Cation-substituted LiFePO4 prepared from the FeSO4·7H2O waste slag as a potential Li battery cathode material

Article abstract of DOI:10.1016/j.jallcom.2010.03.026

The purpose of this study is to utilize the huge FeSO₄·7H₂O waste slag produced by the titanium dioxide industry. FeC₂O₄·2H₂O precursors are synthesized at various pH values by using the waste slag and H₂C₂O₄·2H₂O as raw materials, and without any purifying process. ICP analysis confirms that the impurity content of FeC₂O₄·2H₂O increases with the pH value. Crystalline cation-substituted LiFePO₄ are prepared from the FeC₂O₄·2H₂O precursors. The cation dopants do not obviously change the structure of LiFePO₄, and all the samples are single olivine-type phase and well crystallized. The lattice parameters of LiFePO₄ decrease with the increased dopants contents. The dopants limit the size of LiFePO₄ nanocrystals, LiFePO₄ particles agglomeration and, consequently, improve the electrochemical performance of LiFePO₄. The cation-substituted LiFePO₄ prepared from the waste slag show much better electrochemical properties than the pure LiFePO₄ at high current rates. The optimal pH value for synthesizing FeC₂O₄·2H₂O from the waste slag is about 1.0, with 96.6% iron recovery. The cation-substituted LiFePO₄ prepared from this precursor exhibits the best electrochemical properties, which delivers a capacity of 152, 142 and 126 mAh g^-1 at 1C, 2C and 5C rate, respectively, and shows excellent cycling performance.

Full text of DOI:10.1016/j.jallcom.2010.03.026

Journal of Alloys and Compounds 497 (2010) 278–284
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
Cation-substituted LiFePO4 prepared from the FeSO4·7H2O waste slag as a
potential Li battery cathode material
Ling Wu, Zhixing Wang∗, Xinhai Li, Huajun Guo, Lingjun Li, Xiaojuan Wang, Junchao Zheng
School of Metallurgical Science and Engineering, Central South University, Changsha 410083, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
The purpose of this study is to utilize the huge FeSO4·7H2O waste slag produced by the titanium dioxide
industry. FeC2O4·2H2O precursors are synthesized at various pH values by using the waste slag and
H2C2O4·2H2O as raw materials, and without any purifying process. ICP analysis confirms that the impurity
content of FeC2O4·2H2O increases with the pH value. Crystalline cation-substituted LiFePO4 are prepared
from the FeC2O4·2H2O precursors. The cation dopants do not obviously change the structure of LiFePO4,
and all the samples are single olivine-type phase and well crystallized. The lattice parameters of LiFePO4
decrease with the increased dopants contents. The dopants limit the size of LiFePO4 nanocrystals, LiFePO4
particles agglomeration and, consequently, improve the electrochemical performance of LiFePO4. The
cation-substituted LiFePO4 prepared from the waste slag show much better electrochemical properties
than the pure LiFePO4 at high current rates. The optimal pH value for synthesizing FeC2O4·2H2O from
the waste slag is about 1.0, with 96.6% iron recovery. The cation-substituted LiFePO4 prepared from
this precursor exhibits the best electrochemical properties, which delivers a capacity of 152, 142 and
126 mAh g−1 at 1C, 2C and 5C rate, respectively, and shows excellent cycling performance.
© 2010 Elsevier B.V. All rights reserved.
Received 29 January 2010
Received in revised form 28 February 2010
Accepted 2 March 2010
Available online 6 March 2010
Keywords:
Cathode material
LiFePO4
Titanium dioxide
Waste slag
FeSO4·7H2O
1. Introduction
The common methods for utilization of the waste slag are
preparing iron oxide series productions, such as magnetic iron
Titanium dioxide is widely used in the manufacture of paints,
paper, varnishes, lacquer, printing inks, rubber, ceramics, food, etc.
Titanium dioxide pigment is commercially manufactured by two
main processes, namely the sulfate and the chloride processes. The
sulfate process still produces about 50% of global production [1].
However, large quantity of FeSO4·7H2O waste slag is produced in
the sulfate route, for example, to produce 1 tonne pigment from
ilmenite (containing TiO2 ∼48 wt.%, FeO ∼33 wt.%, Fe2O3 ∼6 wt.%
and MgO ∼6 wt.%), about 3 tonnes FeSO4·7H2O waste slag would be
produced [2]. In China, about 0.6 million tonnes of TiO2 pigment is
manufactured each year, of which more than 90% are operated with
the sulfate route [3]. According to this, it can be estimated that more
than 1.62 million tonnes FeSO4·7H2O waste slag would be produced
each year. Unfortunately, this waste slag is less marketable and
difficult to be utilized because of its high impurity content (a typical
chemical composition is shown in Table 1), which causes not only
severe environmental problems but also the waste of iron resource.
Therefore, the urgent need for proper utilization of waste slag has
attracted great attention of the world.
oxide [4], iron oxide red [5], iron oxide yellow [6], iron oxide black
[7]. However, these methods are usually complicated and consist of
many purifying processes, due to various impurities (i.e., Mg, Mn,
Ca, Al, Ti, etc.) which are harmful to the performance of the mate-
rials. Apart from these methods, little work has been performed to
utilize the FeSO4·7H2O waste slag. Therefore, a simple and effective
way to utilize the waste slag should be researched urgently.
In the present study, we propose a novel, simple, low cost and
effective method to utilize the waste slag. In this method, oxalic
acid was selected as a precipitator and added to the solution of
FeSO4·7H2O waste slag. By controlling the pH value, metallic ions
were precipitated selectively due to the various solubility prod-
ucts of metallic oxalates, then FeC2O4·2H2O with small amounts of
metallic impurities (i.e., Mg, Mn, Ca and Ti) was obtained. Finally,
cation-substituted LiFePO4 was prepared with the as-synthesized
FeC2O4·2H2O as raw material. The whole route requires no further
purification, owing to some impurities (Mg, Mn and Ti, for example)
could benefit the electrochemical performance of LiFePO4 [8–15].
2. Experimental
2.1. Materials and preparation
FeC2O4·2H2O was synthesized by a co-precipitation method with the following
procedures: (1) FeSO4·7H2O waste slag (chemical composition is shown in Table 1)
was dissolved in de-ionized water to obtain 0.25 M (Fe) solution, and the solution
∗
Corresponding author. Tel.: +86 731 88836633; fax: +86 731 88836633.
E-mail address: wuling19840404@163.com (Z. Wang).
0925-8388/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2010.03.026

L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
279
Table 1
perature. Electrochemical impedance spectroscopy was carried out with a CHI660
electrochemical analyzer. The impedance spectra were measured at the stable volt-
age of 2.5 V, and were recorded by applying an AC voltage of 5 mV amplitude in the
1 MHz–0.01 Hz frequency range.
A typical chemical composition (wt.%) of FeSO4·7H2O waste slag.
FeSO4·7H2O
88.52
MgSO4·7H2O
6.04
MnSO4·5H2O
0.35
Al2(SO4)3·18H2O
0.28
CaSO4·2H2O
TiOSO4
0.52
Water insoluble
3.83
Others
0.28
0.18
3. Results and discussion
3.1. FeC2O4·2H2O precursors identification
Table 2
The initial metallic ions concentration ([Mn+]0, mol L−1), solubility-product constant
Ksp and theoretic initial precipitation pH values of the corresponding poorly soluble
metallic oxalate and metallic hydroxide.
The initial metallic ions concentration, solubility-product con-
stant Ksp and theoretic initial precipitation pH of the corresponding
poorly soluble metallic oxalate and metallic hydroxide are shown
in Table 2. The initial precipitation pH values were calculated by
using the Eqs. (1) and (2), respectively, according to the solubility
product principle.
Precipitate
Ksp
[Mn+
]
0
Initial precipitation pH
FeC2O4·2H2O
MgC2O4·2H2O
MnC2O4·2H2O
CaC2O4·H2O
Fe(OH)2
3.2 × 10−7 [16]
4.8 × 10−6 [17]
1.7 × 10−7 [17]
4.0 × 10−9 [16]
8.0 × 10−16 [16]
6.0 × 10−10 [16]
1.9 × 10−13 [16]
1.3 × 10−33[16]
5.5 × 10−6 [16]
1.0 × 10−29 [16]
0.25
0.042
1.187
1.083
0.337
7.979
7.763
7.288
2.014
8.244
2.905
0.0192
0.0011
0.0008
0.25
0.0192
0.0011
0.0007
0.0008
0.0026
ꢀ
Ka1Ka2[H2C2O4]0
pH = −lg
(1)
(2)
Ksp/[M2+
]
0
Mg(OH)2 [fresh]
Mn(OH)2
ꢀ
Al(OH)3 [Al3+, 3OH−
Ca(OH)2
]
lg Ksp
n
pH = 14 −
[Mn+
]
0
TiO(OH)2
where M is the metal elements, n is the valent of M, Ka1 and Ka2
are the first ionization constant and second ionization constant
was filtered to separate the water-insoluble; (2) H2C2O4·2H2O was added to the
solution under vigorous stirring, a yellow color precipitate formed immediately;
(3) then NH3·H2O (2 M) was dropped slowly into the solution to control the pH at
different values (i.e., 0.3, 0.7, 1.0 and 1.3, respectively). (4) After being stirred for
30 min, the precipitates were filtered, washed three times with de-ionized water
and dried at 80 ◦C for 12 h in an oven, and labeled as A0, B0, C0 and D0, according to
the different pH values of 0.3, 0.7, 1.0 and 1.3, respectively.
Cation-substituted LiFePO4 was synthesized by a mechanical activation process
followed by high-temperature calcination. The as-prepared FeC2O4·2H2O, Li2CO3
and NH4H2PO4 (all chemicals of .99% purity) were weighed in the stoichiometric of
LiFePO4. The ingredients were dispersed in ethanol and ground for 4 h by high energy
ball milling (200 rpm) at room temperature. The as-obtained slurries were dried
in an oven at 80 ◦C. Finally the precursors were calcinated in a tubular furnace at
650 ◦C for 12 h with flowing argon (99.999%) and cooled to room temperature. Thus,
cation-substituted LiFePO4 were obtained and labeled as A, B, C and D according to
the different FeC2O4·2H2O of A0, B0, C0 and D0, respectively. For comparison, a pure
LiFePO4 sample (labeled as P) was prepared from analytically pure FeC2O4·2H2O
with the same processes.
of oxalic acid, respectively, [Mn+
] is the initial metallic ions con-
0
centration, and [H2C2O4]0 is the initial oxalic acid concentration
(0.28 mol L−1).
In order to precipitate the metallic ions selectively, we chose the
pH values of 0.3, 0.7, 1.0 and 1.3 to prepare the FeC2O4·2H2O precur-
sors. Table 3 shows the molar ratio of the metal elements in waste
slag and FeC2O4·2H2O, and the iron recovery at various pH values.
According to the theoretic calculations (Table 2), Mg, Mn, Ca, Al and
Ti will not be deposited until the pH reaches 1.187, 1.083, 0.337,
2.014 and 2.905, respectively. However, it can be seen (Table 3)
that small amounts of Mg, Mn and Ti are detected below their the-
oretic precipitation pH values. For example, Mg are detected in all
the samples, nevertheless, the Mg/Fe values of the samples synthe-
sized at pH ≤ 1.0 (sample A0 0.0020/1, B0 0.0022/1 and C0 0.0023/1)
are not affected obviously by the pH values, and they are much
lower than the value of the sample synthesized at pH = 1.3 (sample
D0 0.0082/1), which indicates that the magnesium detected in sam-
ple A0, B0 and C0 should be owing to magnesium ions adsorbed on
the FeC2O4·2H2O particles and difficult to be washed out. Likewise,
a small amount of Ti in sample A0, B0, C0 and D0 (Ti/Fe 0.0018/1,
0.0020/1, 0.0021/1 and 0.0021/1, respectively) is due to the adsorp-
tion of titanium ions. However, Mn is detected in sample C0, which
attributes to the pH (=1.0) is very close to the theoretic initial pre-
cipitation pH (=1.083).
2.2. Characterization
The metal content of samples was analyzed using inductively coupled plasma
emission spectroscopy (ICP, IRIS intrepid XSP, Thermo Electron Corporation). The
SEM image and elemental mapping of the particles were observed with scanning
electron microscopy (SEM, JEOL, JSM-5600LV). The powder X-ray diffraction (XRD,
Rint-2000, Rigaku) using CuK␣ radiation was employed to identify the crystalline
phase of the synthesized materials. XRD Rietveld refinement was performed by
FULLPROF.
2.3. Battery preparation and measurement
It is obvious that the impurity content of FeC2O4·2H2O increases
with the pH increasing from 0.3 to 1.3, corresponding, the iron
recovery increases. The recovery rate of iron is 63.5%, 93.4%, 96.6%
and 98.6% at pH of 0.3, 0.7, 1.0 and 1.3, respectively. The iron recov-
ery of sample A0 is unsatisfactory though it contains the lowest
impurity. When the pH is higher than 1.0, its contribution to the
increase of iron recovery is limited, and the impurity content will
be increased rapidly.
The electrochemical performance was performed using a two-electrode coin-
type cell (CR2025) of Li|LiPF6 (EC:EMC:DMC = 1:1:1 in volume) |LiFePO4. The
working cathode is composed of 80 wt.% LiFePO4 powders, 10 wt.% acetylene black
as conducting agent, and 10 wt.% poly (vinylidene fluoride) as binder. After being
blended in N-methyl pyrrolidinone, the mixed slurry was spread uniformly on a
thin aluminum foil and dried in vacuum for 12 h at 120 ◦C. A metal lithium foil
was used as anode. Electrodes were punched in the form of 14 mm diameter disks,
and the typical positive electrode loadings were in the range of 1.95–2 mg cm−2. A
polypropylene micro-porous film was used as the separator. The assembly of the
cells was carried out in a dry argon-filled glove box. The cells were charged and
discharged over a voltage range of 2.5–4.1 V versus Li/Li+ electrode at room tem-
Fig. 1 shows the XRD patterns of FeC2O4·2H2O synthesized at
different pH values. As shown, all the samples are single phase
Table 3
The molar ratio of Fe, Mg, Mn, Ca, Al and Ti in FeSO4·7H2O waste slag and FeC2O4·2H2O, and the iron recovery at various pH values.
Sample
Fe
Mg
Mn
Ca
Al
Ti
Fe recovery (%)
Waste slag
1
1
1
1
1
0.1122
0.0020
0.0022
0.0023
0.0082
0.0046
0.0033
∼0
0.0004
0.0009
0.0021
0.0026
0.0102
0.0018
0.0020
0.0021
0.0021
–
A0 (pH = 0.3)
B0 (pH = 0.7)
C0 (pH = 1.0)
D0 (pH = 1.3)
∼0
∼0
∼0
∼0
∼0
∼0
63.5
93.4
96.6
98.6
0.0012
0.0036

280
L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
Fig. 1. XRD patterns of FeC2O4·2H2O which are labeled as A0, B0, C0 and D0, accord-
ing to different pH values 0.3, 0.7, 1.0 and 1.3, respectively.
of iron oxalate hydrate (FeC2O4·2H2O), which can be indexed to
an orthorhombic system with the space group Cccm (JCPDS no.
22-0635). Additionally, no impurity phases are detected, which
demonstrates that the adsorbed impurities are amorphous.
3.2. LiFePO4 identification
The XRD patterns of LiFePO4 samples are shown in Fig. 2(a).
As shown, the diffraction lines of all samples are indexed to an
orthorhombic crystal structure (triphylite, space group Pnma),
and no impurity phases are detected. The results indicate that
the metallic dopants (Mg2+, Mn2+, Ti4+ and Ca2+) do not obvi-
ously change the structural characteristics of LiFePO4 during heat
treatment. However, the XRD peaks intensities decrease with the
increased dopants contents, which reveals that the metallic cations
has entered the lattices of LiFePO4 crystal. Previous studies have
shown that metallic ions could be doped not only in Li-site but also
in Fe-site [8–15], therefore, Mg2+, Mn2+, Ti4+ and Ca2+ ions could
be doped in Li-site, Fe-site or both. But in either case, it will induce
the formation of cation-deficient solid solution, which can benefit
the electrochemical performance of LiFePO4.
Fig. 2. XRD patterns of LiFePO4 samples (a) and magnified (1 3 1) peaks of the pat-
terns (b), where P, A, B, C and D are synthesized from pure FeC2O4·2H2O, precursor
A0, B0, C0 and D0, respectively.
However, as a result of the cations co-doping, the lattice parameters
decreased.
Fig. 2(a) also reveals that the diffraction peaks shift to higher
degrees with the doping amount of cations. For clear observation,
the peak positions of (1 3 1) planes are magnified and shown in
Fig. 2(b). The XRD refinement according to the Rietveld method
indicates that the lattice parameters a, b and c decrease with the
increased impurities contents, and the results are shown in Table 4.
As we all know, Mg2+ (0.72 Å), Mn2+ (0.67 Å) and Ti4+ (0.61 Å) ions
have smaller ionic radii than both Li+ (0.76 Å) and Fe2+ (0.78 Å) ions,
while Ca2+ (1.00 Å) ion has a larger ionic radius. It can be expected
that Mg2+, Mn2+ and Ti4+ doping could reduce the lattice constants
of LiFePO4, while Ca2+ doping could expand the lattice constants.
The crystallite size, d, was calculated from the X-ray line width
using the Scherrer formula, d = 0.9ꢀ/ˇ1/2 cos ꢁ, where ꢀ is the X-
ray wavelength, ˇ1/2 is the corrected width of the main diffraction
peak (1 3 1) at half height, and ꢁ is the diffraction angle. As shown in
Table 4, the d value decrease with the dopants contents, indicating
that the dopants limit the size of LiFePO4 nanocrystals.
Fig. 3 shows the SEM images of LiFePO4 samples. It is obvi-
ous that the cation-substituted samples own less agglomeration
and scatter more uniform than the pure sample. This demonstrates
that the metallic dopants inhibit particles to aggregate during cal-
cinations, which is conducive to shorten the lithium-ion diffusion
distance. The elements distributions of cation-substituted LiFePO4
were measured by EDS. As a typical case, the EDS maps of Fe, Mg,
Ca, Mn and Ti for sample C are shown in Fig. 4. As shown, the
distribution areas of all the elements are homogeneous, owing to
the co-precipitation, which results in the atom-scale mixed of var-
ious elements. Furthermore, Al is not detected by EDS, which is
consistent with the result obtained from ICP analysis.
Table 4
Lattice parameters and crystallite size of LiFePO4 samples.
Sample The lattice parameters
Crystallite size
d1 3 1 (nm)
a (Å)
b (Å)
c (Å)
V (Å3)
P
10.3201(3)
10.3135(4)
10.3133(5)
10.3130(9)
10.3122(7)
6.0091(5)
6.0046(8)
6.0024(7)
6.0010(5)
6.0003(6)
4.7018(4)
4.6975(6)
4.6953(9)
4.6934(8)
4.6928(6)
291.580(6)
290.909(7)
290.660(8)
290.467(9)
290.373(7)
44.1
42.9
42.0
41.4
40.2
Fig. 5 shows the AC impedance spectra of the LiFePO4 electrodes.
The impedance spectra are fitted using an equivalent circuit. In
the equivalent circuit, Rs and Rct represent the solution resistance
and charge-transfer resistance, respectively. CPE is related to the
A
B
C
D

L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
281
Fig. 3. SEM images of LiFePO4 samples.
double layer capacitance and passivation film capacitance. W rep-
resents the Warburg impedance. The parameters of the equivalent
circuit are summarized in Table 5. The plot of the real axis Zre versus
the reciprocal square root of the lower angular frequencies ω−0.5
is illustrated in Fig. 6. The straight lines are attributed to the dif-
fusion of the lithium ions into the bulk of the electrode materials,
the so-called Warburg diffusion. This relation is governed by Eq.
(3) [18,19]. According to Fig. 6 and Eq. (3), the slopes of the straight
lines represent the values of Warburg impedance coefficient (ꢂw).
The diffusion coefficient (D) of the lithium ions diffusing into the
bulk electrode materials are calculated using Eq. (4) and recorded
in Table 5.
Zre = Rs + Rct + ꢂw · ω−0.5
(3)
(4)
ꢁ
ꢂ
2
RT
AF2ꢂwC
D = 0.5
Table 5
Impedance parameters of the LiFePO4 electrodes.
Sample
Rs (ꢃ)
Rct (ꢃ)
1228
233
157
146
434
ꢂw (ꢃ cm2/s0.5
)
D (cm2/s)
i0 (mA/cm2)
P
5.1
9.4
4.9
4.1
5.5
204
181
143
110
175
3.59 × 10−13
4.56 × 10−13
7.31 × 10−13
1.23 × 10−12
4.88 × 10−13
2.09 × 10−5
1.10 × 10−4
1.64 × 10−4
1.76 × 10−4
5.92 × 10−5
A
B
C
D

282
L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
Fig. 4. EDS maps of Fe, Mg, Ca, Mn and Ti for sample C.
where Rct, charge transfer resistance; Rs, solution resistance; ω,
angular frequency in the low frequency region; D, lithium-ion dif-
fusion coefficient; R, the gas constant; T, the absolute temperature;
F, Faraday’s constant; A, the area of the electrode surface; and C,
molar concentration of Li+ ions.
As shown in Table 5, the cation-substituted LiFePO4 exhibit
much lower charge-transfer resistance than that of pure LiFePO4,
indicating that cations doping can significantly increase the electri-
cal conductivity of LiFePO4. The exchange current density (i0) was
calculated by the formula, i0 = RT/nFRct. Cation-substituted sam-
ples show higher exchange current densities than the pure sample,
and sample C shows the highest value. Also, the cation-substituted
LiFePO4 electrodes have better ionic conductivity than the pure
LiFePO4 electrode. It is reported that the lithium ions diffusivity
increase with the increased parameters b and c [9,20], thus, the
pure LiFePO4 should possess the highest ionic conductivity. But
the results show that the pure sample has the lowest value, which
should be ascribed to the cation-substituted samples have smaller
crystallite sizes and less particle agglomerations than the pure sam-
ple. Furthermore, sample D exhibits lower ionic conductivity than
sample C. That may be owing to sample D has the lowest parameters
b and c, namely has the narrowest tunnel for Li+ ion across [9,20].
Sample C shows the highest electronic conductivity and ionic con-
ductivity, indicating that an appropriate amount of cation dopant
is important.
Fig. 5. AC impedance spectra of LiFePO4 electrodes at the voltage of 2.5 V.
Fig. 7 displays the initial charge/discharge curves of LiFePO4
samples at different C-rates. All the samples exhibit very long and
plat plateau around 3.4 V at low current rates (0.1C and 0.5C). The
sample P, A, B, C and D show a discharge capacity of 167, 165, 164,
164 and 162 mAh g−1 at 0.1C rate, respectively, which approach the
theoretical capacity of 170 mAh g−1. By increasing the current rate,
the utilization percentage of the active material decreased along
with the polarization of electrodes increased. The pure LiFePO4
electrode exhibits serious polarization at 1C, 2C and 5C rates, and
delivers a capacity of 136, 112 and 67 mAh g−1 at the corresponding
C-rate. While the cation-substituted LiFePO4 show smaller polar-
ization and much better electrochemical properties. Sample A, B,
C and D exhibit a capacity of 149, 150, 152 and 148 mAh g−1 at 1C
rate, 139, 140, 142 and 137 mAh g−1 at 2C rate, and 123, 124, 126
and 122 mAh g−1 at 5C rate, respectively.
Cyclic stability of LiFePO4 samples at room temperature is
shown in Fig. 8. As shown, the capacities waved with the room tem-
Fig. 6. The relationship between Zre and ω−1/2 at low frequencies.

L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
283
Fig. 7. The initial charge/discharge curves of LiFePO4 samples with different C-rates at room temperature.
perature. All the samples exhibit excellent cycling performance at
low current rates (0.1C and 0.5C), and without capacity fading after
10 cycles. But at higher current rates, cation-substituted samples
show much better cyclic stability than the pure LiFePO4. After 50
cycles, pure LiFePO4 only retains 90.4%, 85.7% and 83.6% of its ini-
tial discharge capacity at 1C, 2C and 5C rate, respectively. While
sample A, B, C and D maintain 100%, 100.3%, 99.5% and 98.0% of its
initial discharge capacity at 1C rate, 98.6%, 100%, 99.3% and 94.4%
of its initial discharge capacity at 2C rate, and 98.4%, 99.2%, 100.1%
and 92.6% of its initial discharge capacity at 5C rate, respectively.
Cation-substituted LiFePO4 exhibiting much better electrochemi-
cal properties can be attributed to: (1) cation-doping effectively
restrains the LiFePO4 crystals growth and inhibits the particles
aggregation and, consequently, enhances the lithium-ion diffusion
speed cross the LiFePO4/FePO4 interface; (2) cation-doping reduces
the charge transfer resistance (Rct), and thus increases the electri-

284
L. Wu et al. / Journal of Alloys and Compounds 497 (2010) 278–284
pH value for synthesizing FeC2O4·2H2O is about 1.0.In particular,
it must be pointed out that this route is generally practical in this
field, owing to the FeSO4·7H2O waste slag produced by the tita-
nium dioxide industry possesses relatively stable impurity content.
Moreover, this method can also be used for utilizing those waste
slags which have lower or a little higher impurity content than
the slag discussed in this paper. Based on these results, it can be
concluded that this method is a simple, efficient, economical and
environment-friendly way for both FeSO4·7H2O waste slag utiliza-
tion and LiFePO4 preparation.
Acknowledgement
The project was sponsored by National Basic Research Program
of China (973 Program, 2007CB613607).
References
[1] B. Liang, C. Li, C. Zhang, Y. Zhang, Hydrometallurgy 76 (2005) 173.
[2] M. Xu, M. Guo, J. Zhang, T. Wan, L. Kong, J. Iron Steel Res. Int. 13 (2006) 6.
[3] C. Li, B. Liang, L. Guo, Hydrometallurgy 89 (2007) 1.
[4] J.T. Nong, H. Huang, S.Q. Wei, Chin. Patent 200610018642.X (2007).
[5] C.Q. Cai, Z.Y. Zeng, Q.L. Xu, Chin. Patent 02148428.7 (2008).
[6] C.Q. Cai, Z.Y. Zeng, Q.L. Xu, Chin. Patent 02148429.5 (2008).
[7] P.F. Huang, Chin. Patent 200710130428.8 (2007).
Fig. 8. Cycling performances of LiFePO4 samples at different C-rates. The cells cycled
10 times at 0.1C and 0.5C, respectively, and then cycled 50 times at 1C, 2C and 5C in
turn.
cal conductivity of LiFePO4. Furthermore, sample C shows the most
excellent electrochemical properties, due to the highest electronic
conductivity and ionic conductivity.
[8] J.F. Ni, H.H. Zhou, J.T. Chen, X.X. Zhang, Mater. Lett. 59 (2005) 2361.
[9] S.-H. Wu, M.-S. Chen, C.-J. Chien, Y.-P. Fu, J. Power Sources 189 (2009) 440.
[10] G.X. Wang, S. Bewlay, S.A. Needham, H.K. Liu, R.S. Liu, V.A. Drozd, J.-F. Lee, J.M.
Chen, J. Electrochem. Soc. 153 (2006) A25.
[11] L. Li, X. Li, Z. Wang, L. Wu, J. Zheng, H. Guo, J. Phys. Chem. Solids 70 (2009) 238.
[12] S.Y. Chung, J.T. Bloking, Y.M. Chiang, Nat. Mater. 1 (2002) 123.
[13] N. Meethong, Y.H. Kao, S.A. Speakman, Y.M. Chiang, Adv. Funct. Mater. 19 (2009)
1060.
[14] J. Yao, S. Bewlay, K. Konstantionv, V.A. Drozd, R.S. Liu, X.L. Wang, H.K. Liu, G.X.
Wang, J. Alloys Compd. 425 (2006) 362.
[15] T. Nakamura, K. Sakumoto, M. Okamoto, S. Seki, Y. Kobayashi, T. Takeuchi, M.
Tabuchi, Y. Yamada, J. Power Sources 174 (2007) 435.
[16] J.A. Dean (Ed.), Lange’s Handbook of Chemistry, 13th ed., McGraw-Hill, New
York, 1985.
[17] D.R. Lide, Handbook of Chemistry and Physics, 78th ed., CRC Press, Boca Raton,
New York, 1997–1998.
[18] A.Y. Shenouda, K.R. Murali, J. Power Sources 176 (2008) 332.
[19] A.Y. Shenouda, H.K. Liu, J. Alloys Compd. 477 (2009) 498.
[20] C. Ouyang, S. Shi, Z. Wang, X. Huang, L. Chen, Phys. Rev. B 69 (2004) 104303.
4. Conclusions
FeC2O4·2H2O precursors are prepared from the FeSO4·7H2O
waste slag, without any purifying process. ICP results show that
a small amount of metallic ions precipitated into the precursors.
The impurity content of FeC2O4·2H2O and the iron recovery are
greatly affected by the pH value. Cation-substituted LiFePO4 sam-
ples are synthesized with the as-prepared FeC2O4·2H2O, and all the
cation-substituted samples show excellent electrochemical prop-
erties. However, considering the best electrochemical performance
of LiFePO4 and the satisfactory recovery rate of iron, the optimal