Thermal reactions between delithiated lithium nickelate and electrolyte solutions

Article abstract of DOI:10.1149/1.1452114

Lithium nickelate is a promising electrode material for lithium-ion batteries, however, its low thermal stability is a problem that should be solved. We studied the thermal stability of delithiated lithium nickelate compounds in the presence of electrolyte solvents/solutions. We measured the exothermal heat generation of the reaction system using differential scanning calorimetry. We examined simple systems such as ethylene carbonate with Li_0.2NiO₂, using hermetically sealed pans that allowed us to perform the experiments under constant mass conditions. The data were quantitatively analyzed to propose possible reaction mechanisms for the exothermal behavior. We attributed the main source of the exothermal heat to organic solvent combustion caused by oxygen released from the nickelate. We studied similar reactions using other solvents/solutions and electrochemically or chemically delithiated nickelate samples. We also examined the behavior of lithium cobaltate for comparison. We discuss the effect of partial substitution for nickel on thermal stability.

Full text of DOI:10.1149/1.1452114

Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
A401
0
013-4651/2002/149͑4͒/A401/6/$7.00 © The Electrochemical Society, Inc.
Thermal Reactions Between Delithiated Lithium Nickelate
and Electrolyte Solutions
,z
Hajime Arai,* Masayuki Tsuda, Keiichi Saito, Masahiko Hayashi,
and Yoji Sakurai*
NTT Telecommunications Energy Laboratories, Tokai, Ibaraki 319-1193, Japan
Lithium nickelate is a promising electrode material for lithium-ion batteries, however, its low thermal stability is a problem that
should be solved. We studied the thermal stability of delithiated lithium nickelate compounds in the presence of electrolyte
solvents/solutions. We measured the exothermal heat generation of the reaction system using differential scanning calorimetry. We
examined simple systems such as ethylene carbonate with Li0.2NiO , using hermetically sealed pans that allowed us to perform
2
the experiments under constant mass conditions. The data were quantitatively analyzed to propose possible reaction mechanisms
for the exothermal behavior. We attributed the main source of the exothermal heat to organic solvent combustion caused by oxygen
released from the nickelate. We studied similar reactions using other solvents/solutions and electrochemically or chemically
delithiated nickelate samples. We also examined the behavior of lithium cobaltate for comparison. We discuss the effect of partial
substitution for nickel on thermal stability.
©
2002 The Electrochemical Society. ͓DOI: 10.1149/1.1452114͔ All rights reserved.
Manuscript submitted March 6, 2001; revised manuscript received October 31, 2001. Available electronically February 25, 2002.
Lithium nickelate is an attractive positive electrode material for
lithium-ion batteries because of its comparatively low cost and
large capacity. However, its stability must be improved to meet
ing the cell at 0.1 mA cmϪ2. We estimated the lithium content of the
delithiated compounds ͑x in Li NiO or Li CoO ͒ using the trans-
x
2
x
2
1
-3
ferred charge and the weight of the electrode. For example, when a
Ϫ1
safety standards, because highly delithiated compounds Li NiO
(
charge of 220 mAh ͑g-oxide͒ passed for an LiNiO electrode ͑275
x
2
2
xϹ0.3) decompose exothermally at around 200°C.4 It has re-
-6
mAh g for one-electron reaction͒, we expected the product to be
Ϫ1
cently been reported that there is even greater exothermal heat gen-
eration when the nickelate material is heated in the presence of
organic electrolyte solutions.7 However, the origin of this large
heat generation has remained unclear because of the complexity of
the reaction system, which contains many components such as elec-
trode material, electrolyte solute, and mixed organic solvents. For
Li0.2NiO . We sampled an electrode with a nominal composition of
2
Li0.2NiO2 and conducted the chemical analysis using inductively
coupled plasma emission ͑ICPE͒ spectroscopy. The analysis indi-
cated that the Li/Ni ratio was 0.23 Ϯ 0.03, suggesting that our esti-
mation was valid within the margin of experimental error. After the
charging process was stopped at desired values, we allowed the
electrodes to relax for 1 week in the cell and then removed them. We
washed the electrodes with DMC, and then dried them in a vacuum
at room temperature.
-15
examining such thermal behavior, differential scanning calorimetry
-17
͑
DSC͒ has been widely used.5 As DSC shows the amount of gen-
erated heat quantitatively, a study of simple reaction systems, for
example, oxide with a single solvent, may be the key to discovering
the reaction mechanism.
We prepared chemically delithiated samples as follows. We
Ϫ3
added the LiNiO sample to a 0.6 mol dm aqueous solution of
2
ϩ
In this study we report the thermal behavior of Li NiO in con-
sulfuric acid. The initial molar ratio of H /LiNiO was 5. We stirred
x
2
2
tact with electrolyte solutions using DSC coupled with thermo-
gravimetry ͑TG͒. We examined reactions involving ethylene carbon-
ate ͑EC͒ in particular detail, because EC is often used as the main
component in electrolyte solutions. We tested both electrochemically
and chemically delithiated compounds and compared their behavior
with that of lithium cobaltate. We also discuss the effect of partial
substitution for nickel on the thermal stability of the delithiated
material.
the mixture at 25°C for 5 h and then filtered the product. We washed
the filtrate powder with acetone to remove the water on the surface
and then dried it in a vacuum at room temperature. The sample had
a chemical composition of Li0.1NiO , as determined by ICPE analy-
2
6
sis, and was electrochemically active.
We measured the DSC-TG behavior ͑Rigaku TAS-100͒ in an
Ϫ1
argon atmosphere at a heating rate of 10°C min . For measuring
the amount of heat based on reactant weights, it was necessary to
use hermetically sealed ͑HS͒ stainless steel pans to prevent solvent
volatilization. We were able to perform the experiments under con-
stant mass conditions up to 400°C, with a mixture of a few milli-
grams of the delithiated sample and electrolyte solvent/solution. We
occasionally used crimp sealed ͑CS͒ aluminum pans to test the re-
action under more practical pressure conditions. The HS and CS
pans can withstand internal pressures of 50 and 3 atm, respectively.
The use of the CS pans allowed us to characterize the structure of
the product after DSC-TG measurement, using X-ray diffraction
Experimental
We prepared nearly stoichiometric LiNiO samples by heating a
2
mixture of LiOH•H O, LiNO , and Ni͑OH͒ ͑1:1:1 in molar ratios͒
2
3
2
in air at 500°C for 6 h and then at 700°C for 24 h. We removed
unreacted lithium compounds by washing the heated product with
distilled water. The structural and electrochemical characteristics of
this compound are detailed elsewhere.18 We used LiCoO powder
2
͑
Nippon Chemical Industrial͒ as received. All the electrolyte solu-
͑XRD͒, ͑Rigaku RU-200, RAD-rX͒ with Cu K␣ radiation. We ana-
tions and organic solvents were of battery grade ͑Tomiyama Pure
Chemical Industries͒ and contained less than 20 ppm of water.
lyzed the water evolution from the reaction system using the Karl-
Fischer method ͑Mitsubishi Chemicals CA-06͒. We placed the
sample on a heating boat in a closed tube in which there was a dry
argon gas flow. The argon gas carried the water that evaporated from
the sample to a water detector.
The electrochemical cell we used contained a metallic lithium
Ϫ3
counter electrode and a 1 mol dm LiPF electrolyte solution in
6
equal volumes of EC and dimethyl carbonate ͑DMC͒. The working
electrode mixture consisted of the oxide material ͑70 wt %͒, acety-
lene black ͑AB, 25 wt %͒, and polytetrafluoroethylene ͑PTFE, 5 wt
Results and Discussion
%͒. We prepared electrochemically delithiated compounds by charg-
Reactions between electrochemically obtained Li NiO2 and
x
EC.—Figure 1 shows the DSC behavior of EC and electrochemi-
cally obtained Li0.2NiO that we tested in HS pans. When we heated
only EC, we observed endothermal heat generation at 270-330°C,
*
Electrochemical Society Active Member.
E-mail: arai@atsu.rde.ntt.co.jp
2
z
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A402
Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
Figure 2. Observed amount of heat based on ͑a͒ Li0.2NiO or ͑b͒ EC weight
2
Figure 1. DSC behavior of EC and electrochemically obtained Li0.2NiO2
tested in HS pans: ͑a͒ EC alone, ͑b͒ Li0.2NiO2 alone, and ͑c͒ EC and
for different EC/Li0.2NiO ratios.
2
Li0.2NiO ͑1:1 in weight ratio͒. The inset in the figure shows a benchmark for
2
the amount of heat.
temperature of 200°C, where the exothermal reaction certainly oc-
curs ͑see Fig. 1c͒ and we can minimize the EC evaporation that
occurs above 170°C ͑see below͒. While water was hardly detected
on heating EC alone, we detected a considerable amount of water
which we ascribed to EC vaporization under the HS conditions. The
Li0.2NiO electrode mixture, containing AB and PTFE, had a small
2
when we heated the mixture of EC and Li0.2NiO , suggesting that
2
exothermal peak at 170-230°C. When we mixed these two compo-
nents of approximately equal weights and heated them, much more
exothermal heat was generated than with nickelate alone.
To clarify the origin of the heat generation, we measured the
amount of exothermal heat with various reactant weight ratios of
EC mainly decomposes in accordance with Reaction 3.
Reaction 3 seems plausible also because the reaction heat was
proportional to the nickelate weight. We calculated the expected
amount of heat based on Reaction 3 and compared it with the ex-
perimentally observed value to confirm the validity of this sugges-
EC/Li0.2NiO . An exothermal peak, similar to that in Fig. 1c, was
2
tion. Reaction 1 indicates that when we heat 1 g Li0.2NiO , 0.14 g
2
observed in all the DSC profiles. The amount of the exothermal heat
oxygen evolves. This amount of oxygen can burn 0.15 g of EC,
͑
at 170-230°C͒ was nearly proportional to the nickelate weight
according to Reaction 3. The EC combustion heat is reported to be
rather than the EC weight, as shown in Fig. 2, and the value was
about 1.6 kJ ͑g-oxide͒ . This suggests that the large heat generation
Ϫ1 23
1
2
3 kJ ͑g-EC͒
.
Consequently, 1 g Li NiO generates a heat of
0.2 2
Ϫ1
Ϫ1
.0 kJ. This amount, 2.0 kJ ͑g-oxide͒ , is close to the value that we
shown in Fig. 1c chiefly originates from nickelate-related reactions,
which then induces reactions involving EC. It is therefore important
to understand how the nickelate alone decomposes. According to
observed in the DSC measurements ͑see Fig. 2͒, supporting the idea
that Reaction 3 is appropriate.
There are several possible explainations for the difference be-
tween the expected and experimental values. The experimental data
contain heat caused by the self-decomposition of the oxide ͑exother-
mal͒, oxygen vaporization ͑endothermal͒, and EC vaporization ͑en-
dothermal͒. We used data on combustion heat in the standard ther-
modynamic state ͑1 atm, 25°C͒, which were the only data available.
However, they may be invalid at 200°C under HS conditions. There
would be a smaller amount of oxygen at around 200°C than ex-
pected by Reaction 1, because the oxygen evolves at temperatures
1
,4,5
previous reports,
the decomposition follows Reaction 1, result-
ing in the formation of a more stable nickelate phase with a rock-salt
structure and oxygen
Li NiO → ͑1 ϩ x͒Lix/͑1ϩx͒Ni1/͑1ϩx͒O ϩ ͑1 Ϫ x͒/2 O2 ͓1͔
x
2
Based on this nature of nickelate, there are two possible reactions
that could cause a large amount of heat to be generated in the pres-
ence of EC. Reaction 2 is the catalytic decomposition of EC to form
carbon dioxide and ethylene oxide,19 which could be initiated by the
exothermal heat arising from Reaction 1. Reaction 3 is EC combus-
tion, where the oxygen from the nickelate could oxidize EC to form
carbon dioxide and water
6
higher than 190°C up to 500°C. EC would be directly oxidized by
the delithiated nickelate ͑Reaction 4͒, rather than by the released
oxygen ͑Reaction 3͒
25Li NiO ϩ 4C H O → 30Li1/6Ni5/6O ϩ 12CO ϩ 8H O
x
2
3
4
3
2
2
͓
4͔
C H O → CO ϩ C H O
3
͓2͔
4
3
2
2
4
Nevertheless, we concluded that the exothermal heat generation can
be attributed mainly to Reactions 1 and 3 because they can satisfac-
torily account for the observed data.
5
C H O ϩ O → 3CO ϩ 2H O
3
͓3͔
4
3
2
2
2
2
Nickel compounds are generally catalytic and their surfaces
known to be active with regard to electrolyte decomposition.2
Accordingly, Reaction 3 may also be catalyzed in the presence of
the nickelate surface. In fact, nickelate surface modification is partly
effective in suppressing the exothermal heat generation.13 However,
1,24
Carbon dioxide is commonly observed as a reaction ͑particularly
oxidation͒ product of organic solvents2
0-22
and could be observed in
both Reactions 2 and 3. We examined possible water production
using the Karl-Fischer method. The reaction boat was kept at a fixed
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
A403
Figure 4. DSC behavior of Li0.2NiO2 with ͑a͒ EC, ͑b͒ PC, ͑c͒ gBL, ͑d͒
DEC, and ͑e͒ DMC ͑see Table I for abbreviations͒.
Figure 3. DSC behavior of EC with ͑a͒ electrochemically delithiated
Li0.2NiO and ͑b͒ chemically delithiated Li0.1NiO2 .
2
decomposition product such as PF ͒ with gaseous EC, because EC
there was no exothermal reaction when we heated EC with NiO or
5
evaporation occurs in the HS pan at this temperature ͑see Fig. 1a͒. A
LiNiO . We therefore believe that oxygen released from the nick-
2
17
similar result has recently been reported. A large exothermal peak
elate is principally responsible for the large exothermal heat genera-
tion shown in Fig. 1c.
at 210°C in Fig. 5b can be attributed to the combustion of the or-
ganic solvent because the phenomenon is similar to that in Fig. 1c.
There was additional heat at around 250°C in the presence of both
Li0.2NiO and LiPF ; nevertheless, the combustion of the organic
Reactions between chemically obtained Li NiO and EC.—We
x
2
subsequently observed the thermal behavior of chemically delithi-
2
6
ated Li0.1NiO and EC using the HS pan. This allowed us to clarify
how conductive agent ͑AB͒ and binder ͑PTFE͒ contribute to the
solvent that occurs at a lower temperature is more important in
terms of battery safety.
2
reaction. The DSC behavior was similar to that of electrochemically
Reactions under ambient pressure using CS pans.—Figure 6
shows the thermal behavior of the electrochemically delithiated
delithiated Li0.2NiO , as shown in Fig. 3. The exothermal heat of
2
Ϫ1
1
.8 kJ ͑g-oxide͒ was again close to the expected value of 2.3 kJ
Ϫ1
Li0.2NiO with EC that we tested in the HS and CS pans. While the
2
͑
g-oxide͒ , which we calculated from Reactions 1 and 3 with
weight remained unchanged with the HS pan at temperatures up to
00°C, there was a weight loss with the CS pan at temperatures
x ϭ 0.1. We therefore concluded that these electrode additives have
essentially no effect on the exothermal behavior at around 200°C.
4
above 170°C ͑see TG behavior shown in the figure͒, indicating that
the EC vaporization occurred before the exothermal heat was gen-
erated. However, the amount of heat was similar to that observed
with the HS pans, as shown. When we tested the reaction between
Li0.2NiO2 and DMC in the CS pan, we observed much less exother-
mal heat. This was because there was no DMC left in the pan at
200°C, as deduced from the TG data, and we actually observed the
thermal behavior of Li0.2NiO2 alone, like that shown in Fig. 1b.
These results suggest that EC with a high boiling point ͑248°C at an
ambient pressure͒ can remain in the battery at the nickelate decom-
position temperature, even if the battery is equipped with a safety
vent. Accordingly, reactions involving EC are important in terms of
battery safety.
Reactions of various solvents and solutions.—We examined the
thermal behavior of other organic solvents with electrochemically
delithiated Li0.2NiO using the HS pan. Figure 4 shows the DSC
2
behavior. Although the peak shapes are somewhat different from
each other, the exothermal peak appeared at around 200°C regard-
less of the solvent used, suggesting that the reactions originated
from the nickelate decomposition reaction. Table I summarizes the
amount of generated heat, which was close to that expected for the
corresponding solvent combustion. Although we have no data avail-
able for propylene carbonate ͑PC͒ combustion, the observed value
seems reasonable because PC is structurally similar to EC. It is
Ϫ1
noteworthy that the amount of heat is nearly 2 kJ (g-Li0.2NiO2)
for all the cases we studied. In general, organic compounds consist-
ing of many carbon/hydrogen atoms ͑hence having a large molecular
weight͒ exhibit large combustion heat but require a large amount of
oxygen for combustion. In this study the available amount of oxy-
gen was limited by Reaction 1; accordingly, the heat produced was
almost the same regardless of the solvent.
We analyzed the XRD pattern of solid residues after the DSC-TG
measurement when we used the CS pans. The residue after heating
Li0.2NiO with EC up to 400°C was mostly Li Ni O (z Ϸ 5/6)
with a rock salt structure, as expected. However, when we used a
relatively small amount of EC compared with the nickelate ͑typi-
2
1Ϫz
z
2
8
cally EC/Li0.2NiO Ӷ 1 in the weight ratio͒, there was a by-product
2
To clarify the influence of the electrolyte solute LiPF6 on tϪhe3
of metallic nickel ͑Ni͒ together with the Li1ϪzNi O phase. Figure 7b
z
thermal properties, we measured the DSC behavior of a 1 mol dm
shows the XRD pattern of the residue for EC/Li0.2NiO ϭ 0.5, dem-
2
LiPF solution in equal volumes of EC and DMC with and without
onstrating a few peaks corresponding to Ni. Using the intensities of
these peaks, we estimated the weight ratio in the product to be
6
Li0.2NiO . We set the nickelate-to-solution weight ratio at unity. As
2
shown in Fig. 5, there was an exothermal peak at around 280°C in
Li1ϪzNi O:Ni ϭ 9:1. Figure 8 provides the corresponding DSC
z
both cases. This can be attributed to the reaction of LiPF ͑or its
curves. The observed amount of heat was around 1.8 kJ
6
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A404
Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
Table I. Expected and observed amounts of heat for reactions between Li0.2NiO and organic solvents.
2
Heat of
combustion
/kJ ͑g-solvent͒
Moles of O used
for combustion
Expected heat
/kJ (g-Li0.2NiO2)
Observed heat
/kJ (g-Li0.2NiO2)
2
Ϫ1
Ϫ1
Ϫ1
Solvent
Ethylene carbonate ͑EC͒
Propylene carbonate ͑PC͒
13 ͑Ref. 23͒
2.5
4
4.5
6
2.0
-
1.8
1.8
2.1
1.6
2.0
1.9
2.0
2.1
͑Not available͒
22 ͑Ref. 25͒
21 ͑Ref. 26͒
16 ͑Ref. 27͒
␥
-Butyrolactone ͑gBL͒
Diethyl carbonate ͑DEC͒
Dimethyl carbonate ͑DMC͒
3
(
g-Li0.2NiO2)Ϫ1 in both cases, showing that Ni formation has only a
g Li0.5CoO . This value provides the expected amount of combus-
tion heat of 0.8 kJ ͑g-oxide͒ , which again nearly accords with the
2
Ϫ1
minor effect on the overall reaction mechanism. Nevertheless, when
EC/Li0.2NiO ϭ 0.5, the DSC peak was very sharp as shown in Fig.
experimental data
2
8
b, and this phenomenon probably corresponds to the Ni formation.
1
1
1
The detailed mechanism for this Ni formation is unclear at the mo-
ment. A possible explanation is that when the EC/Li0.2NiO ratio is
Li0.5CoO → LiCoO ϩ Co O ϩ O2
͓5͔
2
2
3
4
2
6
6
2
small, the nickelate-catalyzed reaction ͑Reaction 2͒ proceeds and
unstable ͑explosive͒ ethylene oxide is formed as an intermediate
species, which then reduces the nickelate to Ni. This hypothesis
certainly requires further study. As we never observed the formation
Recent work using accelerating rate calorimetry has also shown that
30
solvent combustion occurs with oxygen released from Li CoO
x
2
3
1
and also from Li Mn O .
x
2
4
of Ni when we heated Li NiO2 alone, it is at least clear that Ni
Figure 10 shows the voltage-lithium concentration profiles of
x
formation requires a reducing agent such as EC.
Li NiO and Li CoO electrodes during delithiation at a constant
x
2
x
2
Ϫ2
It has been reported that Li NiO turns into Li Ni O via an
x
2
1Ϫz
z
current density of 0.1 mA cm . As shown, Li0.2NiO and Li0.5CoO2
were obtained at almost the same potential of 4.2 V vs. Li/Li .
2
intermediate phase with a spinel-related structure.1,4,29 To observe
the effect of this spinel phase formation, we measured the DSC
behavior of Li0.2NiO2 and EC using CS pans and interrupted the
measurement at different temperatures. However, the products were
ϩ
Nevertheless, considerably more heat was generated for the former
than the latter. This is because the nickelate is readily delithiated and
reaches high oxidation states, resulting in more oxygen generation
on heating than with the cobaltate. Moreover, when we adopt
x ϭ 0.5 in Reaction 1, we obtain the following
either LiϪzNi O or remaining reactants, and no spinel-related phase
z
was seen. This is probably because, in the presence of reducing
agents, Li1ϪzNi O is more stable than the spinel-related phase.
z
3
1
Li0.5NiO → 2 Li1/3Ni2/3O ϩ O2
͓6͔
2
Stability of Li NiO and Li CoO .—We examined the thermal
x
2
x
2
4
behavior of EC with delithiated lithium cobaltate comparatively.
Li0.5CoO was prepared by charging the LiCoO electrode. When
2
2
Reactions 5 and 6 show that even if the same amount ͑50%͒ of
lithium is extracted, the cobaltate produces less oxygen, and hence,
less combustion heat, than the nickelate. Figure 9 shows another
important feature of the cobaltate, namely, that it has a higher reac-
we heated the mixture of Li0.5CoO and EC in the HS pan, there was
2
Ϫ1
exothermal heat of 0.9 kJ ͑g-oxide͒ with a heat peak at 250°C, as
shown in Fig. 9. Li CoO is known to decompose in accordance
0
.5
2
4
with Reaction 5, leading to 0.06 g of oxygen being released from 1
Figure 5. DSC behavior of 1 mol dmϪ3 LiPF solution in DMC ϩ EC ͑1:1͒
6
Figure 6. Thermal behavior of EC and Li0.2NiO : ͑a͒ DSC behavior for the
2
͑a͒ without and ͑b͒ with Li0.2NiO2 .
HS pan; ͑b͒ DSC and ͑c͒ TG behavior for the CS pan.
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
A405
Figure 9. DSC behavior of EC with ͑a͒ Li0.5CoO and ͑b͒ Li NiO in the
HS pans.
2
0.2
2
Figure 7. XRD patterns of residues for EC/Li0.2NiO ϭ (a) 1.0 and ͑b͒
2
0
.5: ͑᭺͒ Li1/6Ni5/6O and ͑᭹͒ metallic nickel peaks.
It is thus desirable to improve the insufficient thermal stability of
Li NiO . Partial substitution for nickel seems to be a promising
approach, as several research groups have proposed useful
x
2
tion temperature than the nickelate. This temperature is as high as
the boiling point of EC. In fact, when we heated the cobaltate and
EC using the CS pan, we observed no EC combustion heat but only
a small amount of exothermal heat caused by Reaction 5. This sug-
gests that it is possible to avoid the EC combustion caused by the
7
14
10,15
substituents such as aluminum, cobalt, aluminum-cobalt,
1
2
9
32
cobalt-manganese, magnesium-titanium, and others. Of these,
cobalt substitution has been well studied because the cobaltate is
clearly more stable than the nickelate as discussed previously. We
have recently shown that manganese or titanium-substituted lithium
nickelate exhibits endothermal behavior when dilithiated and heated
Li0.5CoO decomposition if the battery has a safety vent allowing
2
the release of gaseous EC. These characteristics account for the
insufficient safety of nickelate electrodes when compared with co-
3
3,34
4
,5,8,9
without electrolyte solutions.
In our preliminary experiments
baltate electrodes.
these compounds unfortunately exhibited exothermal behavior when
heated with electrolyte solutions, however, some had a smaller exo-
Figure 8. DSC behavior for ͑a͒ EC/Li0.2NiO ϭ 1 and ͑b͒ EC/Li0.2NiO2
ϭ 0.5 in the CS pans.
Figure 10. Profiles of ͑a͒ Li NiO2 and ͑b͒ Li CoO2 electrodes during
delithiation at a constant current density of 0.1 mA cm .
2
x
x
Ϫ2
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A406
Journal of The Electrochemical Society, 149 ͑4͒ A401-A406 ͑2002͒
thermal peak with a higher reaction temperature than the nonsubsti-
tuted nickelate. We are currently unable to explain the effect of these
substituents; nevertheless, we believe it is important to select sub-
stituents that can suppress solvent combustion.
6. H. Arai and Y. Sakurai, J. Power Sources, 81-82, 401 ͑1999͒.
7
. T. Ohzuku, T. Yanagawa, M. Kouguchi, and A. Ueda, J. Power Sources, 68, 131
1997͒.
͑
8
9
. Z. Zhang, D. Fouchard, and J. R. Rea, J. Power Sources, 70, 16 ͑1998͒.
. Y. Gao, M. V. Yokovleva, and W. B. Ebner, Electrochem. Solid-State Lett., 1, 117
͑
1998͒.
0. M. Nagata, S. Komatsu, H. Tsukamoto, and M. Mizutani, GS News Tech. Rep., 57,
5 ͑1998͒.
1. Ph. Biensan, B. Simon, J. P. Peres, A. de Guibert, M. Broussely, J. M. Bodet, and
Conclusions
1
1
Delithiated lithium nickelate generates a large amount of exo-
thermal heat when heated with organic electrolyte. We attributed the
main origin of the heat to the organic solvent combustion caused by
oxygen released from the nickelate, while the electrolyte solute
1
F. Perton, J. Power Sources, 81-82, 906 ͑1999͒.
1
2. K. Honbo, Y. Muranaka, and F. Kita, in Extended Abstracts of the 40th Battery
Symposium in Japan, p. 243, Kyoto ͑Nov 1999͒.
LiPF has a minor effect on the thermal behavior. The amount of
13. H-J. Kweon and D. G. Park, Electrochem. Solid-State Lett., 3, 128 ͑2000͒.
6
1
4. J. Cho, H. Jung, Y. Park, G. Kim, and H. S. Lim, J. Electrochem. Soc., 147, 15
2000͒.
heat is largely independent of the solvent because the amount of
released oxygen is governed by the amount of nickelate. Reactions
involving EC are of particular importance with regard to battery
͑
1
5. M. Kanda, Y. Tatebayashi, M. Sekino, Y. Isozaki, and I. Mitsuishi, The 10th Inter-
national Meeting on Lithium Batteries Abstract 231, Como, Italy, May 28-June 2,
2000.
safety due to its high boiling point. Li NiO2 is reduced to
x
1
6. A. Du Pasquier, F. Disma, T. Bowmer, A. S. Gozdz, G. Amatucci, and J. M.
Tarascon, J. Electrochem. Soc., 145, 472 ͑1998͒.
Lix/(1ϩx)Ni1/(1ϩx)O and occasionally also to metallic nickel. Solvent
combustion also occurs when heated with delithiated lithium cobal-
tate, but with less heat and a high reaction temperature. The insuf-
ficient thermal stability of the nickelate should be improved for use
in practical applications, and trials with partial substitution for
nickel are underway.
1
1
7. G. G. Botte, R. E. White, and Z. Zhang, J. Power Sources, 97-98, 570 ͑2001͒.
8. H. Arai, S. Okada, H. Ohtsuka, M. Ichimura, and J. Yamaki, Solid State Ionics, 80,
261 ͑1995͒.
19. Yozai Handbook ͑Solvent Handbook, in Japanese͒ T. Asakura, N. Tokura, M.
Okawara, J. Kumanotani, and M. Senoo, Editors, p. 627, Kodansha, Tokyo ͑1982͒.
2
2
2
0. M. Arakawa and J. Yamaki, J. Power Sources, 54, 250 ͑1995͒.
1. R. Imhof and P. Nov a´ k, J. Electrochem. Soc., 146, 1702 ͑1999͒.
2. K. Kumai, H. Miyashiro, Y. Kobayashi, K. Takei, and R. Ishikawa, J. Power
Sources, 81-82, 715 ͑1999͒.
Acknowledgments
The authors express their gratitude to Dr. Ichiro Yamada and
Kazuhiko Komatsu for their encouragement during the course of
this research. The authors are also grateful to Noriko Kurusu and
Ryuichi Nishio ͑both of NTT Advanced Technology Company͒ for
their excellent technical support.
23. J. D. Cox and G. Pilcher, Thermochemistry of Organic and Organometallic Com-
pounds, p. 252, Academic Press, London ͑1970͒.
2
2
2
4. D. Aurbach, K. Gamolsky, B. Markovsky, G. Salitra, Y. Gofer, U. Deider, R. Oes-
ten, and M. Schmidt, J. Electrochem. Soc., 147, 1322 ͑2000͒.
5. J. M. Brown, A. D. Conn, G. Pilcher, M. L. P. Leitao, and M. Yang, J. Chem. Soc.
Chem. Commun., 23, 1817 ͑1989͒.
˜
NTT Telecommunications Energy Laboratories assisted in meeting the
publication costs of this article.
6. Kagaku Binran II ͑Chemistry Handbook, in Japanese͒, The Chemical Society of
Japan, p. 316, Maruzen, Tokyo ͑1988͒.
2
2
2
7. Swietoslawski, J. Am. Chem. Soc., 42, 1099 ͑1920͒.
References
8. J. N. Reimers, W. Li, and J. R. Dahn, Phys. Rev. B, 47, 8486 ͑1993͒.
9. R. Kanno, H. Kubo, Y. Kawamoto, T. Kamiyama, F. Izumi, Y. Takeda, and M.
Takano, J. Solid State Chem., 110, 216 ͑1994͒.
1
2
. M. G. S. R. Thomas, W. I. F. David, and J. B. Goodenough, Mater. Res. Bull., 20,
137 ͑1985͒.
. J. R. Dahn, U. von Sacken, M. W. Juzkow, and H. Al-Janaby, J. Electrochem. Soc.,
38, 2207 ͑1991͒.
. T. Ohzuku, A. Ueda, and M. Nagayama, J. Electrochem. Soc., 140, 1862 ͑1993͒.
. J. R. Dahn, E. W. Fuller, M. Obrovac, and U. von Sacken, Solid State Ionics, 69,
1
3
0. D. D. MacNeil and J. R. Dahn, J. Electrochem. Soc., 148, A1205 ͑2001͒.
1
31. D. D. MacNeil and J. R. Dahn, J. Electrochem. Soc., 148, A1211 ͑2001͒.
32. M. Broussely, P. Biensan, and B. Simon, Electrochim. Acta, 45, 3 ͑1999͒.
33. H. Arai, S. Okada, Y. Sakurai, and J. Yamaki, J. Electrochem. Soc., 142, 3117
͑1997͒.
3
4
2
65 ͑1994͒.
5
. H. Arai, S. Okada, Y. Sakurai, and J. Yamaki, Solid State Ionics, 109, 295 ͑1998͒.
34. H. Arai, M. Tsuda, and Y. Sakurai, J. Power Sources, 90, 76 ͑2000͒.
Downloaded on 2015-03-23 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).