Silicon Nanotube Battery Anodes
Mi-Hee Park,
†
Min Gyu Kim,
‡
Jaebum Joo,
§
Kitae Kim,
|
Jeyoung Kim,
|
Soonho Ahn,
|
Yi Cui,*
,⊥
and Jaephil Cho*
,†
School of Energy Engineering, Ulsan National Institute of Science & Technology,
Ulsan, Korea 689-798, Beamline Research DiVision, Pohang Accelerator Laboratory,
Pohang, Korea 790-784, Department of Applied Chemistry, Hanyang UniVersity,
Ansan, Korea 426-791, Battery R&D, LG Chem, Ltd. 104-1, Moonji-dong, Yuseong-gu,
Daejeon, Korea 305-380, and Department of Materials Science and Engineering,
Stanford UniVersity, Stanford, California 94305
Received June 26, 2009; Revised Manuscript Received August 23, 2009
ABSTRACT
We present Si nanotubes prepared by reductive decomposition of a silicon precursor in an alumina template and etching. These nanotubes
show impressive results, which shows very high reversible charge capacity of 3247 mA h/g with Coulombic efficiency of 89%, and also
demonstrate superior capacity retention even at 5C rate ()15 A/g). Furthermore, the capacity in a Li-ion full cell consisting of a cathode of
LiCoO
2
and anode of Si nanotubes demonstrates a 10 times higher capacity than commercially available graphite even after 200 cycles.
The specific energy storage capacity and the charge/discharge
rate of lithium ion batteries are critical for their use in electric
vehicles (EVs) and to store energy produced by intermittent
sources such as solar cells and wind.
1
Despite significant
gains in rate capability and safety through the use of new
materials,
2,3
increasing specific energy capacity remains a
challenge. Replacing graphitic carbon with Si as the anode
material can result in an increase in anode capacity by a
factor of 10 and can considerably increase the energy
capacity of the total battery. However, a rapid loss of the
reversible capacity upon cycling, which is associated with
the large volume expansion of Si, is a challenge.
4
Recent
studies on Si nanowires
5,6
and nanoporous Si
7
show great
promise in overcoming this issue.
When reacting with Li, Si is known to incorporate 4.4 Li
atoms per Si atom. In Li-ion batteries, this results in the
extremely high specific capacity of 4200 mA h/g, which is
10 times higher than the capacity of graphitic carbon (372
mA h/g).
8,9
However, the 300% volume change upon lithium
insertion commonly causes pulverization and thus a loss of
electrical contact between Si and the current collector.
Previous studies on improving the capacity of Si-based
materials have been focused on nanoparticle composites and
carbon prepared by ball-milling and thermal reduction.
10-17
Despite these efforts, these electrodes still suffer from rapid
capacity fading. Recently, Si nanowires
5,6
and 3D porous Si
particles
18
have been demonstrated to exhibit good cycling
performance as anode materials since both types of structures
provide empty space to accommodate Si volume changes
and allow for facile strain relaxation without mechanical
fracture upon lithium insertion. However, these materials
exhibited increased polarization at higher current rates and
some degree of capacity fading over many cycles, which
could possibly be due to the limited surface area accessible
to the electrolyte and the continuous growth of solid
electrolyte interphase (SEI) at the interface between the
silicon and electrolyte. For instance, hollow nestlike silicon
nanospheres showed the first discharge capacity (lithium
dealloy) of 3052 mA h/g at a rate of 2000 mA/g, but its
Coulombic efficiency and capacity retention ratio after 50
cycles were 73% and <25% (∼1000 mA h/g), respectively.
14
Here, we fabricate novel Si nanotube structures and
demonstrate their superior cycling performance for the first
time. The nanotube electrodes have ultrahigh reversible
charge capacities of ∼3200 mA h/g and have outstanding
capacity retention of 89% after 200 cycles at a rate of 1C in
practical Li-ion cells.
In this work, we used Si nanotubes (Figure 1) to increase
the surface area accessible to the electrolyte, which allows
the Li ions to intercalate at the interior and exterior of the
nanotubes. In addition, we deposited a carbon coating on
the surfaceof thenanotubes,which stabilizes theSi-electrolyte
interface and promotes stable SEI formation for long cycle
life. To synthesize carbon-coated Si nanotubes, a method
involving chemical deposition within porous alumina mem-
brane templates was employed. (See Supporting Information
* Corresponding authors, jpcho@unist.ac.kr and yicui@stanford.edu.
†
School of Energy Engineering, Ulsan National Institute of Science &
Technology.
‡
Beamline Research Division, Pohang Accelerator Laboratory.
§
Department of Applied Chemistry, Hanyang University.
|
Battery R&D, LG Chem, Ltd.
⊥
Department of Materials Science and Engineering, Stanford University.
NANO
LETTERS
2009
Vol. 9, No. 11
3844-3847
10.1021/nl902058c CCC: $40.75 2009 American Chemical Society
Published on Web 09/11/2009

for detailed synthesis procedures.)
Figure 2 shows scanning electron microscopy (SEM)
(Figure 2a-c) and transmission electron microscopy (TEM)
(Figure 2d) images of the Si nanotubes after the removal of
the alumina template by treatment with NaOH (also see
Figure S1 in Supporting Information). The carbon coating
protects the Si nanotubes from being etched away by NaOH.
The assemblies of uniform Si nanotubes with outer diameters
of ∼200-250 nm were recovered, and their wall thickness
was measured to be ∼40 nm as indicated by arrows. After
ultrasonication, the Si nanotube bundles were separated (see
Figure S2 in Supporting Information). The side view of the
bundle of nanotubes in Figure 2a shows a tube length of
∼40 µm and the top-view SEM image in Figure 2b shows
the open ends of the nanotubes. Figure 2e shows a high-
resolution TEM (HRTEM) image of the surface of the outer
wall of a nanotube, which indicates that the outermost layer
is covered with nanometer thick amorphous carbon, although
weak lattice fringes with a d spacing of 1.93 Å can also be
observed, corresponding to the Si(220) plane. Figure 2f is
an HRTEM image of the inner surface of a nanotube wall,
and selected area diffraction pattern and lattice fringe images
confirm the dominant presence of a crystalline Si phase
although we cannot rule out possible presence of minor
amorphous Si phase. The X-ray diffraction (XRD) pattern
of the Si nanotubes shows the presence of the diamond cubic
Si phase (Figure S3 in Supporting Information) and crystallite
size was estimated to ∼10 nm.
Figure 3 shows the Raman spectrum of the Si nanotubes.
The sharp peak at ∼516 cm
-1
is related to the Si-Si
stretching mode, which is identical to that of the reference
Si wafer. A small peak at ∼957 cm
-1
is due to the stretching
mode of amorphous Si-Si, which is also observed in the Si
wafer. The two other peaks at ∼1360 and ∼1580 cm
-1
are
assigned to the D band (disordered band) and G band
(graphene band) of carbon, respectively.
19
These peaks
confirm the carbon coating on the surface of the Si nanotubes;
the carbon is deposited by the high-temperature decomposi-
tion of the organic precursors used for the synthesis of the
Si nanotubes. The dimensional ratio of the D and G bands
of the samples can be estimated to be 1.4. This value is much
smaller than the values previously reported for carbon-coated
Si nanoparticles
20
and Sn
0.9
Si
0.1
nanoparticles,
21
which have
a ratio larger than 2. We do not observe a peak in the range
of 1000-1100 cm
-1
for SiO
2
using FT-IR spectrometry
(Figure 3 inset), suggesting that there is little SiO
2
on the Si
nanotube surface. Therefore, the prepared Si nanotubes can
be considered to be at least as pure as the detection limit
(100 ppm) of FT-IR. Overall, these results indicate that the
Si nanotubes are coated with a very thin layer of amorphous
carbon.
Parts a and b of Figure 4 show electrochemical rate
capability and cycle life performance of the Si nanotubes in
coin-type half cells. The first discharge and charge capacities
of the Si nanotubes are 3648 and 3247 mA h/g at the 0.2C
rate, respectively, which demonstrates an excellent Coulom-
bic efficiency of 89%. The high value of the Coulombic
efficiency of the first cycle is believed to be due to the thin
carbon layer, which minimizes the direct contact between
Figure 1. Schematic diagram of Li-ion pathway in Si nanotubes.
Figure 2. (a, b, and c) FE-SEM images of Si nanotubes: (b) top
view and (c) side view. (d, e, and f) TEM images of Si nanotubes:
(e) edge of the outer wall and (f) inner surface of nanotube wall.
The inserted figure in (f) is a selected area diffraction pattern of
(f). Arrows in image d indicate the tubewalls.
Figure 3. Raman spectra of Si nanotubes and Si wafer reference.
The insert is a FT-IR spectrum of Si nanotubes.
Nano Lett.,Vol. 9, No. 11,2009 3845