Improvement of the electrochemical properties of Li [Ni0.5 Mn0.5] O2 by AlF3 coating

Article abstract of DOI:10.1149/1.2956088

An ～10 nm AlF3 layer was coated on the surface of a Li [Ni _0.5Mn_0.5]O₂ positive electrode material for lithium-ion batteries, and the effects of this coating on battery performances and thermal stability of the cathode materials were studied. Although no significant bulk structural differences were observed between the coated and pristine material, a slightly higher capacity was seen for the AlF3 -coated Li [Ni_0.5Mn_0.5]O₂ electrode, and the rate capability was also greatly enhanced by the AlF3 coating. These improvements are mainly attributed to the suppression of the transition metal dissolution benefited from the AlF3 coating. This suppression contributed to the reduction in the charge-transfer resistance. Time-of-flight secondary ion mass spectroscopic analysis showed that insulating LiF, as a product of decomposed LiPF₆, was deposited on the surface of pristine and AlF3 -coated Li [Ni_0.5Mn_0.5]O₂. The deposition of LiF was greatly suppressed by AlF₃ coating on the outer surface of Li [Ni_0.5Mn_0.5]O₂. The protection of the active material by the AlF₃ coating substantially improved the capacity, capacity retention, and rate capability of the batteries. It also enhanced the thermal stabilities of the positive electrode material.

Full text of DOI:10.1149/1.2956088

Journal of The Electrochemical Society, 155 ͑10͒ A705-A710 ͑2008͒
A705
0013-4651/2008/155͑10͒/A705/6/$23.00 © The Electrochemical Society
Improvement of the Electrochemical Properties of
Li†Ni0.5Mn0.5‡O2 by AlF3 Coating
Yang-Kook Sun,a, Seung-Taek Myung,b Byung-Chon Park,a and
,z
*
Hitoshi Yashirob,
*
aDepartment of Chemical Engineering, Hanyang University, Seoul 133-791, South Korea
bDepartment of Chemical Engineering, Iwate University, Morioka, Iwate 020-8551, Japan
An ϳ10 nm AlF3 layer was coated on the surface of a Li͓Ni0.5Mn0.5͔O2 positive electrode material for lithium-ion batteries, and
the effects of this coating on battery performances and thermal stability of the cathode materials were studied. Although no
significant bulk structural differences were observed between the coated and pristine material, a slightly higher capacity was seen
for the AlF3-coated Li͓Ni0.5Mn0.5͔O2 electrode, and the rate capability was also greatly enhanced by the AlF3 coating. These
improvements are mainly attributed to the suppression of the transition metal dissolution benefited from the AlF3 coating. This
suppression contributed to the reduction in the charge-transfer resistance. Time-of-flight secondary ion mass spectroscopic analysis
showed that insulating LiF, as a product of decomposed LiPF6, was deposited on the surface of pristine and AlF3-coated
Li͓Ni0.5Mn0.5͔O2. The deposition of LiF was greatly suppressed by AlF3 coating on the outer surface of Li͓Ni0.5Mn0.5͔O2. The
protection of the active material by the AlF3 coating substantially improved the capacity, capacity retention, and rate capability of
the batteries. It also enhanced the thermal stabilities of the positive electrode material.
© 2008 The Electrochemical Society. ͓DOI: 10.1149/1.2956088͔ All rights reserved.
Manuscript submitted April 11, 2008; revised manuscript received June 13, 2008. Published July 29, 2008.
Since 1990, lithium-ion batteries with high energy and power
output have become an important power source for portable devices,
such as cellular phones and laptop computers. As the field of lithium
battery systems has expanded, extensive research efforts have been
initiated to develop new positive electrode materials with higher
energy and higher power outputs. The fabrication of electrodes us-
ing commercialized LiCoO2 positive electrodes for Li-ion batteries
has been improved to satisfy the power requirements of new de-
vices. However, physical and mechanical improvements of this type
of electrode are limited, and ultimately new power sources will be
required. Furthermore, the structural instability1 and poor thermal
properties2 of Li1−␦CoO2 in a highly delithiated state must be im-
proved if this material is to be adopted in more electronic or vehicle
applications. Therefore, the development of new positive electrode
materials is required to address these issues.
Derivatives of Li͓NixCo1−2xMnx͔O2 ͑x = 0–0.333͒ represent
some of the more recently developed positive electrode materials.
These derivatives have stable electrochemical performance during
the reversible lithium extraction/insertion process.3-5 The structural
stability of Li͓Ni0.5Mn0.5͔O2 is due to the presence of tetravalent
Mn ions in the structure. The Mn4+ ions remain electrochemically
inactive during cycling in the voltage range of 2.7–4.6 V.6 However,
Li͓Ni0.5Mn0.5͔O2 shows a relatively poor rate capability at higher
currents due to its lower electronic conductivity compared to
LiCoO2 and Li͓Ni1/3Co1/3Mn1/3͔O2.7 Therefore, intensive research
efforts have been focused on improving the rate capability of
Li͓Ni0.5Mn0.5͔O2 with a partial cobalt substitution in the metal ion
layer of the host structure.8 Li͓NixCo1−2xMnx͔O2 ͑x = 0–0.333͒
showed a stable cycling performance and improved rate capability
in the conventional voltage window ͑3.0–4.3 V͒. However, when
Co-substituted Li͓NixCo1−2xMnx͔O2 ͑x = 0–0.333͒ was cycled in
the higher voltage range to increase energy density, the structural
instability in the highly delithiated state brought about a drastic ca-
pacity fade during cycling.9
͑TOF-SIMS͒.12,13 This indicates that, although an oxide electrode
material is coated with other amphoteric oxides, they finally trans-
form to stable metal fluorides. For this reason, we intentionally ap-
plied an AlF3 nanolayer coating on Li͓Ni0.5Mn0.5͔O2 and observed
the resulting electrochemical behavior.
Experimental
Spherical Li͓Ni0.5Mn0.5͔O2 was synthesized via coprecipitation.
Details of the synthetic process are reported in our previous paper.14
To prepare the AlF3-modified Li͓Ni0.5Mn0.5͔O2, ammonium fluoride
͑Aldrich͒ and aluminum nitrate nonahydrate ͑Aldrich͒ were first
separately dissolved in distilled water. Li͓Ni0.5Mn0.5͔O2 powders
were immersed in the aluminum nitrate nonahydrate solution, and
the ammonium fluoride solution was then slowly added to the solu-
tion. Ammonia solution was used to adjust the pH during precipita-
tion of the powder. The solution containing the active material was
constantly stirred at 80°C for 5 h. After filtering, the active materials
were washed with deionized water and dried at 120°C. The
Li͓Ni0.5Mn0.5͔O2 powders obtained were heated at 400°C for 5 h
under a nitrogen atmosphere to avoid the formation of Al2O3.
Powder X-ray diffraction ͑XRD, Rigaku Rint-2000͒ employing
Cu K␣ radiation was used to characterize the powders. Transmission
electron microscopy ͑TEM, 200 kV, Hitachi, H-800͒ was used to
observe the coating layer.
Charge–discharge tests were performed with 2032 coin-type
cells. The cell consisted of a positive electrode and a lithium metal
negative electrode separated by a porous polypropylene film. For
fabrication of the positive electrode, a mixture of 20 mg of positive
electrode materials and 5 mg of conducting binder ͑3.3 mg of Te-
flonized acetylene black and 1.7 mg of graphite͒ was pressed on a
2.0 cm2 stainless screen at 500 kg cm−2. The electrolyte used was
a mixture of 1 molar dm−3 LiPF6 in ethylene carbonate and diethyl
carbonate ͑1:1 in volume͒. AC-impedance measurements were per-
formed using a Zahner Elektrik IM6 impedance analyzer over the
frequency range from 1 MHz to 1 mHz with an amplitude of
10 mVrms ͑root mean square͒.
Recently, we reported that the coating of AlF3 on LiCoO2 and
Li͓Ni1/3Co1/3Mn1/3͔O2 effectively improved their electrochemical
properties.10,11 We previously showed that an Al2O3 coating layer on
Li͓Li0.05Ni0.4Co0.15Mn0.4͔O2 particles gradually transformed to AlF3
through an intermediate complex of an Al–O–F compound, as con-
firmed by time-of-flight secondary-ion mass spectroscopy
To measure the extent of Ni and Mn dissolution, cells charged to
4.5 V were carefully disassembled and then active materials were
stored in electrolyte at 90°C for 7 days. The amounts of dissolved
Ni and Mn were measured by atomic absorption spectroscopy
͑AAS, Analytik Jena AG, Vario 6͒.
To confirm the presence of by-products on the surface of the
active materials after extensive cycling, the cycled active materials
were examined using a TOF-SIMS ͑ULVAC-PHI TFS2000, Perkin
Elmer͒ surface analyzer operated at 10−9 Torr equipped with a liq-
*
Electrochemical Society Active Member.
z E-mail: yksun@hanyang.ac.kr
Downloaded on 2015-05-21 to IP 157.82.153.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A706
Journal of The Electrochemical Society, 155 ͑10͒ A705-A710 ͑2008͒
a
b
a
Active
material
10
20
30
40
60
70
80
Cu Kα 50
2θ
20 nm
Figure 1. XRD patterns of ͑a͒ pristine and ͑b͒ AlF3-coated Li͓Ni0.5Mn0.5͔O2.
AlF3 coating
layer
b
uid Ga ion source and pulse electron flooding. During the analysis,
the targets were bombarded with 15 keV Ga beams, with the pulsed
primary ion current varying from 0.3 to 0.5 pA. The total collection
time was 300 s over a 12 ϫ 12 ␮m area.
For differential scanning calorimetry ͑DSC͒ experiments, cells
were charged to 4.3 V vs Li and opened in an Ar-filled dry box. A
stainless steel sealed pan with a gold-plated copper seal ͑which can
withstand 150 atm of pressure before rupturing and has a capacity of
30 ␮L͒ was used to collect 3–5 mg samples. Measurements were
carried out in a Pyris 1 differential scanning calorimeter ͑Perkin
Elmer͒ using a temperature scan rate of 1°C min−1. The weight was
constant in all cases, indicating that there were no leaks during the
experiments.
Active
material
20 nm
Results and Discussion
Figure 1 shows powder XRD patterns of the as-synthesized
Li͓Ni0.5Mn0.5͔O2 and AlF3-coated Li͓Ni0.5Mn0.5͔O2. Both materials
have well-ordered R3m layer structures. The trace of the superstruc-
ture at 21° in 2␪ is also seen for both materials, implying that there
Figure 2. TEM bright-field images of ͑a͒ pristine and ͑b͒ AlF3-coated
Li͓Ni0.5Mn0.5͔O2.
15
15
was Ni2+ ͑0.69 Å͒ and Mn4+ ͑0.535 Å͒ ordering in the transi-
tion metal layer. After heat-treatment at 400°C in a N2 atmosphere,
there was no appearance of aluminum fluoride-induced product ͑Fig.
1b͒. The calculated lattice parameters for the pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2 were similar to each other, indicating
that no significant structural change occurred as a result of surface
modification by AlF3.
Further cycles were performed for both samples at a current den-
sity of 85 mA g−1 ͑0.5 C rate͒ at 30°C. Figure 4 presents the cor-
responding continuous discharge capacities of the pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2. The pristine material retained about
75% of its initial discharge capacity throughout the 100 cycles. The
capacity after 100 cycles was 120 mAh g−1. In contrast, the capac-
ity retention of the coated active materials at the 100th cycle was
about 86% ͑143 mAh g−1͒ of the initial capacity.
To estimate the thickness of the coating layer, bare and coated
materials were examined using TEM. As can be seen in Fig. 2a, the
bare material exhibits a smooth edge line. For the coated sample, a
thin coating layer is observed on the surface of Li͓Ni0.5Mn0.5͔O2, as
seen in Fig. 2b. Based on TEM images, the thickness of the coating
layer was around 10 nm. Due to the lower crystallinity and smaller
amount of the AlF3, this compound was hardly observed in the XRD
pattern in Fig. 1b. Approximately 2 mol % AlF3 was coated on the
surface of Li͓Ni0.5Mn0.5͔O2.
To investigate the effectiveness of the surface modification,
charge–discharge experiments were performed by applying currents
of 34 mA g−1 ͑0.2 C rate͒ and 85 mA g−1 ͑0.5 C rate͒ across the
positive electrode at 30°C. The corresponding results are shown in
Fig. 3. The voltage for the coated sample was slightly lower on
charging compared with that for the pristine sample. Meanwhile, the
corresponding voltage on discharge was slightly higher for the
coated sample. The same tendency was also observed at higher cur-
rents, as seen in Fig. 3. These results indicate that the AlF3 coating
on the surface of Li͓Ni0.5Mn0.5͔O2 reduces cell impedance during
Li+ extraction and insertion. For this reason, the surface-modified
electrode delivers a slightly higher discharge capacity ͑Fig. 3͒. Simi-
lar behavior was also observed in the Al2O3-coated oxide electrode,
which we previously reported.12
4.5
4.0
Pristine, 34 mA g-1
3.5
Pristine, 85 mA g-1
AlF3-coated, 34 mA g-1
AlF3-coated, 85 mA g-1
3.0
0
50
100
150
200
Capacity / mAh g-1
Figure 3. ͑Color online͒ Initial charge and discharge curves of pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2.
Downloaded on 2015-05-21 to IP 157.82.153.40 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, 155 ͑10͒ A705-A710 ͑2008͒
Regeneration, Li[Ni0.5Mn0.5]O2
A707
200
150
100
50
Discharge, Li[Ni0.5Mn0.5]O2
Regeneration, AlF3-coated Li[Ni0.5Mn0.5]O2
Discharge, AlF3-coated Li[Ni0.5Mn0.5]O2
200
150
100
50
a
Pristine
AlF3-coated
0
20
40
60
80
100
Number of cycle
0
10 20 30 40 50 60 70 80 90
Figure 4. Comparison of cyclability of pristine and AlF3-coated
Li͓Ni0.5Mn0.5͔O2.
DOD / %
0.20
0.15
0.10
0.05
0.00
b
Figure 5 shows the discharge capacities of the Li/Li͓Ni0.5
Mn0.5͔O2 and AlF3-coated Li͓Ni0.5Mn0.5͔O2 cells at various currents
͑34–850 mA g−1͒ between 3.0 and 4.5 V vs Li. The cells were
charged galvanostatically with a constant current of 20 mA g−1 be-
fore each discharge test, and were then discharged at different cur-
rent densities ranging from 34 mA g−1 ͑0.2 C rate͒ to 850 mA g−1
͑5 C rates͒. At a low current density ͑34 mA g−1͒, the two elec-
trodes showed similar discharge capacities, although the nominal
discharge voltage was slightly higher for the AlF3-coated
Li͓Ni0.5Mn0.5͔O2, as shown in Fig. 5a. At increased current density,
10 20 30 40 50 60 70 80 90
DOD (%)
Figure 6. ͑Color online͒ Comparison of HPPC test results for pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2: ͑a͒ ASI and ͑b͒ pulse power capability.
Pristine,
AlF3-coated
0.2 C, 34 mA cm-2
0.5 C, 85 mA cm-2
1 C, 170 mA cm-2
2 C, 340 mA cm-2
5 C, 850 mA cm-2
a
4.5
4.0
3.5
3.0
the AlF3-coated Li͓Ni0.5Mn0.5͔O2 electrode exhibited much higher
discharge capacities than the pristine electrode. For example, as
shown in Fig. 5a, the AlF3-coated Li͓Ni0.5Mn0.5͔O2 delivered a dis-
charge capacity of 130 mAh g−1 at higher current ͑850 mA g−1͒,
whereas the pristine electrode delivered around 50 mAh g−1 at the
same current density, which corresponds to 30% of its capacity at
34 mA g−1 ͑0.2 C rate͒. From the above tests, it is clear that
AlF3-coated Li͓Ni0.5Mn0.5͔O2 has a higher capacity, retention, and
rate capability than uncoated Li͓Ni0.5Mn0.5͔O2.
The Hybrid Pulse Power Characterization ͑HPPC͒ test was used
during the discharge procedure.16 The HPPC sequence was carried
out by first applying a 10 s discharge pulse followed by 40 s rest
period. After the rest period, the cell was subjected to a group of
regenerative pulses for 10 s at a 10 C rate ͑1700 mA g−1͒. The cell
was subsequently discharged to 10% of capacity at a 1 C rate fol-
lowed by an hour rest period to allow the cell to return to a charge
equilibrium condition before applying the next HPPC sequence. Fig-
ure 6 compares the area specific impedance ͑ASI͒ values for 10 C
rate pulse discharges as a function of depth of discharge of the
C/pristine and C/AlF3-coated Li͓Ni0.5Mn0.5͔O2 cells as evaluated by
the HPPC test. The ASI values of the AlF3-coated Li͓Ni0.5Mn0.5͔O2
electrode indicate a lower impedance during discharge and regen-
eration ͑Fig. 6a͒. Pulse powers higher by a factor of 2 were obtained
by surface treatment of Li͓Ni0.5Mn0.5͔O2 with AlF3, as shown in
Fig. 6b. Therefore, surface treatment by AlF3 is effective for im-
proving battery performance.
0
50
100
150
Capacity / mAh g-1
200
150
100
50
100 %
b
95.5 %
90.2 %
0.2 C
100 %
82.7 %
0.5 C
93.3 %
74.1 %
1 C
76.6 %
2 C
Pristine
AlF3-coated
5 C
52.3 %
30 %
To understand the improved cell performance of the AlF3-coated
Li͓Ni0.5Mn0.5͔O2, ac impedances for the pristine Li͓Ni0.5Mn0.5͔O2
and AlF3-coated Li͓Ni0.5Mn0.5͔O2 were measured after 20 cycles in
the fully charged state ͑4.5 V vs Li͒, as shown in Fig. 7. The equiva-
lent circuit was proposed in our previous report.10 Cole–Cole plots
0
5
15
Number of c1y0cle
Figure 5. ͑Color online͒ Comparison of rate capability of ͑a͒ pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2 and ͑b͒ the corresponding efficiency.
Downloaded on 2015-05-21 to IP 157.82.153.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A708
Journal of The Electrochemical Society, 155 ͑10͒ A705-A710 ͑2008͒
Li͓Ni0.5Mn0.5͔O2 were approximately 146.4 and 84.6 ppm, respec-
1000
800
600
400
200
0
50
40
30
20
10
0
tively, after storage. In contrast, the AlF3-coated Li͓Ni0.5Mn0.5͔O2
showed greatly reduced Ni and Mn dissolution into the electrolyte,
with amounts of only 39.6 and 18.6 ppm, respectively. Thus, it is
clear that there was a significant suppression of Ni and Mn dissolu-
tion in the AlF3-coated Li͓Ni0.5Mn0.5͔O2 electrode during storage at
90°C.
a
b
Dissolution of transition metal elements has been shown to lead
to capacity fade during cycling.10,11,17 The electrolyte used in this
experiment was a LiPF6 salt-containing composite solvent. The
LiPF6 can decompose, and this reaction is facilitated at higher po-
tentials, such as 4.5 V vs Li, according to the following reactions
suggested by Aurbach et al.18 and Edström et al.19
0
10
20
30
40
50
100 mHz
200
10 mHz
1 mHz
LiPF6 → LiF↓ + PF5
PF5 + H2O → POF3 + 2HF
͓1͔
͓2͔
͓3͔
0
400
600
800
1000
Z'RE / Ω
POF3 + 3Li2O → 6LiF↓ + P2O5↓ ͑or LixPOFy͒
Figure 7. ͑Color online͒ Cole–Cole plots for ͑a͒ pristine and ͑b͒ AlF3-coated
Li͓Ni0.5Mn0.5͔O2 at the 20th charge.
As can be seen from these reactions, the decomposed LiPF6 salt
produces LiF and HF as by-products. The LiF would be deposited
on positive or negative electrodes in the cell, while the HF formed
would continuously attack the oxide electrode material, resulting in
partial dissolution of the active material during cell operation. The
dissolved transition metal ions would adhere onto the surface of the
positive and negative electrodes and affect cell impedance. Once the
dissolved ions are deposited on the surface of the negative electrode,
the ions are reduced to metal. Then, the newly formed metal blocks
Li+ movement, which causes an increase in cell impedance. We
therefore ascribe the relatively higher cell impedance observed to
adhesion of dissolved transition metal ions onto the surface of the
positive electrode ͑Fig. 7͒.
for both electrodes present two semicircles, one in the high to me-
dium frequency region and the other in the low-frequency region.
The semicircle in the high to medium-frequency range is attributed
to contact resistance of the active particle surface ͑Rsf͒, while the
semicircle in the low-frequency range is associated with the charge-
transfer resistance ͑Rct͒ coupled with a double-layer capacitance.
The Rsf component is larger for the AlF3-coated material, as shown
in the inset of Fig. 7. It is likely that the insulating properties of AlF3
could result in an increase in the impedance in the higher frequency
region. In contrast, the charge-transfer resistance ͑Rct͒ due to the
interface between the positive electrode and the electrolyte was sig-
nificantly reduced by the AlF3 coating ͑Fig. 7͒. We expect that the
presence of an insulating AlF3 layer disturbs Li+ movements, result-
ing in a higher Rct for the AlF3-coated Li͓Ni0.5Mn0.5͔O2 because
insulating AlF3 particles can block the continuous connection of
active materials. However, tendencies contrary to what we predicted
were observed, as observed in rate capability tests ͑Fig. 5͒, HPPC
tests ͑Fig. 6͒, and Cole–Cole plots ͑Fig. 7͒.
Ni and Mn dissolution can occur in the higher voltage region. To
estimate the amount of Ni and Mn dissolution, half cells using a Li
metal negative electrode were charged to 4.5 V and were carefully
disassembled in a glove box. The recovered active materials were
stored in the electrolyte at 90°C for 1 week and the electrolytes were
then examined by AAS. Figure 8 shows the amounts of Ni and Mn
that dissolved into the electrolyte from the pristine and AlF3-coated
Li͓Ni0.5Mn0.5͔O2 electrodes after high-temperature storage. As
shown, the amounts of dissolved Ni and Mn from pristine
If decomposition of the active material progresses by HF attack,
by-products are inevitably formed. To confirm the formation of by-
products, TOF-SIMS analysis was carried out for the surfaces of the
cycled positive electrodes ͑100 cycles͒. Ga+ sputtering was carried
out for 3 s, resulting in etching of the top 0.8 Å of the outer surface
to remove the air-contaminated layer during sampling. Figures 9a
and b show Li–F fragments for the extensively cycled pristine
Li͓Ni0.5Mn0.5͔O2 and AlF3-coated Li͓Ni0.5Mn0.5͔O2, respectively. It
is clear that LiPF6 salt decomposed, resulting in LiF deposition on
the surface of the electrode. Direct comparison of the amount of LiF
formed is difficult from the spectra. It is likely, however, that more
LiF ͑as a by-product of LiPF6 salt decomposition͒ is deposited on
the electrode surface of the pristine Li͓Ni0.5Mn0.5͔O2, because the
total count for this material is about five times higher compared to
that of the AlF3-coated Li͓Ni0.5Mn0.5͔O2 ͑Fig. 9a and b͒. This LiPF6
salt decomposition reaction is related to the generation of HF. In-
creased LiF formation indicates increased HF formation, as shown
in Reactions 1 and 2. This consequently results in increased disso-
lution of Ni and Mn elements of the pristine Li͓Ni0.5Mn0.5͔O2, as
shown in Fig. 8, which is confirmed by the TOF-SIMS spectra in
Fig. 9c-f. That is, much higher total counts of Ni–O and Mn–O
fragments are seen on the surface of the pristine Li͓Ni0.5Mn0.5͔O2.
The following reaction is suggested
160
146.04
Pristine
140
AlF3-coated
120
2Li͓Ni0.5Mn0.5͔O2 ͑AlF3-coated͒ + 2HF + Li+ + e−
→ Li1−␦͓Ni0.5Mn0.5͔O2 + NiO + MnO + LiF + H2O ͓4͔
The NiO and MnO formed are energetically unstable because they
100
84.6
80
60
have relatively high Gibbs free energies of formation, NiO
39.6
20
͑−211.7 kJ mol−1
͒
and MnO ͑−369.9 kJ mol−1͒.20 These mate-
40
20
rials consecutively react with HF, and the divalent oxides are par-
tially transformed to NiF2 or MnF2, as shown in Fig. 9g-j. Though
the Li͓Ni0.5Mn0.5͔O2 is coated with AlF3, NiF2 and MnF2 are
formed on the outer surface of AlF3-coated Li͓Ni0.5Mn0.5͔O2 by the
following reactions
18.6
0
Mn
Ni
Figure 8. Amount of Ni and Mn dissolved for delithiated pristine and
AlF3-coated Li͓Ni0.5Mn0.5͔O2 stored at 90°C for 7 days.
NiO + 2HF → NiF2 + H2O
͓5͔
Downloaded on 2015-05-21 to IP 157.82.153.40 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, 155 ͑10͒ A705-A710 ͑2008͒
A709
Li-F (26.02)
a
c
b
15000
10000
5000
0
15000
10000
5000
Li-F (26.02)
0
25.0
25.0
25.5
26.0
26.5
25.5
26.0
26.5
3000
2500
2000
1500
1000
500
3000
2500
2000
1500
1000
500
Ni-O (73.93)
d
f
Ni-O (73.93)
0
73.0
0
73.0
73.5
74.0
74.5
73.5
74.0
74.5
1400
1200
1000
800
600
400
200
0
1400
1200
1000
800
600
400
200
0
e
g
i
Mn-O (70.93)
Mn-O (70.93)
70.0
70.5
71.0
71.5
70.0
70.5
71.0
71.5
500
400
300
200
100
0
500
400
300
200
100
0
Ni2-F (134.90)
h
j
Ni2-F (134.90)
134.0
134.5
135.0
135.5
134.0
134.5
135.0
135.5
100
80
60
40
20
0
100
80
60
40
20
0
Mn2-F (128.86)
Mn2-F (128.86)
128.0
128.5
129.0
129.5
128.0
128.5
129.0
129.5
Mass
/
Z
Mass
/
Z
Figure 9. Comparison of TOF-SIMS fragments for extensively cycled pristine and AlF3-coated Li͓Ni0.5Mn0.5͔O2 electrodes: Li–F fragments of ͑a͒ pristine and
͑b͒ AlF3-coated, Ni–O fragments of ͑c͒ pristine and ͑d͒ AlF3-coated, Mn–O fragments of ͑e͒ pristine and ͑f͒ AlF3-coated, Ni2–F fragments of ͑g͒ pristine and
͑h͒ AlF3-coated, and Mn2–F fragments of ͑i͒ pristine and ͑j͒ AlF3-coated.
Myung et al.12,13 recently revealed that an Al2O3 coating layer on
MnO + 2HF → MnF2 + H2O
͓6͔
Li͓Li0.05Ni0.4Co0.15Mn0.4͔O2 particles gradually transformed to AlF3
via an intermediate stage ͑copresence of mixed chemical bonds of
Al–O–F and Al–F after 300 cycles at 60°C͒ on the surface of the
Al2O3 coating layer. This was proved by TOF-SIMS, which indi-
cated that Al2O3 scavenges HF. In this case, the gradual change of
From the TOF-SIMS analysis, it is apparent that dissolution of tran-
sition metals is related to the formation of insulating by-products on
the surface of the active materials. In particular, the amount of LiF
deposited was significantly reduced. These by-products disturb Li+
diffusion, and are responsible for poor battery performance in terms
of capacity, retention, and rate capability.
Extensively cycled active particles were studied by TEM, and the
results are shown in Fig. 10. As can be seen in Fig. 10a, the un-
coated particle was severely damaged during cycling. In a magnified
image, the smooth edges observed in Fig. 2a are hardly observable
in Fig. 10b. This disruption of particle morphology is mainly due to
the dissolution of metal elements during cycling. In contrast, the
cycled AlF3-coated Li͓Ni0.5Mn0.5͔O2 particles still show smooth
edge lines with clear coating lines ͑Fig. 10c͒, and the AlF3 coating
layer is still present on the surface of Li͓Ni0.5Mn0.5͔O2 in a magni-
fied image ͑Fig. 10d͒.
The thermal stability of the uncoated pristine and AlF3-coated
Li͓Ni0.5Mn0.5͔O2 electrodes was measured by DSC. Figure 11
shows the DSC profiles of the uncoated pristine and AlF3-coated
Li͓Ni0.5Mn0.5͔O2 electrode at 4.5 V. The onset temperature of ther-
mal decomposition was increased by AlF3 coating. The AlF3-coated
Li͓Ni0.5Mn0.5͔O2 electrode showed a major exothermic peak at
297.9°C while the uncoated electrode had a major exothermic peak
at 284.1°C. In addition, the total heat associated with each exother-
mic peak was greatly reduced by AlF3 coating. When the charge
potential was 4.5 V, the uncoated and Al-coated electrode produced
2893 and 1515 J g−1, respectively. This result suggests that coating
Li͓Ni0.5Mn0.5͔O2 with AlF3 retards the exothermic reaction with the
electrolyte.
a
c
b
500 nm
300 nm
100 nm
d
AlF3 coating
layer
Active
material
100 nm
Figure 10. TEM bright-field images of ͑a͒ extensively cycled pristine
Li͓Ni0.5Mn0.5͔O2 particle and ͑b͒ its magnified image, ͑c͒ extensively cycled
AlF3-coated Li͓Ni0.5Mn0.5͔O2 particle, and ͑d͒ its magnified image. The
circled parts are magnified.
Downloaded on 2015-05-21 to IP 157.82.153.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A710
Journal of The Electrochemical Society, 155 ͑10͒ A705-A710 ͑2008͒
transfer resistance. As confirmed by TOF-SIMS, Ni and Mn were
30
20
10
0
somewhat dissolved from the active materials whether the materials
were coated or not. The by-products formed were NiO, MnO, NiF2,
MnF2, and LiF, and significantly more of these by-products were
formed with pristine Li͓Ni0.5Mn0.5͔O2. Because the AlF3 coating
layer covers the active materials, the amount of LiF deposited on the
surface of coated Li͓Ni0.5Mn0.5͔O2 was significantly smaller than on
the uncoated surface. The protection of the active material by AlF3
coating is substantially effective to improve the capacity, capacity
retention, rate capability of the batteries, and the thermal stabilities
of the positive electrode material as well.
a
b
Acknowledgment
This work was supported by University IT Research Center
Project.
100
150
200
250
300
350
Temperature / oC
Hanyang University assisted in meeting the publication costs of this ar-
ticle.
Figure 11. Comparison of DSC traces of delithiated ͑a͒ pristine and ͑b͒
AlF3-coated Li͓Ni0.5Mn0.5͔O2 electrodes.
References
1. T. Ohzuku and A. Ueda, J. Electrochem. Soc., 141, 2972 ͑1994͒.
2. G. G. Amatucci, J. M. Tarascon, and L. C. Klein, J. Electrochem. Soc., 143, 1114
͑1996͒.
3. M. E. Spahr, P. Novak, B. Schnyder, O. Haas, and R. Nesper, J. Electrochem. Soc.,
145, 1113 ͑1998͒.
4. T. Ohzuku and Y. Makimura, Chem. Lett., 30, 744 ͑2001͒.
5. Z. Lu, D. D. MacNeil, and J. R. Dahn, Electrochem. Solid-State Lett., 4, A200
͑2001͒.
6. W.-S. Yoon, C. P. Grey, M. Balasubramanian, X.-Q. Yang, and J. McBreen, Chem.
Mater., 15, 3161 ͑2003͒.
7. K. Shizuka, T. Kobayashi, K. Okahara, K. Okamoto, S. Kanzaki, and R. Kanno, J.
Power Sources, 146, 589 ͑2005͒.
8. B.-C. Park, H. J. Bang, C. S. Yoon, S.-T. Myung, J. Prakash, and Y.-K. Sun, J.
Electrochem. Soc., 154, A520 ͑2007͒.
9. G.-H. Kim, J.-H. Kim, S.-T. Myung, C. S. Yoon, and Y.-K. Sun, J. Electrochem.
Soc., 152, A1707 ͑2005͒.
the Al2O3 layer to AlF3 was effective in preserving the powder
morphology upon cycling. In our work, the AlF3 was intentionally
formed before electrochemical cycling. It is highly probable that the
strong AlF3 coating layer protects the active material from HF attack
during cycling. This, in turn, would lead to suppression of insulating
LiF deposition on the surface of the active material and dissolution
of transition metal elements during cycling, thereby maintaining the
original particle morphology. All these factors simultaneously result
in lowering of the charge-transfer resistance. For the above reasons,
AlF3-coated Li͓Ni0.5Mn0.5͔O2 presented a higher capacity, good cy-
clability, good rate capability, and improved thermal stability com-
pared to pristine Li͓Ni0.5Mn0.5͔O2.
10. Y.-K. Sun, J.-M. Han, S.-T. Myung, S.-W. Lee, and K. Amine, Electrochem. Com-
mun., 8, 821 ͑2006͒.
11. B.-C. Park, H.-B. Kim, S.-T. Myung, K. Amine, I. Belharouak, S.-M. Lee, and
Y.-K. Sun, J. Power Sources, 178, 826 ͑2008͒.
12. S.-T. Myung, K. Izumi, S. Komaba, Y.-K. Sun, H. Yashiro, and N. Kumagai, Chem.
Mater., 17, 3695 ͑2005͒.
Conclusions
The effects of AlF3 coating on Li͓Ni0.5Mn0.5͔O2 subjected to
high cutoff voltage ͑4.5 V͒ cycling were investigated. Although
there was no large difference in the structure of Li͓Ni0.5Mn0.5͔O2
after AlF3 coating, the AlF3-coated Li͓Ni0.5Mn0.5͔O2 showed a
greatly enhanced specific discharge capacity, cycling performance,
and rate capability compared to pristine Li͓Ni0.5Mn0.5͔O2. Also, re-
duced resistance properties and improved rate capability were ob-
served by HPPC testing. The improved electrochemical performance
could be mainly attributed to the reduced Ni and Mn dissolution of
the AlF3-coated active material, giving rise to a lower charge-
13. S.-T. Myung, K. Izumi, S. Komaba, H. Yashiro, H. J. Bang, Y.-K. Sun, and N.
Kumagai, J. Phys. Chem. C, 111, 4061 ͑2007͒.
14. M.-H. Lee, Y.-J. Kang, S.-T. Myung, and Y.-K. Sun, Electrochim. Acta, 50, 939
͑2004͒.
15. R. D. Shannon, Acta Crystallogr. A, 32, 751 ͑1976͒.
16. PNGV Battery Test Manual, Revision 3, DOE/ID-11069 ͑2003͒.
17. S.-T. Myung, K. Hosoya, S. Komaba, H. Yashiro, Y.-K. Sun, and N. Kumagai,
Electrochim. Acta, 51, 5912 ͑2006͒.
18. D. Aurbach, J. Electrochem. Soc., 136, 906 ͑1989͒.
19. K. Edström, T. Gustafsson, and J. O. Thomas, Electrochim. Acta, 50, 397 ͑2004͒.
20. J. A. Dean, Lange’s Handbook of Chemistry, 4th ed., p. 6, McGraw-Hill Inc., New
York ͑1992͒.
Downloaded on 2015-05-21 to IP 157.82.153.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).