Charge-transfer phase transition and ferromagnetism of iron mixed-valence complexes (n-CnH2n+1)4N[FeIIFe III(dto)3] (n = 3-6; dto = C2O 2S2)

Article abstract of DOI:10.1002/ejic.200500560

The iron mixed-valence complex, (n-C₃H₇) ₄N[Fe^IIFe^III(dto)₃] (dto = dithiooxalato) shows a charge-transfer (CT) phase transition at T_CT = 122.4 K. In the vicinity of T_CT, the spin state changes from Fe ^II (S = 2) - Fe^III (S = 1/2) (high-temperature phase: HTP) to Fe^II (S = 0) - Fe^III (S = 5/2) (low-temperature phase: LTP) accompanied by a charge transfer between Fe^II and Fe ^III. This complex also undergoes a ferromagnetic transition at 7 K in the LTP. In order to investigate the mechanism of the CT phase transition and the ferromagnetism, we have systematically synthesized (n-C_nH _2n+1)₄N[Fe^IIFe^III-(dto) ₃] (n = 3-6), and have investigated their physical properties by magnetic susceptibility, powder X-ray diffraction measurements, and ESR spectroscopy. The compounds (n-C_nH_2n+1) ₄N[Fe^IIFe^III(dto)₃] (n = 3-6) display ferromagnetic phase transitions at 7 K, 7 K (& 13 K), 19.5 K, and 22 K, respectively. For n = 3 and 4, the CT phase transitions take place at T _CT ≈ 120 K and T_CT ≈ 140 K, respectively. For n = 5 and 6, on the other hand, the CT phase transition does not occur, and the spin configuration of Fe^II (S = 2) and FeIII (S = 1/2) corresponding to the HTP for n = 3 and 4 is stable between 2 K and 300 K. The cation size of (n-C_nH_2n+1)₄N⁺ (n = 3-6) acts as an effective internal pressure which induces the CT phase transition and the ferromagnetic ordering in the [Fe^IIFe^III(dto) ₃]^-_∞ layer. We also discuss the mechanism of the CT phase transition and the ferromagnetism induced by the charge-transfer interaction between Fe^II and Fe^III. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejic.200500560

FULL PAPER
DOI: 10.1002/ejic.200500560
Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence
Complexes (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6; dto = C2O2S2)
Miho Itoi,*[a] Yuuki Ono,[b] Norimichi Kojima,*[b] Kenichi Kato,[c,d] Keiichi Osaka,[c,d] and
Masaki Takata[c,d]
Keywords: Mixed-valence compounds / Iron / Magnetic properties / Charge transfer
The iron mixed-valence complex, (n-C3H7)4N[FeIIFeIII(dto)3]
(dto = dithiooxalato) shows a charge-transfer (CT) phase
transition at TCT = 122.4 K. In the vicinity of TCT, the spin
phase transitions at 7 K, 7 K (& 13 K), 19.5 K, and 22 K,
respectively. For n = 3 and 4, the CT phase transitions take
place at TCT Ϸ 120 K and TCT Ϸ 140 K, respectively. For n =
state changes from FeII (S = 2) – FeIII (S = 1/2) (high-tempera- 5 and 6, on the other hand, the CT phase transition does not
ture phase: HTP) to FeII (S = 0) – FeIII (S = 5/2) (low-tempera- occur, and the spin configuration of FeII (S = 2) and FeIII (S =
ture phase: LTP) accompanied by a charge transfer between
FeII and FeIII. This complex also undergoes a ferromagnetic
1/2) corresponding to the HTP for n = 3 and 4 is stable be-
tween 2 K and 300 K. The cation size of (n-CnH2n+1)4N+ (n =
transition at 7 K in the LTP. In order to investigate the mecha- 3–6) acts as an effective internal pressure which induces the
nism of the CT phase transition and the ferromagnetism, we
have systematically synthesized (n-CnH2n+1)4N[FeIIFeIII-
(dto)3] (n = 3–6), and have investigated their physical proper-
ties by magnetic susceptibility, powder X-ray diffraction
measurements, and ESR spectroscopy. The compounds (n-
CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6) display ferromagnetic
CT phase transition and the ferromagnetic ordering in the
[FeIIFeIII(dto)3]–ϱ layer. We also discuss the mechanism of the
CT phase transition and the ferromagnetism induced by the
charge-transfer interaction between FeII and FeIII.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
spin states are situated in that region. It is well-known that
tris(dithiocarbamato) iron() complexes show the spin-
crossover transition.[4] In these complexes, the FeIII atom is
coordinated by six S atoms. Taking account of these par-
ticulars, we have synthesized dithiooxalato-bridged iron
mixed-valence complexes. (Hereafter, dithiooxalato is de-
noted as dto.)
Introduction
Transition-metal complexes with d4-d7 configuration
have a possibility of a spin transition between a low-spin
(LS) state and a high-spin (HS) state. The spin-crossover
phenomenon has gained a renewed importance through the
discovery of Light Induced Excited Spin State Trapping
(LIESST) for [Fe(ptz)6](BF4)2 (ptz = 1-propyltetrazole)[1]
and the thermally induced spin-crossover transition with a
large thermal hysteresis around room temperature in a tri-
azole-bridged iron() complex.[2,3]
One expects mixed-valence complexes with spin states
situated in the spin-crossover region to display several new
types of conjugated phenomena, coupled with spin and
charge, between neighboring metal ions, that minimize the
free energy in the whole system. To explore this possibility,
we have synthesized iron mixed-valence complexes whose
¯
Heterometal dto complexes were first synthesized by Ok-
awa et al., who described the ferromagnetism of (n-C3H7)4-
N[MIICrIII(dto)3] (MII = Fe, Co, Ni)[5] and compared their
ferromagnetic properties with those of (n-C4H9)4N[MII-
CrIII(ox)3] (MII = Mn, Fe, Co, Ni, Cu, ox = oxalato).[6] The
structures and magnetic properties of A[MIICrIII(dto)3] [A+
= Ph4P+, (CnH2n+1)4N+, n = 3–5; MII = Mn, Fe, Co] com-
plexes were investigated by J. M. Bradly et al.,[7] who con-
firmed that all the complexes were ferromagnets with Curie
temperatures between 5 and 16 K except for the [MnIICrIII
-
(dto)3] complex, which did not show a long-range ordering
above 2 K.
[a] Institute for Solid State Physics, The University of Tokyo,
Kashiwanoha 5-1-1, Kashiwa 277-8581, Chiba, Japan
Fax: +81-4-7136-3333
Recently, we have discovered a new type of first-order
phase transition at TCT Ϸ 120 K for (n-C3H7)4N[FeIIFeIII
-
E-mail: itoimiho@issp.u-tokyo.ac.jp
[b] Graduate School of Arts & Sciences, The University of Tokyo,
Komaba 3-8-1, Meguro-Ku, Tokyo 153-8902, Japan
Fax: +81-3-5454-6741
(dto)3], where the thermally induced charge-transfer be-
tween the FeII and FeIII is reversible.[8,9] This phase transi-
tion is schematically shown in Figure 1. In the high-tem-
perature phase (HTP), the FeIII (S = 1/2) and FeII (S =
2) sites are coordinated by six S atoms and six O atoms,
respectively. On the other hand, in the low-temperature
E-mail: cnori@mail.ecc.u-tokyo.ac.jp
[c] SPring-8/JASRI, 1-1-1 Kouto,
Mikazuki-cho, Sayo-gun, Hyogo 679-5198, Japan
[d] CREST, JST 4-1-8,
Honcho, Kawaguchi-shi, Saitama 332-0012 Japan
1198
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1198–1207

Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence Complexes
FULL PAPER
Figure 1. Schematic representation of a charge-transfer phase transition in (n-C3H7)4N[FeIIFeIII(dto)3].
phase (LTP), the FeIII (S = 5/2) site is coordinated by six O [MIICrIII(dto)3] (M = Mn, Fe, Co, Ni; n = 3–5) and A[MnII-
atoms, and the FeII (S = 0) site by six S atoms. Moreover, FeIII(dto)3] [A = N(CnH2n+1)4 (n = 3–5) and P(C6H5)4].[7,13]
we have found a ferromagnetic transition at 7 K. The ferro- In support of this suggestion, note that the infrared spectra
magnetic interaction takes place over a long distance in- of all the complexes show dto-ligand vibration bands; the
volving FeIII (S = 5/2) – S2C2O2 – FeII (S = 0) – O2C2S2 – ν(CO) stretching mode of the dithiooxalato molecule ap-
FeIII (S = 5/2). The ferromagnetic ordering is presumably pears around 1500 cm–1, (1490.91 cm–1, 1496.0 cm–1,
induced by the charge-transfer interaction between FeII and 1490.5 cm–1, and 1490.5 cm–1 for n = 3–6, respectively). In
FeIII
.
the case of the parent compounds K2(C2O2S2) and
In order to investigate the mechanism of the CT phase KBa[FeIII(dto)3]·3H2O, the ν(CO) vibration modes appear
transition and the ferromagnetism induced by the charge- at higher energies (1516 cm–1 and 1528 cm–1). The low val-
transfer interaction between FeII and FeIII for (n-C3H7)4N- ues of ν(CO) of the n = 3–6 complexes indicate that the dto
[FeIIFeIII(dto)3], we have synthesized iron mixed-valence molecule acts as a bridging ligand to form the 2-D network.
complexes, (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6), and in-
The
structural
characteristics
of
(n-CnH2n+1)-
[FeIIFeIII(dto)3] (n = 3–6) were investigated by powder X-
ray diffraction (PXD) using synchrotron radiation. Figure 2
shows the X-ray diffraction patterns for n = 3–6 taken over
1 h at room temperature. The profiles are broad, and several
peaks are only observed below 2θ Ͼ 20°. For the title com-
plexes, the thermal vibration of (n-CnH2n+1)4N+ appears to
be quite large. A PXD measurement at low temperatures
was carried out so that thermal vibration of the cation
might be prevented; however, the diffraction profiles were
still broad.
vestigated their crystal structures and physical properties by
powder X-ray diffraction, magnetic susceptibility measure-
ments, and ESR spectroscopy. In this paper, we report our
results concerning the CT phase transition and the ferro-
magnetic phase transition for (n-CnH2n+1)4N[FeIIFeIII
-
(dto)3] (n = 3–6; the complex is hereafter denoted as n = 3–
6). In addition, we discuss the mechanism of the CT phase
transition and the ferromagnetism induced by the charge-
transfer interaction between FeII and FeIII
.
The sharpest peak can be attributed to the (002) reflec-
tion, which generally moves to a higher angle with decreas-
ing temperature. The (002) reflections for (n-CnH2n+1)4N-
[FeIIFeIII(dto)3] systematically shift to lower angle as shown
in Figure 2(a). These results suggest that the interlayer spac-
ing becomes larger as the number of carbon atoms in the
alkyl chains increases.
Results
Structural Characterization
The crystal structure of (n-C3H7)4N[FeIIFeIII(dto)3] has
already been determined by single-crystal X-ray diffraction
analysis.[10] In this compound, FeII and FeIII are alternately
The unit cell parameters of n = 3 and 4 were estimated
bridged by dithiooxalato molecules, which forms the 2-D by the Rietveld method as shown in Figure 2(b) and (c).
honeycomb network structure of [FeIIFeIII(dto)3]–ϱ. The (n- For the refinement of n = 4, we used the coordination of
C3H7)4N+ cation layer is intercalated between two [FeII- (n-C4H9)4N[MnIIFeIII(ox)3] with the space group P63.[12]
FeIII(dto)3]–ϱ layers. This structure is common in heterome- The cell parameters a are 10.06 Å (n = 3) and 10.09 Å (n =
tal oxalato and dithiooxalato complexes.[11,12] The space 4). The lattice parameters c are 16.18 Å (n = 3) and 18.14 Å
group of (n-C3H7)4N[FeIIFeIII(dto)3] is P63.
(n = 4). The c parameter determined by PXD differs from
Single crystals of (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 4– the result of a single-crystal X-ray diffraction (SXD) study,
6) have not yet been obtained. However, crystals of (n- since the h k l reflections obtained by the Rietveld analysis
CnH2n+1)4N[FeIIFeIII(dto)3] (n = 4–6) appear to have honey- cover the range h = 0–3, k = 0–2, l = 0–4, a data range that
comb structures that are restricted to the ab plane. For these is much smaller than that of the single-crystal analysis (h =
crystals, the lattice parameter c increases with the number 0–12, k = –12 to 10, l = –16 to 15). The error in the cell
of carbon atoms in (n-CnH2n+1)4N+, a relationship based parameters in the PXD thus becomes larger than that in
on results from the analogous complexes (n-CnH2n+1)4N- the SXD.
Eur. J. Inorg. Chem. 2006, 1198–1207
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1199

M. Itoi, Y. Ono, N. Kojima, K. Kato, K. Osaka, M. Takata
FULL PAPER
shows the χT and the inverse magnetic susceptibility (χ–1)
as a function of temperature for n = 3–6. The magnetic
susceptibilities for n = 5 and 6 obey the Curie–Weiss law
[χ–1 = (T – θ)/C] in the range 50–300 K. However, for n =
3 and 4, the magnetic susceptibilities do not obey the Cu-
rie–Weiss law over this range. The change of slope is at
around 100 K on the χ–1 curves showing a thermal depen-
dence. The Weiss constants are +12 K (TCT Ͻ T: 120–
300 K) and +20 K (TCT Ͼ T: 30–98.4 K) for n = 3, +18 K
(TCT Ͻ T: 140–300 K), and +27 K (TCT Ͼ T; 39.7–100 K)
for n = 4. Those of n = 5 and 6 are +23 K and +21 K,
respectively.
Figure 2. (a) Powder X-ray diffraction patterns for (n-CnH2n+1)4N-
[FeIIFeIII(dto)3] (n = 3–6) at room temperature. (b) Result of the
Rietveld refinement for n = 3, and (c) for n = 4. Rwp and RI are
expressed as {[Σiwi(yi – yi,calc)2/Σiwiyi2]1/2} and {[ΣK|IK,calc – IK|/
ΣKIK,calc]}, respectively.
In the cases of n = 5 and 6, the cell parameter could
not be determined by the Rietveld method. The estimated
parameters of c are 19.67 Å and 22.71 Å, respectively. The
lattice parameters for (n-CnH2n+1)4N[FeIIFeIII(dto)3] are
listed in Table 1.
Table 1. Lattice parameters for (n-CnH2n+1)4N[FeIIFeIII(dto)3].
(n-CnH2n+1)4N[FeIIFeIII(dto)3] a [Å]
c [Å]
R*, RWP**
n = 3
10.06[a]
10.06[b]
10.09[b]
16.04[a]
16.18[b]
18.14[b]
19.67[c]
22.71[c]
0.091[a]
*
0.051[b]**
0.049[b]**
n = 4
n = 5
n = 6
[a] Value obtained from single-crystal structure analysis. [b] Refined
value obtained by using the Rietveld method. [c] Calculated from
the position of the (002) peak. Rwp is expressed as {[Σiwi(yi – yi,calc)2/
Σiwiyi2]1/2}.
Magnetic Properties
Figure 3. Temperature dependence of the magnetic susceptibility
multiplied by temperature (χT) and the inverse susceptibility (χ–1)
The static magnetic susceptibilities of n = 3–6 were mea-
sured in an external magnetic field of 5000 G. Figure 3 of (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6).
1200
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1198–1207

Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence Complexes
FULL PAPER
All these positive Weiss constants imply the existence of
a ferromagnetic interaction between FeII and FeIII. At room
temperature, the effective moments, µeff, are 5.87 µB (n = 3),
5.58 µB (n = 4), 5.94 µB (n = 5), and 5.48 µB (n = 6).
When kTϾϾ|J|, the effective moment of HTP with FeII
(S = 2) and FeIII (S = 1/2) is calculated as 5.91 µB (g =
2.00).
In order to confirm the ferromagnetic phase transition,
we investigated the field-cooled magnetization (FCM), the
remanent magnetization (RM) and the zero-field cooled
magnetization (ZFCM) for n = 3–6, the results of which are
shown in Figure 4. The FCM curve was obtained on cool-
ing with an external magnetic field of 30 G. After the FCM
measurement, the magnetic field was switched off at 2 K;
then, the RM was measured from 2 K to 35 K. After cool-
ing from 300 K to 2 K in zero external field, an external
field of 30 G was switched on at 2 K. Then, the ZFCM was
measured from 2 K to 35 K.
Figure 4(a) displays the FCM, RM, and ZFCM as a
function of temperature for n = 3. The FCM curve shows
a rapid increase below 8 K, as well as a tendency to saturate
below 6 K. The RM vanishes at about 7 K. Below 7 K, the
ZFCM is smaller than the FCM, a result which is due to
the fact that the applied magnetic field of 30 G is too weak
to move the magnetic domain walls below the ferromag-
netic temperature. The ZFCM and FCM curves overlap at
7 K, where the magnetic hysteresis disappears. From these
results, we estimate the Curie temperature to be 7 K. The
FCM and ZFCM curves for n = 4, provided in Figure 4(b),
are different from those for n = 3. The FCM has a shoulder
at 13 K and the ZFCM has two peaks at 7 K and 13 K.
The RM vanishes at 13 K, where the ZFCM and FCM
meet each other. The ZFCM of n = 4 implies that two ferro-
magnetic phases with Tc = 7 K and 13 K coexist, which we
can regard as a mixture of HTP and LTP at a low tempera-
ture. The coexistence of HTP and LTP is consistent with
the ESR data. Figure 4(c) and (d) show the temperature
dependences of the magnetizations for n = 5 and 6. The
FCM steeply increases at 23 K and 26 K for n = 5 and 6,
respectively, and the RM disappears at 19 K and 25 K for
n = 5 and n = 6, respectively, where the ZFCM and FCM
curves meet each other. The Curie temperatures for n = 3–
6 were evaluated at 7 K, 7 K (&13 K), 19.5 K and 22 K,
respectively, from heat capacity measurements (for n = 3
and 4),[14,15] and by an Arrott plot,[16] i.e. M2 – H/M (1/χ)
plot (for n = 5 and 6). The Curie temperatures for n = 3–6
are listed in Table 2.
Figure 4. Temperature dependence of the field cooled magnetiza-
tion (FCM:᭺), the remnant magnetization (RM:ᮀ), and the zero-
field cooled magnetization (ZFCM:᭡) for (n-CnH2n+1)4N[FeIIFeIII
(dto)3] (n = 3–6). Applied magnetic field H = 30 G.
-
ESR Spectroscopy
Reflecting the CT phase transition, a small bump ap-
In order to investigate the CT phase transition and the
pears in the χT curve and the slope of χ–1 changes around spin state of the Fe atom, we obtained the ESR spectra for
120 K (n = 3) and 140 K (n = 4). We carefully measured n = 3–6 between 4 K and 300 K. The experimental condi-
the magnetic susceptibility for n = 3 and 4 with a sweep tions are as described in ref.[17]
rate of 0.1 Kmin–1 on heating and cooling, the results of
Figure 6 shows the temperature dependence of the ESR
which are shown in Figure 5. A thermal hysteresis loop ap- signal of the n = 3 complex. When the temperature was
pears between 60 K and 130 K for n = 3, and between 50 K lowered from room temperature, an ESR signal corre-
and 150 K for n = 4. The χT curve for n = 4 changes by sponding to the FeIII site with S = 5/2 appeared below
about 4% between 50 K and 150 K. This value is smaller 105 K, and its intensity strengthened with decreasing tem-
than that for n = 3, which drops by about 7% between 60 K perature. On heating, the ESR signal vanished above 130 K.
and 130 K.
The g value of the ESR signal at the center of the 3200-G
Eur. J. Inorg. Chem. 2006, 1198–1207
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1201

M. Itoi, Y. Ono, N. Kojima, K. Kato, K. Osaka, M. Takata
FULL PAPER
Table 2. The parameters related to the change-transfer phase transition and ferromagnetism for (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6).
most disappeared above 145 K. The ∆Hpp for n = 4 has an
almost constant value of 1600 G between 50 and 140 K. In
addition to the ESR signal due to the HS state (6A1g) of
FeIII, a weak signal is observed at 1600 G, whose g value is
4.1. This weak signal should be attributed to the HTP frac-
5
tion, i.e. the HS state (S = 2; T2g) of the FeII and/or the
2
LS state (S = 1/2; T2g) of the FeIII. For n = 5 and 6, a
similar ESR signal appears at about 1600 G. Therefore, the
HTP coexists with the LTP at 20 K for n = 4, a finding
which could account for the existence of the two peaks in
the ZFCM.
Figure 5. Temperature dependence of the magnetic susceptibility
for (n-C3H7)4N[FeIIFeIII(dto)3] and (n-C4H9)4N[FeIIFeIII(dto)3].
Temperature sweep rate is 0.1 Kmin–1; (Ǟ) and (ǟ) denote heating
and cooling, respectively.
field is 2.10, which is consistent with that of the HS state
(6A1g) of FeIII, whose spin-orbit interaction is negligibly
small. The half width (∆Hpp) has an almost constant value
of ∆Hpp = 1470 G between 60 K and 120 K. In the case of
n = 4, as shown in Figure 7, a similar ESR signal corre-
sponding to the HS state (6A1g) of FeIII appeared below
Figure 6. Temperature dependence of the ESR spectra on heating
105 K in the cooling mode. On heating, the ESR signal al- and cooling for (n-C3H7)4N[FeIIFeIII(dto)3].
1202
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1198–1207

Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence Complexes
FULL PAPER
Figure 8. Temperature dependence of the spin susceptibility for (n-
CnH2n+1)N[FeIIFeIII(dto)3] (n = 3 and 4).
Figure 7. Temperature dependence of the ESR spectra on heating
and cooling for (n-C4H9)4N[FeIIFeIII(dto)3]. Inset denotes the ESR
signal corresponding to the FeIII (2T2g, S = 1/2) of the HTP at
160 K and room temperature.
We estimated the spin susceptibilities corresponding to
the HS state (6A1g) of the FeIII atom for n = 3 and 4.
The spin susceptibility is expressed as χspin
=
A(I(H)×{∆HPP}2), where A is a constant value, I(H) is the
signal intensity, and ∆HPP is the half width.[18] The spin
susceptibilities for n = 3 and n = 4 are shown in Figure 8.
The large thermal hysteresis loop in χspin for n = 4 is in
good agreement with that of the dc magnetic susceptibility;
however, the thermal hysteresis loop in χspin for n = 3 is
narrower than that of the dc magnetic susceptibility.
In the cases of KBa[FeIII(dto)3]·3H2O and (n-CnH2n+1)4-
N[FeIIFeIII(dto)3] (n = 5 and 6), the ESR spectra are com-
pletely different from those of n = 3 and 4. For n = 5 and
6, we could not observe a drastic signal change caused by
the CT phase transition such as we noted for n = 3 and 4.
The ESR signals of n = 5 and 6 are shown in Figure 9. The
ESR spectra of the parent complex, KBa[FeIII(dto)3]·3H2O,
is shown in Figure 9(a). The paramagnetic spin of
Figure 9. The ESR spectra at several temperatures for (a)
KBa[FeIII(dto)3]·3H2O; (b) (n-C5H11)4N[FeIIFeIII(dto)3]; and (c)
(n-C6H13)4N[FeIIFeIII(dto)3].
KBa[FeIII(dto)3]·3H2O has been already reported as the the strong signal with g = 4.10 disappeared above 90 K.
FeIII LS state (S = 1/2) on the basis of magnetic suscep- Therefore, we suggest that the HS state (S = 5/2) of the
tibility and Mössbauer spectroscopy results.[11,19,20]
FeIII atom caused by linkage isomerization (FeIIIO4S2 and /
The KBa[FeIII(dto)3]·3H2O complex has two ESR sig- or FeIIIO6) is responsible for this appearance of the ESR
nals: a strong signal at 1600 G, and a weak signal at 3200 G. signal around 3200 G at room temperature. The 57Fe Möss-
The g values of the ESR signals at 1600 G and 3200 G are bauer spectrum for KBa[FeIII(dto)3]·3H2O did not enable
estimated at 4.10 and 2.01 and are attributed to the gЌ and us to distinguish the minor fraction of the FeIII HS state (S
gʈ of FeIII (S = 1/2), respectively. However, the ESR signal = 5/2) caused by linkage isomerization from the LS state (S
with g = 2.01 persisted at room temperature, even though = 1/2), because these Mössbauer spectra appear at almost
Eur. J. Inorg. Chem. 2006, 1198–1207
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1203

M. Itoi, Y. Ono, N. Kojima, K. Kato, K. Osaka, M. Takata
FULL PAPER
the same velocities. Therefore, the ESR signal due to the
FeIII HS state (S = 5/2) is presumably superposed on the
ESR signal with gʈ = 2.01 corresponding to the FeIII LS
state (S = 1/2).
For n = 5 and 6, the ESR spectra were quite similar to
those of KBa[FeIII(dto)3]·3H2O. A strong signal at 1580 G
and a weak signal at 3260 G were observed. The g values
of the ESR signals at 1580 G and 3260 G are estimated at
4.10 and 2.01, respectively, which are attributed to the gЌ
and gʈ of the LS state (S = 1/2) of FeIII. The signal with g
= 2.01 persisted at room temperature, so that an ESR signal
due to the HS state (S = 5/2) of the FeIII atom caused by
linkage isomerization (FeIIIO4S2 and /or FeIIIO6) appeared
around 3260 G. Note that the ratio of the linkage isomer
FeIIO4S2 (S = 2) can be estimated easily by Mössbauer
spectroscopy. The percentages of the linkage isomer of
FeIIO4S2 are about 8% at 60 K for n = 3, 14% at 25 K for
n = 4, 4% at 25 K for n = 5, and 11% at 30 K for n = 6.
The FeIIO4S2 persists at low temperatures without creating
charge-transfer.[21]
In the cases of n = 5 and 6, the ESR signals of the LS
state (S = 1/2) of the FeIII atom decrease with increasing
temperature and then almost disappear above 90 K; the lat-
ter change takes place because of rapid spin-lattice relax-
ation. The ESR signal of the HS state (5T2g) of FeII and the
LS state (2T2g) of FeIII could not be observed in the high-
temperature region. From the analysis of the ESR spectra,
we conclude that the CT phase transition does not take
place for n = 5 and 6, in which the spin configurations of
FeII (S = 2) and FeIII (S = 1/2) are stable between 4 K and
300 K.
Figure 10. Schematic representation of ferromagnetic ordering for
(n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6). (a) The ferromagnetic or-
dering for n = 3 and n = 4 takes place in the LTP with the spin
configuration FeII (S = 0) and FeIII (S = 5/2). (b) The ferromagnetic
ordering for n = 5 and n = 6 takes place in the HTP with the spin
configuration FeII (S = 2) and FeIII (S = 1/2).
the [FeIIFeIII(dto)3] honeycomb ring. The cell parameter a
of the analogous complexes, (n-CnH2n+1)4N[FeIIFeIII(ox)3]
(n = 3–5)[22] tends to expand slightly as n increases. For n
= 4, the slight expansion or distortion of the 2-D honey-
Discussion
The spin state in the paramagnetic-ferromagnetic region comb ring (∆a = 0.03 Å) might be expected to lead to an
and the existence of the charge-transfer phase transition of inhibition of CT between the Fe atoms; thus, the HTP spin
(n -CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6) are revealed by partially remains at low temperature. We should note that
magnetic susceptibility measurements and ESR spec- n = 5 shows a CT phase transition under 0.5 GPa, and the
troscopy. The appearance of the CT phase transition and ferromagnetic transition temperature changes from 19.5 K
the mechanism of the ferromagnetism in an iron mixed-val- to 7 K. In other words, the spin configuration changes from
ence tris-dithiooxalato complex critically depend on the cat- FeII (S = 2) – FeIII (S = 1/2) (HTP) to FeII (S = 0) – FeIII
ion size (n-CnH2n+1)4N+ (n = 3–6). The CT phase transi- (S = 5/2) (LTP).[23] There is a limiting size of the honey-
tions are observed for n = 3 (120 K) and 4 (140 K), and comb ring between n = 4 and n = 5 where the CT does not
then the FeII (S = 0) – FeIII (S = 5/2) spin combination occur. Unfortunately, we could not obtain information on
causes the ferromagnetic transition at 7 K. On the other the crystal structure of the (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n
hand, the CT phase transition is not apparent for n = 5 and = 5, 6) complexes by PXD. A single crystal X-ray analysis
n = 6; FeII (S = 2) – FeIII (S = 1/2) spins are stable between of the n = 4–6 and an X-ray diffraction measurement taken
2 K and 300 K. The spin configuration corresponding to under pressure are necessary to quantitatively clarify the
the HTP of n = 3 and 4 is responsible for the ferromagnetic cation size effects on the [FeIIFeIII(dto)3]– layer.
transitions at 19.5 K (n = 5) and 22 K (n = 6), respectively.
The mechanism of the magnetic interaction in the iron
The ferromagnetically ordered states and several param- tris-dithiooxalato complexes is dramatically changed ac-
eters for n = 3–6 are schematically shown in Figure 10 and cording to whether a CT phase transition exists. Here, we
Table 2.
will discuss the origin of the magnetic interaction in the title
Due to the effect of the cation size, the [FeIIFeIII(dto)3] complexes. The ferromagnetic ordering for n = 3 and 4 is
layers are separated by an increment corresponding to the considered to be established by the charge-transfer interac-
length of the alkyl chain from n = 3 to n = 6. The bulkier tion between FeII and FeIII, the same mechanism as in FeIII
-
4
(n-CnH2n+1)4N+ cations probably expand or distort slightly [FeII(CN)6]3·15H2O (TC = 5.5 K),[24,25] as shown in Fig-
1204
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1198–1207

Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence Complexes
FULL PAPER
ure 11(a). In contrast, for n = 5 and 6, the spin configura- angular interactions in the 2-D plane [Figure 10(a)]. In the
tion of FeII (S = 2) and FeIII (S = 1/2) is the ground state case of the triangular lattice, the number of nearest neigh-
between 2 K and 300 K. The ferromagnetic ordering for n bors of the FeIII (S = 5/2) site is 6. The molar susceptibility
= 5 and 6 is schematically shown in Figure 11(b); each FeIII of the triangular lattice, calculated by a high-temperature
site in the LS state (S = 1/2) accepts a t2 electron with a expansion up to the second order, is given by [Equation
down-spin from the neighboring FeII site in the HS state (S (1)]:[26]
= 2). Therefore, the coupling between the ground configura-
5
4 2
tion φi[FeIII(t2 )]φj[FeII(t2 e )] and the charge-transfer con-
6
3 2
figuration φi[FeII(t2 )]φj[FeIII(t2 e )] stabilizes the ground
(1)
state, a situation favoring ferromagnetic ordering. In the
Here β is 1/kT, N is Avogadro’s number, and µB is the
Bohr magneton. After substituting S = 5/2 and g = 2.00
into (1), we estimate the exchange coupling by the least-
squares method. The fitting curves are shown in Fig-
ure 12(a). The range analyzed in the LTP is from 20 to
122 K for n = 3, and from 20 to 140 K for n = 4. The values
obtained are J = 3.12 K for n = 3, and 2.39 K for n = 4.
5
4 2
ground configuration of φi[FeIII(t2 )]φj[FeII(t2 e )], in ad-
dition to the charge-transfer interaction, the potential ex-
change interaction due to the orbital orthogonality between
the t2 electron with an up-spin FeIII and the FeII electrons
with up-spins (in e orbitals and one t2 orbital) induces fer-
romagnetic ordering, which is presumably responsible for
the ferromagnetic transition with a high TC in n = 5 and 6.
Figure 12. Temperature dependences of χ for n = 3–6 with high-
temperature expansion fitting. (a) dashed line: fitting for n = 3;
short-dashed line: fitting for n = 4 in LTP spin configuration (FeIII
S = 5/2 – FeII S = 0) on triangular lattice. Molar susceptibility for
n = 3 and 4 is indicated by plus (+) and circle (᭺) symbols, respec-
tively. (b) dash-dotted line: fitting for n = 5; solid line: fitting for n
= 6 in HTP spin configuration (FeIII S = 1/2 – FeII S = 2) on
hexagonal lattice. Molar susceptibility for n = 5 and 6 is indicated
by the solid triangle (᭡) and square (ᮀ) symbols, respectively.
Figure 11. (a) Ferromagnetism induced by the charge-transfer in-
teraction between FeII (S = 0) and FeIII (S = 5/2) for (n-CnH2n+1)4-
N[FeIIFeIII(dto)3] (n = 3, 4). (b) Ferromagnetism induced by the
charge-transfer interaction as well as the potential exchange inter-
action between FeII (S = 2) and FeIII (S = 1/2) for (n-CnH2n+1)4N-
[FeIIFeIII(dto)3] (n = 5, 6). Solid line arrow: charge transfer interac-
tion (JCT Ͼ 0); dash-dotted line arrow: kinetic exchange interaction
(JK Ͻ 0); short-dashed line arrow: potential exchange interaction
(JP Ͼ 0).
On the other hand, for n = 5 and 6, there is no CT phase
transition, and the FeIII (S = 1/2) – FeII (S = 2) interaction
takes place on hexagonal rings [Figure 10(b)]. In the case
of a hexagonal lattice, the number of nearest neighbors is
3. For n = 5 and 6, the FeII (S = 2) atom is connected with
three FeIII (S = 1/2) atoms. The high-temperature expansion
of the susceptibility for a hexagonal lattice is calculated by
the same procedure as (1), and we get the molar suscep-
tibility as follows [Equation (2)]:
Next, to quantify the ferromagnetic interaction between
the iron atoms of (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6),
we will estimate the exchange interaction J in a 2-D layer
by means of the simplest Hamiltonian,
Here, h is the external magnetic field. For n = 3 and 4,
the LTP spin state with FeIII(S = 5/2) – FeII(S = 0) appears
via a CT phase transition, and the states produce large tri-
(2)
Eur. J. Inorg. Chem. 2006, 1198–1207
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1205

M. Itoi, Y. Ono, N. Kojima, K. Kato, K. Osaka, M. Takata
FULL PAPER
tion, and KBa[FeIII(dto)3]·3H2O precipitated. The salt was recrys-
tallized from water and dried for one day.[20,29,30] C6H6BaFeKO9S6
(646.8): calcd. C 11.14, H 0.94, S 29.75; found C 11.39, H 1.20, S
29.66.
After substituting SA = 1/2 and SB = 2 into (2), we esti-
mated the exchange coupling and g values by the least-
squares method. The range analyzed is from 31 to 300 K for
n = 5, and from 21 to 300 K for n = 6. Then, we obtained J
= 36.57 K for n = 5 and J = 29.99 K for n = 6, with g1/2 (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6): The title complexes were
synthesized in a way similar to the preparation of (n-C3H7)4N-
=g2 = 4.00 values, which were in agreement with the TC and
g values obtained by the ESR spectra. The fitting curves are
shown in Figure 12(b). The ferromagnetic interactions J for
n = 3 and 4 are considerably lower than those for n = 5 and
6.
[MIICrIII(dto)3] (M = Fe, Co, Ni, Zn).[5]
A solution of KBa[Fe(dto)3]·3H2O (0.5 g, 0.77 mmol) in a 3:2
methanol/water mixture (20 mL) was stirred. To this was added
slowly a solution of FeCl2·4H2O (0.24 g, 0.77 mmol) and (n-
CnH2n+1)4NBr (n = 3–6) (1.15 mmol) in a 3:2 methanol/water mix-
ture (20 ml), whereupon a black powdered crystal precipitated. The
compounds (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 3–6) were separated
as powdered crystals by suction filtration and washed first with a
1:1 methanol/water mixture and then with methanol and diethyl
ether. The compounds (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n = 7, 8)
could not be obtained as crystalline powders. The elemental analy-
sis for the precipitated crystal that had been stirred at 30–40 °C for
12 h was in good agreement with the calculated one; the results are
listed in Table 3.
Conclusion
We have synthesized (n-CnH2n+1)4N[FeIIFeIII(dto)3] (n =
3–6), and have investigated the cation-size effect on the CT
phase transition and the ferromagnetic transition in these
compounds. The result of our PXD measurement con-
firmed that the 2-D honeycomb network layer [FeIIFeIII
-
(dto)3]– is increasingly well separated with increases in the
number n of the (n-CnH2n+1)4N cations. We have also deter-
mined, by magnetic susceptibility and ESR spectroscopy
measurements, the spin states of the FeII and FeIII sites for
n = 3–6 between 2 and 300 K. The (n-CnH2n+1)4N+ cation
size affects the existence or nonexistence of the CT phase
transition and the ferromagnetic interaction J between the
FeII and FeIII sites. For n = 3 and 4, we observed the ther-
mally induced charge transfer between FeII and FeIII
around 120 K (n = 3) and 140 K (n = 4). The spin configu-
ration changes from FeII (S = 2) – FeIII (S = 1/2) (HTP) to
FeII (S = 0) – FeIII (S = 5/2) (LTP) in the vicinity of the CT
phase transition temperature (TCT). Furthermore, the FeIII
(S = 5/2) spin, which is located on a large triangular lattice,
is ferromagnetically ordered at 7 K. On the other hand, for
n = 5 and 6, the CT phase transition did not occur, and the
spin configuration of the HTP for n = 3 and 4 dominated
over the whole temperature range. The FeII (S = 2) – FeIII
(S = 1/2) spin configurations located on a hexagonal lattice
produce the ferromagnetic transitions at 19.5 K (n = 5) and
22 K (n = 6), respectively.
Table 3. Elemental analysis of C, H, N, and S in (n-CnH2n+1)4N-
[FeIIFeIII(dto)3] (n = 3–6).
Measurements: The powder X-ray diffraction data was taken at the
BL02B2 beam line of SPring-8.[31] The as-grown powdered sample
was sealed in a Lindemann capillary having a diameter of 0.3 mm,
which gave a homogeneous intensity distribution in the Debye–
Scherrer ring. The wavelength of the X-ray was 1.001 Å, and the
exposure time at 300 K was 1 h. The static magnetic susceptibility
was measured with a Quantum Design MPMS5 SQUIID magne-
tometer. Powdered crystals (10 mg) were wrapped in a polyethylene
film and held in a plastic straw. The magnetic susceptibility ob-
tained was corrected for the background and the core diamagne-
tism. The diamagnetic correction for constituting atoms was car-
ried out using Pascal’s constants.[32] The diamagnetic Pascal’s
corrections are as follows: –4.18·10–4 cm3 mol–1 (n
= 3),
–4.66·10–4 cm3 mol–1 (n = 4), –5.13·10–4 cm3 mol–1 (n = 5), and
–5.60·10–4 cm3 mol–1 (n = 6). The ESR spectra for the powdered
crystals were obtained with a JEOL JES-TE300 X-band ESR spec-
trometer. The magnetic field was monitored with an Echo Elec-
tronics ES-FC5 field meter.
Experimental Section
Syntheses
K2C2O2S2: Potassium dithiooxalate, K2C2O2S2, was prepared by
the reaction of potassium hydrogensulfide KHS (14.57 g, 0.20 mol)
and di-n-pentyl dithiooxalate (n-C5H11)2(C2O2S2) (0.10 mol) in
methanol. The di-n-pentyl dithiooxalate was obtained by the reac-
tion of oxalyl dichloride (COCl)2 (8.61 mL, 0.10 mol) and 1-pen-
tanethiol CH3(CH2)4SH (25 mL, 0.20 mol). After filtration, the
K2C2O2S2 was washed well with methanol to remove the KHS and,
because K2C2O2S2 is highly hygroscopic,[27,28] it was dried in vacuo.
C2K2O2S2 (198.35): calcd. C 12.11, S 32.33; found C 12.28, S 32.54.
Acknowledgments
The authors wish to thank Prof. K. Asai, Prof. Y. Uwatoko, Prof.
F. Varret, Prof. K. Boukheddaden, Prof. C. Itoi, Dr. Y. Kobayashi,
and Dr. M. Enomoto for many helpful discussions. This work was
supported by a Grant-in-Aid for Scientific Research and a Grant-
in-Aid from the 21st Century COE (Center of Excellence) Program
(Research Center for Integrated Science) of the Japanese Ministry
KBa[FeIII(dto)3]·3H2O: Fe(NO3)3·10H2O (2.10 g, 5.61·10–3 mol) of Education, Science, Sports, and Culture. The synchrotron-radia-
was treated with K2C2O2S2 (3.00 g, 15.1 mmol) in cold water tion X-ray powder experiments were performed at the SPring-8
(25 mL), and the solution was filtered to remove iron sulfide. Then BL02B2 with approval of the Japanese Synchrotron Radiation Re-
BaBr2·2H2O (2.50 g, 0.17 mol) was added to the dark purple solu- search Institute (JASRI). One of the authors (M. Itoi) was sup-
1206
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1198–1207

Charge-Transfer Phase Transition and Ferromagnetism of Iron Mixed-Valence Complexes
FULL PAPER
[17] (1) n = 3: Frequency 9.15307 GHz, Microwave power 1.06 mW,
modulation width 1 mT, Amplitude 63–2000 (from 10 K to
150 K) (2) n = 4: 9.15811 GHz, 1.00 mW, 0.1 mT, 2–1000 (from
10 K to 160 K) (3) n = 5: 9.16210 GHz, 1.00 mW, 0.5 mT, 160–
630 (from 14 K to 206 K) (4) n = 6: 9.16754 GHz, 1.01 mW,
0.02 mT, 250–1000 (from 10 K to 91 K).
[18] C. P. Slichter, Principles of Magnetic Resonance, Springer-Ver-
lag, Berlin, Heidelberg, 2000.
[19] T. Bichall, K. M. Tnn, Inorg. Chem. 1976, 15, 376–380.
[20] F. P. Dwyer, A. M. Sargeson, J. Am. Chem. Soc. 1959, 81,
2335–2336.
[21] Y. Ono, Doctoral Thesis, The University of Tokyo, 2004.
[22] C. Mathonière, C. J. Nuttall, S. G. Carling, P. Day, Inorg.
Chem. 1996, 35, 1201–1206.
[23] Y. Kobayashi, M. Itoi, N. Kojima, K. Asai, J. Magn. Magn.
Mater. 2004, 272–276, 1091–1092.
[24] B. Mayoh, P. Day, J. Chem. Soc., Dalton Trans. 1974, 846–852.
[25] B. Mayoh, P. Day, J. Chem. Soc., Dalton Trans. 1976, 1483–
1486.
ported by Research Fellowships of the Japanese Society for the
Promotion of Science for Young Scientists.
[1]
[2]
S. Decurtins, P. Gütlich, C. P. Köhler, H. Spiering, A. Hauser,
Chem. Phys. Lett. 1984, 105, 1–4.
J. Kröber, E. Codjovi, O. Kahn, F. Grolière, C. Jay, J. Am.
Chem. Soc. 1993, 115, 9810–9811.
[3]
[4]
O. Kahn, C. J. Martinez, Science 1998, 279, 44.
L. Cambi, A. Cagnasso, Atti Accad. Naz. Lincei 1931, 13, 809–
813.
H. Okawa, M. Mitsumi, M. Ohba, M. Kodera, N. Matsumoto,
Bull. Chem. Soc. Jpn. 1994, 67, 2139–2144.
¯
[5]
[6]
H. Tamaki, Z. J. Zhong, N. Matsumoto, S. Kida, M. Koikawa,
¯
N. Achiwa, Y. Hashimoto, H. Okawa, J. Am. Chem. Soc. 1992,
114, 6974–6979.
[7]
[8]
J. M. Bradly, S. G. Carling, D. Visser, P. Day, D. Hautot, G. J.
Long, Inorg. Chem. 2003, 42, 986–996.
N. Kojima, W. Aoki, M. Seto, Y. Kobayashi, Yu. Maeda,
¯
¯
[26] T. Oguchi, Statistical Theory of Magnetism, SYOKABO, 1st
ed., 1970, ch. 4.
Synth. Met. 2001, 121, 1796–1797.
[9] N. Kojima, W. Aoki, M. Itoi, Y. Ono, M. Seto, Y. Kobayashi,
Yu. Maeda, Solid State Commun. 2001, 120, 165–170.
[10] M. Itoi, A. Taira, M. Enomoto, N. Matsushita, N. Kojima, Y.
Kobayashi, K. Asai, K. Koyama, T. Nakano, Y. Uwatoko, So-
lid State Commun. 2004, 130, 415–420.
[27] H. O. Jones, H. S. Tasker, J. Chem. Soc., Trans. 1910, 95, 1904–
1909.
[28] C. S. Robinson, H. O. Jones, J. Chem. Soc., Trans. 1912, 101,
62–76.
[29] The number (n = 3) of the water of crystallization in KBa-
[FeIII(C2O2S2)3]·nH2O was determined by TG analysis.
[30] D. Coucouvanis, D. Piltingsrud, J. Am. Chem. Soc. 1973, 95,
5556–5563.
[11] Y. Ono, M. Okubo, N. Kojima, Solid State Commun. 2003,
126, 291–296.
[12] R. Pellaux, H. W. Schmalle, R. Huber, P. Fischer, T. Hauss, B.
Ouladdiaf, S. Decurtins, Inorg. Chem. 1997, 36, 2301–2308.
[13] S. G. Carling, J. M. Bradley, D. Visser, P. Day, Polyhedron 2003,
22, 2317–2324.
[14] T. Nakamoto, Y. Miyazaki, M. Itoi, Y. Ono, N. Kojima, M.
Sorai, Angew. Chem. Int. Ed. 2001, 40, 4716–4719.
[15] T. Nakamoto, Y. Miyazaki, M. Itoi, Y. Ono, N. Kojima, M.
Sorai, to be submitted.
[31] E. Nishibori, M. Takata, K. Kato, M. Sakata, Y. Kubota, S.
Aoyagi, Y. Kuroiwa, M. Yamamoto, N. Ikeda, The Large De-
bye–Scherrer Camera Installed at SPring-8 BL02B2 for Charge
Density Studies, Nuclear Instruments and Method in Physics
Research A 2001, pp. 467–468, 1045–1048.
[32] O. Kahn, Molecular Magnetism, Wiley-VCH, 1993, ch.1.
Received: June 25, 2005
[16] A. Arrott, Phys. Rev. 1957, 108, 1394–1396.
Published Online: February 2, 2006
Eur. J. Inorg. Chem. 2006, 1198–1207
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1207