Nanocomposite of amorphous Ge and Sn nanoparticles as an anode material for Li secondary battery

Article abstract of DOI:10.1149/1.3073877

Nanocomposites of amorphous Ge and Sn nanoparticles are prepared from the vacuum annealing of the butyl-capped Ge gels and Sn P0.94 nanoparticles at 400°C for 7 h, and their electrochemical properties and nanostructure changes are investigated during electrochemical cycling. The nanocomposite displays a first charge capacity of 1078 mAhg with a capacity retention of 64% after 50 cycles. However, both individual Ge and Sn nanoparticles display rapid capacity fade, showing <20% capacity retention after 50 cycles. A superior electrochemical property of the nanocomposite to the Sn and Ge nanoparticles is believed to be related to the fact that the amorphous Ge phase acts as an intergrain electrical connecter when Sn particles are pulverized.

Full text of DOI:10.1149/1.3073877

Journal of The Electrochemical Society, 156 ͑4͒ A277-A282 ͑2009͒
A277
0
013-4651/2009/156͑4͒/A277/6/$23.00 © The Electrochemical Society
Nanocomposite of Amorphous Ge and Sn Nanoparticles
as an Anode Material for Li Secondary Battery
a
b,z
Min Gyu Kim and Jaephil Cho
a
Pohang Accelerator Laboratory, Beamline Research Division, Pohang 790-784, Korea
Division of Energy Engineering, Ulsan National Institute of Science and Technology,
b
Ulsan 689-805, Korea
Nanocomposites of amorphous Ge and Sn nanoparticles are prepared from the vacuum annealing of the butyl-capped Ge gels and
SnP0.94 nanoparticles at 400°C for 7 h, and their electrochemical properties and nanostructure changes are investigated during
electrochemical cycling. The nanocomposite displays a first charge capacity of 1078 mAh/g with a capacity retention of 64% after
5
5
0 cycles. However, both individual Ge and Sn nanoparticles display rapid capacity fade, showing Ͻ20% capacity retention after
0 cycles. A superior electrochemical property of the nanocomposite to the Sn and Ge nanoparticles is believed to be related to the
fact that the amorphous Ge phase acts as an intergrain electrical connecter when Sn particles are pulverized.
2009 The Electrochemical Society. ͓DOI: 10.1149/1.3073877͔ All rights reserved.
©
Manuscript submitted October 27, 2008; revised manuscript received December 26, 2008. Published February 4, 2009.
Lithium reactive metals including Sn metal have been intensively
investigated as high-capacity anode materials in lithium batteries,
However, when using the composite electrode with a thickness of
Ͼ30 ␮m like our study, achievement of such cycle stability is dif-
ficult due to an increased volume change. In addition, such a uni-
form distribution of each active phase, as in the electroplated elec-
trode, is impossible in the conventional composite.
In order to overcome this problem and, further, to achieve higher
capacity, active–active crystalline Sn and amorphous Ge nanocom-
posite is adopted in this study. The present study reports the signifi-
cantly improved electrochemical property of the nanocomposite of
amorphous Ge and Sn nanoparticles prepared from the thermal de-
composition of the mixture of butyl-capped Ge gels and SnP0.94 at
400°C compared to individual Sn or Ge nanoparticles. Such a nano-
composite consisting of a uniformly dispersed amorphous Ge phase
between the crystalline Sn phase has not been reported in the open
literature. Compared with the capacity retention of the Sn nanopar-
ticles, the nanocomposite demonstrates over 55% capacity enhance-
ment.
1
-10
However, the particle pulverization associated with these materials,
which originates from a large volume change ͑Ͼ200%͒ during
lithium alloying and dealloying, can result in electrically discon-
nected smaller particles, causing the electrode capacity to fade. In
order to reduce this considerable volume change, an inactive phase
that can prevent the aggregation of the particle growth and that acts
as an electrically connecting medium between anode particles and
current collector when the particle is pulverized was first proposed
by Huggins’ group in what they termed a mixed-conductor matrix
11
concept ͑active–inactive composites͒. The matrix should allow
rapid transport of the electroactive species ͑Li͒, act as current col-
lector, and maintain the microscopic morphology of the dispersed
reactant. As an example, Li Si in a matrix of Li Sn was proposed,
x
7
3
in which cycling took place in a potential range wherein Li Sn did
7
3
not react. In this case, the Li Sn phase can act as a conduction
7
3
media.
A similar concept was introduced with the Sn-based amorphous
composite oxide by Fuji, with the first discharge to 0 V. They found
that the amorphous tin-based oxide ͑Sn1.0B0.56P0.4Al0.42O3.6͒ ͑TCO͒
Experimental
Sample preparation.— The SnP0.94 was prepared from thermal
decomposition of Sn acetate, Sn͑C H O ͒ ͑Aldrich, 99%͒, in a
is irreversibly reduced to metallic Sn and Li O, and the other non-
2
2
3
2 2
reacting components in glasses are electrochemically inert with re-
high-temperature solution of trioctyl phosphine ͑TOP͒,
͓CH ͑CH ͒ ͔ P ͑Aldrich 99%͒, and TOP oxide, ͓CH ͑CH ͒ ͔ PO
͑Aldrich 99%͒ at 390°C for 1 h according to a previously reported
1
2
spect to Li. Courtney et al. investigated SnO -based anodes ͑TCO͒
x
3
2 7 3
3
2 7 3
in detail and proposed that the network formers and Li O, which is
2
20
formed during the first charge, hinder smaller Sn clusters to aggre-
method. In order to prepare amorphous Ge viscous gels, sodium
naphthalide solution was prepared from 0.69 g of sodium and 2.93 g
13,14
gate to larger ones.
A greater amount of Li O slows the aggre-
2
3
gation of the Sn, resulting in higher cycling stability but lowering
the gravimetric and volumetric capacities. As a result, the formation
of large tin regions can be suppressed, which are more sensitive to
cracking and crumbling because of the higher absolute volume
changes than small ones. The suppression of Sn aggregation and the
stabilization of the composite microstructure by the use of network
formers are in line with the mixed-conductor matrix concept pro-
posed by Huggins et al. A similar approach was proposed by Mao et
of naphthalene stirred in 70 cm of 1.2-dimethoxyethane. GeCl4
͑1.04 g͒ and 1,2-dimethoxyethane ͑100 mL͒ were thoroughly
mixed, and the mixture was poured into a sodium naphthalide solu-
tion. This mixed solution was stirred for 2 h, and 0.1 g of butyl-
lithium ͑LiC H ͒ was then added ͑butyl groups were replaced by Cl
4
9
in GeCl and Li in LiC H was reacted into LiCl, which was re-
4
4
9
moved by washing with water͒. Instantly, a yellow solution was
obtained, and the solvent and naphthalene were removed using a
rotating evaporator. Finally, the resulting viscous butyl-capped Ge
gel was washed with water six times and was dispersed in n-hexane,
followed by vacuum-drying at 200°C for 24 h. The final product
was homogeneously mixed with SnP0.94 nanoparticles using an ag-
ate mortar and pestle. The final mixture was vacuum-annealed at
00°C for 2 or 7 h. For preparing the amorphous Ge nanoparticles,
the viscous gels were annealed at 400°C for 7 h in a vacuum.
For preparing pure Sn nanoparticles as a reference sample,
1
5,16
al. by using various intermetallic phases of Sn and Fe.
During
lithiation the Sn acts as reactant, forming Li Sn, which is surrounded
x
by the remaining elemental Fe. The fine Fe ͑ϳ10 nm͒ does not
alloy with lithium but acts as an inactive matrix and supports the
intergrain electronic contact in the material. Unlike the TCO, where
the Sn remains after dealloying, parts of the original Sn–Fe interme-
tallic phase are restored during dealloying of the Sn.
4
Active–active composites, such as electroplated multiphase
Sn/SnSb and SnSnAg /SnAg films with
a
thickness of
3
4
1.17 mL of anhydrous SnCl in 40 mL of 1.2-dimethoxyethane was
4
1
7-19
Ͻ3 ␮m,
were reported. These were a composite film and, there-
mixed thoroughly with 0.8 g of hydrobenzamide in 30 mL of 1.2-
dimethoxyethane. As a reducing agent, 0.75 g of NaBH was also
fore, showed stable capacities of ϳ600 mAh/g out to 100 cycles.
4
dissolved in 50 mL of 1.2-dimethoxyethane and poured into the
mixed SnCl solution. The mixture was stirred for 30 min at room
temperature under an argon atmosphere. After the reaction, the mix-
4
z
E-mail: jpcho@unist.ac.kr
Downloaded on 2015-06-20 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).

A278
Journal of The Electrochemical Society, 156 ͑4͒ A277-A282 ͑2009͒
mator was employed to monochromatize the X-ray photon energy.
o
1
800
600
After annealing at 400 C for 7h
Higher-order harmonic contaminations were eliminated by detuning
the monochromator to reduce the incident X-ray intensity by ϳ25%.
The incident X-ray intensity was monitored with pure nitrogen gas-
filled ionization chambers. The spectroscopic data were collected in
transmittance mode. Energy calibration was made using standard
metallic Ge film.
impurity peaks
1
1
200
o
After annealing at 400 C for 2h
Results and Discussion
Figure 1 shows XRD patterns of the as-prepared SnP0.94 and the
nanocomposite of the butyl-capped Ge gels and SnP0.94 nanopar-
ticles after vacuum annealing at 400°C for 2 and 7 h. All of the
peaks of the SnP0.94 can be indexed to the hexagonal cell of the
SnP0.94 phase with space group P3m1 without showing the forma-
tion of other tin phosphide phases. Two unidentified impurity peaks
at approximately 26.5 and 34° are not assigned to SnP0.94, Sn, or
8
00
00
SnP
0
.94
4
0
2
0
25
30
35
40
45
θ
50
55
60
2
1-23
Scattering angle (2
)/degree
SnP.
Upon annealing at 400°C for 2 h, Ge peaks were not detected,
indicating that Ge existed as an amorphous phase. In addition, te-
tragonal ␤-Sn peaks are observed, indicating that SnP0.94 was de-
composed to the Sn phase. Upon further annealing for 7 h, the
SnP0.94 phase completely disappeared and only the Sn phase was
observed. Synthesis of the SnP0.94 is not risky, and it can be easily
prepared with a large quantity in an inert atmosphere. Furthermore,
SnP0.94 particles are stable in open air but start to decompose above
Figure 1. XRD patterns of the as-prepared SnP0.94 and the nanocomposite of
butyl-capped Ge gels and SnP0.94 nanoparticles after thermal annealing at
4
00°C for 2 and 7 h.
ture was washed four times with water and acetone and evaporated
under vacuum to yield tin particles at room temperature.
3
00°C. The purpose of using the SnP0.94 was to get a homoge-
Characterization.— High-resolution transmission electron mi-
croscopy ͑HRTEM͒ samples were prepared by the evaporation of
dispersed nanoparticles in acetone or hexane on carbon-coated cop-
per grids. The field-emission electron microscope used at this stage
was a JEOL 2100F device operating at 200 kV. After sonication and
centrifugation of the composite electrode in acetone, the upper part
of the solution was used for ex situ transmission electron micros-
copy ͑TEM͒ analysis.
Powder X-ray diffraction ͑XRD͒ measurements were carried out
with a Cu-target tube. Inductively coupled plasma-mass spectros-
copy ͑ICP, ICPS-1000IV, Shimadzu͒ was used to determine the Sn
and P contents. The electrode consisted of 80 wt % active materials,
neously mixed nanocomposite of the amorphous Ge and Sn nano-
particles. When separately prepared Sn nanoparticles and viscous
butyl-capped Ge gels were used as precursors, each of them was
found to be severely locally segregated.
Figure 2 shows TEM images of the as-prepared SnP0.94 and the
nanocomposite after vacuum annealing at 400°C for 2 and 7 h. As-
prepared SnP0.94 nanoparticles consist of tear-drop-shaped nanopar-
ticles with a length of Ͻ ϳ0.5 ␮m ͑image a͒. After mixing with
butyl-capped Ge gels and thermal annealing at 400°C for 2 h ͑im-
age b͒, the particle morphology was partially transformed into
spherical particles, as denoted by the white dotted circles in the
figure. After annealing at 400°C for 7 h, all particle morphologies
became spherical with a coating layer ͑image c͒. The HRTEM image
in Fig. 2d confirms the formation of ␤-Sn ͑region 1͒. The energy-
dispersive X-ray ͑EDX͒ pattern of the dispersed phase between the
Sn nanoparticles ͑region 2͒ confirms the formation of Ge ͑image e͒.
Additionally, the electron diffraction pattern of region 2 confirms the
formation of the amorphous Ge phase ͑inset of image e͒. EDX
analysis in Fig. 2e shows a small oxygen peak at ϳ1.5 keV, corre-
sponding to Ͻ3 wt % oxygen content, and we believe this contami-
10 wt % Super P carbon black, and 10 wt % polyvinylidene fluo-
ride. A mixture of ethylene carbonate/diethylene carbonate with 1 M
LiPF salt was used as the electrolyte. The cell was cycled at a rate
6
of 0.2C ͑=440 mA/g͒ between 0 and 1.2 V. Ge K-edge X-ray ab-
sorption spectra ͑XAS͒ of Ge-coated Sn for the first discharge–
charge process were taken on a BL7C1 ͑Electrochemistry͒ beamline
in a storage ring of 2.5 GeV with a ring current of 130–190 mA at
the Pohang Light Source ͑PLS͒. A Si͑111͒ double-crystal monochro-
Figure 2. TEM images of ͑a͒ the as-
prepared SnP0.94, ͑b͒ the nanocomposite of
butyl-capped Ge gels and SnP0.94 nanopar-
ticles after thermal annealing at 400°C for
2
h, ͑c͒ the nanocomposite of butyl-
capped Ge gels and SnP0.94 nanoparticles
after thermal annealing at 400°C for 7 h,
͑
d͒ an HRTEM image of region 1 in c, ͑e͒
EDX spectrum of region 2 in c ͑inset is
selected area diffraction pattern of region
2
͒, and ͑f͒ an HRTEM image of region 3
in c.
Downloaded on 2015-06-20 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, 156 ͑4͒ A277-A282 ͑2009͒
a) 2.5
A279
(
2
1
1
0
0
.0
.5
.0
.5
.0
C rate = 0.2 C ( = 440 mA/g)
50
40 30 2010 1
0
200 400 600 800 1000 1200 1400 1600 1800
Capacity (mAh/g)
Figure 3. ͑Color online͒ P and Sn elemental line scans across the particle
along the line of the as-prepared SnP0.94 nanoparticle and the nanocomposite
of butyl-capped Ge gels and SnP0.94 nanoparticles after thermal annealing at
(b)
1
400
1400
o
Amorphous Ge (obtained at 400 C)
Nanocomposite
4
00°C for 7 h.
1200
1200
1000
800
600
400
200
0
1
000
00
600
00
8
nation may be from partial surface oxidation of the Ge during
SnP0.94 decomposition. An expanded TEM image of region 3 shows
that part of the amorphous Ge is coated on the Sn nanoparticles
Physical mixture of
Sn and Ge
4
͑
image f͒. The Sn particle size estimated at ϳ100 nm is much
smaller than the original SnP0 nanoparticles, which may be related
200
0
.94
Sn nanoparticles
to the decomposition of P from the particles, as SnP0.94 started to
decompose above 300°C. Because the melting point of P is ϳ44°C,
0
5
10 15 20 25 30 35 40 45 50
Cycle number
decomposed P from the SnP0 is susceptible to evaporation during
.94
vacuum annealing at 400°C. Actually, the cooling part of the inside
of the quartz tube of the furnace was observed precipitation of the
phosphor powder.
Figure 4. ͑a͒ Cycling curves of the nanocomposite of butyl-capped Ge gels
and SnP0.94 nanoparticles after thermal annealing at 400°C for 7 h between 0
and 1.2 V at a rate of 0.2C ͑=440 mA/g͒ in a coin-type half-cell, and ͑b͒
charge capacity vs the cycle number of the nanocomposite of butyl-capped
Ge gels and SnP0.94 nanoparticles after thermal annealing at 400°C for 7 h,
physical mixture of Sn and amorphous Ge nanoparticles, Sn nanoparticles,
and amorphous Ge nanoparticles prepared at 400°C for 7 h.
Figure 3 shows line scans of the Sn and P elements that enable
observation of the change of Sn and P contents before and after
annealing at 400°C for 7 h. After thermal annealing, the P content
appears to decrease significantly, while the Sn content remains con-
stant. This result indicates that genuine P loss occurs. ICP analyses
of Sn and P confirmed that the sample annealed for 7 h exhibited
SnP0.05. This means that during thermal annealing, small amounts of
P may reside on particle surfaces, although most of the P in SnP0.94
has been evaporated in vacuum annealing. Accordingly, TEM and
ICP results imply that P may reside on the Sn particle surface. The
weight percent of Sn and Ge was estimated as 70 and 30 after
thermal annealing at 400°C for 7 h, respectively; this ratio shows
the best electrochemical cycling result, as discussed below. Even
though there is a possible formation of solid solution between Sn
and Ge, EDX spectroscopy of the region of amorphous Ge does not
show the Sn peaks ͑Fig. 2e͒. Moreover, line scans of the annealed
sample in Fig. 3 do not show Ge, indicating no possible solid-
solution formation. If pure Sn nanoparticles and amorphous Ge can
be directly annealed at 400°C, there is great possibility of solid-
solution formation. However, the decomposition reaction of SnP0.94
is expected to hinder the solid-solution reaction between Sn and Ge.
Figure 4a shows voltage profiles of the nanocomposite obtained
after annealing at 400°C for 7 h in 1, 10, 20, 30, 40, and 50 cycles
between 1.2 and 0 V at a rate of 0.2C ͑=440 mA/g͒ in coin-type
half-cells. A small bump around 1.5 V during the first discharge may
be due to the decomposition of the surface oxide layer of the nano-
composite. The nanocomposite shows the first discharge and charge
capacities of 1638 and 1078 mAh/g, respectively, with an irrevers-
ible capacity ratio of 34%. Such a large first discharge capacity is
due to the severe side reaction of the enlarged active surface area
to the oxides decompositions. The capacity-retention ratio of the
nanocomposite is found to be 66% after 50 cycles. When the indi-
vidual amorphous Ge and Sn nanoparticles ͑30 wt % Ge and
70 wt % Sn͒ were physically blended, we could not get good dis-
persion of them because the nanoparticles severely aggregated each
other. The electrochemical result of this mixture is much inferior to
that of the nanocomposite ͑Fig. 4a͒, and the capacity retention is
only 28% of the first charge capacity ͑1400 mAh/g͒ after 22 cycles.
The first charge capacity of Sn nanoparticles was 920 mAh/g, and
the capacity retention was 5% after 30 cycles. Moreover, the capac-
ity retention of the amorphous Ge nanoparticles obtained from ther-
mal annealing of the viscous butyl-capped Ge gels at 400°C for 7 h
shows 21% ͑from 1450 to 300 mAh/g after 50 cycles͒. Based upon
the charge capacities of the individual Sn and Ge, which were 920
and 1450 mAh/g, respectively, the estimated charge capacity of the
composites is 1079 mAh/g. This value agrees well with the experi-
mental charge capacity of the composite ͑Sn:Ge = 70:30 wt %͒,
which is 1078 mAh/g ͑Fig. 4a͒. Because both Sn and Ge are elec-
trochemically active and physically dispersed in the composites, the
lithium alloy–dealloy process occurs simultaneously according to
the following equations: 4.4Li + Sn ↔ Li4.4Sn and 4.4Li
+ Ge ↔ Li4.4Ge.
Pure amorphous Ge nanoparticles with a size of Ͻ20 nm, pre-
pared at 200°C, showed relatively good dispersions between the
24
with the electrolyte. In the side reaction, solvent molecules and
salt anions are reduced on the active surface, forming insoluble Li
salts that precipitate to form a passivating film surface. In addition,
the relatively large contact area of the electrode/electrolytes of the
nanocomposite leads to intensified surface reactions and hence to
their high irreversible capacity. Another source that may contribute
2
5
nanoparticles and good capacity retention with Ͼ1000 mAh/g. In
our case, however, the composite should be annealed at 400°C, and
amorphous Ge nanoparticles appear to be severely aggregated. This
is
a
huge difference, compared to pure amorphous Ge
25
nanoparticles, but the particle size is unchanged ͑Fig. 1e͒. Actu-
to the irreversible capacity is the formation of MO , and the decom-
ally, aggregated nanoparticles have a higher chance for earlier par-
x
2
6
position plateau of MO to M ͑M = Sn or Ge͒ and Li O. As can be
ticle pulverization compared to well-dispersed analogues. As the
x
2
seen in Fig. 4a, there is a small bump at ϳ1.6 V, which may be due
particle size of the Sn nanoparticles ͑Fig. 5a͒ is similar to that of the
Downloaded on 2015-06-20 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).

A280
Journal of The Electrochemical Society, 156 ͑4͒ A277-A282 ͑2009͒
Figure 5. TEM images of ͑a͒ the as-prepared Sn nanoparticles and ͑b͒ Sn
nanoparticles after 30 cycles.
nanocomposite, the considerable capacity difference between these
samples is associated with the role of amorphous Ge during cycling.
In the case of the composite consisting of Sn nanoparticles with an
average particle size of 30 nm and carbon composite, the amount of
2
7
carbon estimated from the capacity was Ͼ40 wt %. The capacity
contribution of the carbon was 22.5 mAh/g, and carbon therefore
acts as an inactive matrix. This composite shows a reversible capac-
ity of 500 mAh/g with a coulombic efficiency of 70% ͑for the first
cycle͒ out to 200 cycles.
Figure 6 shows normalized Ge K-edge X-ray absorption near-
edge spectra ͑XANES͒ and Fourier-transformed ͑FT͒ extended
X-ray absorption fine structure ͑EXAFS͒ spectra for the nanocom-
posite for the first discharge and charge. The peak positions of all
the Ge in the XANES spectra of the nanocomposite are consistent
with that of reference metallic Ge film, indicating maintenance of
Ge metallic phase throughout the electrochemical reaction. The
characteristic FT peak position at ϳ2.2 Å corresponding to pure
metallic Ge–Ge bond are constant for the first cycling, which means
there is no solid solution between Sn and Ge in the nanocomposite.
The FT peak intensity of Ge–Ge bonding shows a systematic de-
crease during the first discharge process. In the fully discharged
nanocomposite, the FT peak decreases abruptly, meaning Ge–Ge
metallic bond–cleavage or formation of a smaller Ge metallic clus-
ter. The following charge to 1.2 V leads to a relatively reversible
increase of FT peak intensity, which is due to the formation of Ge
nanocrystals by reformation of Ge–Ge metallic bonding from the
small Ge cluster. This structural variation of the nanocomposite for
the electrochemical reaction remains in the second discharge–charge
process, as shown in Fig. 6. Finally, it is expected that with initial
lithiation into the nanocomposite, the Li-ion can at first be incorpo-
rated into Ge–Ge chemical bonding in the amorphous Ge phase. The
volume expansion by alloying with the Sn phase can also accelerate
the pulverization of the amorphous Ge phase. The smaller Ge cluster
is rearranged and aggregated in the subsequent delithiation process,
acting as a structural buffer phase. Therefore, it can be deduced that
the pulverized amorphous Ge phase during lithiation and delithiation
lends the percolation path for the electrical conduction, which re-
sults in better electrochemical performance of the nanocomposite
with respect to the pure Ge and Sn nanoparticles.
Figure 6. ͑Color online͒ Ge K-edge XANES and FT EXAFS data of the
nanocomposite for the first and second cycling.
cluster starts to become pulverized, as supported by the TEM image
in Fig. 7c. After 20 cycles, the large particle has completely disap-
peared and spherical particles with a size of ϳ100 nm are dispersed
between the Ge matrix, as confirmed by the EDX spectroscopy
marked by dotted circles in Fig. 7d. After 50 cycles, Sn nanopar-
ticles Ͻ10 nm are observed ͑Fig. 7e and f͒, indicating that 100 nm
sized nanoparticles continue to be pulverized into much smaller
nanoparticles.
A decreasing Sn particle size was also observed in ex situ XRD
patterns of the cycled nanocomposite electrodes after 5, 20, and
50 cycles, as shown in Fig. 8. The crystallite size of the Sn nano-
particles after 5, 20, and 50 cycles was determined from the width of
In order to investigate the origin of the difference in capacity
retention between the nanocomposites and pure Sn, the nanocom-
posite after 5, 20, and 50 cycles were extracted from the cell and a
TEM analysis was carried out. Figure 7a is a TEM image after five
cycles ͑the electrode was detached from the cell at 1.2 V charged
state͒ in which uniformly dispersed nanoparticles in a large particle
with a size of 0.8 ␮m can be observed. High-resolution electron
microscopy analysis of the large particle confirms the tetragonal
2
9
the peaks according to the Scherrer equation and was estimated to
be 180, 100, and 10 nm, respectively. These results are in good
agreement with TEM images of the cycled electrodes, as shown in
Fig. 7. As discussed above, the significant capacity-retention differ-
ence between the pure Sn and the nanocomposite is related to the
dispersed amorphous Ge particles, which act as a percolation path
for electrical conduction. A similar result was reported for Sn nano-
particles in Li–P–O matrix phase, which showed no capacity fade
out to 100 cycles, although Sn nanoparticles showed severe
␤
-Sn particle ͑Fig. 7b͒. Compared with a TEM image before cycling
͑
Fig. 2c͒, its size after five cycles showed an eightfold increase,
indicating that a larger Sn cluster forms at the expense of smaller
clusters. The formation of the larger Sn cluster is due to the mini-
2
8
30
mization of the surface-interfacial energy among the particles. Sn
particles aggregate into larger clusters until a maximum cluster size
is reached. However, upon reaching the maximum cluster size, the
pulverization. This improvement was due to the Li–P–O matrix
providing conduction paths between the Sn nanoparticles. In con-
trast to the observation of pulverized Sn nanoparticles which have
Downloaded on 2015-06-20 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, 156 ͑4͒ A277-A282 ͑2009͒
A281
Figure 7. TEM images of the nanocom-
posite of butyl-capped Ge gels and SnP0.94
nanoparticles after thermal annealing at
00°C for 7 h after 5, 20, and 50 cycles
a, c, and e, respectively͒, ͑b͒ HRTEM im-
age of a, ͑d͒ EDX spectrum of the dotted
circle in ͑c͒, and ͑f͒ HRTEM image of ͑e͒.
4
͑
electrically isolated each other ͑Fig. 5b͒, amorphous Ge particles
were shown to be dispersed between the Sn particles, providing
percolation pathways for the conduction ͑Fig. 7e and f͒.
Cu foil
Conclusions
A nanocomposite consisting of amorphous Ge and Sn nanopar-
ticles was prepared via thermal annealing of the mixture of butyl-
capped Ge gels and SnP0.94 nanoparticles at 400°C under a vacuum.
They showed a first charge capacity of 1078 mAh/g with a capacity
retention of 64% after 50 cycles. However, pure Sn nanoparticles
displayed a fast capacity fade, showing 5% capacity retention after
only 20 cycles. The results indicated that while Sn particles within a
cluster became electrically isolated and became inactive as a lithium
metal alloy material, amorphous Ge phase dispersed in the Sn nano-
particles provided conduction paths. However, one of the critical
problems of nanomaterials like this composite is a too-high
electrolyte/electrode surface area that may lead to more significant
side reactions with electrolyte. This causes high irreversible capacity
and low capacity retention. We cannot underestimate contributions
After 50 cycles
After 20 cycles
After 5 cycles
20
25
30
35
40
45
θ
50
55
60
3
1
of the nanomaterials for high-rate and high-capacity applications.
Scattering angle (2 )/degree
However, coulombic efficiency of the nanocomposites may be im-
proved by introducing inactive carbon matrix even though it sacri-
fices the capacity decrease.
Figure 8. Ex situ XRD patterns of the nanocomposite electrodes after 5, 20,
and 50 cycles.
Downloaded on 2015-06-20 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).

A282
Journal of The Electrochemical Society, 156 ͑4͒ A277-A282 ͑2009͒
1
1
1
1. B. A. Boukamp, G. C. Lesh, and R. A. Huggins, J. Electrochem. Soc., 128, 725
1981͒.
2. Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, and T. Miyasaka, Science, 276,
997 ͑1997͒.
3. I. A. Courtney, W. R. McKinnon, and J. R. Dahn, J. Electrochem. Soc., 146, 59
͑1999͒.
Acknowledgments
͑
This work was supported by the Information Technology Re-
search and Development program of MKE/IITA ͑Core Lithium Sec-
ondary Battery Anode Materials for Next Generation Mobile Power
Module, 2008-F-019-01͒. The authors acknowledge PLS for the
XAS measurements ͑BL7C1 beam line͒.
1
1
1
4. I. A. Courtney and J. R. Dahn, J. Electrochem. Soc., 144, 2943 ͑1997͒.
5. O. Mao, R. A. Dunlap, I. A. Courtney, and J. R. Dahn, J. Electrochem. Soc., 145,
4
195 ͑1998͒.
1
1
1
1
2
6. O. Mao and J. R. Dahn, J. Electrochem. Soc., 146, 423 ͑1999͒.
7. M. Wachtler, J. O. Besenhard, and M. Winter, J. Power Sources, 94, 189 ͑2001͒.
8. J. Yang, M. Winter, and J. O. Besenhard, Solid State Ionics, 90, 281 ͑1996͒.
9. M. Winter and J. O. Besenhard, Electrochim. Acta, 45, 31 ͑1999͒.
0. Y. Kim, H. Hwang, M. G. Kim, C. S. Yoon, and J. Cho, Adv. Mater. (Weinheim,
Ger.), 19, 92 ͑2007͒.
Ulsan National Institute of Science and Technology assisted in meeting
the publication costs of this article.
References
1
2
. J. Yin, M. Wada, S. Tanase, and T. Sakai, J. Electrochem. Soc., 151, A583 ͑2004͒.
. M. Noh, Y. Kim, M. G. Kim, H. Lee, H. Kim, Y. Kwon, Y. Lee, and J. Cho, Chem.
Mater., 17, 3320 ͑2005͒.
2
2
2
2
1. P. C. Donohue, Inorg. Chem., 9, 335 ͑1970͒.
2. O. Olofsson, Acta Chem. Scand. (1947-1973), 24, 1153 ͑1970͒.
3. P. Eckerlin and W. Kischio, Z. Anorg. Allg. Chem., 363, 1 ͑1968͒.
4. D. Aurbach, A. Nimberger, B. Markovsky, E. Levi, E. Sominsky, and A. Gedan-
ken, Chem. Mater., 14, 4155 ͑2002͒.
3
. J. Graetz, C. C. Ahn, R. Yazami, and B. Fultz, J. Electrochem. Soc., 151, A698
͑
2004͒.
4
5
. J. Yang, M. Winter, and J. O. Besenhard, Solid State Ionics, 90, 281 ͑1996͒.
. I.-S. Kim, G. E. Blomgren, and P. N. Kumta, Electrochem. Solid-State Lett., 7,
A44 ͑2004͒.
2
5. H. Lee, M. G. Kim, C. H. Choi, Y. K. Sun, C. S. Yoon, and J. Cho, J. Phys. Chem.
B, 109, 20719 ͑2005͒.
6. Y. Kwon and J. Cho, Chem. Commun. (Cambridge), 2008, 1109.
7. G. Derrien, J. Hassoun, S. Panero, and B. Scrosati, Adv. Mater. (Weinheim, Ger.),
2
2
6
7
. J. Yang, M. Wachtler, M. Winter, and J. O. Besenhard, Electrochem. Solid-State
Lett., 2, 161 ͑1999͒.
. N. Tamura, R. Ohshita, M. Fujimoto, M. Kamino, and S. Fujitani, J. Electrochem.
1
9, 2336 ͑2007͒.
28. C. Kim, M. Noh, M. Choi, J. Cho, and B. Park, Chem. Mater., 17, 3297 ͑2005͒.
29. L. S. Birks and H. Friedman, J. Appl. Phys., 17, 687 ͑1946͒.
30. E. Kim, M. G. Kim, and J. Cho, Electrochem. Solid-State Lett., 9, A311 ͑2006͒.
31. P. G. Bruce, B. Scrosati, and J.-M. Tarascon, Angew. Chem., Int. Ed., 47, 2930
͑2008͒.
Soc., 150, A679 ͑2003͒.
8
9
. S. Lai, J. Electrochem. Soc., 123, 1196 ͑1976͒.
. R. A. Sharma and R. N. Seefurth, J. Electrochem. Soc., 123, 1763 ͑1976͒.
1
0. R. A. Huggins, J. Power Sources, 22, 341 ͑1988͒.
Downloaded on 2015-06-20 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).