Electrochemical performances of lithium ion battery using alkoxides of group 13 as electrolyte solvent

Article abstract of DOI:10.1016/j.electacta.2007.07.005

Tris(methoxy polyethylenglycol) borate ester (B-PEG) and aluminum tris(polyethylenglycoxide) (Al-PEG) were used as electrolyte solvent for lithium ion battery, and the electrochemical property of these electrolytes were investigated. These electrolytes, e

Full text of DOI:10.1016/j.electacta.2007.07.005

Electrochimica Acta 53 (2007) 549–554
Electrochemical performances of lithium ion battery using
alkoxides of group 13 as electrolyte solvent
Fuminari Kaneko, Yuki Masuda, Masanobu Nakayama, Masataka Wakihara∗
Department of Applied Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo 152-8552, Japan
Received 13 April 2007; accepted 7 July 2007
Available online 12 July 2007
Abstract
Tris(methoxy polyethylenglycol) borate ester (B-PEG) and aluminum tris(polyethylenglycoxide) (Al-PEG) were used as electrolyte solvent for
lithium ion battery, and the electrochemical property of these electrolytes were investigated. These electrolytes, especially B-PEG, showed poor
electrochemical stability, leading to insufficient discharge capacity and rapid degradation with cycling. These observations would be ascribed to
the decomposition of electrolyte, causing formation of unstable passive layer on the surface of electrode in lithium ion battery at high voltage.
However, significant improvement was observed by the addition of aluminum phosphate (AlPO4) powder into electrolyte solvent. AC impedance
technique revealed that the increase of interfacial resistance of electrode/electrolyte during cycling was suppressed by adding AlPO4, and this
suppression could enhance the cell capabilities. We infer that dissolved AlPO4 components formed electrochemically stable layer on the surface
of electrode.
© 2007 Elsevier Ltd. All rights reserved.
Keywords: Lithium ion battery; Lithium polymer battery; Plasticizer; AlPO4 dispersion; LiFePO4
1. Introduction
ductivity [4–9]. Although many investigations have reported
about polymer electrolyte itself, there exist limited numbers
of examinations on the battery performance in the form of
practical LPB style. This would be due to the technical dif-
ficulty to make desirable solid–solid (electrode–electrolyte)
connection. Thus, the effect of adding plasticizer for the prac-
tical battery usage is still uncertain despite increase in ionic
conductivity.
In present paper, we constructed more simple and conven-
tional battery form, or Li|liquid electrolyte|cathode material
where the 13 group alkoxides was used as liquid electrolyte
instead of assembling LPBs, and investigated their electro-
chemical property. Through the study on the electrochemical
behavior of above cell, the practical and potential problems
that could occur in LPB using the plasticizers would be
clarified.
The present commercial lithium ion battery contains organic
solutions, for example ethylene carbonate (EC) and diethyl car-
bonate (DEC), as electrolyte. Because of flammability of these
solutions, the risk of ignition is unavoidable for battery. Many
efforts have been carried out, and one of the conclusion is the
utilization of all-solid-state lithium polymer battery (LPB). The
LPBs use the lithium ion conductive polymer instead of organic
solutions. This polymer electrolyte is not flammable unlike
the organic solutions. Moreover, polymer electrolyte causes no
electrolyte-leakage and is superior to organic solution or inor-
ganic solid electrolytes in handing and formability. However,
polymer electrolyte has poor ionic conductivity beside organic
solutions [1], preventing the practical use of LPB. Therefore,
intensive R&D of polymer electrolyte has been carried out
mainly to improve the ionic conductivity [2,3].
We have reported that the addition of 13 group alkoxides
into matrix polymer as plasticizers increased the ionic con-
2. Experimental
B-PEG and Al-PEG were synthesized as shown in Fig. 1
by using methoxy poly(ethyleneglycol) (n = 3, 9) and boric
acid or aluminum isopropoxide, respectively [5,9]. LiClO4
(Soekawa chemicals) was dissolved into B-PEG and Al-PEG to
∗
Corresponding author.
E-mail address: mwakihar@o.cc.titech.ac.jp (M. Wakihara).
0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2007.07.005

550
F. Kaneko et al. / Electrochimica Acta 53 (2007) 549–554
Fig. 1. Scheme of the reaction for (a) tris(methoxy polyethyleneglycol) borate ester and (b) aluminum tris(polyethyleneglycoxide).
obtain a 1.0 M concentration. In additional examination, AlPO4
(Aldrich) dispersed into these liquid electrolytes with several
weight ratios to improve the cell performance. Moreover, the
electrolyte solution removing the precipitated AlPO4 was pre-
pared by a cyclone separator. All the procedure described above
was done in Ar filled dry box.
LiCoO2 (Aldrich) and LiFePO4 (Aldrich) were used as
cathode materials. These cathode sheets were prepared as fol-
low; the mixture of active material, polyvinylidene fluoride
(PVdF) (Kureha Kagaku-Kogyo) and acetylene black (AB)
whose composition is 8:1:1 in weight ratio was dissolved into
N-methyl-2-pyrrolidinone (Kanto Kagaku) to obtain the slurry.
Then, the slurry was applied on aluminum foil, and the sheet
was dried for 24 h at 120 ◦C. Coin type cell which included the
1.0 M LiClO4 B-PEG or Al-PEG as electrolyte was assembled
with the cathode sheet and metal lithium anode.
Fig. 2. DSC curves of liquid electrolytes.
The grass transition temperature (Tg) of the electrolyte was
determined by differential scanning calorimetry (DSC) using
DSC220C (Seiko Instruments Inc.) under N2 atmosphere. The
scanning rate was set at 10 ◦C min−1, and an empty pan was
used as a reference.
The measurement of ionic conductivity of the electrolytes
was carried out by AC impedance method using Hewlett-
Packard 4192A LF impedance analyzer over the frequency range
from 5 Hz to 13 MHz. The two stainless steel electrodes were
used in this measurement.
The electrochemical stability of electrolyte was evaluated by
cyclic voltammetry (CV) using Solartron SI 1287 electrochemi-
cal interface. The scan ratio was set as 1 mV s−1 over the voltage
range from 0 to 5.0 V at 60 ◦C, and the same cell as described
above was used.
The cyclic performance tests of the cells were performed
galvanostatically at a current density of 0.1 C and at a regulated
cut-off voltage between 3.5 and 4.1 V (in the case of LiCoO2) or
between 2.5 and 3.8 V (in the case of LiFePO4) at 60 ◦C using
HJR-110m SM6 electrochemical interface (Hokuto Denko).
The resistance of electrode/electrolyte interface of
LiFePO4/1 M LiClO4 B-PEG (n = 9)/Li cell was measured by
AC impedance method using Solartron SI 1287 electrochemical
interface over the frequency range from 10 to 1 MHz. The
measurement was carried out by constant voltage impedance
measurement method [10] at 3.8 V.
3. Results and discussion
Fig. 2 shows DSC curves of prepared liquid electrolytes. The
peaks due to the grass transition appeared in all sample (see
arrows in figure), and obtained grass transition temperature Tg
was summarized in Table 1. As indicated in the table, the Tg
decreases with increasing the number of ethylene oxide (EO)
unit in electrolyte. However, the difference of Tg was relatively
small regardless of the length of EO unit in the case of Al-PEG.
Therefore, the difference of corresponding ionic conductivity in
Al-PEG is expected to be small in the viewpoint of segmental
motion.
The ionic conductivity of each liquid electrolyte is shown
in Fig. 3. The results roughly corresponded to the Tg. Despite
the no marked change in Tg observed in Al-PEG, their conduc-
tivity changes by the length of EO unit. Thus, the lithium ions
Table 1
Tg of liquid electrolyte
The length of EO unit
Tg (◦C)
3
9
−75.6
−52.1
B-PEG
3
9
−60.5
Al-PEG
−57.5

F. Kaneko et al. / Electrochimica Acta 53 (2007) 549–554
551
Fig. 3. Arrhennius plots of ionic conductivity for 1 mol/L LiClO4 B-PEG and
Al-PEG. X and n in X-PEGn indicate the centered atom of the oligomer (B or
Al) and the number of EO units of the oligomer, respectively.
transported not only by segmental motion, but also by the move-
ment of electrolyte molecules which associated with lithium
ion. According to the measurement of ionic conductivity, prefer-
able electrochemical performance will be obtained by using the
electrolyte with the short length of EO unit.
The results of CV measurement are shown in Fig. 4. The
net current flow was observed around 3 V and increased with
scanning voltage, indicating decomposition of the electrolyte
components. As can be seen in the figure, the electrolyte with
shorter EO unit showed decomposition current at lower volt-
Fig. 5. Cyclic performance of LiCoO2/electrolyte/Li cell at 60 ◦C: (a) n = 3 and
(b) n = 9.
Fig. 6. Cyclic performance of LiFePO4/electrolyte/Li cell at 60 ◦C: (a) n = 3 and
(b) n = 9.
Fig. 4. Cyclic voltammograms of B-PEG and Al-PEG at 60 ◦C.

552
F. Kaneko et al. / Electrochimica Acta 53 (2007) 549–554
Fig. 7. Cyclic performance of LiFePO4/electrolyte + AlPO4/Li cell at 60 ◦C: (a) B-PEG n = 3, (b) B-PEG n = 9, (c) Al-PEG n = 3 and (d) Al-PEG n = 9; 3% indicates
addition of 0.003% of AlPO4 powder into B or Al-PEG. CL indicates the usage of clear layer of electrolyte after removing insoluble AlPO4 by centrifuging.
age. It is noted that the electrochemical window shown in Fig. 3
was obtained by Li/electrolyte/SUS cell, so that it is not nec-
essary to apply the case of actual battery construction with
Li/electrolyte/cathode. However, these results would give a rel-
ative electrochemical stability of each electrolyte solution.
The variation of the discharge capacity with cycle number
for LiCoO2/electrolyte/Li cell is displayed in Fig. 5. In the case
of LiCoO2 cathode, obtained capacity is quite small comparing
with theoretical one for LiCoO2 (∼140 mAh/g). Especially, dis-
charge capacities are negligibly small using the electrolyte with
shorter EO units. Such behaviors would arise from the electro-
chemical stability rather than ionic conductivity of electrolytes.
Thus, the operating voltage of LiCoO2 cathode (3.5–4.1 V)
would be much higher than the electrochemical window of
electrolytes. Although the ionic conductivity decreases with
increasing the number of EO unit in electrolyte, improved cycle
performance was obtained for the electrolytes with longer EO
units which showed higher electrochemical stability (Fig. 4).
To avoid the decomposition of electrolyte, LiFePO4 which
works at lower voltage (2.5–3.8 V) was used as cathode mate-
rial instead of LiCoO2. As seen from Fig. 6, improved cycle
performance was observed even for the electrolytes with shorter
EO chain. Thus, the operation voltage would be critical fac-
tor for using B-PEG and Al-PEG as plasticizer in practical LPB
construction. However, the performance is severely poor in com-
parison with currently used general electrolytes; for example
EC/DEC. Hence, it indicates that the improvement of electro-
chemical stability of the plasticizer is necessary for realizing
LPB with plasticized polymer electrolytes.
To improve the relatively poor cycle performance, attempts
were made by adding AlPO4 ceramics into electrolyte solution.
It have been reported that ceramic coating on the electrode mate-
rial seemed to protect the surface of electrode and prevent the
side reaction with electrolytes [11,12]. Addition of the ceramic
filler to polymer electrolyte is another way to obtain the supe-
rior property [13,14]. Expanding these ideas, AlPO4 ceramics
powder was dispersed into electrolyte solutions.
Various electrolytes containing AlPO4 with several weight
ratios and the electrolytes removing the insoluble AlPO4
were prepared. In glance view, AlPO4 ceramics powders were
scarcely dissolved into the electrolyte solutions. Fig. 7 shows
the cyclic performance of LiFePO4/electrolyte(+AlPO4)/Li cell
at 60 ◦C. As seen in the figure, better cyclic performance was
observed by adding AlPO4 powder, especially in electrolyte
with longer EO units. The effect of AlPO4 dispersion showed
slightly small difference in various amount of AlPO4. Moreover,
it is noted that improvement of cycle performance was observed
even for the electrolytes in which AlPO4 powder was removed
by centrifuging (Fig. 7 (a and b)). Thus, it would not be the
reason of improved cycle performance that the dispersed AlPO4

F. Kaneko et al. / Electrochimica Acta 53 (2007) 549–554
553
represent the observed resistance divided by the initial resis-
tance without application of constant voltage. As seen in the
figure, both the interfacial resistance, R1 and R2 significantly
increased during first 5 h for the electrolyte without AlPO4. It
was supposed that the decomposition of electrolyte forms a
passive layer on the surface of electrode, and the increase of
resistance is occurred by the progressive growth of the passive
layer. On the other hand, no marked increase was observed
in case of the electrolyte with AlPO4 dispersion. Therefore,
AlPO4 dispersion suppressed the increase of interfacial resis-
tances between electrolyte and both positive and negative
electrode interfaces, and this would lead to the improvement
of cycle performance. However, exact mechanism is still
uncertain, and further study regarding interfacial reaction is
required.
Fig. 8. Example of impedance spectra of LiFePO4/B-PEG9/Li cell at 60 ◦C.
4. Conclusion
The coin type liquid cell using B-PEG or Al-PEG as
electrolyte solvent was prepared in order to investigate the
electrochemical property of these plasticizers for the future
usage of LPB. The decomposition of electrolyte would occur
at high voltage region, and severe degradation was observed
during the cycling. This would be associated with the rel-
atively poor electrochemical stability for these electrolytes.
However, considerable improvement of cycle performance was
made by addition of AlPO4 ceramics powder into electrolyte
solution. From the impedance measurement, we confirmed
the suppression of the increase in interfacial resistance. Con-
sequently, the addition of AlPO4 into plasticizer would be
one of the promising techniques for developing the practical
LPB.
powder adhered to the surface of electrode materials. Instead,
we infer following two processes: (1) dispersed AlPO4 trapped
traces of residual impurities, such as quite small amount of H2O,
or (2) slightly dissolved ions of Al3+ and (PO4)3− formed stable
ceramic layer at the surface of positive and/or negative elec-
trodes during the first charging, and this layer prevented the side
reaction with electrolytes.
The impedance measurement of LiFePO4/electrolyte
(+AlPO4)/Li cell was carried out at 60 ◦C. B-PEG was chosen
as electrolyte solvent, because of specific improvement of
cycle performance. Typical impedance plots were shown in
Fig. 8. Two semicircles, R1 and R2, were observed in higher
and lower frequency region for both samples. According to
the preliminary impedance measurement using symmetrical
cell with Li/electrolyte/Li construction, R1 and R2 were
related to the resistance of interface of anode/electrolyte and
cathode/electrolyte, respectively. (Detailed procedure for the
assignment of semicircles was described in Ref. [10].) Fig. 9
shows the time evolution of both the interfacial resistances
at constant voltage state (3.8 V). In this figure, R/R(OCV)
Acknowledgement
This project has been funded by Industrial Technology
Research Grand Program from the New Energy and Industrial
Technology Development Organization (NEDO) of Japan.
References
[1] P.P. Prosini, T. Fujiede, S. Passerini, M. Shikano, T. Sakai, Electrochem.
Commun. 2 (2002) 44.
[2] H. Tsutsumi, A. Matuo, K. Takase, S. Doi, A. Hisanaga, K. Onimura, T.
Oishi, J. Power Sources 90 (2000) 33.
[3] Y. Matoba, S. Matsui, M. Tabuchi, T. Sakai, J. Power Sources 137 (2004)
284.
[4] Y. Kato, S. Yokoyama, H. Ikuta, Y. Uchimoto, M. Wakihara, Electrochem.
Commun. 3 (2001) 128.
[5] Y. Kato, K. Hasumi, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M.
Wakihara, Solid State Ionics 150 (2002) 355.
[6] Y. Kato, K. Hasumi, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M.
Wakahara, J. Therm. Anal. Calorim. 69 (2002) 889.
[7] Y. Kato, K. Suwa, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M.
Wakihara, Solid State Ionics 152–153 (2002) 155.
[8] Y. Kato, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M. Wakihara,
Electrochim. Acta 50 (2004) 281.
Fig. 9. Applying voltage time dependence of interfacial resistance of
LiFePO4/1 M LiClO4 B-PEG9 + AlPO4/Li cell at 60 ◦C. Solid and dotted lines
indicate R1 and R2, respectively.
[9] Y. Masuda, M. Seki, M. Nakayama, M. Wakihara, H. Mita, Solid State
Ionics 177 (2006) 843.

554
F. Kaneko et al. / Electrochimica Acta 53 (2007) 549–554
[10] S. Seki, Y. Kobayashi, H. Miyashiro, A. Yamanaka, Y. Mita, T. Iwahori, J.
[13] G. Jiang, S. Maeda, Y. Saito, S. Tanase, T. Sakai, J. Electrochem. Soc. 152
Power Sources 146 (2005) 741.
(2005) A767.
[11] T.J. Kima, D. Son, J. Cho, B. Park, H. Yang, Electrochim. Acta 49 (2004)
[14] F. Croce, L. Settimi, B. Scrosati, Electrochem. Commun. 8 (2006)
364.
4405.
[12] S. Seki, Y. Kabayashi, H. Miyashiro, Y. Mita, T. Iwahori, Chem. Mater. 17
(2005) 2041.