Slow magnetic relaxations in manganese(III) tetra(meta -fluorophenyl) porphyrin-tetracyanoethenide. Comparison with the relative single chain magnet ortho compound

Article abstract of DOI:10.1021/ic3014927

Mn(III) tetra(meta-fluorophenyl)porphyrin-tetracyanoethenide coordination polymer (abbreviated meta-F) was synthesized and crystallographically and magnetically characterized. The compound crystallizes in the space group C2/c with four equivalent molecules in the unit cell arranged along two symmetry related nonparallel linear chain directions. Magnetic properties were studied by SQUID dc magnetization and ac susceptibility techniques and high field-high frequency electron spin resonance (HF-ESR). Glassy transition to a ferromagnetic-like state is observed at 10 K accompanied by slow magnetic relaxations. The glassiness is interpreted as due to 3D domain wall pinning. In a bias dc magnetic field the width of the relaxation time distribution decreases and the relaxations become similar to the relaxations of the single chain magnet Mn(III) tetra(ortho-fluorophenyl)porphyrin-tetracyanoethenide (abbreviated ortho-F), for which comparative HF-ESR studies were also conducted in this work. Magnetic properties of these two compounds are compared, and the nature of magnetic relaxations in meta-F is discussed.

Full text of DOI:10.1021/ic3014927

Article
pubs.acs.org/IC
Slow Magnetic Relaxations in Manganese(III) Tetra(meta-fluorophenyl)-
porphyrin-tetracyanoethenide. Comparison with the Relative Single
Chain Magnet ortho Compound
Z. Tomkowicz,*,† M. Rams,† M. Bałanda,‡ S. Foro,§ H. Nojiri,∥ Y. Krupskaya,⊥ V. Kataev,⊥ B. Buchner,⊥
̈
S. K. Nayak,#,¶ J. V. Yakhmi,× and W. Haase*,#
†Institute of Physics, Jagiellonian University, Reymonta 4, 30-059 Krakow
‡H. Niewodniczan
ski Institute of Nuclear Physics PAN, Radzikowskiego 152, 31-342 Krakow
§Clemens-Schopf-Institut fur Organische Chemie und Biochemie, Technische Universitat Darmstadt, Petersenstrasse 22,
́
, Poland
́
́
, Poland
̈
̈
̈
D-64287 Darmstadt, Germany
∥Institute for Materials Research, Tohoku University, Katahira 2-1-1, Sendai 980-8577, Japan
⊥Institute for Solid State Physics, IFW Dresden, Helmholtzstrasse 20, D-01069 Dresden, Germany
#Eduard-Zintl-Institut fur Anorganische und Physikalische Chemie, Technische Universitat Darmstadt, Petersenstrasse 20,
̈
̈
D-64287 Darmstadt, Germany
×Technical Physics & Prototype Engineering Division, Bhabha Atomic Research Centre, Mumbai -400 085, India
S
* Supporting Information
ABSTRACT: Mn(III) tetra(meta-fluorophenyl)porphyrin-tetracyanoethenide coordination polymer (abbreviated meta-F) was
synthesized and crystallographically and magnetically characterized. The compound crystallizes in the space group C2/c with four
equivalent molecules in the unit cell arranged along two symmetry related nonparallel linear chain directions. Magnetic
properties were studied by SQUID dc magnetization and ac susceptibility techniques and high field-high frequency electron spin
resonance (HF-ESR). Glassy transition to a ferromagnetic-like state is observed at 10 K accompanied by slow magnetic
relaxations. The glassiness is interpreted as due to 3D domain wall pinning. In a bias dc magnetic field the width of the relaxation
time distribution decreases and the relaxations become similar to the relaxations of the single chain magnet Mn(III) tetra(ortho-
fluorophenyl)porphyrin-tetracyanoethenide (abbreviated ortho-F), for which comparative HF-ESR studies were also conducted
in this work. Magnetic properties of these two compounds are compared, and the nature of magnetic relaxations in meta-F is
discussed.
various substitutions, and their magnetic properties may be
tuned in a broad range, as shown by Miller et al., who syn-
thesized many compounds of this family and characterized
them structurally and magnetically.2−5 Haase et al. synthesized
Mn-porphyrin−TCNE compounds (TCNE-tetracyanoethe-
nide) with long alkoxy chains attached at the porphyrin ring
periphery at the para position of the phenyl ring, see Figure 1.
1. INTRODUCTION
It is well-known that 1-dimensional systems do not display long-
range magnetic order. Nevertheless, ferro- or ferrimagnetic chains
comprising in their structure highly anisotropic ions of transition
elements are of great interest due to slow magnetic relaxations.
These systems, known as single chain magnets (SCM), are
intensively investigated since 2001.1
One of the rich families of 1-dimensional magnetic systems
are coordination polymers composed of Mn porphyrin−radical
molecule pairs. Plenty of such compounds may be obtained by
Received: July 10, 2012
Published: September 5, 2012
© 2012 American Chemical Society
9983
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
sight, just differs in the last point from the behavior observed in
the present work. It should be noted that slow relaxations
below critical temperature Tc were confirmed by reports of
SCM-like behavior in an ordered magnetic phase of
compounds consisting of antiferromagnetically coupled chains
of Mn(III)2−Ni(II)15 or Mn(III)−Ni(II) units16 as well as of
ferromagnetically coupled chains of Co(II)−radical units.11,17
The question arises about the character of relaxations
observed in this work. As known, spin dynamics of quasi-1D
magnetic systems above Tc is governed in general by spin
solitons, breathers and spin waves. Solitons are bound states of
spin waves, which may have energy lying below the spin wave
band.18 But in contradistinction to spin waves, they are
localized excitations, which can freely move in the crystal lattice.
Below Tc a weak interchain coupling leads to a 3D ordered
structure with static domains. Nonetheless, narrow solitons,
called kinks, were experimentally observed both above and
below phase transition temperature in quasi-1D systems.19 It
was also theoretically shown that, for weakly coupled chains,
static solitons may occur in the ordered phase, ordering of
which is induced by pairing of solitons belonging to the same
chain.20 Morgunov et al.21 observed spin soliton and standing
spin wave resonances, which simultaneously appeared in ESR
spectra of some chiral 2D and 3D molecular crystals in their
ferromagnetic state. A powerful method, which allows direct
study of a broad spectrum of magnetic excitations, is the high
field, high frequency ESR (HF-ESR). We included HF-ESR
spectroscopy for investigations of our Mn porphyrin-TCNE
compounds. We were interested to compare excitation gaps for
meta- and ortho-F compounds, the values of which should
reflect difference in electronic parameters and a character of
magnetic excitations. It is worth noting that HF-ESR was
already used to study a Mn(III) porphyrin complex MnTPPCl
(Mn axially ligated by one Cl ligand).22 As a result of this study
the anisotropy gap Δ was determined to be 6.9 cm−1, which
gives the value of the single ion zero field parameter D = −Δ/3 =
−2.29 cm−1 (−3.3 K).
Figure 1. Molecular structure of Mn(III) tetra(fluorophenyl)-
porphyrin-tetracyanoethenide. R, R′, R″ labels denote para, ortho
and meta positions of the phenyl ring, respectively.
For these compounds interesting magnetic properties were
reported at low temperatures6,7 and even liquid crystalline
properties8 above room temperature. In spite of the long
interchain distance ∼30 Å, a transition into magnetically
ordered state was observed at about 10−20 K for all these
substitutions. In the course of further studies special attention
was paid to the fluorine atom attached at different positions of
the phenyl ring. It was interesting to see how the strongly
electronegative fluorine ion influences magnetic exchange and
anisotropy of Mn(III) ion. Compounds with para and ortho
substitutions of F were studied also by Miller and co-workers.
They reported that para-F compound exhibits a magnetic phase
transition at Tc = 28 K and that the intrachain value of
exchange integral J for this compound is large, above 300 K.4
(J is defined by the Hamiltonian / = −J Si·Sj.) For ortho-F an
anomaly of ac magnetic susceptibility at 12.5 K was observed
and interpreted as antiferromagnetic transition.5 Consistent
results for para-F compound were obtained by Balanda et al.7,
but as for the ortho-F substitution, the antiferromagnetic
transition was not confirmed by these researchers. Instead, the
compound was reported to be a SCM.9 The character of
magnetic relaxations in ortho-F was investigated, and it was
shown that they have the soliton character.10 In continuation of
these studies the fluorine atom was substituted at the meta
position. As we will show in this report, the compound
obtained shows 3D glassy transition at Tc = 10 K, which is
typical for this family of compounds. The nature of the
glassiness observed may be connected with the 3D- domain
wall pinning11 or with growing of fractal clusters as proposed by
Etzkorn et al.12 Slow relaxations were observed near transition
and down to the lowest temperature attainable (1.8 K). With
increasing applied dc magnetic field these relaxations became
similar to those observed for the ortho-F compound. Quite
recently a paper of Pini et al. has appeared13 (see also ref 14)
describing the effects of an applied magnetic field and finite size
of chains for SCM. Their χac(T) curves have two maxima: the
low temperature one is frequency dependent, and the high
temperature one is not. This behavior, although similar at first
The crystallographic and magnetic characterization of the
meta-F compound, HF-ESR investigations of both meta- and
ortho-F compounds and their comparative study is the aim of
this work. We hope to shed light on the problem of coexistence
of phase transition and slow relaxations in quasi-1D systems.
2. EXPERIMENTAL SECTION
2.1. Synthesis of Manganese(III) tetrakis(3-fluorophenyl)-
porphyrin-tetracyanoethenide (Scheme 1). 2.1.1. Synthesis of 1.23
Freshly distilled pyrrole (13.42 g, 0.2 mol) and 3-fluorobenzaldehyde
(24.82 g, 0.2 mol) were added dropwise simultaneously to a boiling
propionic acid (750 mL). Refluxing was continued for 30 min, and
the mixture was allowed to stand overnight at ambient temperature.
The dark solution was filtered, and the solid was washed with
methanol and water and dried. The crude solid was purified by column
chromatography (silica, eluant CH2Cl2−hexane = 8:2) to get pure
tetrakis(3-fluorophenyl) porphyrin (1) as a purple powder (5.5 g,
1
16%). H NMR (CDCl3): δ 8.92 (s, 8H, pyrrole β-H), 8.04 (m, 8H),
7.72 (m, 4H), 7.53 (m, 4H), −2.82 (bs, 2H). IR (film): 3434, 1644,
1612, 1215, 757 cm−1.
2.1.2. Synthesis of 2.24 A solution of 1 (2.75 g, 4 mmol) and Mn
(OAc)2·4H2O (9.8 g, 40 mmol) in dry N,N-dimethylformamide (30
mL) was refluxed under argon (4 h) when thin layer chromatography
(TLC) showed complete consumption of 1. The mixture was cooled
to ambient temperature and poured into 1400 mL of ice-cooled brine
solution when a solid precipitated. The solid was filtered, washed with
distilled water (1500 mL) and dried. The green solid was dissolved in a
minimum amount of methanol (350 mL), and the solution was poured
9984
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
Scheme 1
into aqueous 6 N HCl (350 mL). The solution was allowed to stand
overnight, filtered, washed with water and dried in air. The brown
colored solid was crystallized from a toluene−hexane mixture to afford
pure tetrakis(3-fluorophenyl)porphyrin manganese(III) chloride (2)
as a green crystalline solid (2.10 g, 68%).
up to 15 T using a homemade spectrometer based on a Millimeter-
wave Vector Network Analyzer (AB Millimetre) 26
́
.
The ortho-F compound was studied in Sendai. The experimental
setup is described in ref 27. Measurements were performed in pulse
magnetic fields. The time of the pulse was ∼10 ms.
2.1.3. Synthesis of 3.25 Using a Schlenck apparatus the compound
2 (1.55 g, 2 mmol) was dissolved in dry methanol and dry pyridine,
and to that NaBH4 was added. The mixture was then refluxed (30
min) and cooled to ambient temperature, argon purged methanol was
added and the mixture was kept standing overnight. The mixture was
filtered under inert atmosphere, washed with dry methanol and finally
dried under vacuum to get a pink colored solid. The solid was
dissolved in argon purged toluene, and a solution of tetracyano-
ethylene (0.512 g, 4 mmol) in argon purged toluene was added to it.
The mixture was briefly stirred and kept standing for 4 days under
argon without stirring. The precipitated dark solid was washed with
argon purged toluene and dried under vacuum for 2 h to get a dark
pink solid (0.560 g, 32%). IR (Nujol): 2198 (m), 2156 cm−1.
Two kinds of material were obtained from synthesis: finely
dispersed polycrystalline material and small crystallites. The crystallites
were very brittle and broke down under a slight touch. For these
studies mainly powdered crystallites were used.
3. CRYSTAL STRUCTURE DESCRIPTION
The crystal structure of the studied meta-F compound is shown
in Figure 2; see also CIF file in Supporting Information. In Table 1
the crystallographic parameters are given. This structure belongs to
the monoclinic space group C2/c with four crystallographically
equivalent pairs of the porphyrin and TCNE molecules in
the elementary crystal cell. These porphyrin-TCNE units are
alternatively ordered along two symmetry related chains
running parallel to diagonals of the a,b crystal cell wall. Thus,
the angle between Mn−Mn bonds of two chains, as being equal
to the angle between diagonals, equals 66°. It should be noted
that the porphyrin planes are not exactly perpendicular to the
Mn−Mn direction: the declination angle is 6.05°. Manganese
atoms are located in the inversion center in a distorted octa-
hedral coordination by 2N1 and 2N2 nitrogen atoms in the
porphyrin plane and 2N3 atoms perpendicular to the plane.
Because of the elongated Mn−N3 distance, the N3−Mn−N3
direction is expected to be the magnetic easy direction. The
angle between the linear N3−Mn−N3 bonds of two chains is
53.9°. The two symmetry independent phenyl rings show
statistical disorder. One phenyl ring is statistically distributed
between two equilibrium orientations differing by 180°, and the
second phenyl ring is disordered, as is well characterized via the
high thermal parameters of some carbon atoms.
2.2. Methods of Investigation. 2.2.1. Crystal Structure. The
crystal structure was determined from X-ray data collection (2Θmax
=
52.74°) on an Oxford Xcalibur diffractometer. The lattice parameters
were refined by 5830 reflections (Θ = 2.59−22.34°). The structure
was solved by the direct method using SHELXS-97, and refinement
was done with SHELXL-97. The H atoms were positioned with idealized
geometry using a riding model with C−H = 0.93 Å, and Uiso(H) values
were set equal to 1.2Ueq(parent atom).
2.2.2. Magnetic Ac and Dc Measurements. These measurements
were performed with a Quantum Design SQUID magnetometer,
model MPMS-XL5. The sample for measurements was powdered and
pressed into a pellet in order to avoid the reorientation of grains in the
external magnetic field. The data were corrected for the diamagnetic
contribution as deduced with the use of a table of Pascal constants.
One good shaped crystallite was chosen for magnetization measure-
ments and secured in melted eicosane.
2.2.3. High-Field/High-Frequency Electron Spin Resonance (HF-
ESR). These measurements were performed in the transmission Faraday
geometry for two compounds, meta-F and ortho-F, for comparison.
Measurements were done on loose powder samples. The small powder
particles were self-oriented in the magnetic field during the measurements.
The meta-F compound was studied in Dresden. The measurements
were done in a frequency range 332−528 GHz in static magnetic fields
Two chain directions is the most distinguishing feature of the
meta-F crystal structure. The ortho-F compound, as other
compounds of this family, has only one chain direction. Also
the character of disorder in ortho-F differs from that in meta-F.
For the ortho compound the [TCNE]•− molecule is orienta-
tionally disordered with the minor form (20.5%) being rotated
by 90° (in the same plane) with respect to the major form
(79.5%). Besides, F in ortho-F is disordered over two sites on both
sides of the phenyl ring in a similar proportion.28
9985
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
Figure 2. Crystallographic structure of Mn(III) tetra(meta-fluorophenyl)porphyrin-tetracyanoethenide. The chain in the middle is inclined to the
plane of the drawing in the opposite direction than the side chains.
Table 2. Selected Interatomic Distances (Å) and Bond
Angles (deg) in meta-F and ortho-F Compounds
Table 1. Crystallographic Parameters of Mn(III) Tetra(meta-
fluorophenyl)porphyrin-tetracyanoethenide
distance or bond angle
meta-F
2.08
ortho-F5
molecular formula
formula weight
temp (K)
crystal system
space group
a (Å)
C60H24F4MnN8
987.81
Mn−N1 in plane
Mn−N2 in plane
Mn−N3 out of plane
Mn−Mn along chains
Mn−Mn between chains
Mn−F
2.000
2.016
2.013
299(2)
a
2.327
(2.309; 2.33) 2.313
10.185
monoclinic
10.400
14.900
8.07
C2/c
11.081
11.3331(7)
5.32
b (Å)
17.4406(9)
N1−Mn−N3
N1−Mn−N2
C23−C24−C24′
N3−C23−C24
Mn−N3−C23
90.40; 89.60
90.03; 89.97
119.78
178.75
172.10
76.3
92.9; 87.1
90.55; 89.45
117.3; 117.3
c (Å)
25.1790(10)
90.00
α (deg)
β (deg)
102.710(4)
a
(179.6; 173) 178.8
(150.3; 140) 148.6
γ (deg)
vol (Å3)
90.00
a
4854.8(4)
b
a
MnN4−[TCNE]
(55.6; 53.9) 55.4
Z
4
density (calcd) (Mg/m3)
1.351
a
Three values for ortho-F are for the major, minor and the occupancy
b
abs coeff (mm−1
)
0.337
weighted average, respectively.5 Dihedral angle between mean planes.
Numeration of atoms is according to CIF file in Supporting
Information.
F(000)
2004
crystal size (mm3)
theta range for data collection (deg)
h k l ranges
0.46 × 0.30 × 0.26
4.13 to 26.37
−14, 7, −21, 21, −31, 31
24219
4. MAGNETIC STUDIES
reflns collected
4.1. Static Properties. In Figure 3 the temperature depen-
indep reflns
4935 [R(int) = 0.0653]
semiempirical from equivalents
0.9175 and 0.8605
4935/15/403
1.092
dence of dc magnetic susceptibility is presented in the form of
abs corr
the effective magnetic moment defined as μeff = (8χmolT)1/2
.
max and min transm
data/restraints/params
goodness-of-fit on F2
final R indices [I > 2sigma(I)]
R indices (all data)
The measurement was done in a magnetic field of 1000 Oe. Ac
susceptibility data measured at the frequency of 1 Hz and the
amplitude of the driving magnetic field of 3 Oe (see the next
section) are also shown for comparison. The high maximum of
ac data is seen at the temperature of 7.5 K, a little lower than
the position of dc maximum 8.0 K. At higher temperatures
μeff(T) dependence shows a minimum, as expected for a ferri-
magnetic chain. This minimum (shallow, not seen in the scale
of the figure) falls at the temperature ∼130 K, much lower than
observed for the ortho-F compound, for which it falls at about
R1 = 0.0782, wR2 = 0.2160
R1 = 0.1002, wR2 = 0.2347
0.543 and −0.354
largest diff peak and hole (e Å−3
)
Another structural difference between meta-F and ortho-F is a
considerable anisotropy in the porphyrin plane for meta-F as can
be seen from the comparison of Mn−N1 and Mn−N2 distances
(Table 2).
9986
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
Figure 3. Temperature dependence of μeff = (8χmolT)1/2 for meta-F
obtained in static magnetic field of 1000 Oe (open circles). Ac data are
shown for comparison (open triangles connected by the dashed line).
Inset: the same data presented as ln(χmolT) vs reciprocal temperature.
Solid lines are fits (see the text).
Figure 4. Field cooling (monotonous dependences) and zero field
cooling magnetization (curves with maximum) plotted as a function of
temperature for meta-F in various static magnetic fields, strengths of
which are given in oersteds.
5.8 to 3.0 K. For ortho-F by the same field change this shift is
much smaller: from 4.8 to 4.0 K.9
room temperature. In order to determine the exchange integral
J, the Seiden model29 was used, which is based on the isotropic
Hamiltonian,
As seen, there are some differences in FC vs ZFC behavior
for meta-F in comparison with ortho-F. There exists the com-
mon point of the divergence of all curves for various fields at 10 K.
For small fields ZFC curves are not falling sharply down below
the bifurcation point but they deviate still going up with
decreasing temperature. This diverse FC vs ZFC behavior may
be interpreted as domain wall pinning or by the presence of
strongly interacting spin clusters.32
N−1
/ = − J
[S
+ SMn,i+1]sTCNE,i
∑
Mn,i
(1)
i=1
where SMn,i = 2 is the spin of Mn(III) ion and sTCNE,i = 1/2 is
the spin of the tetracyanoethenide radical.
In Figure 5 ZFC magnetic hysteresis loops are shown, which
were recorded at 1.8 K. One of the loops (smaller) is for the
By fitting the data in the temperature range 30−300 K the
value of J equal to −100 8 K was obtained in comparison
with the value −217 K for the ortho-F compound. This con-
siderable difference is consistent with the greater value of the
bonding Mn−N−C angle and dihedral MnN4−[TCNE] angle
between mean planes for the meta-F compound (see the last
two lines in Table 2), according to magnetostructural correla-
tions established by Miller et al. for this family of compounds.30
In the inset to Figure 3 the same data are presented in another
form ln(χmol·T) vs reciprocal temperature. The influence of the
applied magnetic field is seen below 10 K. The linear part,
observed in some temperature range, is in accord with the
anisotropic Heisenberg model, which predicts the following
relation:
χ T ∝ eΔ /kT
ξ
(2)
valid for the parallel susceptibility, which, however, is dominant
in this model. Δξ is the creation energy of a domain wall in the
chain.31 The value of Δξ, obtained from the linear fit, is equal to
25 2 K in comparison with the value of 60 K obtained for
ortho-F.
Figure 4 presents field cooling (FC) and zero field cooling
(ZFC) magnetization data (normalized to the susceptibility)
plotted as a function of temperature. At 10 K a sharp upraise of
M/H is seen by decreasing temperature, which is the manifesta-
tion of a phase transition, and a bifurcation point appears at
some lower temperature for every value of the superimposed dc
magnetic field. The bifurcation point moves to lower tempera-
ture with increasing strength of the magnetic field. The falling
low temperature part of the ZFC curve is common for all fields
(compare χ″ behavior, Figure 9b). By the change of field from
50 up to 5000 Oe the bifurcation point for meta-F shifts from
Figure 5. ZFC magnetic hysteresis loops of meta-F recorded at 1.8 K.
The smaller loop is for polycrystalline sample. The red line is extra-
polation to intersection with the ordinate of 3 μB. The greater loop is for
a crystallite oriented to obtain the maximum value of the highest field
magnetization. For the crystallite the minor loop was measured instead
of the virgin curve. In the inset the virgin magnetization curve is shown
together with its derivative.
polycrystalline material, the second one (greater) for a cry-
stallite. The orientation of the crystallite was chosen to obtain a
maximum value of magnetization at the highest magnetic field.
As seen, the coercive field is 9.0 kOe, more than two times
lower than observed for the ortho-F compound. No saturation
is observed at 50 kOe, even for the crystallite at the chosen
9987
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
orientation. The virgin magnetization curve is shown in the
inset to Figure 5 together with its derivative. Two peaks of the
derivative are seen: one at 7.5 kOe and the second at 15 kOe.
The relation of fields, being 1:2 at 1.8 K changes with increas-
ing temperature. This two-step magnetization curve seems
to be related to the presence of two families of chains in the
meta-F compound. The lower field reverses one family and then
a higher external field is required to reverse the second family,
because the external field is partially compensated by molecular
field.
The extrapolation of the high field magnetization of the
polycrystalline sample to its expected saturation value33 gμBST
gives the anisotropy field Ha= 85 kOe, at g = 2.0 and ST = 3/2
(Figure 5). This Ha value gives the estimation of the anisotropy
energy per molecule,
good fit. The presence of this term means constant in time and
homogeneous activation energy distribution. The factor a has
meaning of magnetic viscosity, which is a spin glass property.
Here, the glassiness may originate from the domain wall
pinning.34 There is no waiting time dependence (as checked for
5 and 30 min, see the curve at 3 K), being a characteristic
property of spin glasses.35
Let us see now how the superimposed dc field influences the
time dependences. In this experiment the field was not
completely off. The corresponding plots obtained in the bias
field of 1000 Oe are shown in Figure 7. All these dependences
can be nicely fit with eq 4 without the ln(t) term.
1
2
KA = (gMnSMn − gTCNE
s
TCNE)μBHa
(3)
equal to 8.5 K. By formally equating KA = DSMn2 the zero field
splitting parameter of the single Mn(III) ion |D| = 2.1 K
is obtained. It should be noted that eq 3 was used for the case
of two directions of anisotropy, which, however, for a powder
sample at saturation should be correct. For ortho-F the analog-
ous procedure gives Ha = 117.6 kOe, thus KA= 11.8 K and
|D| =2.95 K.9
4.2. Dynamic Properties. 4.2.1. Dc Measurements.
Magnetization, when observed in a long time scale, is not the
static property. In Figure 6 the time dependence of thermo-
Figure 7. Magnetization of meta-F as a function of time measured at
various temperatures and in the superimposed dc magnetic field of
1000 Oe. Measurement started after switching the field from 50 to
1 kOe. Solid lines are fits, see text.
From these fits the average relaxation time τ was obtained.
Reciprocal temperature dependence of ln τ is shown in Figure
13 together with the corresponding ac data. From the slope of
this plot the activation energy Ea can be determined according
to the Arrhenius equation:
τ = τ0eE /kT
a
(5)
The fitted parameters are Ea = 64 5 K and τ0 = (1.9 0.6) ×
10−8 s.
4.2.2. Ac measurements. Figure 8a presents temperature
dependence of the ac susceptibility of meta-F, measured for
various frequencies with no superimposed dc magnetic field.
Here, it is seen that both χ′ and χ″ maxima move to higher
temperature with increasing frequency and, while the maximum
value of the χ′ component decreases with increasing frequency,
the maximum value of χ″ remains almost constant. When dc
magnetic field is superimposed, both ac susceptibility com-
ponents are quickly suppressed, which is shown in Figure 8b for
the constant frequency of 10 Hz.
How frequency characteristics change in a superimposed dc
magnetic field is separately shown in Figure 9. As is seen, these
characteristics differ remarkably from those obtained in zero dc
magnetic field. It looks like as if both χ′ and χ″ were composed
of two constituents. The first one, with its χ′ maximum near
8.5 K in small field, is weakly (or not) dependent on frequency.
It slightly moves with increasing dc field to higher temperatures
(9.0 K in 1000 Oe) and being strongly suppressed in a field of
1000 Oe is expected to disappear with further field increas-
ing, see Figure 9c (its χ″ companion disappeared earlier). The
Figure 6. Thermoremanent magnetization of meta-F measured as a
function of time at various temperatures. The time dependence at 3 K
was measured with the waiting time 5 and 30 min. Solid lines are fits,
see text. The values of M have no relation to those in Figure 7 because
a new sample weight was used.
remanent magnetization is plotted vs logarithm of time. In this
experiment the sample was first magnetized in field of 50 kOe,
then the field was off and the magnetization was measured as a
function of time at various temperatures. The curves in Figure 6
look differently than the analogous data for the ortho-F com-
pound.9 The data for the meta-F compound can be fitted using
the following relation:
−(t/τ)1−n
Mtrm = σ0 + M0e
+ a ln t
(4)
where n ≈ 0.5, a ≈ −0.04.
The above relation contains many fitting parameters (σ0, M0,
n, τ, a), but the additional a·ln (t) term is necessary to obtain a
9988
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
field (b). As seen in panel a there is a broad distribution (one
can discern two distributions) of relaxation times, however, in
Figure 8. Temperature dependence of both components χ′ and χ″ of
the ac magnetic susceptibility of meta-F compound measured for
various frequencies at zero static magnetic field (a) and for various
static superimposed magnetic fields at the constant frequency of 10 Hz
(b). The amplitude of the driving magnetic field Hac is 3 Oe.
Figure 10. Frequency dependence of ac susceptibility for meta-F
measured at various temperatures. (a) Data obtained in zero dc bias
field. (b) Data obtained in dc bias field of 300 Oe. Solid lines are
guides for eye.
the bias field of 300 Oe (see panel b) the distribution drastically
narrows and relaxation has a Debye character.
Narrowing of the distribution of relaxation times in field
can be also shown by means of Argand diagrams. In Figure 11
Figure 9. Temperature dependence of both components χ′ and χ″ of
the ac magnetic susceptibility of meta-F compound measured for
various frequencies at four static magnetic fields, 100 Oe (a), 300 Oe
(b), 500 and 1000 Oe (c). Note, that panels a and b have the same
legend. The amplitude of the driving magnetic field Hac is 3 Oe.
Figure 11. Argand plots for meta-F obtained at various temperatures
with no superimposed dc magnetic field.
Argand diagrams for meta-F obtained with no superimposed
magnetic field are shown for various temperatures. As is seen,
the distribution of relaxation times is very broad. Again, two
distributions can be differentiated, as is seen, e.g., for T = 7.5 K.
The width of distribution can be characterized by the parameter
α, obtained from fitting arcs of the (1 − α)π size to the χ″ vs χ′
data. When dc magnetic field is applied, the distribution
drastically narrows and only one distribution is present, as
shown in Figure 12a for T = 5.5 K. In dc field of 1000 Oe α
equals 0.24. The values of α obtained for meta-F can be
compared with the α value for the ortho-F compound. The
latter one equals 0.12, see Figure 12b, and is completely
independent of temperature. The value of α expected for
the ideal single chain magnet should be equal to zero (single
relaxation time), but for the real single chain magnets such
values as ∼0.12 were reported in literature for powder
samples.37,38
frequency dependence is associated mainly with the second
constituent, which separates from the first χ′ constituent mov-
ing together with the accompanying frequency dependent χ″
constituent to lower temperatures. It can be seen that this χ″
constituent (at least in fields below 500 Oe) increases with
increasing frequency, which is at variance with the case of zero
dc field.
It is worth noting that poorly crystallized powder samples
(with grain size close to one-domain nanoparticles) from
another synthesis batch show ac behavior in zero dc field, which
is like ac behavior in nonzero field described above. It means
their ac characteristics in zero field are like this in Figure 9b.36
It is also interesting to see relaxations in the frequency
domain. Figure 10 presents frequency dependences of χ′ and χ″
measured for various temperatures without (a) and with dc bias
9989
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
much lower and all are equal to 84 2 K. The values of τ0 are
∼1.2 × 10−9 s.
4.3. HF-ESR Study. HF-ESR measurements were under-
taken to determine excitation gaps for meta-F and ortho-F com-
pounds and thereby to get more insight into the relaxation
character of these compounds. It should be noted, however, that
in large magnetic fields, typical for HF-ESR, chains are effectively
decoupled, thus only properties of individual chains are probed.
The ESR spectra were measured at various frequencies and vari-
ous temperatures.
4.3.1. meta-F Compound. Typical ESR spectra of the
powder sample of meta-F compound measured at T = 4 K and
various frequencies are shown in Figure 14a.
Figure 12. Comparison of Argand plots for meta-F compound
obtained at 5.5 K in various dc magnetic fields (a) and for ortho-F
obtained at various temperatures in 0 Oe dc field (b). The arcs are fits
to the data.
Using the data shown above, temperature dependence of
the relaxation time can be determined. Taking the temperature
for the maximum of χ″(T)f,H curves measured by fixed f and
bias field H and assuming that the relaxation time at the
temperature of maximum is equal to 1/(2πf), the diagram
presenting ln(τ) vs reciprocal temperature can be constructed.
In Figure 13 four linear dependences obtained from ac high
Figure 14. HF-ESR spectra of meta-F compound obtained at the
temperature of 4 K for various frequencies (a) and at the frequency of
372 GHz for various temperatures (b). Panel a additionally contains
frequency vs resonance magnetic field dependence (blue squares,
experimental points; red line, linear fit).
The spectra comprise one intense transmission line, which
narrows by decreasing excitation frequency ν. Variation of ν
with field yields a linear shift of the resonance field Hres. This
dependence is plotted in Figure 14a. The linear extrapolation to
zero magnetic field gives the zero field splitting (gap) ΔESR
336 GHz (16 K).
=
A representative set of the ESR spectra of the meta-F sample
measured at constant frequency ν = 372 GHz and various
temperatures is presented in Figure 14b. Here, a shift of the
spectral weight to higher magnetic fields at higher temperatures
takes place. Such a shift is a result of the thermal activation of the
higher energy spin states with rising temperature and indicates a
negative sign of the axial magnetic anisotropy of the chain.39,40
4.3.2. ortho-F Compound. ESR spectra measured for ortho-
F are presented in Figure 15. Figure 15a shows spectra obtained
at various frequencies at the temperature of 4.2 K. The small
oscillations seen in the figure are extrinsic and due to some
interference of the sample cell. Except for that, two character-
istic features are seen. The first of them is the line width
broadening at decreased resonance field. This is often caused by
anisotropy in the plane normal to the principal axis of D; here,
however, this anisotropy is lower than for the meta-F com-
pound (see crystallographic data), which does not show the
broadening but narrowing effect. The second feature is a
small splitting of the line, which is well seen in Figure 15 at the
lowest temperature of 4.2 K. Figure 15b shows the spectra at a
constant frequency of 357 GHz obtained for various temper-
atures. Here it is seen that the line becomes broader and dis-
appears with increasing temperature and no clear absorption is
Figure 13. Natural logarithm of the relaxation time vs reciprocal
temperature dependence for meta-F compound. The high temperature
data (red color) were obtained from ac susceptibility measurements.
The left (most steep) line was obtained in zero bias magnetic field, and
the other three lines were obtained for nonzero bias magnetic fields.
Dc data are also shown (see section 4.2.1). Solid lines are fits.
temperature data (above 4 K) are shown and compared with
the low temperature dc SQUID data obtained below 3 K in dc
field of 1000 Oe. One of the ac dependences refers to zero bias
dc field, and the other ones refer to the case of the bias dc field
equal to 100, 200, and 300 Oe.
As seen, the slope of ac curves changes drastically with the
bias field increasing from 0 to 100 Oe. Fitting ac dependences
to the Arrhenius equation the values of the activation energy Ea
are obtained. For zero bias field the value of Ea equals 265 5 K.
The very low value of τ0 = 7.8 × 10−20 s is nonphysical. The
values of Ea obtained for three nonzero values of bias field are
9990
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
time distributions, namely, to the longer time distribution. In
the bias dc magnetic field the glassy character is lost when with
increasing field the domain walls are swept away. The only
existing relaxations are now those of SCM having the narrow
time distribution.
In order to get insight into the spin dynamics of meta-F it is
helpful to inspect the activation energies obtained from mea-
surements. In general, the activation time for magnetization
reversal can be described by three activation energies, EA, Δξ
and ΔA, which obey the equation
EA = 2Δξ + ΔA
(6)
assuming that chains are infinite or
EA = Δξ + ΔA
(7)
Figure 15. HF-ESR spectra of ortho-F compound obtained at various
frequencies at the temperature of 4.2 K (a) and at the frequency 357
GHz for various temperatures (b). Panel a additionally contains
frequency vs resonance magnetic field dependence (blue squares,
experimental points; red line, linear fit).
when chains are finite. Chains may be treated as finite when
they are shorter than their correlation length, what may happen
with decreasing temperature. These equations are well proved
for Ising chains, where ΔA is the activation energy of a single
anisotropic spin inside a narrow domain wall, where it sees no
local field.37 In the anisotropic Heisenberg model this value is
still not well-defined. It should be related to anisotropy energy
of a broad 1D domain wall.
If we assume that for meta-F at higher temperatures, in the ac
regime, eq 6 is valid, then we have EA = 84 K, Δξ = 2 × 25 K, so
ΔA = 34 K. At lower temperatures, in the dc regime when eq 7
is valid, then we have EA = 64 K, Δξ =25 K, so ΔA = 39 K.
For ortho-F we had the finite chains only, then EA = 117 K, Δξ =
60 K9 and ΔA = 57 K. Assuming that the observed phase
transition in meta-F is soliton induced, all π kinks should be
paired below Tc. It would mean that for meta-F Δξ values near Tc
(in the ac regime) should be taken with the factor 2.
The meta- and ortho-F compounds differ much with the
exchange parameter J, which equals −100 K for meta-F and
−217 K for ortho-F. Treating the ferrimagnetic chain as a ferro-
magnetic chain composed of identical units with the effective
spin Sef = 3/2 one can estimate the exchange constant Jef of
such a chain. We used the following relation (eq 8) derived
from the isotropic Hamiltonian (eq 9) of the ferromagnetic
Heisenberg chain:42
found at 50 K. The observed splitting of the line seems to be
connected with the crystallographic structure in which the
[TCNE]− group has two orientations with the probability ratio
4:1.5 The linear extrapolation of the mode gives the zero field
splitting of ΔESR = 314 3 GHz or 15.0 K (average of two
peaks), see Figure 15a.
5. DISCUSSION
Both chain compounds investigated in this work show slow
relaxations. While the meta-F compound shows the transition
to the glassy ferromagnetic state at Tc = 10 K and in the applied
dc field the SCM behavior, the ortho-F compound is a typical
SCM. The question arises, what is the reason for such a dif-
ference, the more that intrachain magnetic interaction in meta-F
is much weaker than in ortho-F. In response we would like to
note that, first, in ortho-F we had finite chains in both ac and dc
regimes,9 while in meta-F the chains might be infinite in the
ac regime. Second, meta-F has a structure with two different
chain directions for which stronger dipolar superspin interac-
tions may develop.41
The glassy character of meta-F compound at zero dc field was
revealed by time dependent dc magnetization measurements
and confirmed by ac susceptibility measurements. Both methods
demonstrated a broad relaxation time distribution. For [MnTPP]-
[TCNE]·2(1,3-C6H4Cl2), belonging to the same compound
family, Etzkorn et al. proposed the fractal cluster glass model, in
which slow relaxations near the glass transition Tg are due to
magnetic viscosity caused by interchain interactions.12 We
repeated the scaling analysis of Etzkorn et al., which led them to
the above-mentioned conclusion. This analysis was based on
the critical slowing down power law. However, we found results
of scaling analysis not convincing because scaling with other
laws gave also good overlapping of corresponding curves. In
our opinion we are dealing with a true phase transition. Inclina-
tions of spins due to different easy magnetic directions may favor
ferromagnetic dipolar interchain interactions being responsible
for 3D magnetic ordering.
2
2
⎛
⎝
⎞
J S
2
2
ef
χmol = N g2μ
⎜
⎜
⎟
⎟
A
B
3J k T
⎠
B
ef
(8)
(9)
/ = − J
SS
i j
∑
ef
⟨i,j⟩
The 1/T2 behavior of eq 8 is right at low temperatures only.
In fact, the Seiden model shows also 1/T2 behavior at low
temperatures, but we use eq 8 because it was derived from
the simple Hamiltonian (eq 9) for which many other useful
formulas were derived. In Figure 16 molar susceptibilities of
meta- and ortho-F compounds as a function of 1/T2 are pre-
sented. The interval of fitting was chosen so that its higher limit
is below Tmin and the lower limit is the same as used by fitting
of the Seiden model. The following values of Jef were obtained:
28 4 K for meta-F and 51 4 K for ortho-F, assuming g factor
equal to 2. Using these values and assuming that the single ion
anisotropy is comparable for both compounds, we see that for
both compounds |D/Jef| ≪ 1.
The glassiness observed may be explained as due to 3D-
domain wall pinning resulting from interchain interactions and
lattice defects.11,34 The domains that form consist of coupled
spins belonging to some number of adjacent chains. Domain
relaxations contribute to one of the two observed relaxation
Being beyond the Ising limit |D/Jef| = 4/337 both compounds
cannot be SCM according to the original Glauber model.43,44
9991
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
respectively), with KA value obtained for [MnIII2Ni] compound
from the saturation magnetization.37 It is also worth noting a
good fulfillment of eq 6 and 7 for the Heisenberg system MnIII-
TCNQ, however, the ΔA value in this case is much greater than
the KA value.
Oshima et al. performed HF-ESR study of [MnIII2Ni] single
chain magnet47 and determined the value of the excitation gap
at T = 1.5 K. It equals 321 GHz (15.4 K). It is lower than the
KA value for this compound and is near the gap values of our
more strongly coupled systems. Because in the frequency-field
diagram no series of lines with multiple value of g = 2 appeared,
the authors concluded that the observed mode is neither spin
cluster excitation nor spin cluster resonance. This mode was
attributed to the spin wave excitation obeying the ESR selection
rule ΔSz = 1. This conclusion was supported by theoretical
analysis, which proved that spin wave excitation is lower in
energy than spin cluster excitation if the single ion zero field
splitting parameter |DMn| is lower than 5.7 K. However, by
increasing the |DMn| value, spin wave excitation becomes greater
in energy. This theoretical result of Oshima et al. is in agree-
ment with an older paper of Johnson and Bonner,48 where a
linear ferromagnet with spin 1/2 in a small magnetic field was
analytically treated in the whole anisotropy range between Ising
and Heisenberg models and it was shown that for anisotropy
lower than some specific value the dominant excitations are
spin waves. For anisotropy greater than this specific value the
dominant excitations are bound spin wave complexes.
The excitation gaps obtained from HF-ESR study for meta-F
and ortho-F are comparable, equal to ∼16 K. This value is
greater than the anisotropy energies KA per molecule estimated
from magnetization measurements 8.5 and 11.8 K for meta-F
and ortho-F, respectively. However, it is much lower than the
anisotropy gap ΔA obtained from magnetic ac/dc measure-
ments. So large values of ΔA follow from the fact that both
ortho-F and meta-F are anisotropic Heisenberg systems, which
have spin clusters with broad domain walls (solitons).
Therefore many spins may be involved in the dynamics of
walls and no wonder that the activation energy ΔA of τ0 is
greater than DS2 (see also MnIII-TCNQ in Table 3).
Figure 16. Molar susceptibility as a function of 1/T2 plotted for meta-
F and ortho-F compounds. Solid lines are linear fits.
However, it was shown that SCM behavior also for |D/J| ≪ 4/3
is possible.37,45 In the case of extremely small anisotropy, Δξ
can be independently estimated using a simple expression:46
Δξ ≈ 2Seff 2 |2J ·D|1/2
(10)
ef
Taking the earlier estimated values of D equal to 2.1 and 3.0 K
this gives Δξ ≈ 48.8 and Δξ ≈ 78.7 K for meta- and ortho-F
respectively. Although Jef is much lower than the real J, the
values of Δξ obtained are considerably greater than Δξ values
found from measurements. In Table 3 Δξ and other activation
Table 3. Comparison of Activation Energies Obtained from
Ac/Dc Susceptibility and HF-ESR for meta-F and Selected
Single Chain Magnets
a
a
b
compd syst
EA (K) n × Δξ (K) ΔA (K) KA (K)
Δ
ESR (K)
c
meta-F anis. Heisenberg 84.
2 × 25
1 × 25
1 × 60
34
39
57
8.5
16.1
15.0
d
64.
ortho-F9 anis.
11.8
cd
117. ,
Heisenberg
[MnIII2Ni]37,44,47 Ising
74.
2 × 28
1 × 28
2 × 26.5
1 × 26.5
18
c
22.1
9.8
15.4
d
55.
27
MnIII-TCNQ45 anis.
Heisenberg
94.1
67.7
41.1
41.2
c
Another reason that the ΔESR value is much lower than the
ΔA value obtained from ac/dc measurements is that selection
rules of ESR can choose the lowest energy excitations. In fact,
Johnson and Bonner48 showed that the effective magnetic
excitation gap for the low temperature susceptibility at small
magnetic fields corresponds to the bound spin complexes in the
whole anisotropy range.
For the time being, the question, why three different single
chain magnet systems, meta-F, ortho-F and Mn2Ni, have nearly
the same value of the ΔESR gap, is still open. This value,
however, seems to be close to the single ion anisotropy of Mn3+
ion.
d
a
ΔA = EA − n × Δξ; for infinite chains the factor n equals 2, otherwise
b
it equals 1. KA: anisotropy energy obtained from saturation
magnetization and eq 3. Ac data. Dc data.
c
d
energies obtained are given for comparison. In addition, litera-
ture values for the Ising system [Mn2(saltmen)2Ni(pao)2-
(py)2](ClO4)2 (abbreviated [MnIII2Ni])37,44,47 and anisotropic
Heisenberg system [Mn(5-TMAMsaltmen)(TCNQ)](ClO4)2
(abbreviated MnIII-TCNQ) are enclosed,45 where saltmen =
N,N′-(1,1,2,2-tetramethylethylene) bis(salicylideneiminate),
pao = pyridine-2-aldoximate, py = pyridine, 5-TMAMsaltmen =
N,N′-1,1,2,2-tetramethylethylene bis(5-trimethylammoniome-
thylsalicylideneiminate, TCNQ = 7,7,8,8-tetracyano-p-quinodi-
methane.
6. SUMMARY
The compound Mn(III) tetra(meta-fluorophenyl)porphyrin-
tetracyanoethenide (meta-F) was synthesized and its crystal
structure was determined. This structure is composed of ferri-
magnetic chains. There are two chain directions, which corres-
pond to two magnetic easy axis directions. The compound was
studied by magnetic measurements and, together with the
relative ortho compound, by HF-ESR. Magnetic properties of
both compounds were compared. While the ortho-F compound
is the single chain magnet, the compound meta-F shows a glassy
ferromagnetic transition at 10 K. In applied dc magnetic field
The situation is the simplest and well established for [MnIII Ni].
2
Because [MnIII2Ni] is a SCM system with |D/J| > 4/3, the con-
sidered spin clusters (1D domains) have extremely narrow walls
of the one spin width. In such a case the activation energy ΔA
(for spin reversal in the wall) should be equal to the single ion
anisotropy energy DS2. Looking in Table 3, in fact, we see
rather good agreement of the average ΔA value, obtained from
two temperature regimes (with finite and infinite chains,
9992
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
(14) Bogani, L.; Caneshi, A.; Fedi, M.; Gatteschi, D.; Massi, M.;
Novak, M. A.; Pini, M. G.; Rettori, A.; Sessoli, R.; Vindigni, A. Phys.
Rev. Lett. 2004, 92, 207204.
(15) Coulon, C.; Clerac, R.; Wernsdorfer, W.; Colin, T.; Miyasaka,
́
H. Phys. Rev. Lett. 2009, 102, 167204.
the domain walls, responsible for glassy behavior, are swept
away, so magnetic relaxations acquire a narrow distribution
of relaxation times and look as relaxations of a single chain
magnet. Activation energies for soliton spin reversals and spin
wave excitations were determined using ac susceptibility and
HF-ESR techniques, respectively. It was concluded that meta-F
and ortho-F compounds are Heisenberg systems with aniso-
tropy, for which soliton excitations are higher in energy than
spin wave excitations.
(16) Miyasaka, H.; Takayama, K.; Saitoh, A.; Furukawa, S.;
Yamashita, M.; Cler
(17) Ishida, T.; Okamura, Y.; Watanabe, I. Inorg. Chem. 2009, 48,
́
ac, M. Chem.Eur. J. 2010, 16, 3656−3662.
7012.
(18) Izyumov, Yu.A.; Chernoplekov, N. A. Neutron Spectroscopy;
Plenum Pub. Corp.: New York, 1994.
ASSOCIATED CONTENT
* Supporting Information
Crystallographic data in CIF format. This material is available
free of charge via the Internet at http://pubs.acs.org.
(19) Elmassalami, M.; de Jongh, L. J. Physica 1989, B 154, 254.
(20) Hołyst, J. A.; Sukiennicki, A. Phys. Rev. 1988, B 38, 6975. Hołyst,
J. A.; Sukiennicki, A. Phys. Rev. 1983, B 28, 4062.
(21) Morgunow, R.; Kirman, M. V.; Inoue, K.; Tanimoto, Y.;
Koshine, J.; Ovchinnikov, A. S.; Kazakova, O. Phys. Rev. 2008, B77,
184419.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: z.tomkowicz@uj.edu.pl; haase@chemie.tu-darmstadt.
■
(22) Krzystek, J.; Telser, J.; Pardi, L. A.; Goldberg, D. P.; Hoffman, B.
M.; Brunel, L.-C. Inorg. Chem. 1999, 38, 6121−6129.
(23) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.;
Assour, J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.
(24) Jones, R. D.; Summerville, D. A.; Basolo, F. J. Am. Chem. Soc.
1978, 100, 4416−4424.
de.
Notes
The authors declare no competing financial interest.
¶On leave from Bio-Organic Division, Bhabha Atomic Research
Centre, Trombay, Mumbai-400 085, India.
(25) Summerville, D. A.; Cape, T. W.; Johnson, E. D.; Basolo, F.
Inorg. Chem. 1978, 17, 3297−3300.
(26) Golze, C.; Alfonsov, A.; Klingeler, R.; Buchner, B.; Kataev, V.;
̈
Mennerich, C.; Klauss, H.- H.; Goiran, M.; Broto, J.-M.; Rakoto, H.;
Demeshko, S.; Leibeling, G.; Meyer, F. Phys. Rev. 2006, B 73, 224403.
(27) Nojiri, H.; Ajiro, Y.; Asano, T.; Boucher, J.-P. New J. Phys. 2006,
8, 218.
(28) Brandon, E. J.; Arif, A. M.; Burkhart, B. M.; Miller, J. S. Inorg.
Chem. 1998, 37, 2792−2798.
(29) Seiden, J. J. Phys. (Paris) Lett. 1983, 44, L947.
(30) Rittenberg, D. K.; Miller, J. S. Inorg. Chem. 1999, 38, 4838−
4848. Brandin, E. J.; Kollmar, C.; Miller, J. S. J. Am. Chem. Soc. 1998,
120, 1822−1826. Ribas-Arino, J.; Novoa, J. J.; Miller, J. S. J. Mater.
Chem. 2006, 16, 2600−2611. Her, J.-H.; Stephens, P. W.; Bagnato, J.
D.; Miller, J. S. J. Phys. Chem. C 2010, 114, 20614−20620.
(31) Nakamura, K.; Sasada, T. Solid State Commun. 1977, 21, 891.
(32) Etzkorn, S. J.; Hibbs, W.; Miller, J. S.; Epstein, A. J. Phys. Rev.
2004, B 70, 134419.
ACKNOWLEDGMENTS
W.H. thanks the Internationales Buro des BMBF for supporting
■
̈
the bilateral German-Indian cooperation project IND 02/007
and the Deutsche Forschungsgemeinschaft DFG for the
support via the project HA782/85. Z.T. thanks grant of
Ministry of Science and Higher Education, Poland, Nr. N
N202 103238. H.N. acknowledges the support by Coordination
Program in Grant-in-Aid for Scientific Research on Innovative
Areas by MEXT Japan. The work in Dresden was supported by
the DFG through project FOR 1154 “Towards molecular
spintronics”.
REFERENCES
■
(33) The justification of this procedure in the case of the
unidirectional anisotropy was given by Miyasaka, H.; Takayama, K.;
(1) Caneschi, A.; Gatteschi, D.; Lalioti, N.; Sangregorio, C.; Sessoli,
R.; Venturi, G.; Vindigni, A.; Rettori, A.; Pini, M. G.; Novak, M. A.
Angew. Chem. 2001, 113, 1810.
(2) Brandon, E. J.; Rittenberg, D. K.; Arif, A. M.; Miller, J. S. Inorg.
Chem. 1998, 37, 3376.
(3) Fardis, M.; Diamantopoulos, G.; Papavassiliou, G.; Pokhodnya,
K.; Miller, J. S.; Rittenberg, D. K.; Christides, C. Phys. Rev. 2002, B 66,
064422.
(4) Rittenberg, D. K.; Miller, J. S. Inorg. Chem. 1999, 38, 4838−4848.
(5) Brandon, E. J.; Arif, A. M.; Burkhart, B. M.; Miller, J. S. Inorg.
Chem. 1998, 37, 2792.
(6) Mascarenhas, F.; Falk, K.; Klavins, P.; Schilling, J. S.; Tomkowicz,
Z.; Haase, W. J. Magn. Magn. Mater. 2001, 231, 172.
(7) Bałanda, M.; Falk, K.; Griesar, K.; Tomkowicz, Z.; Haase, W. J.
Magn. Magn. Mater. 1999, 205, 14 −26.
Saitoh, A.; Furukawa, S.; Yamashita, M.; Cler
2010, 16, 3656−3662. See also Miyasaka, H.; Saitoh, A.; Yamashita,
M.; Clerac, R. Dalton Trans. 2008, 2422−2427.
(34) Sampaio, L. C.; da Cunha, S. F. J. Magn. Magn. Mater. 1990, 88,
151.
́
ac, R. Chem.Eur. J.
́
(35) Mydosh, J. A. Spin Glasses: An Experimental Introduction; Taylor
& Francis: London, 1993.
(36) Bałanda, M.; Tomkowicz, Z.; Haase, W.; Rams, M. J. Phys.: Conf.
Ser. 2011, 303, 012036.
(37) Miyasaka, H.; Julve, M.; Yamashita, M.; Clerac, R. Inorg. Chem.
2009, 48, 3420−3437.
(38) Kajiwara, T.; Nakano, M.; Kaneko, Y.; Takaishi, S.; Ito, T.;
Yamashita, M.; Igashira- Kamiyama, A.; Nojiri, H.; Ono, Y.; Kojima, N.
J. Am. Chem. Soc. 2005, 127, 10150−10151.
(8) Griesar, K.; Athanassopoulou, M. A.; Soto Bustamante, E. A.;
Tomkowicz, Z.; Zaleski, A. J.; Haase, W. Adv. Mater. 1997, 9, 45.
(9) Bałanda, M.; Rams, M.; Nayak, S. K.; Tomkowicz, Z.; Haase, W.;
Tomala, K.; Yakhmi, J. V. Phys. Rev. 2006, B 74, 224421.
(10) Pikin, S. A.; Tomkowicz, Z.; Pikina, E. S.; Haase, W. JETP Lett.
2007, 85, 639−643. Pikin, S. A.; Tomkowicz, Z.; Pikina, E. S.; Haase,
W. Ferroelectrics 2007, 361, 1−12.
(39) Kahn, O., Molecular Magnetism; Wiley-VCH: Weinheim, 1993.
(40) Krupskaya, Y.; Alfonsov, A.; Parameswaran, A.; Kataev, V.;
Klingeler, R.; Steinfeld, G.; Beyer, N.; Gressenbuch, M.; Kersting, B.;
Buchner, B. ChemPhysChem 2010, 11, 1961−1970.
̈
(41) Ostrovsky, S.; Haase, W.; Drillon, M.; Panissod, P. Phys. Rev.
2001, B 64, 134418. Wynn, C. M.; Gîrtu, M. A.; Brinckerhoff, W. B.;
Sugiura, K.-I.; Miller, J. S.; Epstein, A. J. Chem. Mater. 1997, 9, 2156−
2163.
(11) Ishii, N.; Okamura, Y.; Chiba, S.; Nogami, T.; Ishida, T. J. Am.
Chem. Soc. 2008, 130, 24.
(42) Suzuki, F.; Shibata, N.; Ishii, C. J. Phys. Soc. Jpn. 1994, 63, 1539.
(43) Glauber, R. J. J. Math. Phys. 1963, 4, 294.
(12) Etzkorn, S. J.; Hibbs, W.; Miller, J. S.; Epstein, A. J. Phys. Rev.
Lett. 2002, 89, 207201.
(44) Coulon, C.; Cler
H. Phys. Rev. 2004, B 69, 132408.
́
ac, R.; Lecren, L.; Wernsdorfer, W.; Miyasaka,
(13) Pini, M. G.; Rettori, A.; Bogani, L.; Lascialfari, A.; Mariani, M.;
Caneshi, A.; Sessoli, R. Phys. Rev. 2011, B 84, 04444.
9993
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994

Inorganic Chemistry
Article
(45) Miyasaka, H.; Madanbashi, T.; Sugimoto, K.; Nakazawa, Y.;
́
Wernsdorfer, W.; Sugiura, K.; Yamashita, M.; Coulon, C.; Clerac, R.
Chem.Eur. J. 2006, 12, 7028.
(46) ElMassalami, M.; Smit, H. H. A.; de Grot, H. J. M.; Thiel, R. C.;
de Jongh, L. J. In Magnetic Excitations and Fluctuations II, Proceedings
of an International Workshop; Balucani, U., Lovesey, S. W., Rasetti, M.
G., Tognetti, V., Eds.; Turin, Italy, May 25−30, 1987; Springer
Proceedings in Physics, Vol. 23; Springer-Verlag: Berlin, 1987; p 178.
(47) Oshima, Y.; Nojiri, H.; Asakura, K.; Sakai, T.; Yamashita, M.;
Miyasaka, H. Phys. Rev. 2006, B 73, 214435.
(48) Johnson, J. D.; Bonner, J. C. Phys. Rev. 1980, B 22, 251.
9994
dx.doi.org/10.1021/ic3014927 | Inorg. Chem. 2012, 51, 9983−9994