Synthesis and electrochemical reaction with lithium of mesoporous iron oxalate nanoribbons

Article abstract of DOI:10.1021/ic8008927

Mesoporous FeC₂O₄ was prepared by dehydration of bulk monoclinic- and micellar orthorhombic FeC₂O₄ ·2H₂O precursors at 200°C. The micellar material shows nanoribbon shaped particles, which are preserv

Full text of DOI:10.1021/ic8008927

Inorg. Chem. 2008, 47, 10366-10371
Synthesis and Electrochemical Reaction with Lithium of Mesoporous
Iron Oxalate Nanoribbons
Mar ´ı a Jos e´ Aragón, Bernardo León, Carlos Pérez Vicente, and Jos e´ L. Tirado*
Laboratorio de Qu ´ı mica Inorg a´ nica, UniVersidad de C o´ rdoba, Edificio C3, Campus de Rabanales,
1
4071 C o´ rdoba, Spain
Received May 16, 2008
2 4 2 4 2
Mesoporous FeC O was prepared by dehydration of bulk monoclinic- and micellar orthorhombic FeC O ·2H O
precursors at 200 °C. The micellar material shows nanoribbon shaped particles, which are preserved after dehydration.
These solids are used as high-capacity lithium storage materials with improved rate performance. The mesoporous
nanoribbons exhibit higher capacities close to 700 mA h/g after 50 cycles at 2C (C ) 1 Li h-1 mol ) rate between
-1
0
and 2 V.
Introduction
Recently, we were able to extend this process to submi-
crometric particles of manganese carbonate with the calcite
In recent years, nanomaterials have burst upon the scene
of lithium battery research. New electrode materials have
been recently found which provide higher capacities at higher
rates by using the advantage of dispersion at the nanoscale
and the enhanced stability of one-dimensional nanopar-
9
structure. This study revealed that MnCO
3
can be used
directly as a conversion electrode versus lithium. The
discharge of lithium test cells takes place by a different
conversion reaction than that observed for the oxide produced
during the thermal decomposition of the carbonate, MnO.
Similar values of reversible capacity and better capacity
retention were observed for the carbonate as compared with
the oxide.
1-3
ticles.
In parallel, a significant advance in the understand-
ing of conversion electrode materials for lithium batteries
3
-6
has taken place. Transition metal oxides
and, more
7
,8
recently, transition metal fluorides are known examples
of solids involved in conversion reactions versus lithium.
The electrochemical conversion process leads to the reduction
of the metal ions to the metallic state together with the
formation of lithium oxide or lithium fluoride, respectively.
The partial reversibility of these processes makes these solids
potential candidates for the negative or positive electrode of
advanced lithium-ion batteries, respectively.
In this work, the study of oxysalts as conversion electrode
material is investigated for a common and inexpensive
compound: iron oxalate prepared by a novel synthesis
procedure. The use of dehydrated nanoparticles of this
compound provides interesting values of capacity at high
rates.
Experimental Section
Two samples of iron(II) oxalate dihydrate were studied. Com-
mercial iron oxalate (Panreac, Barcelona) was used as received.
*
To whom correspondence should be addressed. E-mail: iq1ticoj@
uco.es.
(
1) Armstrong, A. R.; Armstrong, G.; Canales, J.; Garc ´ı a, R.; Bruce, P. G.
The synthesis of FeC
2
O
2
·2H O nanoparticles was carried out by a
4
AdV. Mater. 2005, 17, 862–865.
reverse micelles procedure from water-in-oil microemulsions. First,
two microemulsions (I and II) were obtained under an argon
atmosphere as follows. Microemulsion I contained cetyl-trimethy-
lammonium bromide (CTAB) as the surfactant, hexanol as the
cosurfactant, isooctane as the hydrocarbon phase and 0.3 M iron(II)
sulfate solution as the aqueous phase. Microemulsion II has the
same constituents as above except for having 0.3 M ammonium
oxalate instead of iron sulfate as the aqueous phase. The weight
(
2) Chan, C. K.; Peng, H.; Liu, G.; McIlwrath, K.; Zhang, X. F.; Huggins,
R. A.; Cui, Y. Nature Nanotech. 2008, 3, 31–35.
(
(
3) Li, Y.; Tan, B.; Wu, Y. Nano Lett. 2008, 8, 265–270.
4) Poizot, P.; Laruelle, S.; Grugeon, S.; Tarascon, J. M. J. Electrochem.
Soc. 2002, 149, A1212-A1219.
(
(
(
(
5) Alc a´ ntara, R.; Jaraba, M.; Lavela, P.; Tirado, J. L. Chem. Mater. 2002,
1
4, 2847–2848.
6) Lavela, P.; Otiz, G.; Tirado, J. L.; Zhecheva, E.; Stoyanova, R.;
Ivanova, S. J. Phys. Chem. C 2007, 111, 14238–14246.
7) Badway, F.; Mansour, A. N.; Pereira, N.; Al-Sharab, J. F.; Cosandey,
F.; Plitz, I.; Amatucci, G. G. Chem. Mater. 2007, 19, 4129–4141.
8) Liao, P.; MacDonald, B. L.; Dunlap, R. A.; Dahn, J. R. Chem. Mater.
(9) Arag o´ n, M. J.; P e´ rez-Vicente, C.; Tirado, J. L. Electrochem. Commun.
2007, 9, 1744–1748.
2
008, 20, 454–461.
10366 Inorganic Chemistry, Vol. 47, No. 22, 2008
10.1021/ic8008927 CCC: $40.75  2008 American Chemical Society
Published on Web 10/11/2008

Mesoporous Iron Oxalate Nanoribbons
Figure 1. XRD patterns of iron oxalate samples: (a) micellar FeC2O4 ·2H2O, Miller indices for orthorhombic Cccm, (b) dehydrated (200 °C), FeC2O4
product from the micellar sample, (c) commercial FeC2O4 ·2H2O, Miller indices for C2/m, and (d) dehydrated (200 °C), FeC2O4 product from the commercial
sample.
Figure 2. 57Fe M o¨ ssbauer spectra of iron oxalate samples: (a) micellar FeC2O4 · 2H2O, (b) dehydrated FeC2O4 from micellar sample, (c) commercial
FeC2O4 ·2H2O (d) dehydrated FeC2O4 from commercial sample, (e, f) used electrodes after first discharge at 0 V and recharge at 3 V, respectively.
fractions of the various constituents in these microemulsions are
as follows: 16.76% of CTAB, 13.9% of hexanol, 59.29% of
isooctane, and 10.05% of the aqueous phase. In a second step, both
microemulsions were slowly mixed and stirred overnight on a
magnetic stirrer. The resulting precipitate was separated from the
apolar solvent and surfactant by centrifugation followed by rinsing
with a 1:1 mixture of methanol and chloroform. The different iron
oxalate particles were subjected to a careful thermal decomposition
X-ray diffraction (XRD) patterns were recorded on a Siemens
57
D5000, using Cu KR radiation and a graphite monochromator. Fe
M o¨ ssbauer spectra were recorded under the classical constant
acceleration working mode using a conventional spectrometer. The
source was 57Co(Rh). The velocity scale was calibrated using the
sextet of a high-purity iron foil. All spectra were recorded at room
temperature. Spectra were fitted to Lorentzian profiles by the least-
squares method. Scanning electron microscopy (SEM) and trans-
mission electron microscopy (TEM) images were obtained in a
2 4
at 200 °C under vacuum to yield anhydrous FeC O nanoparticles.
Inorganic Chemistry, Vol. 47, No. 22, 2008 10367

Arag o´ n et al.
a
Table 1. Fitting Results of the M o¨ ssbauer Spectra in Figure 2
sample
IS
QS
LW
%
X2
commercial
1.2011
1.7376
0.2652
991
11
952
65
1001
913
97
633
367
1009
0.683
0
.092
1.1942
.223
micellar
1.7276
0.485
1.631
1.6218
0.724
0.873
0.686
0.943
0.2575
0.269
0.3867
0.41410
0.326
0.514
0.51
0.550
0
commercial 200 °C
micellar 200 °C
1.1827
1.1814
0.935
0.618
0
.263
0.223
.004
0.292
0
3
.0 V
0.546
0
.0 V
a
0.885
0.571
IS, Isomer shift; QS, quadrupole splitting; LW, line width in mm/s.
JEOL JSM63000 microscope and JEOL 200CX microscope,
respectively. Thermogravimetric (TG) and differential thermal
analysis (DTA) curves were obtained in a SHIMADZU DTG-60AH
instrument. N adsorption isotherms at 77 K were obtained with a
2
Quantachrome instrument.
For electrochemical tests, two-electrode Swagelok-type lithium
test cells were used. The electrodes were prepared by blending the
powdered active material (60%) with carbon black (30%) and poly
vinylidene fluoride (10%) dissolved in N-methyl-pyrrolidone. The
slurry was cast onto a Cu foil and vaccuum-dried at 120 °C. As
the negative electrode, a lithium foil was used. The electrolyte was
a solution of ethylene carbonate-diethyl carbonate in 1:1 weight
proportion, including 1 M LiPF
disk. The cells were assembled in an Ar-filled glovebox (H
1 ppm). Galvanostatic charge/discharge cycles were carried out
6
, supported by a porous glass-paper
2
O, O
2
<
-1
-1
at 2C and 5C rates (C ) 1 Li h mol ).
Results and Discussion
According to early literature, FeC
O
2 4
2
·2H O crystallizes
in two allotropic forms, R monoclinic C2/c and ꢀ orthor-
1
0,11
hombic Cccm.
The two samples studied in this work
showed significant differences in their XRD data (Figure 1).
Commercial iron oxalate showed a complex XRD pattern
that could be interpreted in terms of a high-purity R form,
with a ) 12.023
1
Å, b ) 5.5550
6
Å, c ) 9.913 Å and ꢀ )
1
11
1
28.563 °, in good agreement with Sledzinska et al. In
5
contrast, nanocrystalline synthetic oxalate showed the reflec-
Figure 3. TEM images of (a) micellar raw material, and (b-d) after thermal
decomposition at 200 °C under vacuum.
tions ascribable to the orthorhombic ꢀ phase with a )
1
2.290
5
Å, b ) 5.542
3
4
Å and c ) 15.460 Å, also in good
1
0
consistent with the structure description given to the mono-
clinic solid as infinite chains of the ferrous-oxalate ions linked
together by hydrogen bonds engaging the oxygen atoms from
the water molecules and the oxygen atoms belonging to the
agreement with literature values.
57Fe M o¨ ssbauer spectra (Figure 2) were congruent with
the data reported in the literature. For the monoclinic sample,
a single quadrupole split signal was mainly observed with
isomer shift IS ) 1.201
1
1
oxalate ions. The M o¨ ssbauer spectra of the orthorhombic,
micellar sample shows similar parameters for the main phase,
apart from a slightly larger impurity content (ca. 6%), with
1
mm/s and quadrupole splitting QS
2
+
)
1.737 mm/s (Table 1), which is characteristic of Fe
6
ions in an environment with electric field gradient. Because
of the existing dispersion of the M o¨ ssbauer data reported in
the literature for this compound, the data obtained in this
study agree with several previous reports but contrasts with
3
+
3
IS ) 0.22 mm/s closer to Fe resulting from unavoidable
oxidation during the synthesis in aqueous medium. The
absence of marked differences in quadrupole splitting
between the iron (II) oxalate dihydrate samples with a R- or
1
2-15
the results in others.
In any case, the results are
ꢀ
-structure is in agreement with an ionic structure. In oxalates
(
(
10) Deyrieux, R.; Peneloux, A. Bull. Soc. Chim. Fr. 1969, 8, 2675–2681.
11) Sledzinska, I.; Murasik, A.; Piotrowski, M. Physica 1986, 138B, 315–
of divalent metals, the covalency of the metal-oxygen bond
is dependent on the electronegativity of the metal. As
16
3
22.
(
12) Brady, P. R.; Duncan, J. F. J. Chem. Soc. 1964, 653–656.
compared with other divalent first-row transition metal ions,
(13) Gabal, M. A.; El-Bellihia, A. A.; Ata-Allah, S. S. Mater. Chem. Phys.
2+
Fe has a higher ionicity, and the main contribution to the
2
003, 81, 84–92.
14) Takashima, Y.; Tateishi, Y. Bull. Chem. Soc. Jpn. 1965, 38, 1688–
693.
15) Ravi, N.; Jagannathan, R. Hyperfine Interact. 1982, 12, 167–173.
(
(
1
(16) Devillers, M.; Ladri e` re, J.; Apers, D. Inorg. Chim. Acta 1987, 126,
71–77.
10368 Inorganic Chemistry, Vol. 47, No. 22, 2008

Mesoporous Iron Oxalate Nanoribbons
Figure 4. Cycling experiments at (a, c) 2C and (b, d) 5C for micellar (c, d) and commercial (a, b) dehydrated samples.
quadrupole splitting comes from the electronic configuration;
thus, the lattice contribution can be neglected.
The two oxalate samples crystallized with two water
molecules. Thermogravimetric curves were recorded to locate
the temperature at which the two water molecules are
surface iron atoms are affected by a larger electric field
1
6,17
gradient than inner atoms. The average IS and QS values
are collected in Table 1, which are also indicative of little
changes in the oxidation state of iron during the dehydration
reaction.
released from FeC
O
2 4
2
·2H O. The observed weight loss up
The electrochemical reaction of FeC
place with a first extended plateau (ca. 1200 mA h g at
C rate) at about 1 V versus Li (Figure 4) leading to an
2 4
O with lithium took
to 200 °C (20.1% for the commercial sample and 21.2% for
the micellar sample) agrees perfectly with the theoretical
value (20.02%). At temperatures larger than 200 °C, a second
weight loss effect is visible in the TG curves, which is
-1
5
X-ray amorphous product. As commonly found in conversion
1
electrodes, the initial irreversible capacity is large. Further
18
ascribable to oxalate decomposition to magnetite. Irrespec-
tive of the origin of the oxalate sample and the crystal-
lographic structure of the starting solid, the XRD patterns
of the dehydrated products (Figure 1b,d) showed a small
number of highly broadened reflections which agreed well
-1
cycles displayed up to 750 mA h g reversible capacities,
which doubles the theoretical capacity of graphite (372 mA
-1
h g ).
To gain further insight on the electrochemical reaction that
takes place in these electrode materials, the first step was to
examine the nature of the reduced products by an iron-selective
18
2 4
with a poorly crystalline solid with a FeC O composition.
The dehydrated product resulting from the commercial
sample consisted of particles of several micrometers with
heterogeneous particle size and shape. TEM micrographs of
the dehydrated micellar oxalate showed ultrafine particles
57
spectroscopy. The Fe M o¨ ssbauer spectroscopy study of used
electrodes was carried out by interrupting the cycling experi-
ments at selected voltages. The spectrum at 0 V (Figure 2e)
shows a complex profile in which both the average isomer shift
and QS approaches to that R-iron. However the value is above
(
Figure 3) with elongated shapes. Nanoribbons about 150
nm wide are observed for the sample obtained using reverse
micelles. Both samples dehydrated at 200 °C showed a
complex porous texture (see, for example, Figure 3d) with
pores probably created during water release. This observation
is consistent with the Brunauer-Emmett-Teller (BET)
0
mm/s, which agrees with a noncrystalline phase, observed
by XRD. Thus these results can be interpreted as the result of
iron reduction to the metallic state in the form of metal
nanoparticles in contact with Li C O . The iron particles show
2 2 4
superparamagnetic behavior. In an attempt to fit the observed
spectra to a simplified model, two doublet signals were obtained
surface values obtained from the N
(
2
adsorption isotherms
ca. 70 m g for both samples). The mesoporous material
is then obtained by a top-down approach, which has been
2
-1
20,21
(Table 1). As referred to recent reports on iron oxides,
these
1
9
recently found in conversion electrodes.
doublets could account for two different environments of partly
reduced iron: surface iron (with higher IS and QS) and bulk
metallic atoms (with hyperfine parameters closer to those of
R-Fe). The contribution of the former is expected to grow
markedly on decreasing particle size. The percentage of surface
atoms in regular octahedral nanoparticles could be close to 63%
for particles containing few hundreds of atoms.
The M o¨ ssbauer spectra of the dehydrated samples (Figure
b,d) evidence a highly broadened signal that could be fitted
2
to multiple quadrupole split doublets, indicative of multiple
closely related environments of iron atoms. This behavior
can be explained by assuming ultrafine particles in which
(
(
(
17) Aramu, F.; Maxia, V.; Muntoni, C. Lett. NuoVo Cimento 1975, 12,
25–227.
18) Brown, R. A.; Bevant, S. C. J. Inorg. Nucl. Chem. 1966, 28, 387–
91.
19) Hu, Y. S.; Guo, Y. G.; Sigle, W.; Hore, S.; Balaya, P.; Maier, J. Nat.
Mater. 2006, 5, 713–717.
2
(20) Alc a´ ntara, R.; Jaraba, M.; Lavela, P.; Tirado, J. L.; Jumas, J. C.;
Olivier-Fourcade, J. Electrochem. Commun. 2003, 5, 16–21.
(21) Bomio, M.; Lavela, P.; Tirado, J. L. ChemPhysChem 2007, 8, 1999–
2007.
3
Inorganic Chemistry, Vol. 47, No. 22, 2008 10369

Arag o´ n et al.
Figure 5. FTIR spectra of electrolyte, FeC2O4, Li2C2O4, and charged and
discharged electrodes.
Figure 6. Cycling experiments at (a) 5C and (b) 2C for micellar (triangles)
and commercial (circles) dehydrated samples. Disharge (empty symbols)
and charge (filled) capacities are shown.
After cell discharge to 0 V followed by a charge process
to 3 V, the MS of the pristine sample is not recovered (Figure
2
f). Instead, a highly broadened spectrum resulting from
amorphous. Second, the MS spectrum (Figure 2f and Table
multiple doublets is visible, with average IS and QS values
1) shows a doublet signal with an IS value (0.29 mm/s) which
3
+
close those expected for Fe ions. The ultrafine nature of
the oxidized domains prevents again a ferromagnetic order-
ing. The origin of the complex signal with different QS but
differing less that in the case of the metal nanoparticles may
be a consequence of a different spin state of the atoms in
the structure of the oxalate nanoparticles.
is intermediate between that of Fe(0) and that of Fe(III).
Fe(II) is well-known to occur in the 1.00-1.50 mm/s IS
2
2,14
range in oxides and carboxylates.
The QS value also
differs from the original oxalate (Figure 2b,d). In conse-
quence, the presence of Fe(II) compounds such as FeO or
FeC O can be discarded in the cycled electrode. Finally,
2 4
Additional experiments were performed to check the
integrity of oxalate ions in the used electrodes. FTIR data
the FTIR data obtained in transmission mode (Figure 5)
support the presence of oxalate counterions in a partly
oxidized amorphous product different from the conventional
oxide products reported by other authors in conversion oxide
electrodes. The initial oxalate ions are not decomposed during
the charge-discharge processes, as shown by the absence
of gaseous products that could have caused overpressure and/
or breakage in the test cells.
(
Figure 5) confirms that oxalate ions are preserved on
-
1
cycling. Particularly, the band at about 1320 cm , which is
characteristic of both FeC and Li , is absent in the
O
2 4
2 2 4
C O
electrolyte and visible in charged and discharged electrodes.
Thus, it can be concluded that the system is significantly
different from conventional conversion oxide electrodes, as
the matrix formed during the discharge is lithium oxalate
instead of lithium oxide.
The electrochemical behavior in lithium test cells of
FeC O
2 4
2
· 2H O polycrystalline samples in both allotropic
2 4
To deepen in the possible formation of FeC O or FeO
forms was poor. In contrast, the cyclability of both dehy-
drated oxysalts was significantly improved. The results for
both FeC O samples are compared in Figure 6. Even cycling
2 4
on charge, the information obtained in this study by looking
at the recharged electrodes, XRD, MS, and FTIR, is
summarized as follows. The results unequivocally show that
the original iron(II) oxalate is not recovered after charge and
that no iron(II) oxide is formed. First, the product was X-ray
(
22) Smith, M. G.; Goodenough, J. B. J. Solid State Chem. 1993, 103,
25–29.
10370 Inorganic Chemistry, Vol. 47, No. 22, 2008

Mesoporous Iron Oxalate Nanoribbons
at higher rate, iron oxalate shows higher capacity than the
synthesis of these materials makes them an inexpensive
option for this purpose.
9
previously reported manganese carbonate. Capacities are
-
1
always above 600 mA h g during 75 cycles when iron
oxalate is cycled at 2C rate.
Acknowledgment. The authors are indebted to MEC
In conclusion, first-row transition metal oxysalts are in-
teresting candidates for the active material of the negative
electrode of lithium-ion batteries. The low temperature
(MAT2005-00374) and Junta de Andaluc ´ı a (FQM288).
IC8008927
Inorganic Chemistry, Vol. 47, No. 22, 2008 10371