challenges in XMCD analysis. We have shown here that the
XMCD spectrum of a tetrairon(III) SMM at Fe-L2,3 edges
contains fingerprints of its spin structure. Upon isomorphous
replacement of the central iron(III) with a chromium(III) ion,
the intensity and shape of the signal change significantly, in
agreement with the field-opposing magnetic contribution of the
central ion. Substitution has the same effect in both bulk and
monolayer samples, confirming that Fe4 complexes have a
ferrimagnetic spin structure on surfaces. It also follows that,
although much weaker than crystal field splittings, superexchange
interactions between metal ions can maintain the same sign in
bulk samples and on surfaces.7d The herein-demonstrated spe-
cificity of Fe4 XMCD response is now expected to have a far-
reaching usefulness in the development of more demanding
deposition methods, including thermal evaporation in high
vacuum and deposition on more reactive substrates, like ferro-
magnetic metals.
Since the Fe3þ ions in 1 and 3 have very similar coordination
environments, a meaningful comparison can be made between
the XMCD features shown by monolayers of the two compounds
at the Fe-L2,3 edges (Figure 5). At the same T and H values, the
amplitude of the dichroic component at 709 eV increases from
39% to 63% when replacing the central Fe3þ with Cr3þ (a similar
trend was observed on bulk samples, with percentages of 40%
and 76%; see Supporting Information, Figure S4). If the con-
tribution of Fe4 complexes is subtracted, assuming that their
molar fraction in the monolayer is the same as in the bulk, the
XMCD signal indeed reaches 70%. We checked whether such a
large increase in the magnetic polarization at iron sites can be
’ ASSOCIATED CONTENT
S
Supporting Information. Additional crystallographic
b
diagrams for 3 (showing thermal ellipsoids), crystallographic
2915
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917

Inorganic Chemistry
ARTICLE
data in CIF format, details of AC magnetic characterization, of
XAS and XMCD data treatment and of LFM calculations. This
material is available free of charge via the Internet at http://pubs.
acs.org.
Gupta, S. S.; Sarma, D. D.; Bencok, P.; Stanescu, S.; Clair, S.; Pons,
S.; Lin, N.; Seitsonen, A. P.; Brune, H.; Barth, J. V.; Kern, K. Nat.
Mater. 2009, 8, 189. (c) Stepanow, S.; Honolka, J.; Gambardella, P.;
Vitali, L.; Abdurakhmanova, N.; Tseng, T.-C.; Rauschenbach, S.;
Tait, S. L.; Sessi, V.; Klyatskaya, S.; Ruben, M.; Kern, K. J. Am. Chem.
Soc. 2010, 132, 11900. (d) Corradini, V.; Moro, F.; Biagi, R.; De
Renzi, V.; del Pennino, U.; Bellini, V.; Carretta, S.; Santini, P.;
Milway, V. A.; Timco, G.; Winpenny, R. E. P.; Affronte, M. Phys. Rev.
B 2009, 79, 144419.
(8) (a) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y.
J. Am. Chem. Soc. 2003, 125, 8694. (b) Branzoli, F.; Filibian, M.; Carretta,
P.; Klyatskaya, S.; Ruben, M. Phys. Rev. B 2009, 79, 220404(R). (c)
Branzoli, F.; Carretta, P.; Filibian, M.; Zoppellaro, G.; Graf, M. J.; Galan-
Mascaros, J. R.; Fuhr, O.; Brink, S.; Ruben, M. J. Am. Chem. Soc. 2009,
131, 4387.
(9) (a) Gregoli, L.; Danieli, C.; Barra, A.-L.; Neugebauer, P.;
Pellegrino, G.; Poneti, G.; Sessoli, R.; Cornia, A. Chem.—Eur. J. 2009,
15, 6456. (b) Accorsi, S.; Barra, A.-L.; Caneschi, A.; Chastanet, G.;
Cornia, A.; Fabretti, A. C.; Gatteschi, D.; Mortalꢁo, C.; Olivieri, E.;
Parenti, F.; Rosa, P.; Sessoli, R.; Sorace, L.; Wernsdorfer, W.; Zobbi,
L. J. Am. Chem. Soc. 2006, 128, 4742. (c) Barra, A.-L.; Bianchi, F.;
Caneschi, A.; Cornia, A.; Gatteschi, D.; Gorini, L.; Gregoli, L.;
Maffini, M.; Parenti, F.; Sessoli, R.; Sorace, L.; Talarico, A. M. Eur.
J. Inorg. Chem. 2007, 4145.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: acornia@unimore.it. Phone: þ39-059-2055032. Fax:
þ39-059-373543.
’ ACKNOWLEDGMENT
We thank F. Scheurer, J. P. Kappler, B. Muller, P. Imperia, F.
Nolting, and N. Brookes for their contribution to the develop-
ment of the XMCD setup employed in this work. The research
was funded by the European Community (ELISA grant agree-
ment no. 226716, MolSpinQIP FP7-ICT-2007-C-211284 and
ERANET “NanoSci-ERA: Nanoscience in European Research
Area” projects), by Italian CNR (Commessa PM.P05.012), by
Italian MIUR (PRIN 2008) and by Ente CRF.
(10) Mannini, M.; Pineider, F.; Sainctavit, Ph.; Joly, L.; Fraile-
Rodrìguez, A.; Arrio, M.-A.; Cartier dit Moulin, Ch.; Wernsdorfer, W.;
Cornia, A.; Gatteschi, D.; Sessoli, R. Adv. Mater. 2009, 21, 167.
(11) Tancini, E.; Rodriguez-Douton, M.-J.; Sorace, L.; Barra, A.-L.;
Sessoli, R.; Cornia, A. Chem.—Eur. J. 2010, 16, 10482.
(12) (a) Wu, C.-H. S.; Rossman, G. R.; Gray, H. B.; Hammond,
G. S.; Schugar, H. J. Inorg. Chem. 1972, 11, 990. (b) Herwig, W.; Zeiss,
H. H. J. Org. Chem. 1958, 23, 1404.(c) Jeffery, J. H.; Bassett, J.;
Mendham, J.; Denney, R. C. Vogel Analisi Chimica Quantitativa; Casa
Editrice Ambrosiana: Milano, Italy, 1995; p 378.
(13) (a) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.
J. Appl. Crystallogr. 1993, 26, 343. (b) Sheldrick, G. M. Acta Crystallogr.
2008, A64, 112. (c) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837.(d)
Fleischer, H.; Schollmeyer, D. Acta Crystallogr., Sect. E (Struct. Rep.
Online) 2001, 57, o330 (QIBJOA entry of CCDC). (e) Jacobsen,
C. J. H.; Pedersen, E.; Villadsen, J.; Weihe, H. Inorg. Chem. 1993,
32, 1216.
’ REFERENCES
(1) (a) Sanvito, S. Nat. Phys. 2010, 6, 562. (b) Sanvito, S. Nature
2010, 467, 664. (c) Rocha, A. R.; García-Suꢀarez, V. M.; Bailey, S. W.;
Lambert, C. J.; Ferrer, J.; Sanvito, S. Nat. Mater. 2005, 4, 335.
(2) Gatteschi, D.; Cornia, A.; Mannini, M.; Sessoli, R. Inorg. Chem.
2009, 48, 3408.
(3) (a) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000,
408, 541. (b) Bogani, L.; Wernsdorfer, W. Nat. Mater. 2008, 7, 179.
(c) Misiorny, M.; Weymann, I.; Barnas, J. Europhys. Lett. 2010,
89, 18003.
(4) (a) Mannini, M.; Sainctavit, Ph.; Sessoli, R.; Cartier dit Moulin,
C.; Pineider, F.; Arrio, M.-A.; Cornia, A.; Gatteschi, D. Chem.—Eur. J.
2008, 14, 7530.(b) Saywell, A.; Magnano, G.; Satterley, C. J.; Perdig~ao,
L. M. A.; Britton, A. J.; Taleb, N.; del Carmen Gimꢀenez-Lꢀopez, M.;
Champness, N. R.; O’Shea, J. N.; Beton, P. H. Nat. Commun. 2010, 1:75
doi: 10.1038/ncomms1075.
(14) Pineider, F.; Mannini, M.; Danieli, C.; Armelao, L.; Piras, F. M.;
Magnani, A.; Cornia, A.; Sessoli, R. J. Mater. Chem. 2010, 20, 187.
(15) (a) Goulon, J.; Brookes, N. B.; Gauthier, C.; Goedkoop,
J. B.; Goulon-Ginet, C.; Rogalev, M. Phys. B 1995, 208/209, 199.
(b) Englisch, U.; Rossner, H.; Maletta, H.; Bahrdt, J.; Sasaki, S.;
Senf, F.; Sawhney, K. J. S.; Gudat, W. Nucl. Instrum. Meth. A 2001,
467-468, 541. (c) Flechsig, U.; Nolting, F.; Fraile Rodriguez, A.;
Krempasky, J.; Quitmann, C.; Schmidt, T.; Spielmann, S.; Zimoch,
D. AIP Conf. Proc 2010, 1234, 319. (d) Nakajima, R.; Stohr, J.;
Idzerda, Y. U. Phys. Rev. B 1999, 59, 6421.(e) Sessoli, R.; Mannini,
M.; Pineider, F.; Cornia, A.; Sainctavit, Ph. In Magnetism and
Synchrotron Radiation; Beaurepaire, E., Bulou, H., Scheurer, F.,
Kappler, J. P., Eds.; Spinger-Verlag: Berlin, Germany, 2010; Spring-
er Proceedings in Physics, Vol. 133, p 279.
(16) Crystal structure data for 3: chemical formula C96H168O20S2-
Fe3Cr, Mr = 1925.97, crystal system monoclinic, space group C2/c (No.
15), a = 27.4032(15) Å, b = 20.8832(13) Å, c = 18.9981(12) Å, β =
97.929(2)ꢀ, V = 10768.0(11) Å3, Z = 4, T = 120(2) K, λ = 0.71073 Å,
F(000) = 4152, Dcalcd = 1.188 g cm-3, μ = 0.593 mm-1, 2θmax = 55.1ꢀ,
54658/12377 reflections collected/unique (Rint = 0.0334), final R
indices R1 = 0.0651 (I > 2σ(I)), wR2 = 0.1988 (all data), GoF =
1.023, largest diff. peak and hole 0.915/-0.902 eÅ-3. CCDC 793297
contains the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystal-
lographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
(17) Shannon, R. D.; Prewitt, C. T. Acta Crystallogr., Sect. B 1970,
26, 1046.
(5) (a) Mannini, M.; Pineider, F.; Sainctavit, Ph.; Danieli, C.;
Otero, E.; Sciancalepore, C.; Talarico, A. M.; Arrio, M.-A.; Cornia,
A.; Gatteschi, D.; Sessoli, R. Nat. Mater. 2009, 8, 194. (b) Mannini,
M.; Pineider, F.; Danieli, C.; Totti, F.; Sorace, L.; Sainctavit, Ph.;
Arrio, M.-A.; Otero, E.; Joly, L.; Cesar, J. C.; Cornia, A.; Sessoli, R.
Nature 2010, 468, 417.
(6) (a) Coronado, E.; Gimꢀenez-Lꢀopez, M. C.; Korzeniak, T.;
Levchenko, G.; Romero, F. M.; Segura, A.; García-Baonza, V.; Cezar,
J. C.; de Groot, F. M. F.; Milner, A.; Paz-Pasternak, M. J. Am. Chem.
Soc. 2008, 130, 15519. (b) Prinz, M.; Kuepper, K.; Taubitz, C.;
Raekers, M.; Khanra, S.; Biswas, B.; Weyherm€uller, T.; Uhlarz, M.;
Wosnitza, J.; Schnack, J.; Postnikov, A. V.; Schr€oder, C.; George, S. J.;
Neumann, M.; Chaudhuri, P. Inorg. Chem. 2010, 49, 2093. (c)
Khanra, S.; Kuepper, K.; Weyhermuller, T.; Prinz, M.; Raekers, M.;
Voget, S.; Postnikov, A. V.; de Groot, F. M. F.; George, S. J.; Coldea,
M.; Neumann, M.; Chaudhuri, P. Inorg. Chem. 2008, 47, 4605. (d)
Lorusso, G.; Corradini, V.; Candini, A.; Ghirri, A.; Biagi, R.; del
Pennino, U.; Carretta, S.; Garlatti, E.; Santini, P.; Amoretti, G.;
Timco, G.; Pritchard, R. G.; Winpenny, R. E. P.; Affronte, M. Phys.
Rev. B 2010, 82, 144420. (e) Poneti, G.; Mannini, M.; Sorace, L.;
Sainctavit, Ph.; Arrio, M. A.; Otero, E.; Cezar, J. C.; Dei, A. Angew.
Chem., Int. Ed. 2010, 49, 1954.
(7) (a) Wende, H.; Bernien, M.; Luo, J.; Sorg, C.; Ponpandian, N.;
Kurde, J.; Miguel, J.; Piantek, M.; Xu, X.; Eckhold, Ph.; Kuch, W.;
Baberschke, K.; Panchmatia, P. M.; Sanyal, B.; Oppeneer, P. M.;
Eriksson, O. Nat. Mater. 2007, 6, 516. (b) Gambardella, P.; Stepanow,
S.; Dmitriev, A.; Honolka, J.; de Groot, F. M. F.; Lingenfelder, M.;
2916
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917

Inorganic Chemistry
ARTICLE
(18) Barra, A.-L.; Brunel, L. C.; Gatteschi, D.; Pardi, L.; Sessoli, R.
Acc. Chem. Res. 1998, 31, 460.
(19) (a) Brice-Profeta, S.; Arrio, M.-A.; Tronc, E.; Menguy, N.;
Letard, I.; Cartier dit Moulin, C.; Noguꢁes, M.; Chanꢀeac, C.; Jolivet,
J.-P.; Sainctavit, Ph. J. Magn. Magn. Mater. 2005, 288, 354. (b)
Carvallo, C.; Sainctavit, Ph.; Arrio, M.-A.; Menguy, N.; Wang, Y.;
Ona-Nguema, G.; Brice-Profeta, S. Am. Mineral. 2008, 93, 880.
(20) Carlin, R. L. Magnetochemistry; Springer-Verlag: Berlin, Germany,
1986.
2917
dx.doi.org/10.1021/ic102184n |Inorg. Chem. 2011, 50, 2911–2917