A1108 Journal of The Electrochemical Society, 160 (8) A1108-A1111 (2013)

0013-4651/2013/160(8)/A1108/4/$31.00 © The Electrochemical Society

Effects of Electrode Density on the Safety of NCA Positive

Electrode for Li-Ion Batteries
Gu-Yeon Kim and J. R. Dahn∗,z
Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia B3H 3J5, Canada

The effect of electrode density on safety was investigated with Li[Ni0.8 Co0.15 Al0.05 ]O2 (NCA) electrodes in the presence of 1.0M
LiPF6 EC:DEC (1:2 v/v ratio) using accelerating rate calorimetry (ARC). The densities of the electrodes were 1.6 g/cm3 , 2.0 g/cm3 ,
and 2.3 g/cm3 , respectively. Denser electrodes showed inferior thermal stability compared to porous electrodes. ARC tests with
ground electrode powder in the presence of EC:DEC (1:2 v/v) solvent, with and without LiPF6 salt, show that the denser electrode
has lower thermal stability because it contains less LiPF6 . A charged NCA electrode in 0.7M LiBOB EC:DEC (1:2 v/v) electrolyte
shows worse thermal stability than in 1.0M LiPF6 EC:DEC (1:2 v/v) electrolyte. These results may be useful for researchers and
engineers interested in designing safe electrodes using NCA positive electrode material.
© 2013 The Electrochemical Society. [DOI: 10.1149/2.032308jes] All rights reserved.

Manuscript submitted March 19, 2013; revised manuscript received April 23, 2013. Published May 10, 2013.

LiCoO2 has been widely used as a cathode material, but its prac-
tical capacity is much less than its theoretical capacity of 270 mAh/g
because of the structural instability when it is charged above about
4.4 V.1,2 The attainable capacity is not enough to extend the op-
erational life of recent mobile electronics with larger displays and
more functions. Therefore, many battery manufacturers begin to
use Li[Ni0.8 Co0.15 Al0.05 ]O2 (NCA) as an alternative positive elec-
trode material, and it has been investigated widely by researchers.3–9
NCA shows a higher reversible specific discharge capacity of over
175 mAh/g,3 which leads to state-of-the-art Li ion batteries with high
capacity.
However, the results of safety tests such as nail penetration and
external short circuit show that thermal stability of typical cells,
like 18650 size cells, becomes worse as the amount of stored en-
ergy increases.10–12 An 18650-size cell with higher capacity reached
higher surface temperature after a nail penetration test, as reported by
Yamauchi.10 He showed, using nail penetration simulation modeling,
that when cells with different capacity were nailed under the same
conditions, the same short-circuit current flowed through the nail and
more time was needed for the cell with higher capacity to discharge
totally, which lead to more heat generation. Nevertheless, cells with
high capacity are in great demand. To achieve both high safety and
high capacity, many researchers and engineers in battery manufactur-
ing companies and research institutes have made great efforts.13–16
In order to design cells with high capacity, many factors have to
be considered. Among them, electrode dimensions such as thickness,
length, density, and porosity play an important role in the performance
and capacity of commercial Li ion cells. It is generally known that
thick and dense electrodes are needed for high capacity cells, and thin
and porous electrodes are needed for high power cells.4 The relation
between thermal stability and these electrode properties, however, has
not been yet investigated, as far as we know.
In this paper, we investigated the relation between thermal
stability and electrode density using accelerating rate calorimetry
(ARC). The effects of salts (LiPF6 and Lithium bis(oxalato)borate
(LiBOB)) on thermal stability was also explored. Commercial
Li[Ni0.8 Co0.15 Al0.05 ]O2 positive electrode material with an initial dis-
charge capacity of 170 mAh/g (at 0.1 C rate) was used for these
experiments.
Experimental
Li[Ni0.8 Co0.15 Al0.05 ]O2 was obtained from Toda Kogyo Corp.
(Japan). Cathode electrodes were coated on Al foil with NCA (88%
by mass), Super-S carbon black (6%), and polyvinylidene difluoride
(PVDF, 6%) in N-methyl pyrolidinone (NMP). The electrodes were
dried in a vacuum chamber at 120◦ C overnight before use and pressed
∗
Electrochemical Society Fellow.
z
E-mail: jeff.dahn@dal.ca
with a calendar roller with pressures of 1020 atm and 2078 atm.
Weight, thickness and density of these positive electrodes are shown
in Table I. 2325-size coin-type cells with these positive working elec-
trodes and a Li foil reference and counter electrode were assembled in
an Ar-filled glove box. The electrolyte used was 1M LiPF6 in EC:DEC
(1:2 v/v ratio). Then, the coin cells were charged to 4.2 V using the
protocol described in our previous work.17 The cycle life of the coin
cells was also tested with a computer controlled charge/discharge sys-
tem supplied by E-One Moli Energy Canada Ltd. Cells were tested
between cutoff voltages of 2.5 V and 4.2 V using 0.1 C rate current at
30◦ C ± 0.1◦ C.
The charged cells were opened in the Ar-filled glove box and
the positive electrodes were recovered to make ARC samples. The
weights of the electrodes were measured as soon as they were recov-
ered to determine the weight of electrolyte contained in the electrodes,
which is also shown in Table I. Then the electrodes (including the Al
foil) were immediately put into ARC tubes without additional elec-
trolyte. The ARC tubes are made from stainless steel (type 304) of
6.35 mm outer diameter, 0.015 mm wall thickness, and 39.1 mm
length (Micro-group, Medway, MA). ARC sample preparation pro-
cedures were same as described in our earlier work in detail.17,18 The
ARC starting temperature was set at 70◦ C or 150◦ C. Exothermic re-
actions were monitored under adiabatic conditions only when the self
heating rate (SHR) was higher than 0.03◦ C/min, and ARC tests were
stopped when the temperature of the samples exceeded 350◦ C or the
SHR was higher than 20◦ C/min.
Coin cells using pellet-type electrodes with the same composition
of NCA, Super-S carbon black and PVDF were also made and charged
using the same protocol to make ARC samples with powder, not
electrodes. The charged electrodes were recovered in the Ar-filled
glove box. Then the electrodes were ground and rinsed four times with
dimethyl carbonate (DMC) solvent to remove residual electrolyte from
the surface of ground powder and dried in the vacuum antechamber of
the glove box overnight to remove the DMC. Finally the dried powder
was put into the ARC tubes with 1M LiPF6 in EC:DEC (1:2 v/v),
0.7M LiBOB in EC:DEC (1:2 v/v), or EC:DEC (1:2 v/v) solvent and
then the ARC samples were welded closed.
Results and Discussion
Some physical and electrochemical properties of the electrodes
compressed with 0 atm, 1020 atm, and 2078 atm are shown in Table I.
The electrode densities which correspond to pressures of 0 atm, 1020
atm, and 2078 atm are 1.6 g/cc, 2.0 g/cc, and 2.3 g/cc, respectively, and
the amount of electrolyte contained in the electrode pores decreases
from 20 mg to 10 mg as the electrode density increases.
Fig. 1 shows the cycling test results for the various electrodes
between 2.5 V and 4.2 V using 0.1 C rate current at 30◦ C ± 0.1◦ C. The
initial discharge capacity increased from 165 mAh/g to 172 mAh/g and
the 1st cycle coulombic efficiency increased from 86.7% to 89.2% as

Downloaded on 2015-04-22 to IP 193.0.65.67 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, 160 (8) A1108-A1111 (2013) A1109

Table I. Basic physical and electrochemical properties of the Li[Ni0.8 Co0.15 Al0.05 ]O2 electrodes. The electrodes had an area of 1.27 cm2 .

an amount of initial initial specific 1st cycle

electrode electrode electrode electrolyte in specific charge discharge coulombic
weight (mg) thickness (um) density (g/cc) electrode (mg) capacity (mAh/g) capacity (mAh/g) efficiency
0 atm sample 48.2 226 1.6 20 191 166 86.7%
1020 atm sample 48.6 183 2.0 13 196 171 87.1%
2078 atm sample 49.3 163 2.3 10 193 172 89.2%

the electrode density increased.19 After about 10 cycles, the capacity Table II. The total heat generated during the ARC experiments
retention of the 2078 atm electrodes was slightly better than those of shown in Fig. 2b.
the 0 atm and 2078 atm electrodes.
Fig. 2 shows the ARC results for the various electrodes in the estimated Ctot mass of NCA heat
presence of 1M LiPF6 EC:DEC (1:2, v/v) at ARC starting tempera- Experimental Figure T (K) (J/K) (g) (J/g)
tures of: (a) 70◦ C and (b) 150◦ C. Fig. 2a shows that the SHR vs T
electrode of 0 atm 2 (b) 95 0.501 0.0472 1000
between 70◦ C and 150◦ C and the onset temperatures for a significant
electrode of 1020 atm 2 (b) 108 0.491 0.0468 1130
exothermic reaction of about 150◦ C are almost same for all the sam-
electrode of 2078 atm 2 (b) 90 0.486 0.0478 920
ples. All SHR results exceed the tracking rate of the ARC (20◦ C/min)
at about 230◦ C. These results indicate that the thermal stability of the
electrodes having different electrode density is almost the same when experiments, the heat capacity can be calculated as:
the starting temperature is 70◦ C.
Fig. 2b shows ARC results for identical electrodes to those used Ctot =  ci mi = 1.0 J/(gK) mcat + 0.46 J/(gK) 0.9 g + 1.5 J/(gK) mel
in Fig. 2a except that the ARC starting temperature was 150◦ C (du-
plicate cells are shown together). Increasing the starting temperature + 0.91 J/(gK) 0.0073 g
to 150◦ C results in a higher initial SHR of exothermic reactions be- where mcat , 0.9 g, mel , and 0.0073 g are the mass of positive electrode,
cause the “mini exotherms” observed below 150◦ C add up basically ARC tube, electrolyte, and Al current collector, respectively. The
to make a new large ARC peak, as shown in other reports.20,21,22 specific heats, 1.0 J/gK, 0.46 J/gK, 1.5 J/gK, and 0.91 J/gK correspond
Fig. 2b shows dramatically different behavior from Fig. 2a. The SHR to that of the positive electrode, ARC tube, electrolyte, and Al current
of the exothermic reactions between about 150◦ C and 220◦ C increases collector, respectively.23 The total heat generations listed in Table II
as the electrodes become denser. are all about 1000 J/g, which agrees with the results of Fig 2b that
Table II shows the total heat produced during the reaction which shows the exothermic reactions of all the samples complete at around
is given by:

h = Ctot T

where h is the total heat, T is the temperature rise, and the Ctot is
the total heat capacity of the entire ARC sample (electrode material,
solvent, Al current collector, plus stainless steel ARC tube). In our

Figure 1. Capacity versus cycle number for electrodes of 0 atm, 1020 atm, Figure 2. ARC results for the 0 atm, 1020 atm, and 2078 atm electrodes in
and 2078 atm in 1.0M LiPF6 EC:DEC (1:2 v/v ratio) between 2.5 V and 4.2 V 1.0M LiPF6 EC:DEC (1:2 v/v ratio) electrolyte with starting temperatures of:
using a 0.1 C rate current at 30◦ C ± 0.1◦ C. (a) 70◦ C and (b) 150◦ C.

Downloaded on 2015-04-22 to IP 193.0.65.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 A1110 Journal of The Electrochemical Society, 160 (8) A1108-A1111 (2013)
Figure 3. Enlarged initial SHR peaks for the experiments in Figure 2b.
260◦ C. Figure 2b indicates that cells with higher electrode density will
reach a high temperature more rapidly than those with lower electrode
density.
Fig. 3 shows the initial SHR peaks for the experiments in Fig 2b.
These initial peaks also increase as the electrode density increases.
These results suggest that cells with higher electrode density have
lower thermal stability vs. abuse conditions such as external/internal
short circuit and/or nail penetration above 150◦ C. Based on the results
in Fig. 2b and 3, it appears that the amount of the electrolyte contained
in the electrodes is related to the reactivity, so ARC samples were
fabricated with ground electrode powder and the same electrolyte
(1M LiPF6 EC:DEC (1:2, v/v)) in order to investigate the effects of
different amounts of electrolyte.
Fig. 4 shows ARC results that probe the effects of electrolyte
amount on thermal stability. The amount of ground electrode powder
used was the same as used in the experiments of Fig. 2 (about 49 mg),
and the amount of the electrolyte used was 10 mg or 20 mg, which
corresponds to electrodes pressed at 2078 atm or 0 atm, respectively,
as shown in Table I. Fig. 4 shows that the powder with 10 mg of elec-
trolyte is much more reactive between about 150◦ C and 220◦ C than the
powder with 20 mg of electrolyte and the powder without electrolyte.
This result suggests that a denser electrode is more reactive because it
contains less electrolyte. There are earlier reports about the effects of
electrolyte and LiPF6 salt on the reactivity of charged positive elec-
Figure 4. ARC results for ground electrode powder (49 mg) sealed in 0 mg,
10 mg, or 20 mg of 1.0M LiPF6 EC:DEC (1:2 v/v ratio) electrolyte with an
ARC starting temperature of 150◦ C.
Downloaded on 2015-04-22 to IP 193.0.65.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
Figure 5. ARC results for ground electrode powder (49 mg) sealed in 0 mg,
10 mg, or 20 mg of EC:DEC (1:2 v/v ratio) solvent with an ARC starting
temperature of 150◦ C.
trode material. MacNeil, Jiang, and Dahn reported that a larger amount
of electrolyte reduced reactivity because: 1) the heat capacity of ARC
sample increased with the mass of electrolyte21,23 and 2) the LiPF6
in the electrolyte suppressed the rate of reaction because polymeric
reaction products created by decomposition of LiPF6 formed on the
surface of the cathode material at elevated temperature and hindered
the release of the oxygen from the cathode material.20,23,24,25 There-
fore the SHR shown in Fig. 4 should decrease due to the addition of
solvent (increases heat capacity) and salt (slows reaction rate). Which
factor is the dominant one? To study this, ARC measurements were
made with only solvent added. 10 mg or 20 mg of EC:DEC (1:2 v/v)
solvent with the same amount of ground electrode powder (49 mg)
was used for these ARC tests.
Fig. 5 shows that samples without LiPF6 are very reactive and
exceed the ARC tracking rate at about 200◦ C, which is about 40◦ C
lower compared to experiments where LiPF6 was added. The slight
difference between the SHR for duplicate experiments with 10 mg
or 20 mg of EC:DEC (1:2 v/v) solvent is due to small differences in
the amount of solvent added. By comparing the ARC results of the
samples with 20 mg of solvent or 20 mg of electrolyte with LiPF6 salt,
one concludes that denser electrodes are more reactive because they
have less electrolyte; to be precise, less LiPF6 salt.
Additionally, the effects of another salt, LiBOB, on the reactivity
of NCA electrodes were investigated. Fig. 6 shows the ARC results for
Figure 6. ARC results for ground electrode powder (49 mg) sealed in 0 mg,
10 mg, or 20 mg of 0.7M LiBOB EC:DEC (1:2 v/v ratio) electrolyte with an
ARC starting temperature of 150◦ C.
 Journal of The Electrochemical Society, 160 (8) A1108-A1111 (2013)
electrodes in the presence of 10 mg or 20 mg of 0.7M LiBOB EC:DEC
(1:2 v/v) electrolyte. LiBOB makes the electrodes more reactive than
LiPF6 due to the high stability of the LiBOB.25 The results measured
for NCA in the presence of LiBOB are very similar to those measured
in Fig. 5 for NCA in the presence of pure solvent.
Conclusions
ARC tests were performed on NCA electrodes with different den-
sity; 1.6 g/cc, 2.0 g/cc, and 2.3 g/cc. The results show that the reac-
tivity becomes greater as the electrode density increases. The ARC
results using ground electrode powder suggest that this is because
a denser electrode contains less electrolyte, in particular less LiPF6 .
The amount of LiPF6 in the electrolyte plays a very important role
in reducing the reactivity because of the polymeric reaction products
formed on the surface of the cathode material by the reactions between
LiPF6 and solvent at high temperature.
The results shown here could be useful for researchers and en-
gineers who want to design and develop safe high capacity Li-ion
batteries. As an example, it is expected that higher molarities of LiPF6
will lead to safer cells with NCA positive electrodes, at least from a
cathode reactivity perspective.
References
1. R. V. Chebaim, F. Prado, and A. Manthiram, Chem. Mater., 13, 2951 (2001).
2. S. Venkatraman, Y. Shin, and A. Manthiram, Electrochem. Solid-State Lett., 6, A9
(2003).
Downloaded on 2015-04-22 to IP 193.0.65.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
A1111
3. S. H. Kang, W. S. Yoon, K. W. Nam, X. Q. Yang, and D. P. Abraham, J. Mater. Sci.,
43, 4701 (2008).
4. H. Y. Tran, G. Greco, C. Taubert, M. Wohlfahrt-Mehrens, W. Haselrieder, and
A. Kwade, J. Power Sources, 210, 276 (2012).
5. I. K. Hwang, C. W. Lee, J. C. Kim, and S. H. Yoon, Mater. Res. Bull., 47, 73
(2012).
6. D. J. Lee, B. Scrosati, and Y. K. Sun, J. Power Sources, 196, 7742 (2011).
7. Y. Wang, J. Jiang, and J. R. Dahn, Electrochem. Commun., 9, 2534 (2007).
8. C. Taubert, M. Fleischhammer, M. Wohlfahrt-Mehrens, U. Wietelmann, and
T. Buhrmester, J. Electrochem. Soc., 157, A721 (2010).
9. H. Y. Tran, C. Taubert, M. Fleischhammer, P. Axemann, L. Kuppers, and
M. Wohlfahrt-Mehrens, J. Electrochem. Soc., 158, A556 (2011).
10. T. Yamauchi, K. Mizushima, Y. Satoh, and S. Yamada, J. Power Sources, 136, 99
(2004).
11. T. Sato, K. Banno, T. Maruo, and R. Nozu, J. Power Sources, 152, 264 (2005).
12. K. Kitoh and H. Nemoto, J. Power Sources, 81, 887 (1999).
13. M. Kise, S. Yoshioka, and H. Kuriki, J. Power Sources, 174, 861 (2007).
14. T. Tsujikawa, K. Yabuta, M. Arakawa, and K. Hayashi, J. Power Sources, In Press.
15. G. H. Kim, K. Smith, J. Ireland, and A. Pesaran, J. Power Sources, 210, 243
(2012).
16. M. L. Lee, Y. H. Li, J. W. Yeh, and H. C. Shih, J. Power Sources, 214, 251
(2012).
17. J. Jiang, K. W. Eberman, L. J. Krause, and J. R. Dahn, J. Electrochem. Soc., 152,
A556 (2005).
18. J. Jiang and J. R. Dahn, Electrochim. Acta, 49, 4599 (2004).
19. T. Marks, S. Trussler, A. J. Smith, D. Xiong, and J. R. Dahn, J. Electrochem. Soc.,
158, A51 (2011).
20. D. D. MacNeil and J. R. Dahn, J. Electrochem. Soc., 150, A21 (2003).
21. J. Jiang, J. Chen, and J. R. Dahn, J. Electrochem. Soc., 151, A2082 (2004).
22. J. Jiang and J. R. Dahn, Electrochem. Commun., 6, 39 (2004).
23. D. D. MacNeil and J. R. Dahn, J. Electrochem. Soc., 148, A1205 (2001).
24. D. D. MacNeil and J. R. Dahn, J. Electrochem. Soc., 149, A912 (2002).
25. J. Jiang and J. R. Dahn, Electrochim. Acta, 49, 2661 (2004).