SnHPO4: A promising precursor for active material as negative electrode in Li-ion cells

Article abstract of DOI:10.1016/S0013-4686(01)00735-6

SnHPO₄ was synthesized by precipitation and characterized by X-ray diffraction. The discharge curve of electrochemical cells using SnHPO₄ as active cathode material and metallic lithium as anode showed a first reduction step at ca. 1.5 V. X-ray diffraction patterns of lithiated samples showed very weak bands ascribable to metallic tin. For the voltage values lower than 0.8 V the well-known Li-Sn alloy formation takes place. An intermediate step appears in the range from 1.2 to 0.8 V. The capacity retention is stabilized after 10 cycles at ca. 200 mAh/g.

Full text of DOI:10.1016/S0013-4686(01)00735-6

Electrochimica Acta 47 (2001) 489–493
www.elsevier.com/locate/electacta
SnHPO4: a promising precursor for active material as negative
electrode in Li-ion cells
M.L. Elidrissi Moubtassim, J.I. Corredor, J.L. Tirado, C. Pe´rez Vicente *
Laboratorio de Qu´ımica Inorga´nica, Uni6ersidad de Co´rdoba, Campus de Rabanales, Edificio C-3, 14071 Cordoba, Spain
Received 14 May 2001; received in revised form 17 July 2001
Abstract
SnHPO4 was synthesized by precipitation and characterized by X-ray diffraction. The discharge curve of electrochemical cells
using SnHPO4 as active cathode material and metallic lithium as anode showed a first reduction step at ca. 1.5 V. X-ray
diffraction patterns of lithiated samples showed very weak bands ascribable to metallic tin. For the voltage values lower than 0.8
V the well-known Li–Sn alloy formation takes place. An intermediate step appears in the range from 1.2 to 0.8 V. The capacity
retention is stabilized after 10 cycles at ca. 200 mA h/g. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Tin phosphate; Electrode material; Lithium cell
2. Experimental
1. Introduction
The compound SnHPO4 was obtained by a precipita-
tion procedure. H3PO4 (20 ml, 85%) was added to a
At present tin-based compounds are intensively stud-
ied as negative electrode materials for lithium-ion bat-
teries. This special interest has been developed since the
communication by members of the Fuji research labo-
ratories on the outstanding performances of tin-con-
tin(II) solution (prepared by dissolving 20
g of
SnCl2·2H2O in propylene carbonate (PC) up to 100 ml),
with continuous stirring until the precipitation was
completed. The precipitate was separated by centrifuga-
tion at 10 000 rpm for 10 min, washed several times
with absolute ethanol, dried at 60 °C in air for 2 h, and
then at 110 °C under vacuum for one day (sample A).
Another sample was only dried at 80 °C for one day
(sample B). The dissolution of tin chloride in distilled
water leads to the precipitation of Sn2PO4Cl [5]. X-ray
diffraction (XRD) was carried out with a Siemens
D5000 diffractometer working with Cu Ka radiation
and a graphite monochromator.
The electrochemical reactions with lithium were stud-
ied in two-electrode cells. The cathode was prepared by
mixing the pure compound (70% in weight), poly(te-
trafluoro ethylene) PTFE (15%), and carbon-black
(15%). The final pellets of 7 mm diameter were ob-
tained by pressing the mixture (10–15 mg) at 250 MPa.
The anode consists of a metallic lithium disk. Both
electrodes were held in a nickel grid and immersed in
the electrolyte (1 M solution of LiPF6 in ethylene
carbonate–dimethyl carbonate (EC:DMC=1:1). The
taining glasses [1]. However,
a
major problem
associated to their use results from the formation of
large-size metallic-tin clusters during prolonged cycling.
Thus, the complete lithium–tin alloying process is hin-
dered, leading to a strong loss of capacity. Tin-based
crystalline and glassy oxides have been widely studied,
as well as some oxysalts. The presence of atoms indi-
rectly involved in the electrochemical reactions (such as
phosphorus or boron), creating a matrix where tin
atoms are embedded, seems to have a beneficial effect
on the electrochemical performance of these materials.
Thus, besides oxides and oxyglasses some tin phos-
phates were studied such as glassy Sn2BPO6 [2], crys-
talline Sn2P2O7 and SnMn0.5PO4 [3], Sn2P2O7 [4] and
more recently Sn2PO4Cl [5]. In this way, we report a
preliminary electrochemical study of tin(II) hydrogen
phosphate: SnHPO4.
* Corresponding author. Fax: +34-957-218606.
E-mail address: iq3pevic@uco.es (C. Pe´rez Vicente).
0013-4686/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(01)00735-6

490
M.L. Elidrissi Moubtassim et al. / Electrochimica Acta 47 (2001) 489–493
Fig. 3. IR spectra of SnHPO4, indicating the assignation of the
different observed bands. Sample A, solid line; sample B, doted line.
using a multi-channel microprocessor-controlled system
(Mac Pile) [6]. The cells were thermostated with water
circulation controlled to a temperature of 90.05 °C,
using a Haake F3K apparatus.
Fig. 1. X-ray diffraction patterns of sample A (top) and sample B
(bottom). Rietveld refinement (solid line); experimental (dots) and
difference are given for the sample B.
3. Results
3.1. Structural characterization
The X-ray diffraction patterns of the pristine materi-
als shown in Fig. 1 agree well with those reported for
SnHPO4 [7–9]. It has a monoclinic structure, space
group P21/c, with all atoms located in 4e sites. The
structure can be described as a double layer of isolated
PO4 tetrahedra, linked by Sn atoms, in the (010) plane,
the double layers being held together by hydrogen
bonding (Fig. 2). A Rietveld refinement of the data
obtained for sample B was carried out to confirm its
crystallographic structure. The refined unit cell parame-
ters are included in Table 1.
The IR spectra of both samples were also coincident,
as shown in Fig. 3. The normal modes of vibration of
a regular PO4 tetrahedron are: A1, E (both modes
Raman active), and 2 F2 (Raman and IR active), which
are shown in Fig. 3. The PO4 stretching (PO4 st) mode
appears at ca. 570 cm−1, and the PO4 bending (PO4 l)
mode at ca. 1020 cm−1, for free undistorted PO4 tetra-
Fig. 2. Projection in the (100) plane of the structure of SnHPO4,
showing the double layer of PO4 units linked by Sn atoms, where the
double layers are held together by hydrogen bonding.
galvanostatic discharge–charge curves were obtained at
different temperatures, at C/10 (C=1 Li/mol/h) by
Table 1
Unit cell parameters of SnHPO4
Sample A
Sample B
From Ref. [6]
From Ref. [7]
From Ref. [8]
,
a (A)
4.658(1)
13.667(3)
5.883(1)
98.77(1)
4.614(4)
13.63(1)
5.834(5)
98.77(1)
4.576
13.548
5.785
98.7
4.608
13.603
5.823
98.8
4.596
13.625
5.824
98.8
,
b (A)
,
c (A)
i (°)

M.L. Elidrissi Moubtassim et al. / Electrochimica Acta 47 (2001) 489–493
491
hedra. If the Td symmetry decreases to give C1 symme-
try, each F2 mode split in 3 A modes, all IR active.
Thus, the PO4 st mode appears as three bands at ca.
580, 555 and 530 cm−1, and the PO4 l at 1080, 1025
and 975 cm−1. A similar effect was observed for
Sn2ClPO4 [5], but in that case the splitting was higher,
thus indicating a higher distortion from the theoretical
Td symmetry. Additionally, the presence of the ꢀOH
group in HPO24− is confirmed by the bands ascribable
to OꢁPꢀOH bending, POꢀH combined, POꢀH stretch-
ing and OH stretching, as shown in Fig. 3.
was found to justify this 1 Li/mol excess consumption.
It could be tentatively attributed to a possible H+ to H2
reduction, according to:
SnHPO4+3 Li“Sn +(1/2)H2+Li3PO4
Unfortunately, attempts to extract information from
PMR and FTIR spectra of partially lithiated samples
were unsuccessful due to the presence of the solvent
used to support the electrolyte in cells (the mixture
DMC and EC). Especially, in the case of the PMR
spectra, a large signal was obtained corresponding to H
atoms of solvents, which masks the possible signals of
the active phase.
3.2. Electrochemical characterization
Because a preliminary study showed us that the
voltage of the first plateau is highly temperature sensi-
tive, all the results presented here have been carried out
under thermostatic conditions. Thus, the voltage of the
first plateau varies from 1.49 V at 5 °C to 1.735 V at
40 °C, as shown in the inset of Fig. 4-top). The end of
the plateau at ca. 1.5 V is best represented by the
derivative curves −dV/dx versus x (Fig. 4-middle).
Sample B shows a simple peak at ca. 2 F/Sn (marked
B1), corresponding to the end of the SnII into Sn0
reduction, followed by a complex profile. Concerning
sample A, the end of the plateau is evidenced by a split
peak, centered at ca. 3 F/Sn (marked A1 and A2). On
the contrary, the plot of −dV/dx versus V shows
similar profiles for both samples (Fig. 4-bottom), with a
first peak at ca. 1.3 V (marked A1 and B1, for samples
A and B, respectively), and a second peak at ca. 0.9 V
(marked A2 and B2). Then, a large band of low ampli-
tude is observed.
Fig. 4-top shows the first discharge of the cell using
SnHPO4 (samples A and B) as active electrode materi-
als in lithium cells. The first plateau for sample B
involves the well-documented reduction of SnII to
metallic Sn, with the theoretical consumption of 2
Li/mol. Concerning sample A, this first plateau extends
up to 3 Li/mol. Two Li/mol can be attributed to
SnII/Sn reduction. The third Li/mol cannot be inter-
preted with similar arguments. No simple explanation
3.3. XRD of lithiated samples
The X-ray diffraction patterns of selected partially
lithiated samples are shown in Fig. 5a and b. At 2 F/Sn,
after the SnII/Sn0 reduction, sample B is completely
X-ray amorphous. Very weak bands corresponding to
metallic tin are observed. At the same lithiation amount
for sample A, the peaks of the pristine product are still
visible. It is worth noting that when the sample B
reaches the end of the plateau (2 F/mol), for sample A
the plateau is still present. Sample A becomes com-
pletely amorphous at the end of this plateau, which
corresponds to 3 F/mol (ca. 1 V). At the beginning of
the initial expected voltage of Li–Sn alloy formation
(ca. 0.8 V), the products are always amorphous and
metallic tin remains almost undetected by X-ray diffrac-
tion. Finally, at the end of the discharge (at 0 V) typical
reflections of Li–Sn alloys are absent.
Fig. 4. Cell discharge using SnHPO4 as the cathode material and
metallic Li as the anode material. Top: comparison between samples
A and B (the inset shows the evolution of the voltage of the first
plateau of sample A as a function of the temperature); middle,
derivative curve −dV/dx of the cell discharge curves vs. x; bottom,
−dV/dx vs. the voltage.
3.4. Cycling results
Fig. 6-top shows the evolution of the capacity during
cycling at two different temperatures: 5 and 40 °C. In

492
M.L. Elidrissi Moubtassim et al. / Electrochimica Acta 47 (2001) 489–493
4. Discussion
The Li–Sn alloying process was earlier characterized
in the 360–590 °C temperature range [10], which
clearly showed the different plateaus and voltage drops
in the discharge curve, corresponding to the alloys
LiSn, Li7Sn3, Li5Sn2, Li13Sn5, Li7Sn2 and Li22Sn5. At
room temperature, the Li–Sn alloying process has been
described to take place below 0.8 V [11]. The LiySn
alloys versus Li gives voltages of 0.660 V (0.4ByB
0.7), 0.530 V (0.7ByB2.33), 0.485 V (2.33ByB2.63),
0.420 V (2.63ByB3.5) and 0.380 V (3.5ByB4.4)
[12]. Thus, taking into account these previous results,
the discharge curve of sample B can be decomposed in
three steps. The first step at ca. 1.5 V, which ends at ca.
2 F/Sn (peak marked B1 in the derivative curves),
corresponds to the SnII to Sn0 well-known metallic
reduction, as observed in other SnII phosphates [4,5].
The third one, below 0.8 V, is assigned to Li–Sn alloys
formation. The second step takes place in the range
from 1.2 to 0.8 V (marked B2), which is too low to be
SnII/Sn0 reduction, and too high for Li–Sn alloys
formation.
Besides, the electrochemical discharge curve of sam-
ple A can be interpreted in terms of two steps: the first
one corresponds to a reduction of SnII into Sn0, accom-
panied by a possible H+ reduction, which extends from
0 to ca. 3 F/Sn, according to Fig. 4-top; the second
which extends from ca. 3 to 7 F/mol, in the range 0.8–0
V. The capacity reached at this step agrees fairly well
Fig. 5. (a) X-ray diffraction patterns of selected partially lithiated
samples (sample A); reference lines position of Sn and Li22Sn5 are
shown. (b) X-ray diffraction patterns of selected partially lithiated
samples (sample B); reference lines position of Sn and Li22Sn5 are
shown.
both cases, the capacity is almost stabilized after 10
cycles, around 200 mA h/g, being higher for the experi-
ment carried out at higher temperature (ca. 180 mA h/g
at 5 °C, 220 mA h/g at 40 °C). For a more prolonged
cycling, the relative capacity retention (Ci+1/Ci) is
practically 100%.
The X-ray diffraction patterns of the active material
after 10 cycles are shown in Fig. 6-bottom, where tin
reflections are clearly visible. The broadened profiles
evidence that the growth of Sn domains took place in a
limited extension.
Fig. 6. Top: evolution of the cell capacity as a function of the number
of cycles (triangles, at 5 °C; circles, at 40 °C; open, discharge; solid,
charge). Bottom: X-ray diffraction pattern of the active material after
the 10th cycle.

M.L. Elidrissi Moubtassim et al. / Electrochimica Acta 47 (2001) 489–493
493
metallic tin are visible, indicating that the clusters for-
mation is developed during the successive discharge/
charge cycling. A good design of the cathode pellets
would probably hinder this cluster formation, and thus
enhance the retention capacity.
with the formation of typical Li–Sn alloys. The maxi-
mum theoretical capacity is 4.4 Li/Sn, which corre-
sponds to Li22Sn5. When both samples are compared, a
good correspondence is observed in: (a) the process
assigned to alloying process (Fig. 4-top); (b) the peaks
(A1 and B1, and A2 and B2) in the derivative curves in
Fig. 4-bottom. Thus, the first step of the discharge in
sample A can be assigned to steps one and two in
sample B. While in sample A both processes take place
simultaneously, with a single plateau, in sample B these
two processes are clearly differentiated. The only evi-
dence of these two processes in sample A appears at the
end of the plateau, with a complex voltage drop, which
makes appearing two peaks in the derivative curve
(marked A1 and A2). Basically, both samples have the
same behavior, but the voltage evolution of the two
first processes seems to be very sensitive to the way of
drying the initial precipitate.
Acknowledgements
The authors are grateful to CICYT for financial
support (contracts MAT99-0741 and contract SAB99-
0128).
References
[1] Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka,
Science 276 (1997) 1395.
[2] I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2943.
[3] K.B. Wan, S.F.Y. Li, Z.Q. Gao, K.S. Siow, J. Power Sources 75
(1998) 9.
[4] Y.W. Xiao, J.Y. Lee, A.S. Yu, Z.L. Liu, J. Electrochem. Soc.
146 (1999) 3623.
The X-ray diffraction patterns show very weak bands
assigned to metallic Sn, while no Li–Sn alloys signals
are detected. It implies that metallic tin is highly dis-
persed in the phosphate matrix, thus large clusters
(either metallic tin or Li–Sn alloys) are not formed.
This result strongly differs to that observed in other tin
phosphates, where reflections were clearly observed at
the end of the tin reduction (metallic tin) or/and the
end of the discharge (Li–Sn alloys) [4,5]. Since the
formation of large metallic-tin clusters during cycling
partially prevents the complete lithium–tin alloying
process, the absence of these reflections makes this
compound an interesting candidate as precursor of
active anodic material. After 10 cycles, reflections of
[5] F.J. Ferna´ndez Madrigal, C. Pe´rez Vicente, J.L. Tirado, J.
Electrochem. Soc. 147 (2000) 1663.
[6] Y. Chabre, J. Electrochem. Soc. 138 (1991) 329.
[7] A.F. Berndt, R. Lamberg, Acta Crystallogr. 27 (1971) 1092.
[8] L.W. Schoeder, E. Prince, Acta Crystallogr. 32 (1976) 3309.
[9] R.C. McDonald, K. Eriks, Inorg. Chem. 19 (1980) 1237.
[10] C.J. Wen, R.A. Huggins, J. Electrochem. Soc. 128 (1981) 1181.
[11] Both experimental and theoretical extended bibliography can be
found. See, for example: I.A. Courtney, J.S. Tse, O. Mao, J.
Hafner, J.R. Dahn, Phys. Rev. B 58 (1998) 15583.
[12] J. Wang, I.D. Raistrick, R.A. Huggins, J. Electrochem. Soc. 133
(1986) 457.