Compatibility of Li7La3Zr2 O12 solid electrolyte to all-solid-state battery using Li metal anode

Article abstract of DOI:10.1149/1.3474232

Electrochemical properties of Li₇La₃Zr₂ O₁₂ (LLZ) were investigated to reveal its availability as a solid electrolyte for all-solid-state rechargeable batteries with a Li metal anode. After calcination at 1230°C, a well-sintered LLZ pellet with a garnet-like structure was obtained, and its conductivity was 1.8 × 10^-4 S cm_-1 at room temperature. The cyclic voltammogram of the Li/LLZ/Li cell showed that the dissolution and deposition reactions of lithium occurred reversibly without any reaction with LLZ. This indicates that a Li metal anode can be applied for an LLZ system. A full cell composed of a LiCoO₂ /LLZ/Li configuration was also operated successfully at expected voltage estimated from the redox potential of Li metal and LiCoO₂. Simultaneously, an irreversible behavior was observed at the first discharge and charge cycle due to an interfacial problem between LiCoO₂ and LLZ. The discharge capacity of the full cell was 15 μA h cm^-2. These results reveal that LLZ is available for all-solid-state lithium batteries.

Full text of DOI:10.1149/1.3474232

A1076
Journal of The Electrochemical Society, 157 ͑10͒ A1076-A1079 ͑2010͒
0
013-4651/2010/157͑10͒/A1076/4/$28.00 © The Electrochemical Society
Compatibility of Li La Zr O Solid Electrolyte
7
3
2
12
to All-Solid-State Battery Using Li Metal Anode
a,z
a,
a,
b
Masashi Kotobuki, Hirokazu Munakata, * Kiyoshi Kanamura, * Yosuke Sato,
b,
and Toshihiro Yoshida *
a
Department of Applied Chemistry, Graduate School of Urban Environmental Science, Tokyo Metropolitan
University, Tokyo 192-0397, Japan
b
NGK Insulators, Limited, Aichi 467-8530, Japan
Electrochemical properties of Li La Zr O ͑LLZ͒ were investigated to reveal its availability as a solid electrolyte for all-solid-
7
3
2
12
state rechargeable batteries with a Li metal anode. After calcination at 1230°C, a well-sintered LLZ pellet with a garnet-like
−
4
−1
structure was obtained, and its conductivity was 1.8 ϫ 10 S cm at room temperature. The cyclic voltammogram of the
Li/LLZ/Li cell showed that the dissolution and deposition reactions of lithium occurred reversibly without any reaction with LLZ.
This indicates that a Li metal anode can be applied for an LLZ system. A full cell composed of a LiCoO /LLZ/Li configuration
2
was also operated successfully at expected voltage estimated from the redox potential of Li metal and LiCoO . Simultaneously, an
2
irreversible behavior was observed at the first discharge and charge cycle due to an interfacial problem between LiCoO and LLZ.
2
−
2
The discharge capacity of the full cell was 15 ␮A h cm . These results reveal that LLZ is available for all-solid-state lithium
batteries.
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3474232͔ All rights reserved.
Manuscript submitted April 21, 2010; revised manuscript received July 2, 2010. Published August 17, 2010.
All-solid-state lithium batteries consisting of solid electrodes and
a Li-ion conductive solid electrolyte have been expected to over-
come the safety problem of present lithium-ion batteries including
after each heat-treatment in air ͑first, at 900°C for 6 h and second, at
1125°C for 6 h͒. The obtained powder was finally pressed into a
pellet under 50 MPa for 5 min and then annealed at 1230°C for 36
h in air at a heating rate of 1°C min . The crystal structure of LLZ
was confirmed by X-ray diffractometer ͑XRD, Rigaku RINT-
UltimaII͒ using Cu K␣ radiation after each heat-treatment. The
morphology of the LLZ pellet was observed with a scanning elec-
tron microscope ͑SEM, JEOL͒.
Both surfaces of the LLZ pellet after heat-treatment were pol-
ished to obtain flat surfaces and to control its thickness before elec-
trochemical measurements. The conductivity of LLZ was measured
by the ac impedance method using an SI 1260 impedance/gain-
phase analyzer ͑Solartron Analytical͒ in the frequency range from
10 Hz to 1 MHz and an alternating voltage signal of 10 mV. In this
measurement, Au was sputtered on each surface of the LLZ pellet to
ensure electrical contact between the Cu current collector and the
LLZ pellet. A symmetric cell with a Li/LLZ/Li configuration was
constructed and evaluated by cyclic voltammetry and chronopoten-
tiometry using ALS-660B ͑BAS Inc.͒. Cu foils were used as current
collectors. The cell was heated at 140°C for 1 h to obtain better
contact between the LLZ pellet and the Li foil.
1
−
1
flammable nonaqueous solvent. High lithium-ion conductivity, low
electronic conductivity, and good chemical stability against elec-
trode materials are required for electrolytes in rechargeable lithium-
2
ion batteries. Lithium-ion conductivity of ceramic electrolytes is
lower than those of nonaqueous electrolytes, but some oxides, for
3
-5
6,7
example, Li0.35La0.55TiO ͑LLT͒ and LiTi ͑PO ͒ ͑LTP͒, pos-
3
2
−4
4 3
−
3
−1
sess high ionic conductivities of ϳ10 to 10 S cm , which are
acceptable for an all-solid-state battery. Therefore, many research
groups have investigated their applications to all-solid-state re-
chargeable lithium-ion batteries. However, anode materials with a
lower redox potential than 1.8 V vs Li/Li cannot be utilized in an
LLT and LTP system due to the redox reaction of titanium, which
limits the energy and power densities of all-solid-state rechargeable
+
8
lithium-ion batteries.
In the past several years, a series of garnet-like structural com-
pounds with a nominal chemical composition of Li La M O ͑M
5
3
2
12
=
Nb, Ta͒ has been investigated as a novel family of fast lithium-ion
9
conductors by Thangadurai and co-workers. These compounds con-
tain La and M in the cubic and octahedral environments of the
3
+
5+
A LiCoO2 cathode was prepared on the LLZ pellet by the sol–gel
11
garnet structure. Recently, Li La Zr O ͑LLZ͒ also has been syn-
method. The precursor sol for LiCoO2 was prepared from
7
3
2
12
thesized by Murugan et al. as a new garnet-like compound with
CH COOLi, Co͑CH COO͒ ·4H O, i-C H OH, CH COOH, and
3
3
2
2
3
7
3
10
12
lithium-ion conductivity. This material has received much atten-
tion as a ceramic electrolyte due to its chemical stability against
lithium metal and relatively high lithium-ion conductivity ͑3
H O ͑molar ratio = 1.1:1:20:10:70͒ according to a previous paper.
2
50 ␮L of precursor sol was dropped on the LLZ pellet. Then, the
pellet was calcinated at 450°C for 15 min. This procedure was re-
−
4
−1
ϫ 10 S cm ͒. However, the electrochemical properties of LLZ
have hardly been investigated. In this study, we examined the elec-
trochemical properties of LLZ minutely to reveal its availability as a
solid electrolyte in all-solid-state rechargeable lithium-ion batteries.
peated two times to obtain thick LiCoO film. Finally, the pellet was
2
calcined at 800°C for 1 h. A Li metal anode was prepared by the
same procedure as that of the Li/LLZ/Li cell. A galvanostatic charge
and discharge test of the LiCoO /LLZ/Li cell was performed by
2
using a battery charge/discharge unit ͑HJ1001SM8A, Hokuto
−
2
Experimental
Denko͒ under a constant current density of 2 ␮A cm , and cutoff
voltages were 2.5 and 4.3 V for discharge and charge, respectively.
LLZ was prepared by a solid-state method according to the pro-
10
cedure reported elsewhere. The powders of Li CO , La O , and
2
3
2
3
Results
ZrO were used as starting materials. They were mixed and ball-
2
milled with zirconia balls ͑␾ = 0.2 mm͒ for 6 h in 2-propanol. 10
The XRD patterns for LLZ obtained after each heat-treatment are
displayed in Fig. 1. The structure of LLZ was not cubic at 900°C,
and it turned to a cubic phase after heat-treatment at 1230°C. The
peak intensity of the XRD patterns increased with heat-treatment
temperature. The diffraction peaks observed after heat-treatment at
wt % excess of Li CO was added to compensate the loss of lithium
2
3
during heat-treatments, resulting in the molar ratio of Li:La:Zr
7.7:3:2 in the starting mixture. The milling process was repeated
=
1
230°C were assigned to a well-crystallized garnet-like structure, as
9 10
previously reported by Thangadurai et al. and Murugan et al. The
*
Electrochemical Society Active Member.
E-mail: masakoto@tmu.ac.jp
z
formation of pyrochlore La Zr O as an impurity phase was also
2
2
7
Downloaded on 2014-01-20 to IP 138.251.14.35 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 157 ͑10͒ A1076-A1079 ͑2010͒
A1077
-
-
20
15
7 3 2 12
Li La Zr O
(
c)
(
(
b)
a)
-10
-
5
0
1
0
20
30
40
50
60
2
θ [CuKα] / degree
0
5
10
15
20
Z’ / kΩ
Figure 1. XRD pattern of LLZ powder calcined at ͑a͒ 900 and ͑b͒ 1125°C
and pellet calcined at ͑c͒ 1230°C.
Figure 3. Impedance spectrum of the prepared LLZ pellet measured at room
temperature in the frequency range of 0.1–1 MHz.
confirmed on the XRD patterns of the LLZ powder sintered at lower
than 1125°C ͑Fig. 1a and b͒. After heat-treatment at 1230°C, the
diffraction peaks derived from the impurity disappeared, suggesting
that the LLZ pellet prepared in this study had a single phase of
Li La Zr O with a garnet-like structure. Figure 2 shows typical
LLZ electrolyte without degradation of LLZ. The chronopotentio-
grams of the Li/LLZ/Li cell obtained under different current densi-
ties from 10 to 50 ␮A cm are depicted in Fig. 5. The overpoten-
−2
7
3
2
12
tials for the dissolution and deposition reactions of lithium increased
with current density. The time–potential curves for the lithium dis-
solution and deposition collected at the same current density showed
cross-sectional SEM images of the LLZ pellet. The pellet was sin-
tered well and grain boundary was not observed. Some voids existed
in the cross-sectional view, but it can be confirmed that they were
isolated sufficiently to prevent electrical short-circuit in rechargeable
lithium batteries. It is expected that the void formation can be re-
duced by using smaller LLZ particles for the pellet preparation. An
impedance spectrum of the prepared LLZ pellet measured at room
temperature is exhibited in Fig. 3. Clear semicircles were not ob-
served; only a line with slope of 45º appeared, probably due to
almost no grain-boundary resistance. The obtained impedance spec-
trum was not consistent with that reported by Murugan et al., which
−2
the same absolute overvoltage. Up to 10 ␮A cm , the dissolution
and deposition curves gave the mirrored relationship at least until
6
00 s. This behavior also suggests that the lithium dissolution and
deposition took place reversibly.
2
0
1
0
showed a nearly vertical tail at the low frequency range. This
difference may be attributed to the starting materials. They used
10
0
LiOH as a Li source and La O3 after drying at 900°C as a La
2
source. Meanwhile, Li CO and La O without any treatment were
2
3
2
3
used in this research. This difference may cause the different imped-
ance behavior. The conductivity of LLZ was calculated to be 1.8
-
10
−
4
−1
ϫ 10 S cm at room temperature by using an apparent thickness
2
͑
0.1 cm͒ and surface area ͑0.5 cm ͒ of the pellet.
To investigate the potential use of the lithium metal as anode for
-20
-0.3
all-solid-state rechargeable batteries with an LLZ electrolyte, cyclic
voltammetry and chronopotentiometry were carried out by using the
electrochemical cell with a Li/LLZ/Li configuration. Figure 4 re-
-0.2
-0.1
0
0.1
E / V vs. Li / Li+
Figure 4. CV of Li/LLZ/Li cell at room temperature taken at 10 mV min−1
0.2
0.3
0.4
0.5
−
1
veals the cyclic voltammogram ͑CV͒ taken at 10 mV min . The
dissolution and deposition reactions of lithium were observed re-
versibly, indicating that lithium-ion could be transferred through the
.
0
0
.4
5
0 μA cm-2
20 μA cm-2
.2
0
1
0 μA cm-2
-
-
0.2
0.4
0
100
200
300
400
500
600
Time / sec
1
0μm
Figure 5. ͑Color online͒ Chronopotentiograms of Li/LLZ/Li cell obtained
2
Figure 2. Cross-sectional SEM image of LLZ pellet calcined at 1230°C.
under different current densities from 10 to 50 ␮A cm−
.
Downloaded on 2014-01-20 to IP 138.251.14.35 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

A1078
Journal of The Electrochemical Society, 157 ͑10͒ A1076-A1079 ͑2010͒
-
2
1
0
LiCoO2
LLZ
-
0
1
2
1
0
20
30
40
50
60
2θ [CuKα] / degree
Z’ / M
ΩΩ
Figure 6. XRD pattern of LiCoO cathode prepared on LLZ pellet.
Figure 8. Impedance spectrum of LiCoO /LLZ/Li cell measured at room
2
2
temperature in the frequency range of 0.1–1 MHz.
The LiCoO2 cathode was prepared on the LLZ pellet by the
sol–gel method. Figure 6 shows the XRD pattern of the LiCoO2
cathode on the LLZ pellet. Diffraction peaks attributed to LiCoO2
nopotentiograms was ca. 5500 ⍀, which is much higher than that of
the LLZ pellet ͑1100 ⍀͒. This difference may be due to the large
interfacial resistance between the LLZ pellet and lithium metal, de-
rived from an inhomogeneous current distribution. Therefore, thin
and dense LLZ electrolyte with a uniform surface is needed for
operation under higher current densities. The increment in the con-
tact area between LLZ and lithium metal is also one of the solutions
to reduce the internal resistance of the cell. We have reported that a
solid electrolyte with a three-dimensionally ordered macroporous
͑3DOM͒ structure, which possesses a large surface area, is very
were observed clearly, indicating that the LiCoO cathode was pre-
2
pared successfully. All of the diffraction peaks were assigned to
LiCoO or LLZ, and no impurity phase was observed. The thickness
2
and weight of the LiCoO cathode were 6 ␮m and 12 mg, respec-
2
tively. Figure 7 reveals the galvanostatic charge and discharge
curves of the LiCoO /LLZ/Li cell. The cell worked successfully at
2
the expected operation voltage estimated from the redox potential of
the Li metal and LiCoO . A large irreversible capacity was ob-
2
1
3,14
served, and the discharge capacity of the all-solid-state cell was
useful to cause a decline in the internal resistance.
Therefore,
−
2
1
5 ␮A h cm .
using LLZ with a 3DOM structure, cell operation under a higher
current density would become possible.
Discussion
The electrochemical properties of the LLZ solid electrolyte were
The LiCoO /LLZ/Li cell demonstrated clearly reversible charge
2
and discharge behavior, indicating that an all-solid-state battery with
investigated to confirm its compatibility with all-solid-state re-
chargeable Li batteries with a Li metal anode. The well-sintered
LLZ pellet was obtained by calcination at high temperature as well
as at 1230°C. The pellet was sintered well and grain boundary could
not be found. The conductivity of the pellet was relatively high
a lithium metal anode and a LiCoO cathode can be fabricated by
2
using an LLZ solid electrolyte. However, the discharge capacity was
−
2
only 15 ␮A h cm . This was only 0.2% of the full capacity
−
2
͑8.4 mA h cm ͒ estimated from the weight of the cathode mate-
rial. This low capacity is attributed to the high interfacial resistance
between the LLZ pellet and electrodes, as shown in the impedance
spectrum of the LiCoO2/LLZ/Li cell ͑Fig. 8͒. A large semicircle was
observed in Fig. 8. This semicircle did not appear in the impedance
spectrum of the LLZ pellet. Therefore, the semicircle is attributed to
the interfacial resistance between the LLZ pellet and the electrode.
The resistance of the Li/LLZ/Li cell estimated from the chronopo-
tentiograms was ca. 5500 ⍀, but the impedance of the semicircle
was much larger, nearly 1 M⍀. Therefore, most of the contribution
of the semicircle may be the interfacial resistance between the
LiCoO cathode and LLZ. The XRD pattern of LiCoO /LLZ dis-
−
4
−1
͑
1.8 ϫ 10 S cm ͒, which is acceptable for an all-solid-state bat-
tery.
The CV of the Li/LLZ/Li cell revealed that the dissolution and
deposition reactions of lithium occurred reversibly without side re-
actions with LLZ. This result indicates that the Li metal anode can
be applied for the LLZ system. Although the CV was measured only
at the first cycle, it was confirmed that the LLZ pellet was stable
against contact with a large excess of molten lithium for 72 h. The
Li/LLZ interface is thought to be stable for some period. However, a
long time stability of the LLZ against the Li metal is not clear at the
moment. The chronopotentiograms of the Li/LLZ/Li cell showed the
mirrored relationship. However, the mirrored relationship collapsed
2
2
played no impurity formation ͑Fig. 6͒. However, a small amount
and/or amorphous impurity may be formed at the interface and pro-
vide high resistance to the cell. Optimization of the preparation pro-
−
2
over 10 ␮A cm . A resistance of the cell estimated from the chro-
cess for the LiCoO cathode on LLZ and screening of a more suit-
2
able cathode are needed.
4
3
2
.5
4
These results proved that LLZ has a very high potential to pro-
duce all-solid-state lithium batteries with high power and energy
densities. However, some studies such as long time stability and rate
capability of the LiCoO /LLZ/Li cell as well as the Li/LLZ/Li are
2
1
-3cycle
.5
3
needed. These tests are under way, and it will be reported in due
course.
Conclusions
.5
The electrochemical properties of LLZ solid electrolyte were in-
vestigated for application to all-solid-state lithium batteries using a
lithium metal anode. The LLZ pellet prepared in this study was
chemically and electrochemically stable against lithium metal. Ac-
cordingly, the cell with a Li/LLZ/Li configuration was successfully
0
25
50
75
100
125
Capacity / µA h cm-2
Figure 7. Galvanostatic charge and discharge curves of LiCoO /LLZ/Li cell
2
measured at 2 ␮A cm−2
.
operated. Moreover, in the LiCoO /LLZ/Li cell, reversible charge
2
Downloaded on 2014-01-20 to IP 138.251.14.35 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 157 ͑10͒ A1076-A1079 ͑2010͒
A1079
4
5
6
. Y. Inaguma, C. Liquan, M. Itoh, T. Nakamura, T. Uchida, H. Ikuta, and M. Waki-
hara, Solid State Commun., 86, 689 ͑1993͒.
. T. Abe, M. Ohtsuka, F. Sagane, Y. Iriyama, and Z. Ogumi, J. Electrochem. Soc.,
and discharge behavior was observed, indicating that the all-solid-
state battery with lithium metal anode was operated successfully.
These results revealed that LLZ is available for all-solid-state
lithium batteries with high power and energy densities. However,
1
51, A1950 ͑2004͒.
. H. Aono, E. Sugimoto, Y. Sadaoka, N. Imanaka, and G. Adachi, J. Electrochem.
Soc., 137, 1023 ͑1990͒.
the discharge capacity of the LiCoO /LLZ/Li cell was only
2
−
2
1
5 ␮A h cm . Further study to reduce interfacial resistance is re-
7. J. Fu, Solid State Ionics, 96, 195 ͑1997͒.
8
9
. P. Knauth, Solid State Ionics, 180, 911 ͑2009͒.
. V. Thangadurai, H. Kaack, and W. Weppner, J. Am. Ceram. Soc., 86, 437 ͑2003͒.
quired.
Tokyo Metropolitan University assisted in meeting the publication costs
of this article.
10. R. Murugan, V. Thangadurai, and W. Weppner, Angew. Chem., Int. Ed., 46, 7778
͑2007͒.
1
1. M. Kotobuki, Y. Suzuki, H. Munakata, K. Kanamura, Y. Sato, K. Yamamoto, and T.
References
Yoshida, J. Electrochem. Soc., 157, A493 ͑2010͒.
2. T. H. Rho, K. Kanamura, and T. Umegaki, J. Electrochem. Soc., 150, A107 ͑2003͒.
3. H. Nakano, K. Dokko, M. Hara, Y. Isshiki, and K. Kanamura, Ionics, 14, 173
͑2008͒.
1
1
1
. Y. Suzuki, H. Munakata, K. Kajihara, K. Kanamura, Y. Sato, K. Yamamoto, and T.
Yoshida, ECS Trans., 16͑26͒, 37 ͑2009͒.
. V. Thangadurai, J. Schwenzel, and W. Weppner, Ionics, 11, 11 ͑2005͒.
. X. Yu, J. B. Bates, G. E. Jellison, Jr., and F. X. Hart, J. Electrochem. Soc., 144, 524
2
3
14. M. Hara, H. Nakano, K. Dokko, S. Okuda, A. Kaeriyama, and K. Kanamura, J.
Power Sources, 189, 485 ͑2009͒.
͑
1997͒.
Downloaded on 2014-01-20 to IP 138.251.14.35 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use