Enhanced stability of LiCoO2 cathodes in lithium-ion batteries using surface modification by atomic layer deposition

Article abstract of DOI:10.1149/1.3258274

Ultrathin atomic layer deposition (ALD) coatings enhance the performance of lithium-ion batteries (LIBs). Previous studies have demonstrated that LiCoO₂ cathode powders coated with metal oxides with thicknesses of ～100 to 1000 A grown using wet chemical techniques improved LIB performance. In this study, LiCoO₂ powders were coated with conformal Al₂O₃ ALD films with thicknesses of only ～3 to 4 A established using two ALD cycles. The coated LiCoO₂ powders exhibited a capacity retention of 89% after 120 charge-discharge cycles in the 3.3-4.5 V (vs Li/Li⁺) range. In contrast, the bare LiCoO₂ powders displayed only a 45% capacity retention. Al₂O₃ ALD films coated directly on the composite electrode also produced improved capacity retention. This dramatic improvement may result from the ultrathin Al₂O₃ ALD film acting to minimize Co dissolution or reduce surface electrolyte reactions. Similar experiments with ultrathin ZnO ALD films did not display enhanced performance.

Full text of DOI:10.1149/1.3258274

Journal of The Electrochemical Society, 157 ͑1͒ A75-A81 ͑2010͒
A75
0
013-4651/2009/157͑1͒/A75/7/$28.00 © The Electrochemical Society
Enhanced Stability of LiCoO Cathodes in Lithium-Ion
2
Batteries Using Surface Modification by Atomic
Layer Deposition
a, ,e
b
c,
Yoon Seok Jung, * Andrew S. Cavanagh, Anne C. Dillon, *
d
e
a, ,z
Markus D. Groner, Steven M. George, and Se-Hee Lee *
a
Department of Mechanical Engineering, University of Colorado at Boulder, Boulder, Colorado
8
0309-0427, USA
b
e
Department of Physics, Department of Chemistry and Biochemistry, and Department of Chemical and
Biological Engineering, University of Colorado at Boulder, Boulder, Colorado 80309-0215, USA
c
National Renewable Energy Laboratory, Golden, Colorado 80401, USA
ALD NanoSolutions Incorporated, Broomfield, Colorado 80020, USA
d
Ultrathin atomic layer deposition ͑ALD͒ coatings enhance the performance of lithium-ion batteries ͑LIBs͒. Previous studies have
demonstrated that LiCoO2 cathode powders coated with metal oxides with thicknesses of ϳ100 to 1000 Å grown using wet
chemical techniques improved LIB performance. In this study, LiCoO powders were coated with conformal Al O ALD films
2
2
3
with thicknesses of only ϳ3 to 4 Å established using two ALD cycles. The coated LiCoO powders exhibited a capacity retention
2
+
of 89% after 120 charge–discharge cycles in the 3.3–4.5 V ͑vs Li/Li ͒ range. In contrast, the bare LiCoO powders displayed only
a 45% capacity retention. Al O ALD films coated directly on the composite electrode also produced improved capacity retention.
2
2
3
This dramatic improvement may result from the ultrathin Al O ALD film acting to minimize Co dissolution or reduce surface
2
3
electrolyte reactions. Similar experiments with ultrathin ZnO ALD films did not display enhanced performance.
2009 The Electrochemical Society. ͓DOI: 10.1149/1.3258274͔ All rights reserved.
©
Manuscript submitted July 1, 2009; revised manuscript received October 5, 2009. Published November 18, 2009.
Efficient and durable electrical energy storage is one of the major
factors limiting the widespread adoption of renewable energy. Since
lithium-ion batteries ͑LIBs͒ were first commercialized in the early
trodes. The results reveal that ultrathin Al O ALD films can dra-
2 3
matically enhance the stability of LiCoO cathodes.
2
1
990s, LIBs have emerged as an important energy storage device for
Experimental
1-3
portable electronics. LIBs are very desirable because of their high
energy storage per volume and per mass. However, LIBs with higher
stability are needed for their use in plug-in hybrids or all-electric
vehicles.
ALD on LiCoO powders.— Al O ALD films were grown di-
2
2
3
rectly on the LiCoO powders. The precursors utilized for Al O
2
2
3
ALD were trimethylaluminum ͑TMA͒ and H O, as shown in Fig. 1.
2
The two self-limiting surface reactions that define Al O ALD
2
3
1
8-20
Li1−xCoO is the most commercialized cathode material. Unfor-
2
growth are
tunately, the practical use of Li1−xCoO is limited because the sta-
2
ء
ء
3 2
AlOH + Al͑CH ͒ → AlO–Al͑CH ͒ + CH
͓1͔
͓2͔
bility rapidly deteriorates at potentials higher than 4.2–4.3 V ͑vs
3 3
4
+
4
Li/Li ͒. Cobalt dissolution, structural changes, and oxidative de-
ء
ء
AlO–Al͑CH ͒ + H O → AlO–Al–OH + CH
Al2O3 ALD films are amorphous and pinhole-free with a density
3
2
2
4
composition of the electrolyte produce a dramatic increase in the
4
capacity fade at the higher potentials. These instabilities can be
−
3 19,21
addressed by coating the LiCoO powders with metal oxide coatings
of ϳ3.0 g cm .
The typical growth rate for Al2O3 ALD is
2
1
9,20
5,6
1
.1–1.2 Å per ALD cycle.
However, purging H O may be diffi-
with thicknesses of ϳ100 to 1000 Å. Examples of the metal ox-
2
cult for ALD on high surface area powders. The presence of H O
ides that have been explored include Al O , ZrO , ZnO, SiO , and
2
2
3
2
2
5
,6
7
8
during the TMA reaction may lead to a slightly larger growth per
TiO2. Metal phosphates ͑e.g., AlPO ͒ and metal fluorides ͑e.g.,
4
22,23
cycle resulting from some chemical vapor deposition.
AlF ͒ have also been studied as coatings.
3
To compare Al O ALD with other ALD materials, ZnO ALD
2
3
The majority of the coating strategies have been based on solu-
5-9
was also grown on LiCoO2 powders. The ZnO ALD surface chem-
tion techniques such as the sol–gel method. These wet chemical
coating methods require large amounts of solvent and precursor. A
istry employs Zn͑CH CH ͒ ͑diethylzinc, DEZ͒ and H O as the
2
3 2
2
5
-9
reactants. In similarity with Al2O3 ALD, the two self-limiting reac-
post-heat-treatment is also necessary after the sol–gel coating. In
contrast, atomic layer deposition ͑ALD͒ is a gas-phase method of
thin-film growth using sequential, self-limiting surface
24,25
tion sequences are
ء
ء
3
ZnOH + Zn͑CH CH ͒ → ZnO–ZnCH CH + CH CH ͓3͔
2
3 2
2
3
3
1
0,11
reactions.
ALD requires only a minimal amount of precursor,
ء
ء
and ALD coatings are conformal and offer atomic thickness control.
ALD could be a promising alternative method to coat electrode ma-
terials for LIBs. Although ALD films have been employed in a va-
ZnCH2CH3 + H2O → ZnOH + CH3CH3
͓4͔
26
The typical growth rate for ZnO ALD is 2.0 Å per ALD cycle.
ZnO ALD may also display larger growth rates on powders because
1
2-16
riety of application areas,
the use of ALD films for LIB elec-
of incomplete purging of H O.
2
17
trodes has not been pursued extensively.
ALD on LiCoO powders was performed using a rotary ALD
2
2
2
In this paper, the electrochemical performance is reported for
LiCoO coated with ultrathin conformal Al O and ZnO films by
reactor. A schematic of the rotary reactor is shown in Fig. 2. To
perform ALD on powders, the powders were placed in a porous
stainless steel cylinder ͑A͒ in the reaction chamber. The cylinder
was positioned on a magnetically coupled shaft via a load lock door
͑B͒. A rotor turned the cylinder to agitate the powder ͑C͒. A capaci-
tance manometer ͑D͒ was used to measure the pressure in the reac-
tion chamber. The introduction of precursor and purge gases was
controlled via a series of pneumatic ͑E͒ and needle valves ͑F͒. To
evacuate the chamber, a gate valve ͑G͒ was opened to connect the
chamber to a vacuum pump ͑H͒.
2
2
3
ALD. The experiments examine the effects of the coating material
and the coating thickness on cycle performance and rate capability.
The viability of ALD is also explored directly on composite elec-
*
Electrochemical Society Active Member.
Present address
E-mail: sehee.lee@colorado.edu
e
z
Downloaded on 2014-06-24 to IP 134.117.10.200 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A76
Journal of The Electrochemical Society, 157 ͑1͒ A75-A81 ͑2010͒
Figure 3. ͑Color online͒ Atomic fraction of Al and Co for Al O ALD-
2
3
coated LiCoO powders vs the number of ALD cycles. Atomic fraction was
2
determined from XPS spectra.
was 3 ϫ 10−10 Torr. During the data acquisition, the constant ana-
lyzer energy mode was employed at pass energies between 58.7 and
9
3.9 eV and a step size between 0.25 and 0.4 eV. Using the Al 2s
and the Co 2p3 peaks, the growth rate for the Al O ALD on
/2
2
3
LiCoO was determined from thicknesses calculated using a model
for flat surfaces.
2
Figure 1. ͑Color online͒ Schematic representation of Al O ALD on LiCoO
2
powders.
27,28
2
3
A LiCoO -embedded Ag foil for obtaining XPS spectra was pre-
2
pared by pressing ͑ϳ1.5 GPa͒ the LiCoO2 powders that were
spread on a piece of Ag foil with a thickness of 0.5 mm. The com-
posite electrodes were used for the ex situ XPS analysis. Cells were
disassembled, and electrodes were rinsed with dimethyl carbonate
͑DMC͒ and dried in an Ar-filled dry box. The conductivity of
The Al O ͑ZnO͒ ALD reaction sequence was: ͑i͒ TMA ͑DEZ͒
dose to 1.0 ͑1.0͒ Torr, ͑ii͒ TMA ͑DEZ͒ reaction time, ͑iii͒ evacuation
2
3
of the reaction products and excess TMA ͑DEZ͒, ͑iv͒ N dose to
2
2
9
2
0.0 Torr, ͑v͒ N static time, ͑vi͒ evacuation of N and any entrained
LiCoO powders was measured by the van der Pauw method using
2
2
2
gases, ͑vii͒ H O dose to 1.0 Torr, ͑viii͒ H O reaction time, ͑ix͒
120 MPa.
2
2
evacuation of reaction products and excess H O, ͑x͒ dose N , ͑xi͒ N
2
2
2
static time, and ͑xii͒ evacuation of N and any entrained gases. This
2
Electrochemical characterization.— For
the
galvanostatic
sequence constitutes one cycle of Al O ͑ZnO͒ ALD. Both Al O
2
3
2
3
charge–discharge cycling, a composite electrode was prepared by
spreading a slurry mixture of LiCoO2 powder ͑7–10 ␮m, L106,
LICO Technology͒, acetylene black ͑AB͒, and poly͑vinylidene fluo-
ride͒ ͑PVDF͒ ͑83.0:7.5:9.5 weight ratio͒ on a piece of Al foil. The
cells were assembled in an Ar-filled dry box and tested in a
temperature-controlled oven. The galvanostatic charge–discharge
cycling was performed with a two-electrode 2032-type coin cell in
ALD and ZnO ALD were conducted at 180°C.
Material characterization.— X-ray photoelectron spectroscopy
XPS͒ measurements were performed on a PHI 5600 X-ray photo-
͑
electron spectrometer using a monochromatic Al K␣ source
1486.6 eV͒. The base pressure in the spectrometer during analysis
͑
+
the potential range of 3.3–4.5 V ͑vs Li/Li ͒ at a current density of
−
1
0
.1 C rate ͑14 mA g ͒ for the first two cycles and 1 C rate for the
subsequent cycles at room temperature. Li metal foil was used as the
counter electrode.
1
.0 M LiPF dissolved in a mixture of ethylene carbonate, and
6
DMC ͑1:1 v/v͒ was used as the electrolyte. A porous 20 ␮m thick
polypropylene ͑PP͒/polyethylene/PP trilayer film was used as the
separator. The electrochemical impedance spectroscopy ͑EIS͒ study
was performed using a 1280C Solartron instrument. The ac imped-
ance measurement was recorded using a signal with an amplitude of
5
mV and a frequency range from 20 kHz to 5 mHz. After the
+
LiCoO /Li cells were charged to 4.5 V ͑vs Li/Li ͒ with a current
2
−
1
density of 0.1 C rate ͑14 mA g ͒ and stabilized by resting for 6 h,
the ac impedance spectra were recorded at the open-circuit voltage.
Results and Discussion
Al O ALD on LiCoO powders.— Figure 3 shows the Al and
2
3
2
Co atomic fraction on the LiCoO powders as determined by XPS vs
2
the number of Al O ALD cycles. The rapid attenuation of the Al
2
3
signal is evidence that the Al O ALD is conformally coating the
2
3
LiCoO powders. The conformality of Al O ALD films on particles
2
2
3
2
3
has also been verified with transmission electron microscopy. If
the attenuation of the Co signal is modeled as Al O grown on a flat
2
3
2
7
Figure 2. Schematic diagram of the rotary ALD reactor.
LiCoO surface, XPS analysis indicates a growth rate of 2.2 Å per
2
Downloaded on 2014-06-24 to IP 134.117.10.200 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, 157 ͑1͒ A75-A81 ͑2010͒
A77
Figure 4. Specific discharge capacity of Al O ALD-coated LiCoO elec-
2
3
2
trodes vs the number of ALD cycles.
Figure 6. ͑Color online͒ Electronic conductivity of bare and Al O ALD-
2
3
coated LiCoO powders.
2
ALD cycle. A higher Al O ALD growth rate than expected is at-
2
3
tributed to the incomplete purging of H O from the LiCoO powders
2
2
22,23
during Al O ALD.
slower charge/discharge kinetics. The Al O ALD film could also
2 3
2
3
Figure 4 plots the specific discharge capacity for a current den-
reduce lithium-ion conductivity. However, some previous studies
−
1
sity of 1 C rate ͑140 mA g ͒ during the third charge–discharge
have reported faster Li diffusion in the Al O -coated LiCoO result-
2
3
2
3
0,31
cycle vs the number of Al O ALD cycles deposited on the LiCoO
ing from a thin LiCo1−xAlxO2 solid solution layer.
Likewise,
2
3
2
powders. The capacity during the third charge–discharge cycle does
not change considerably for up to four ALD cycles. After six ALD
cycles, the capacity starts to decrease significantly and shows a neg-
ligible value of ϳ20 mAh g−1 after the 10th ALD cycle. The loss
of capacity is attributed to the large overpotential required for the
other investigations report increases in electrical conductivity for
3
2
Al O -coated LiCoO . These reports of a faster Li diffusion and a
2
3
2
higher electrical conductivity for Al O -coated LiCoO are surpris-
2
3
2
ing given that the LiCoO powders have been coated by films with
2
thicknesses of ϳ100 to 1000 Å using wet chemical methods.
The final step is a post-heat-treatment for many of the wet
chemical methods. This heat-treatment may lead to interdiffusion
between the Al O layer and the LiCoO core, resulting in an amor-
LiCoO powders coated with more than six ALD cycles. These large
2
overpotentials for Ͼ6 ALD cycles are shown in Fig. 5.
The loss of capacity results mainly from the electronically insu-
2
3
2
21
9,33
lating character of the Al O ALD film. The conductivities mea-
phous LiAl Co1−xO2 solid solution. + This LiAl Co1−xO2 alloy
2
3
x
x
sured using the van der Pauw method are shown in Fig. 6. The bare
may be responsible for the enhanced Li -ion transport properties and
increased electronic conductivity. To examine the effect of the heat-
treatment, LiCoO powders coated with 20 cycles of Al O ALD
LiCoO powders have an electronic conductivity of
2
2
−
4
−1
ϫ 10 S cm . After only two ALD cycles, the electronic conduc-
2
2
3
−
5
−1
tivity is significantly reduced to 5 ϫ 10 S cm . The conductiv-
ity continuously decreases with increasing number of ALD cycles.
An Al O ALD film with a thickness of Ͼ10 Å after six ALD
were heat-treated at 450°C for 10 h in air. The discharge capacity
before the heat-treatment was ϳ8 mAh g− at 1 C rate. After the
heat-treatment, the discharge capacity was slightly increased to
1
2
3
−
1
cycles on the LiCoO powders significantly reduces the electronic
15 mAh g , as shown in Fig. 7. Because the heat-treatment did not
significantly enhance the discharge capacity to values
2
conductivity. The reduction in electron conductivity could result in
Figure 5. ͑Color online͒ The second and
third charge–discharge voltage profiles of
electrodes fabricated using bare and Al O
2
3
ALD-coated LiCoO2 powders. The ALD
was performed on the bare LiCoO pow-
2
ders. The current densities for the second
and third charge–discharge cycles were
0
.1 and 1 C rate, respectively.
Downloaded on 2014-06-24 to IP 134.117.10.200 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A78
Journal of The Electrochemical Society, 157 ͑1͒ A75-A81 ͑2010͒
Figure 7. ͑Color online͒ Charge–discharge voltage profiles at the third
charge–discharge cycle ͑1 C rate͒. ͑a͒ Electrode prepared from the bare
LiCoO2 powder. Electrodes prepared using LiCoO2 powders after ͑b͒ 20
Al O ALD cycles and ͑c͒ 20 Al O3 ALD cycles and subsequent heat-
2
3
2
treatment at 450°C for 10 h in air.
−
1
Ͼ100 mAh g , the effect of the ALD coatings is attributed to the
high quality conformal insulation provided by the Al O ALD coat-
2
3
Figure 9. ͑Color online͒ Charge–discharge voltage profiles of electrodes
ing compared with previous Al O films grown using wet chemical
2
3
fabricated with the bare and the Al O ALD-coated LiCoO powders at the
2
3
2
techniques.
3
rd, 10th, and 50th cycles when cycled at a current density of 1 C rate
Figure 8 compares the cycle performance for several electrodes
−1
͑
140 mA g ͒.
+
when cycled in the range of 3.3–4.5 V ͑vs Li/Li ͒ with a current
−
1
density of 0.1 C rate ͑14 mA g ͒ for the first two cycles and 1 C
rate for the subsequent cycles. The initial capacity of LiCoO pow-
2
coated LiCoO powders using two ALD cycles. The voltage profiles
2
ders coated with Al O ALD using two ALD cycles is similar to the
2
3
reveal the changing behavior of the bare LiCoO powders. The po-
2
initial capacity of bare LiCoO powders. Based on the XPS results,
2
larization increases and the specific capacity drops rapidly vs the
charge/discharge cycle number for the electrodes prepared using the
the thicknesses of the Al O ALD film after two ALD cycles are ϳ3
2
3
to 4 Å. The initial capacity decreases for 6 and 10 ALD cycles as
the current density increases from 0.1 to 1 C rate. This reduction in
capacity is attributed to the restricted electron transport and possibly
bare LiCoO powders. In contrast, the voltage profile and specific
2
capacity do not change significantly for the electrodes fabricated
using the Al2O3 ALD-coated LiCoO2 powders.
+
to the slower Li diffusion kinetics in the Al O ALD layer.
2
3
EIS analyses were also utilized to evaluate the performance of
the electrodes prepared using bare LiCoO2 powders and Al O
The LiCoO powders coated with Al O ALD show a dramati-
2
2
3
2
3
cally improved retention of capacity vs the charge/discharge cycle
number regardless of the ALD coating thickness. The coated
ALD-coated LiCoO powders prepared using two ALD cycles. Re-
2
sults showing the Nyquist plots vs the number of charge–discharge
cycles are shown in Fig. 10. ZЈ ͑⍀ g͒ is the magnitude of the real
impedance and −ZЉ ͑⍀ g͒ is the magnitude of the imaginary im-
pedance.
LiCoO powders exhibited a capacity retention of 89% after 120
2
+
charge–discharge cycles in the 3.3–4.5 V ͑vs Li/Li ͒ range. In con-
trast, the bare LiCoO powders displayed only a 45% capacity re-
2
tention. The stability is highest after 10 ALD cycles although these
electrodes have severely restricted capacity because of the insulating
The impedance spectra for bare LiCoO powders follow the typi-
2
cal spectra of LiCoO , which comprise two semicircles and a 45°
Al O ALD layer.
2
2
3
34,35
inclined line.
The first semicircle in the higher frequency zone
Figure 9 shows the charge–discharge voltage profiles for the
͑
#͒ is related to the solid electrolyte interphase, while the second
electrodes fabricated with bare LiCoO powders and Al O ALD-
2
2
3
ء
semicircle in the lower frequency zone ͑ ͒ is a charge-transfer reac-
tion at the electrolyte/electrode interface.
3
4,35
The charge-transfer
resistance at the Al O /LiCoO interface may also contribute to the
2
3
2
3
4,35
overall charge transfer.
Although one superficial semicircle is
observed for Al O –LiCoO , the shape of the semicircle is not sym-
2
3
2
metric. One semicircle for Al O –LiCoO includes a small charge-
2
3
2
transfer resistance at the Al O /LiCoO and electrolyte/electrode in-
2
3
2
terfaces. The first semicircle ͑#͒ does not change much during
ء
cycling. In contrast, the radius of the second semicircle ͑ ͒ increases
dramatically with the number of charge–discharge cycles. The in-
creases occur both for the real and imaginary impedances, indicating
increases in charge-transfer resistance.
For the electrode prepared using Al O ALD-coated LiCoO2
2
3
powders, the interface with the electrolyte is very stable. Figure 10
shows that the semicircle for Al O ALD-coated LiCoO holds its
2
3
2
overall radius and shape even after 50 charge–discharge cycles.
There is no indication of increases in charge-transfer resistance. This
electrode prepared with Al O ALD-coated LiCoO particles is ex-
2
3
2
ceptionally stable.
There are several possible mechanisms for the enhancement
Figure 8. ͑Color online͒ Charge–discharge cycle performance of electrodes
6
,36
fabricated using the bare LiCoO2 powders and the Al O ALD-coated
caused by the Al O ALD coatings.
2
3
First, the Al O ALD coating
2 3
2
3
LiCoO powders using 2, 6, and 10 ALD cycles.
may help suppress the structural instabilities related to lithium in-
2
Downloaded on 2014-06-24 to IP 134.117.10.200 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, 157 ͑1͒ A75-A81 ͑2010͒
A79
Figure 11. ͑Color online͒ Charge–discharge voltage profiles of electrodes
fabricated using Al O ALD and ZnO ALD-coated LiCoO powders at the
2
3
2
3
rd, 10th, and 50th charge–discharge cycles.
powders in Fig. 8. In contrast to the results for Al O ALD, there is
2
3
no improvement in the cycle performance for the ZnO ALD-coated
LiCoO particles compared with electrodes fabricated with the bare
2
LiCoO powders.
2
The ZnO ALD layer may not be stable on the LiCoO particles.
2
The atomic fraction of Zn͓Zn/͑Zn + Co͔͒ and Al͓Al/͑Al + Co͔͒ be-
fore cycling and after 10 charge–discharge cycles was obtained us-
ing XPS. The atomic fraction of Zn dramatically decreased from
Figure 10. ͑Color online͒ Series of mass-normalized impedance spectra for
electrodes fabricated with bare and Al O ALD-coated LiCoO powders for
2
3
2
various charge–discharge cycle numbers.
0
.49 before any charge–discharge cycling to 0.01 after 10 charge–
discharge cycles. In contrast, the atomic fraction of Al maintained
the same value within experimental error. The initial atomic fraction
of Al was 0.55 before cycling and 0.53 after 10 charge–discharge
cycles.
5,37,38
sertion and desertion.
However, the ultrathin Al O ALD coat-
2 3
ing would not have the mechanical strength to withstand lattice ex-
pansions, as originally suggested for much thicker Al O
2
3
5
,37,38
These XPS measurements indicate that the Al2O3 ALD film is
more electrochemically stable than the ZnO ALD film. The Al O
coatings.
Second, the Al O ALD coating may act as a solid
2 3
2
3
electrolyte and may prevent the direct contact between the cathode
6,36
ALD film or the resulting LiAl Co O solid solution remains on
x 1−x 2
surface and the electrolyte.
The Al O ALD-covered surface may
2 3
the LiCoO powders after 10 charge–discharge cycles. In contrast,
2
be less effective for electrolyte decomposition at higher potentials
the ZnO ALD film either diffuses into the LiCoO powders or dis-
2
compared with a bare LiCoO surface. The direct attack of HF that
2
solves into the electrolyte after 10 charge–discharge cycles. The loss
of Zn signal in the XPS spectrum is consistent with no improvement
in the performance of the electrodes fabricated using the ZnO ALD-
coated LiCoO2 powders.
results from the reaction of the trace amounts of water and LiPF in
6
3
9,40
the electrolyte
can also be protected by Al O ALD coating.
2 3
Consequently, Co dissolution from the LiCoO particle by the HF
2
30,41,42
attack may be effectively suppressed.
The Al O ALD film
2 3
43
can serve as a scavenger for HF.
XPS was performed on the fabricated electrodes with 4 cycles of
Al O ALD both before and after 10 charge–discharge cycles. Be-
ALD films on composite electrodes.— ALD films can also be
grown directly on the composite electrode. The internal surfaces of
these composite electrodes could be porous and accessible to the
ALD precursors. ALD reactants such as TMA diffuse easily into
2
3
fore charge–discharge cycling, the binding energy of the Al 2s peak
was 118.7 eV ͓full width at half-maximum ͑fwhm͒ = 2.2 eV͔.
After 10 charge–discharge cycles, the binding energy of the Al 2s
peak was at 119.2 eV ͑fwhm = 2.8 eV͒. The Al 2s peak for Al O
4
6
polymers. A similar high diffusion for TMA may be observed in
the electrolyte and PVDF polymeric binder. The ALD films cover
2
3
could fall in the range of 116.25–119.5 eV, and the Al 2s peak for
44
AlF could be 121.0 eV. The shift in the binding energy and peak
3
broadening of the Al 2s peak after 10 charge–discharge cycles may
indicate the formation of some AlF3.
ZnO ALD on LiCoO powders.— Earlier studies of LiCoO2
2
powders coated with various metal oxides reported that the capacity
retention of coated LiCoO was independent of the specific metal
2
4
5
oxide. To determine if the effect of Al O ALD is unique to Al O ,
2
3
2
3
LiCoO powders were also coated using ZnO ALD. ZnO ALD was
2
deposited on LiCoO powders using four ALD cycles. The stability
2
of the electrodes prepared using the ZnO ALD-coated LiCoO pow-
2
ders was then compared with electrodes fabricated using the Al O
2
3
ALD-coated LiCoO powders.
2
Figure 11 shows that the electrodes fabricated with ZnO ALD-
coated LiCoO powders displayed larger overpotentials after 10 and
2
5
0 charge–discharge cycles compared with the results for Al O
2 3
ALD shown in Fig. 9. Figure 12 also reveals that the ZnO ALD-
coated LiCoO powders displayed a pronounced reduction in capac-
ity with the number of charge–discharge cycles. These results are
2
Figure 12. ͑Color online͒ Cycle performance of electrodes fabricated using
bare, Al O ALD, and ZnO ALD-coated LiCoO powders. Four ALD cycles
2
3
2
very similar to the charge–discharge cycle results for bare LiCoO2
were performed on the LiCoO powders.
2
Downloaded on 2014-06-24 to IP 134.117.10.200 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A80
Journal of The Electrochemical Society, 157 ͑1͒ A75-A81 ͑2010͒
sulted from an increase in the conductivity between particles. The
TiN ALD film could have also formed a passivating layer that pro-
tected the LTS particles. Other investigators have observed similar
performance to the electrodes fabricated using the TiN ALD-coated
4
9
LTS particles without any coatings.
Future prospects.— ALD shows great potential to enhance the
performance of LIBs. Although the results with ultrathin Al O
2
3
ALD films are already exceptional, other ALD materials, such as
TiO and Ta O , may also prove to be useful. ALD can be used to
2
2
5
coat the anode and cathode powder materials used to fabricate elec-
trodes. ALD may also be applied directly to the composite elec-
trodes. In contrast, wet chemistry methods are not able to coat the
fabricated electrodes. There may be performance or convenience
advantages associated with coating either the powders or fabricated
electrodes.
5
0
ALD is also able to deposit nanolaminates and functionalized
5
1
films that may serve various purposes. Thin-film nanoengineering
is possible using the layer-by-layer control provided by ALD. Mo-
lecular layer deposition methods that are similar to ALD and can
Figure 13. ͑Color online͒ Charge–discharge cycle performance of bare and
Al O ALD-coated LiCoO2 electrodes. Two Al O ALD cycles were per-
2
3
2
3
formed on the composite LiCoO electrode.
2
5
2,53
54-56
deposit organic
or hybrid organic–inorganic
polymers are
also available. These organic and organic–inorganic polymers allow
for the tuning of thin-film mechanical properties. This capability
may be useful to handle the expansion and contraction during Li
insertion and removal.
any exposed surface of the LiCoO powders. The only exceptions
2
would be the contact points between the LiCoO particles and other
2
LiCoO particles or AB conducting aids.
2
Figure 13 compares the results for the electrodes prepared using
the bare LiCoO2 powders and the electrodes prepared using the
LiCoO powders coated with Al O ALD using two ALD cycles.
Conclusions
Ultrathin Al O ALD films were grown on LiCoO powders used
as cathodes in LIBs. The ALD coatings dramatically improved the
performance of electrodes fabricated with the Al O ALD-coated
2
3
2
2
2
3
The capacity retention for Al O ALD directly on the composite
2
3
2
3
electrodes shown in Fig. 13 is slightly inferior to the capacity reten-
LiCoO powders. The Al O ALD-coated LiCoO powders using
2
2
3
2
tion for ALD on the LiCoO powders displayed in Fig. 8. However,
2
two ALD cycles showed a capacity retention of 89% after 120
charge–discharge cycles with respect to the reversible capacity at the
third charge–discharge cycle. This behavior was observed when cy-
cling in the range of 3.3–4.5 V ͑vs Li/Li ͒. In contrast, the bare
LiCoO powders maintained only 45% of the initial capacity after
there is a clear enhancement of capacity vs charge–discharge cycle
number compared with the electrodes fabricated using the bare
LiCoO powders. These results confirm the expectation that ALD on
2
+
the composite electrode has coated the exposed LiCoO2 surface
area.
2
the 120 charge–discharge cycles. Similar experiments with ultrathin
ZnO ALD films did not display an enhanced performance. Al O
ALD films directly on the composite electrodes fabricated using the
2
3
Comparison with previous results.— Electrodes fabricated using
the Al O ALD-coated LiCoO powders show a greatly enhanced
2
3
2
bare LiCoO powders also led to an improvement in performance.
2
stability of capacity compared with electrodes fabricated using the
bare LiCoO2 powders. The enhancement in capacity stability is
similar to the improvements observed using much thicker metal ox-
The underlying enhancement mechanism may result from the Al O
2
3
ALD film minimizing Co dissolution or reducing surface/electrolyte
5
reactions. These promising results for Al O ALD-coated LiCoO
2
3
2
ide coatings deposited using wet chemical methods. However, the
powders may lead to other opportunities for ALD to improve the
performance of LIBs.
Al O ALD coating of only 3–4 Å can provide the enhancement
2
3
with a much lower added mass of metal oxide. This lower mass
loading would be particularly important to obtain lighter weight Li-
ion batteries when using nanoparticle electrodes.
Acknowledgments
This work was funded by a subcontract from a DOE SBIR grant
to ALD NanoSolutions ͑Broomfield, CO͒. A.S.C. received addi-
tional support from the iMINT DARPA Center at the University of
Colorado. Y.S.J. also acknowledges a Korea Research Foundation
grant ͑KRF-2008-357-D00066͒.
While ZnO coating by chemical methods improves the cycle
4
7
performance, the ZnO ALD coating does not survive repeated
charge–discharge cycling and displays rapid capacity fading. The
weight fraction of the coating layer produced by a few ALD cycles
is much smaller than the weight fraction produced by chemical
methods. The improved cycle performance of ZnO coatings on
University of Colorado assisted in meeting the publication costs of this
article.
LiCoO obtained using chemical methods can be explained by a
2
4
3
larger quantity of ZnO that is available to scavenge HF.
References
ALD may also be a much more efficient method to obtain the
metal oxide coatings compared with the solution phase techniques.
As a gas-phase method, ALD can deposit the Al O coating without
1. J. M. Tarascon and M. Armand, Nature (London), 414, 359 ͑2001͒.
2
. M. S. Whittingham, Chem. Rev. (Washington, D.C.), 104, 4271 ͑2004͒.
2
3
3
. M. Winter, J. O. Besenhard, M. E. Spahr, and P. Novak, Adv. Mater., 10, 725
the use of solvent and with very high reactant efficiency. Reactant
efficiencies of nearly 100% can be obtained by using static expo-
͑1998͒.
4
5
. G. A. Nazri and G. Pistoia, Lithium Batteries: Science and Technology, Kluwer
Academic, Boston ͑2004͒.
. J. Cho, Y. J. Kim, T. J. Kim, and B. Park, Angew. Chem., Int. Ed., 40, 3367 ͑2001͒.
22,23
sures in rotary ALD reactors
or by monitoring for a reactant
breakthrough” using a mass spectrometer in fluid bed ALD
“
4
8
6. C. Li, H. P. Zhang, L. J. Fu, H. Liu, Y. P. Wu, E. Ram, R. Holze, and H. Q. Wu,
reactors.
Electrochim. Acta, 51, 3872 ͑2006͒.
. J. Cho, Y. W. Kim, B. Kim, J. G. Lee, and B. Park, Angew. Chem., Int. Ed., 42,
1618 ͑2003͒.
. Y. K. Sun, J. M. Han, S. T. Myung, S. W. Lee, and K. Amine, Electrochem.
Commun., 8, 821 ͑2006͒.
. J. Cho, Y. J. Kim, and B. Park, Chem. Mater., 12, 3788 ͑2000͒.
In a previous study, TiN ALD was also used to enhance the
interparticle conductivity and cycle rate of the electrodes fabricated
7
8
9
17
using lithium titanate spinel ͑LTS͒ particles. An enhanced capacity
and cyclability was observed compared with electrodes fabricated
using uncoated LTS particles. However, the mechanism for the en-
hanced performance was unclear. The enhancement could have re-
1
0. S. M. George, A. W. Ott, and J. W. Klaus, J. Phys. Chem., 100, 13121 ͑1996͒.
11. M. Ritala and M. Leskela, Atomic Layer Deposition, Academic, San Diego, CA
Downloaded on 2014-06-24 to IP 134.117.10.200 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, 157 ͑1͒ A75-A81 ͑2010͒
A81
͑
2001͒.
34. B. Kim, C. Kim, T. G. Kim, D. Ahn, and B. Park, J. Electrochem. Soc., 153,
A1773 ͑2006͒.
1
1
1
2. H. Kim, J. Vac. Sci. Technol. B, 21, 2231 ͑2003͒.
3. M. Knez, K. Niesch, and L. Niinisto, Adv. Mater., 19, 3425 ͑2007͒.
4. T. M. Mayer, J. W. Elam, S. M. George, P. G. Kotula, and R. S. Goeke, Appl. Phys.
Lett., 82, 2883 ͑2003͒.
35. M. D. Levi, G. Salitra, B. Markovsky, H. Teller, D. Aurbach, U. Heider, and L.
Heider, J. Electrochem. Soc., 146, 1279 ͑1999͒.
36. L. J. Fu, H. Liu, C. Li, Y. P. Wu, E. Rahm, R. Holze, and H. Q. Wu, Solid State Sci.,
8, 113 ͑2006͒.
1
1
1
5. L. Niinisto, J. Paivasaari, J. Niinisto, M. Putkonen, and M. Nieminen, Phys. Status
Solidi A, 201, 1443 ͑2004͒.
3
7. J. Cho, Y. J. Kim, and B. Park, J. Electrochem. Soc., 148, A1110 ͑2001͒.
6. M. J. Pellin, P. C. Stair, G. Xiong, J. W. Elam, J. Birrell, L. Curtiss, S. M. George,
C. Y. Han, L. Iton, H. Kung, et al. Catal. Lett., 102, 127 ͑2005͒.
7. M. Q. Snyder, S. A. Trebukhova, B. Ravdel, M. C. Wheeler, J. DiCarlo, C. P.
Tripp, and W. J. DeSisto, J. Power Sources, 165, 379 ͑2007͒.
8. A. C. Dillon, A. W. Ott, J. D. Way, and S. M. George, Surf. Sci., 322, 230 ͑1995͒.
9. M. D. Groner, F. H. Fabreguette, J. W. Elam, and S. M. George, Chem. Mater., 16,
3
8. Z. X. Wang, L. J. Liu, L. Q. Chen, and X. J. Huang, Solid State Ionics, 148, 335
͑2002͒.
3
4
4
9. D. Aurbach, J. Electrochem. Soc., 136, 906 ͑1989͒.
0. K. Edstrom, T. Gustafsson, and J. O. Thomas, Electrochim. Acta, 50, 397 ͑2004͒.
1. L. J. Liu, Z. X. Wang, H. Li, L. Q. Chen, and X. J. Huang, Solid State Ionics,
1
1
1
52–153, 341 ͑2002͒.
6
39 ͑2004͒.
0. A. W. Ott, J. W. Klaus, J. M. Johnson, and S. M. George, Thin Solid Films, 292,
35 ͑1997͒.
1. M. D. Groner, J. W. Elam, F. H. Fabreguette, and S. M. George, Thin Solid Films,
13, 186 ͑2002͒.
4
4
4
2. S. Verdier, L. El Ouatani, R. Dedryvere, F. Bonhomme, P. Biensan, and D. Gon-
beau, J. Electrochem. Soc., 154, A1088 ͑2007͒.
3. S. T. Myung, K. Izumi, S. Komaba, Y. K. Sun, H. Yashiro, and N. Kumagai, Chem.
Mater., 17, 3695 ͑2005͒.
4. A. J. Kropf, H. Tostmann, C. S. Johnson, J. T. Vaughey, and M. M. Thackeray,
2
2
2
2
1
4
2. J. A. McCormick, B. L. Cloutier, A. W. Weimer, and S. M. George, J. Vac. Sci.
Electrochem. Commun., 3, 244 ͑2001͒.
Technol. A, 25, 67 ͑2007͒.
4
4
4
4
5. Z. H. Chen and J. R. Dahn, Electrochem. Solid-State Lett., 6, A221 ͑2003͒.
6. C. A. Wilson, R. K. Grubbs, and S. M. George, Chem. Mater., 17, 5625 ͑2005͒.
7. T. Fang, J. G. Duh, and S. R. Sheen, J. Electrochem. Soc., 152, A1701 ͑2005͒.
8. D. M. King, J. A. Spencer, X. Liang, L. F. Hakim, and A. W. Weimer, Surf. Coat.
Technol., 201, 9163 ͑2007͒.
3. J. A. McCormick, K. P. Rice, D. F. Paul, A. W. Weimer, and S. M. George, Chem.
Vap. Deposition, 13, 491 ͑2007͒.
2
2
2
2
2
4. A. Yamada, B. S. Sang, and M. Konagai, Appl. Surf. Sci., 112, 216 ͑1997͒.
5. E. B. Yousfi, J. Fouache, and D. Lincot, Appl. Surf. Sci., 153, 223 ͑2000͒.
6. J. W. Elam and S. M. George, Chem. Mater., 15, 1020 ͑2003͒.
7. P. J. Cumpson, Surf. Interface Anal., 29, 403 ͑2000͒.
8. C. J. Powell and A. Jablonski, NIST Electron-Effective-Attenuation-Length Data-
base, NIST SRD 82, Version 1.1, National Institute of Standards and Technology,
Gaithersburg, MD ͑2003͒.
4
5
5
9. K. Nakahara, R. Nakajima, T. Matsushima, and H. Majima, J. Power Sources, 117,
1
31 ͑2003͒.
0. F. H. Fabreguette, R. A. Wind, and S. M. George, Appl. Phys. Lett., 88, 013116
2006͒.
͑
1. C. F. Herrmann, F. H. Fabreguette, D. S. Finch, R. Geiss, and S. M. George, Appl.
Phys. Lett., 87, 123110 ͑2005͒.
2
3
9. L. J. van der Pauw, Philips Tech. Rev., 20, 220 ͑1958͒.
0. Y. J. Kim, H. Kim, B. Kim, D. Ahn, J. G. Lee, T. J. Kim, D. Son, J. Cho, Y. W.
Kim, and B. Park, Chem. Mater., 15, 1505 ͑2003͒.
52. N. M. Adamczyk, A. A. Dameron, and S. M. George, Langmuir, 24, 2081 ͑2008͒.
53. Y. Du and S. M. George, J. Phys. Chem. C, 111, 8509 ͑2007͒.
54. A. A. Dameron, D. Saghete, B. B. Burton, S. D. Davidson, A. S. Cavanagh, J. A.
Bertand, and S. M. George, Chem. Mater., 20, 3315 ͑2008͒.
55. S. M. George, B. Yoon, and A. A. Dameron, Acc. Chem. Res., 42, 498 ͑2009͒.
56. B. Yoon, J. L. O’Patchen, D. Seghete, A. S. Cavanagh, and S. M. George, Chem.
Vap. Deposition, 15, 112 ͑2009͒.
3
3
3
1. Y. J. Kim, T. J. Kim, J. W. Shin, B. Park, and J. P. Cho, J. Electrochem. Soc., 149,
A1337 ͑2002͒.
2. S. Oh, J. K. Lee, D. Byun, W. I. Cho, and B. W. Cho, J. Power Sources, 132, 249
͑
2004͒.
3. N. Kosova, E. Devyatkina, A. Slobodyuk, and V. Kaichev, Solid State Ionics, 179,
745 ͑2008͒.
1
Downloaded on 2014-06-24 to IP 134.117.10.200 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).