PROGRESS REPORT

Lithium-Ion Batteries www.advenergymat.de

Challenges and Recent Progress in the Development of Si

Anodes for Lithium-Ion Battery
Yan Jin, Bin Zhu, Zhenda Lu, Nian Liu,* and Jia Zhu*

energy density because every six carbon

Silicon, because of its high specific capacity, is intensively pursued as one of atoms can hold only one lithium atom.
the most promising anode material for next-generation lithium-ion batteries. Silicon is widely considered as one
of the most promising anode mate-
In the past decade, various nanostructures are successfully demonstrated
rials.[8–10] Each silicon atom can bind
to address major challenges for reversible Si anodes related to pulverization to four lithium ions, which gives it ten
and solid-electrolyte interphase. However, the electrochemical performance times more capacity than graphite anode
is still limited by challenges that stem from the use of nanomaterials. In this (4212 mA h g−1 for Li4.4Si at high tem-
progress report, the focus is on the challenges and recent progress in the peratures and 3579 mA h g−1 for Li15Si4
development of Si anodes for lithium-ion battery, including initial Coulombic observed in room-temperature electro-
chemical cells,[11,12] versus 372 mA h g−1 of
efficiency, areal capacity, and material cost, which call for more research
graphite anode). Besides, silicon is attrac-
effort and provide a bright prospect for the widespread applications of silicon tive because of its natural abundance (the
anodes in the future lithium-ion batteries. second-most abundant element in the
earth’s crust), environmentally benignity
and low electrochemical potential.
1. Introduction Unfortunately, ≈300% volume change of silicon during lithi-
ation and delithiation leads to severe particle pulverization,
As the emerging markets of portable electronics and electric unstable solid-electrolyte interphase (SEI) formation and loss
vehicles create tremendous demand for advanced lithium-ion of electrical contact in the electrode level, resulting in capacity
batteries (LIBs),[1,2] there is growing interest in developing bat- fading and limited cycle life (Figure 1a–c). A variety of nano-
tery electrodes with high gravimetric and volumetric capacity to sized silicon structures[13,14] such as nanoparticles,[15] nano­
surpass the energy density of the current LIBs.[3–5] wires,[16] nanotubes,[17] and nanoporous networks[18,19] have
Rechargeable lithium-ion batteries mainly consist of four demonstrated enhanced mechanical integrity, which can effec-
parts: two electrodes, a separator and electrolyte. The two elec- tively address this issue by accommodating the large volume
trodes are divided by a separator and liquid electrolyte that per- change. The electrochemical cycling performance of these
mits lithium ions to shuttle back and forth between the two nanostructures is stable above 1000 mA h g−1 in a large number
electrodes (Figure 1). When a battery is charged, lithium ions of previous studies. Even though fundamental materials guide-
released from the positive electrode (cathode) will move toward lines for designing nanostructures provide possible solution
the negative electrode (anode). When a battery is discharged, for improved gravimetric capacity and capacity retention, for
lithium ions will move from the anode to the cathode. Mean- practical applications, there are still many other important
while, electrons released from lithium atoms in the anode will figures of merit to improve, including initial Coulombic effi-
travel through the external circuit to the cathode, which provide ciency (CE), areal capacity, volumetric capacity, and materials
electric power from chemical energy.[1] cost, which will be discussed one at a section in the rest of the
In the case of anode, graphite is commercially used as article. Due to the length of this article, other factors such as
the electrode material because of its high conductivity, good binders, electrolyte, safety issues, electrode composition will be
reversibility, and relatively low cost. But graphite has limited briefly mentioned instead of extensively discussed. The former
four figures of merit are often impaired by the use of nanoma-
Y. Jin, B. Zhu, Prof. Z. Lu, Prof. J. Zhu terials to overcome fracture problem. First, high electrode and
National Laboratory of Solid State Microstructures electrolyte contact area leads to low initial Coulombic efficiency
College of Engineering and Applied Sciences (Figure 1d), consumption of the lithium from the cathode and
and Collaborative Innovation Center of Advanced Microstructures electrolyte, which seriously affect full cell performance. Second,
Nanjing University
Nanjing 210093, China tap density of nanomaterials can be limited due to much inter-
E-mail: jiazhu@nju.edu.cn particle space (Figure 1e), consequently generating low mass
Prof. N. Liu loading, low areal capacity, and even low volumetric capacity.
School of Chemical & Biomolecular Engineering Only when the highly promising electrochemical properties can
Georgia Institute of Technology still be maintained at higher areal capacity, these considerations
Atlanta, GA 30332, USA
of electrode structure design are meaningful. Third, complex
E-mail: nian.liu@chbe.gatech.edu
synthesis procedures or expensive raw materials hinder the
DOI: 10.1002/aenm.201700715 widespread application (Figure 1f).

Adv. Energy Mater. 2017, 7, 1700715 1700715  (1 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

There are several classical reviews about silicon anodes

or alloy anodes,[20] focusing on various nanostructured Yan Jin received her bachelor
materials, including 0D (nanoparticles), 1D (nanowires and degree at School of Materials
nanotubes), and 2D (thin film) for stable long cycle life and Science and Engineering of
increased rate performance because of superior advantages Nanjing University of Science
of nano­materials.[8,14,21,22] Different from all the previous and Technology, June 2014.
reviews, we prepare and organize this progress report by She is currently pursuing
summarizing the barriers and progress toward the wide- her Ph.D. degree under the
spread applications of silicon anodes for next-generation supervision of Professor Jia
lithi­um-ion battery. Zhu at Nanjing University.
In this progress report, we will first briefly discuss the fun- Her research interests focus
damental challenges and solutions of silicon anodes, then focus on high energy density anode
on the issues and recent progress toward its commercializa- of lithium-based batteries.
tion, including initial Coulombic efficiency, areal capacity, and
material cost. Specific examples are provided to illustrate these
fundamental design principles for commercialization. In addi- Nian Liu is an Assistant
tion, we also discuss full cell performance of silicon anode and Professor at Georgia Institute
expect reliable prospect. of Technology, School of
Chemical and Biomolecular
Engineering (since Jan. 2017).
His research group focuses
2. Fundamental Challenges and Solutions
on materials engineering
of Si Anodes for reversible high-energy
2.1. Material Pulverization chemical reactions, and
understanding their
Silicon is an alloy anode and each silicon atom can host four fundamental science using
lithium atoms. This mechanism of lithiation is associated with advanced light microscopy.
massive volume changes due to their intake of large amounts Liu received his B.S. from Fudan University (China), and
of lithium. When transforming from Si to Li4.4Si, the volume Ph.D. from Stanford University, where he worked with
expansion is about 420%. This large volume expansion/contrac- Prof. Yi Cui on the structure design for Si anodes for
tion during lithium insertion/extraction induces large stresses, high-energy Li-ion batteries. He then worked with
which can cause cracking and pulverization of Si, leading to Prof. Steven Chu at Stanford University as a postdoc,
loss of electrical contact and capacity fading. where he developed in situ optical microscopy to probe
In order to solve the material pulverization problem, it is beam-sensitive battery reactions.
found that silicon nanowire can circumvent the issue as they
can accommodate large strain without pulverization.[16] Besides,
it also shortens lithium diffusion distances, and provides good Jia Zhu is a Professor at
electronic contact and conduction, as shown in Figure 2a.[16] College of Engineering and
When lithium ions moved into and out of the silicon nano­ Applied Sciences, Nanjing
wires, the nanowires suffered little fracture. The theoretical University, China (since Sep.
charge capacity for silicon anodes is achieved with 75% capacity 2013). He obtained Ph.D.
retention after 10 cycles. Shape and volume changes upon first from Stanford University and
lithiation on a single nanostructure level and critical size of the then worked as a postdoctoral
Si nanowires is investigated in detail.[23–27] fellow at University of
Meanwhile, silicon particles in the nanoscale are also free California, Berkeley and
of pulverization. Huang and co-workers reported a strong Lawrence Berkeley National
size dependence of fracture by studying lithiation behavior Lab. His current research
of individual spherical Si nanoparticles.[28] It was found that focuses on the development
a critical particle diameter (≈150 nm) also existed. When the of nanomaterials based energy conversion and storage,
diameter of particles is below this value, the particles will including solar steam generations, solar cells, and thermo-
neither crack nor fracture upon initial lithiation. When the electric energy conversions.
diameter of particles is above this value, the particles will ini-
tially form surface cracks and then fracture because of lithi­
ation-induced stress, presented in Figure 2b. The lithiation batteries with size optimization. Besides, various struc-
behavior was carefully studied using in situ transmission elec- tures with porosity or void space control, such as yolk–shell
tron microscopy (TEM). Electrochemical lithiation of a small nanostructures and porous networks have been successfully
Si nanoparticle around the critical size showed no fracture.[28] employed to provide stable interphases and structures despite
Particle-size-dependent fracture behaviors of Si nanoparticles the large volume expansion, leading to improved cycling
provide guidance for building high-performance lithium-ion performance.[30–33]

Adv. Energy Mater. 2017, 7, 1700715 1700715  (2 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
Figure 1.  Overview of silicon-based anode for lithium-ion battery: a–c) Fundamental challenges. d–f) Challenges and recent progress toward com-
mercialization. d,e) Reproduced with permission.[42] Copyright 2014, Nature Publishing Group. f) Reproduced with permission.[7] Copyright 2015,
National Academy of Sciences.
2.2. Unstable Solid-Electrolyte Interphase
Apart from pulverization, large volume change during electro-
chemical cycling results in another severe issue: unstable solid-
electrolyte interphase.[17] When the potential of the anode is
below ≈1 V versus Li/Li+, the anode surface becomes covered
by a SEI film, due to the reductive decomposition of the organic
electrolyte. In order to prevent further side chemical reactions
from occurring, this layer should be dense and stable, elec-
tronically insulating but ionically conducting. However, the SEI
crushed due to the volume change of Si anodes during cycling.
The exposed fresh electrode surface leads to very thick SEI
layer, resulting in low Coulombic efficiency, low electronic con-
ductivity and higher resistance to ionic transport of the whole
electrode. So, it is a big challenge to maintain a stable SEI layer
over cycling.
Nanotechnology again provides new opportunities to address
the unstable SEI issue. Wu et al. have successfully designed a
double-walled silicon nanotube structure to stabilize SEI and
achieve long cycle life.[17] Double-walled Si-SiOx nanotube was
prepared by a templating method, in which Si nanotubes are
confined in SiOx outer shells. The void space in the center and
the outer SiOx clamping layer force the silicon tube to expand
inward during lithiation, leading to the intact and thin SEI
layer over thousands of cycles (Figure 2c). The average CE of
the double-walled silicon nanotubes electrode from the 2nd
Adv. Energy Mater. 2017, 7, 1700715 1700715  (3 of 17)
www.advenergymat.de
to 6000th cycles is 99.938% and 88% retained capacity was
achieved over 6000 cycles. SEI control has also been realized
on silicon nanoparticles (SiNPs), which are a promising can-
didate because they are industrially scalable and commercially
available. In the respect of silicon nanoparticles, artificial SEI
layer is introduced to overcome unstable SEI layer. Liu et al.
also designed a “yolk–shell” structure to achieve a stable and
scalable Si anode, in which silicon nanoparticle (≈100 nm)
is “the yolk” and amorphous carbon (5−10 nm thick) is “the
shell”, as shown in Figure 2d.[29] The outer carbon shell can
conduct electron and Li ion. The rationally designed void space
between the inner particles and outer shell can permit the
volume expansion of Si particle, leading to stable carbon shell
and SEI layer on the outer surface without breaking. In another
effort, mechanically strong and ductile graphene layer can also
be used to maintain stabilized SEI layer.[34] Recently, in situ and
operando atomic force microscopy was used for mechanical
property of SEI, of which the fracture energy is estimated to be
≈13 J m−2.[35]
Although nanosized materials can achieve stable long cycling
electrochemical performance by addressing materials pulveri-
zation and unstable SEI problem through elegant structure
designs, there are also challenges related to nanosized mate-
rials, including low initial Coulombic efficiency, low tap density
leading to low electrode mass loading and low areal capacity,
and high cost. In the following sections, we will summarize
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 2.  Solutions for fundamental problems of silicon anodes: a) Schematic of silicon nanowire and b) size-dependent fracture of silicon nanoparticle
for material pulverization problem. c) Schematic of SEI formation on silicon nanowire, nanotube, and double-walled silicon nanotube and d) yolk–shell
silicon nanoparticle for unstable solid electrolyte interphase. a) Reproduced with permission.[16] Copyright 2008, Nature Publishing Group. b) Repro-
duced with permission.[28] Copyright 2012, American Chemical Society. c) Reproduced with permission.[17] Copyright 2012, Nature Publishing Group.
d) Reproduced with permission.[29] Copyright 2012, American Chemical Society.

these three main issues and recent progress toward commer-
cialization of Si anodes for lithium-ion battery.
3. Initial Coulombic Efficiency
Coulombic efficiency refers to the ratio of discharge capacity
and charge capacity, and is an important metric to evaluate the
reversibility of battery reaction. As mentioned above, when the
potential of the anode is below 1 V during charging, electro-
lyte decomposes on the interphase of anode surface and elec-
trolyte to form SEI. The ideal SEI layer can permit lithium ion
transport, prevent further electrolyte decomposition on the
electrode surface, and increase the electrochemical cyclability
and stability. However, in the first cycle, active Li+ is consumed
due to the SEI film formation, resulting in irreversible capacity
and extra cathode electrode consumption. In recent years, a lot
of research focused on the initial cycle of silicon electrodes to
investigate the formation mechanism at the very starting of bat-
tery operation, which can accelerate the understanding of the
usually observed large irreversible capacity loss.[36–40]
Edström and co-workers demonstrated the interfacial mecha-
nisms during charge and discharge, including surface silicon
oxide change transformation, Li−Si alloying reaction, and for-
mation of SEI layer.[41] As shown in Figure 3a, at the very begin-
ning of discharge (0.5 V vs Li+/Li), a thin SEI layer has already
formed but lithium has not yet reacted with silicon. Lithium
started to react with the silicon nanoparticle after further dis-
charge to 0.1 V vs Li+/Li. The formation of Li2O and LixSiOy
(most probably Li4SiO4) is resulted by reaction with the surface
SiO2 layer. Further reaction with the silicon leads to the forma-
tion of the Li−Si alloy. After full lithiation (0.01 V vs Li+/Li), the
Li−Si alloying process is almost finished. After full delithiation
(0.9 V vs Li+/Li), lithium has been moved out of the particle,
Li2O has disappeared and the thickness of the SEI has lightly
decreased. Figure 3b shows a summary of the composition
on the surface of the silicon nanowire samples at different
working potentials as determined from the X-ray photelectron
spectroscopy (XPS) high-resolution scans of the initial two
cycles.[6] SEI layer is composed of hydrocarbons, polyethylene
oxide (PEO)-type oligomers, LixPFy, LiF, and LixPFyOz products
(0.5 V); hydrocarbons and PEO-oligomers (0.1 V); Li2CO3 and
LiF (0.01 V), respectively, with Li2CO3 as a major element.
While nanosized silicon has been effective to address pul-
verization problem for long cycling performance, large surface
area of nano-Si leads to low initial Coulombic efficiencies due
to the formation of SEI, resulting in irreversible capacity loss.
Figure 3c presents the calculated surface area in contact with
the electrolyte for secondary particles composed of 80 nm pri-
mary nanoparticles, with different size of secondary particles.
The specific SEI area, which is the surface area in contact with
the electrolyte divided by the mass of silicon, increases from

Adv. Energy Mater. 2017, 7, 1700715 1700715  (4 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 3.  a) Schematic of the reactions occurring at the surface of the silicon nanoparticles during the first cycle. The inset shows Si 2p XPS spectra
of the Si/C/CMC composite electrodes upon lithiation at three different voltages. b) Molecular species found on the surface of the SiNW samples
at different potentials as determined from the XPS high-resolution scans. c) Calculated surface area in contact with electrolyte (specific SEI area)
versus secondary particle diameter. a) Reproduced with permission.[41] Copyright 2012, American Chemical Society. b) Reproduced with permission.[6]
Copyright 2009, Elsevier B.V. c) Reproduced with permission.[42] Copyright 2014, Nature Publishing Group.

1.5 m2 g−1 for 10 µm secondary particles to 15 m2 g−1 for 1 µm
secondary particles and to 90 m2 g−1 for single yolk–shell par-
ticles.[42] Large surface area leads to severe SEI formation in
the first cycle and results in low initial Coulombic efficiency,
and consumption of the lithium from the cathode and electro-
lyte during cycling. Besides, we should concern more about
increased and accelerated native oxide surface derived from
nanosized silicon materials with large surface area, which can
affect the electrochemical performance and induce the inevi-
table use of hydrofluoric acid (HF).[43,44]
The first-cycle Coulombic efficiency of silicon anode is typi-
cally in the range of 65–85%, far below that of commercial
graphite anodes (90–94%). Below we will summarize the recent
progress to achieve high initial Coulombic efficiency from three
aspects, including secondary structure design, prelithiation,
and electrolyte additive.
3.1. Secondary Structure Design
As mentioned above in Figure 3c, the electrode/electrolyte
surface area can be tuned by forming secondary structures
with nanoscale building blocks. Higher initial Coulombic effi-
ciency is achieved by designing microstructures with electro-
lyte blocking layers in the outer surface, which decrease side
reaction between electrode surface and electrolyte.[45,46] We pre-
sent four notable designs, including a pomegranate-inspired
nanoscale design, porous silicon with graphene coating,
micrometer-sized silicon with graphene cage, and silicon-
nanolayer-embedded graphite.
In the pomegranate-inspired structure designed by Liu
et al. (Figure 4a), secondary silicon particle of micrometer
size is made up with yolk–shell silicon nanoparticle through
a self-assembly process, where single silicon nanoparticle is

Adv. Energy Mater. 2017, 7, 1700715 1700715  (5 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 4.  Secondary structure design to improve initial Coulombic efficiency: Schematic of a) the pomegranate-inspired design. b) Graphene cage
on micrometer-sized silicon particles. c) Schematic, fabrication process, and cross-sectional SEM image of silicon-nanolayer-embedded graphite.
a) Reproduced with permission.[42] Copyright 2014, Nature Publishing Group. b) Reproduced with permission.[49] Copyright 2016, Nature Publishing
Group. c) Reproduced with permission.[47] Copyright 2016, Nature Publishing Group.

encapsulated by a conductive carbon layer with void space for
volume expansion and the whole microsized particle is then
encapsulated by a thicker carbon layer as an electrolyte bar-
rier.[42] The pomegranate design affords remarkable high ini-
tial Coulombic efficiency because of decreased surface area
and outer electrolyte blocking layer. Besides, the self-sup-
porting conductive carbon framework can hold electrochemical
activity of silicon nanoparticle and facilitate lithium transport
throughout the whole particle. At the same time, a high tap
density is achieved for high areal capacity, which will be dis-
cussed later.
Besides electrolyte blocking layer of carbon shell, the gra-
phene coated porous silicon particles[48] have also been used
to enable high Coulombic efficiency. The porous silicon par-
ticles were produced by two steps: ball milling the mixture of
low grade silicon and water for silicon oxide (SiOx) particles,
and then convert SiOx particles to mixed crystalline (Si) and
amorphous (SiO2) phases by thermal disproportionation.[48]
This structure offers several unique features. Porous silicon
with continuous and precise porosity control can offer a
stable structure despite 400% volume change. Also the gra-
phene shell can offer a stable interphase layer for stabilizing
SEI formation, while increasing the electrical conductivity
of whole structures. In addition, different from amorphous
carbon layer, graphene shell will not trap the lithium ions,
which is beneficial for high initial Coulombic efficiency.
Therefore, porous/graphene composite can offer high CE of
the first cycle, stable cycling performance, as well as superior
rate performance, ideal for the next-generation lithium-ion
batteries.
Instead of building secondary structure with nano­
particle building block, there are also some progress to use
micrometer-sized Si as source materials. As discussed above,
there are huge challenges associated with micrometer-sized
Si due to particle pulverization. During repeated electro-
chemical cycling, Si microparticles rupture due to volume
change, resulting in electrical contact failure. Besides,
freshly exposed surfaces of silicon microparticle repeatedly
react with the electrolyte, leading to a thick and ionically
insulating SEI film. The destruction of electronic and ionic
pathways causes critical cell decay. Li et al. introduced the
conformal conductive graphene cage as an encapsulation
layer in order to stabilize the Si microparticles electrodes
over battery operations (Figure 4b).[49] The mechanically
strong and flexible graphene cage with a predesigned void
space can confine all the broken Si pieces within and provide

Adv. Energy Mater. 2017, 7, 1700715 1700715  (6 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

undamaged electrically pathways to the ruptured pieces. 3.2. Prelithiation

Most importantly, graphene cage can act as the electrolyte
blocking layer, leading to SEI formation mainly on the sur-
face of outer graphene cages. Efficient SEI formation on the
microscale and graphite-like surface of the graphene cage
reduces irreversible lithium ion consumption, and enhances
initial Coulombic efficiency.
Si-containing graphite composites have been intensively pur-
sued as one of the most promising anode materials for prac-
tical applications. Figure 4c shows the schematic and detailed
structure of Si-nanolayer-embedded graphite/carbon hybrids
(SGC).[47] The Si nanolayer through chemical vapor deposition
(CVD) process is both beneath inner pores of graphite par-
ticle and on the surface of graphite. The thickness of silicon
layer is less than 20 nm, which can relieve the stress induced
by volume change and permit fast lithium ion transportation,
while homogeneous carbon coating layer from the last syn-
thetic steps can further promote stable SEI formation. The
unique Si-nanolayer-embedded graphite/carbon hybrids play a
beneficial role for improved reversible specific capacity and ini-
tial Coulombic efficiency.
Prelithiation presents exciting opportunities for next-genera-
tion high-capacity anodes with an alternative lithium source for
high initial Coulombic efficiency, which prevents the lithium
in the cathode from being consumed during side reactions of
SEI formation. There are three main ways to prestore lithium
in anodes. The first approach is to incorporate a temporary cell
to realize electrochemical prelithiation by contacting lithium
foil with electrolyte-wetted anodes.[50,51] The second approach
is to prelithiate silicon with microscale stabilized lithium metal
powder through mechanical activation.[52,53] The third one is to
apply LixSi nanoparticles from chemically synthesis as a poten-
tial prelithiation reagent.[54] We will discuss the three main
methods one by one.
Electrochemical prelithiation stores lithium in advance to
form a SEI layer through electrochemical reactions. As shown
in Figure 5a, in the prelithiation process, silicon anodes were
in contact with a piece of Li metal foil in liquid environment
of electrolyte through a temporary cell under inert atmos-
phere. Potential difference between the two electrodes can

Figure 5.  Prelithiation improves initial Coulombic efficiency: Schematic of a) electrochemical prelithiation, b) microscale stabilized lithium metal
powder, c) ambient-air stable prelithiation with polymer coating, and d) LixSi nanoparticles. a) Reproduced with permission.[50] Copyright 2016, American
Chemical Society. b) Reproduced with permission.[55] Copyright 2014, American Chemical Society. c) Reproduced with permission.[56] Copyright 2016,
American Chemical Society. d) Reproduced with permission.[54] Copyright 2014, Nature Publishing Group.

Adv. Energy Mater. 2017, 7, 1700715 1700715  (7 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
sponsor spontaneous prelithiation for SEI formation, because
of an external short circuit built between lithium metal foil and
c-SiOx electrode. Proper electrochemical prelithiation requires
a delicate control, since insufficient lithiation cannot improve
the initial Coulombic efficiency enough due to remaining Li
trapping sites, while overlithiation reduces the lithium ion
accommodation in the real first lithiation process. Choi et al.
introduced a roll-to-roll scalable and delicate prelithiation pro-
cess based on electrochemical reactions with lithium metal. The
degree of prelithiation can be precisely controlled via a voltage
monitoring process. The roll-to-roll electrochemical prelithia-
tion process is consistent with the current manufacturing route
and have the possibility to be employed in the next-generation
lithium-ion battery with higher initial Coulombic efficiency.
Microscale stabilized lithium metal powder (SLMP) can serve
as another sacrificial lithium source. It is necessary to mechani-
cally activate the SLMP to fully employ lithium, after depositing
the SLMP onto the Si anode. The pressure-activation process of
SLMP through rolling compression is demonstrated as a car-
toon in Figure 5b. By compressing the SLMP onto the surface
of silicon with mechanical forces, the contact area between the
Si surface and the SLMP is maximized.[55] This step guarantees
proper lithium diffusion into the silicon, which promotes stable
SEI formation on the surface of silicon anode. The whole pro-
cess is consistent with the present battery fabrication.
The whole prelithiation process should be finished under
inert atmosphere because of the low potential and high
chemical reactivity of lithium metal with H2O and O2. Com-
bining above two prelithiation methods, Yang and co-workers
proposed an active material/polymer/lithium anode trilayer
structure with uniform distribution of lithium source, which
is stable in ambient air. As illustrated in Figure 5c, lithium is
deposited on a Cu foil, and then lithium is coated with a protec-
tive polymer layer (polymethyl methacrylate, PMMA) to make it
ambient-air-stable.[56] Then, the anode materials, deposited on
the top of the polymer layer, can be integrated with electrolyte
and conventional cathode materials to make up a real battery.
As the polymer coating layer dissolves in the electrolyte, the
anode materials contact lithium directly. As a result of pressure
inside the cell caused by cell assembly process, lithium metal is
activated and forms an in situ lithiated silicon anode.
The third excellent prelithiation reagent is LixSi nanoparti-
cles, proposed by Zhao et al. for the first time.[54] Compared
with microscale stabilized lithium metal powder, LixSi nanopar-
ticles show multiple excellent properties: First, nanoscale LixSi
particles with size around 100–200 nm can distribute homoge-
neously in the entire electrode structure for a uniform and fast
prelithiation. Second, the surface of LixSi nanoparticle can be
modified to realize ambient air compatibility. Third, LixSi nano-
particle is compatible with the existing battery manufacturing
environment and can be mixed with various anode materials
during slurry processing and serves as an excellent prelithia-
tion reagent. Figure 5d demonstrated LixSiLi2O core–shell
nanoparticles synthesized from a one-step thermal alloying
process with simple experimental setups. Si nanoparticles react
with molten Li to form LixSi nanoparticles. As the nanoparti-
cles reacts with limited oxygen, a dense Li2O passivation layer
is formed on the surface of LixSi. As-synthesized LixSiLi2O
core–shell nanoparticles can be used as a localized prelithiation
Adv. Energy Mater. 2017, 7, 1700715 1700715  (8 of 17)
www.advenergymat.de
reagent in the mixture of electrode slurry and is activated at the
presence of carbonate electrolyte, resulting lithium diffusion
into anode materials. In the follow-up study, concept of artifi-
cial SEI layer is built to improve air stability. A continuous and
uniform artificial SEI coating consisting of Li alkyl carbonate
and LiF with long hydrophobic carbon chains is formed on
the surface of LixSi nanoparticles by using 1-fluorodecane.[57]
The coated LixSi nanoparticles showed improved air stability
at a low-humidity level (<10% relative humidity). Further-
more, ambient air stable LiSiO anode is demonstrated with
low-cost SiO and SiO2 starting materials.[58] As-synthesized
LixSi/Li2O composites provided a unique structure combining
homogeneously dispersed active LixSi nanodomains embedded
in a robust Li2O matrix. The Li2O matrix is highly crystalline
and contact area between LixSi nanodomains and Li2O matrix
is increased, which is beneficial for stability in dry air and even
ambient air compatibility (≈40% RH). Besides, LixSi/graphene
composite can serve as an alternative for air stable and free-
standing lithium metal anode.[59]
3.3. Electrolyte Additive
Electrolyte additive plays a significant role in passivating the
Si/electrolyte interphase during the initial lithiation process.
Aiming to form a stable SEI layer on the surface and improve
the cycle life of Si electrodes, additives/cosolvents including
fluoroethylene carbonate (FEC), propylene carbonate (PC),
vinylene carbonate (VC), lithium bis(oxalate)borate (LiBOB),
lithium difluoro(oxalate) borate (LiDFOB), and succinic anhy-
dride have been introduced into the electrolyte solutions for
anodes materials.
Among all the additives, FEC shows the great promise.[60]
There are two main degradation routes for FEC additive:
defluorination and ring-opening.[61] FEC degrades at a higher
reduction potential, prior to ethylene carbonate (EC)/diethyl
carbonate (DEC), to form an initial SEI layer combining LiF
and other organic species (CHFOCO2-type compounds).
Furthermore, the initially formed SEI can substantially limit
the degradation of EC/DEC and thus prevent the excess for-
mation of an inhomogeneous SEI layer, and enhance the ini-
tial Coulombic efficiency. Density functional theory simulation
based molecular dynamics shows lithium fluoride (LiF), the
most plausible product of the FEC reductive decomposition,
adsorbs the organic SEI film components through the FLi
bindings, which demonstrate the role of glue within the SEI
film for stabilizing SEI.[62] However, much attention should be
paid to optimize amount of FEC because residual FEC results
in gas evolution and HF formation, which causes significant
dissolution of cathode materials in full cell, especially at ele-
vated temperatures.[63,64]
We summarize the initial Coulombic efficiency in Figure 6
by comparing the three main improve methods with original
nanostructured silicon anodes. The enhancement of initial
Coulombic efficiency is quite obvious. The first-cycle Cou-
lombic efficiency of silicon anode is typically in the range of
65–80%, such as 65% for core–shell silicon nanoparticle[65]
and 73% for silicon nanowire,[16] far below that of commercial
graphite anodes (90–94%). FEC electrolyte additive can increase
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
Figure 6.  Initial Coulombic efficiency of various Si-based anodes.
the initial Coulombic efficiency to 81% of silicon film[66] and
87% of silicon particle[61] due to the prior formed SEI layer by
FEC and the limited degradation of EC/DEC. The secondary
structure design can further improve the initial Coulombic effi-
ciency comparable to the level of commercial graphite (90%).
For example, because of reduced surface area between silicon
and electrolyte and the outer electrolyte blocking layer, as pre-
sented in the Figure 6, pomegranate Si microparticle,[42] porous
Si with graphene coating,[48] silicon-nanolayer-embedded
graphite,[47] multiscale hyperporous Si flake,[67] and micrometer-
sized Si particles of graphene cages[49] show initial Coulombic
efficiencies of 82%, 89.5%, 92%, 92.7%, and 93.2%, respec-
tively. Furthermore, initial Coulombic efficiency with prelithia-
tion method is above the commercial standard. The three main
approaches of prelithiation can prestore Li and realize initial
Coulombic efficiency of 90% (Li metal powder),[55] 94.9% (elec-
trochemical prelithiation),[50] >100% (Si/polymer/lithium),[56]
and >100% (LixSi nanoparticles).[54]
In summary, the main three methods indeed improve the
initial Coulombic efficiency of silicon anode. Through structure
design and sacrificial lithium, the initial Coulombic efficiency
can reach or exceed the commercial level of state-of-the-art
graphite anodes (90%) step by step. The prelithiation proce-
dure, such as LixSi nanoparticle, is compatible with the existing
battery manufacturing process. The challenge of initial Cou-
lombic efficiency can be solved by combining the above method
considering the scalable production and feasible operations.
4. Areal Capacity
Practical application requires high areal capacity, the capacity
on a unit area of electrode, to minimize the weight percentage
of metal foil current collectors in the battery. The large volume
changes of silicon anode makes it challenging to achieve stable
cycling at high areal capacity, even though it cycles stably at
low areal capacity. In thick electrode, silicon spheres expand
Adv. Energy Mater. 2017, 7, 1700715 1700715  (9 of 17)
www.advenergymat.de
vertically during lithiation, resulting in the
increase of electrode thickness. During deli­
thiation, as lithium extracts, silicon particles
contract resulting in destruction of electrical
pathways to their surrounding particles
and current collectors. As exemplified in
Figure 7a, the original electrode consisted
6 × 6 grid of cubic close-packed Si spheres.[68]
After 270% expansion, silicon spheres
are displaced vertically to accommodate
the horizontal expansion. This drastic elec-
trode morphology change can lead to elec-
trical isolation of active material, electrode
failure, and capacity fading (Figure 1c). Thus,
it is urgently needed to maintain the elec-
trical connection at the electrode level and
the integrity of the morphology electrodes
over cycling. Tarascon et al. use a simple
mode with geometrical considerations to sug-
gest that in order to avoid particle rearrange-
ment due to volume change of the silicon,
the volumetric percentage of silicon in single
layer sphere is only 14%, as presented in Figure 7b, as enough
void space is necessary to accommodate the volume change.
In the bulk-nanoporous-silicon negative electrode, there are
many nanopores surrounding the Si nanoligaments, which can
accommodate large volume changes. The accommodation limit
up to porosity is quite important since low porosity can lead
to fractures the Si nanoligament (Figure 7c).[69] Therefore, the
areal mass loading is severely limited by the low silicon propor-
tion and low electrode thickness for all the reasons listed above.
Although nanostructure silicon materials have been effec-
tive to solve silicon pulverization problem, the reduced size cre-
ates more interparticle space and surface, giving rise to low tap
density and high interparticle resistance. The low tap density
causes low volumetric capacity and long electron pathway for
the same mass loading, which makes it difficult to maintain
electrical and ionic pathways during cycling. The large interpar-
ticle resistance also creates barriers for the electron transport
between current collectors and active materials. Low areal mass
loading and low tap density together hinder the achievement
of high areal capacity. In this section, we will summarize the
recent effort to improve areal capacity from the two aspects of
higher mass loading and higher tap density through structure
designs and employing conductive binders.
4.1. Structure Design
As discussed in section 3.1, silicon particle with large size
(micrometer) can reduce surface area. In addition, the tap den-
sity of micrometer-sized particles is generally higher than that
of nanoparticles due to the reduced interparticle space. Higher
tap density can decrease the electrode thickness and shorten
the electron pathway for the same mass loading.
To follow up with the pomegranate-inspired design, indus-
trially mature mechanical based approach is used to densely
pack primary building blocks (Figure 8a).[70] A high tap den-
sity of 0.91g cm−3 and stable cycling with over 2 mg cm−2 of
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 7.  a,b) A theoretical model of a 6 × 6 grid of cubic close-packed Si spheres and after 270% expansion to present electrode volume change.
c) Schematic of the electrode design of bulk 3D NP-Si: volume accommodation limit and porosity. a,b) Reproduced with permission.[68] Copyright 2008,
The Electrochemical Society. c) Reproduced with permission.[69] Copyright 2014, American Chemical Society.

areal mass loading (≈3.5 mA h cm−2) is achieved (Figure 8d).
Besides, mesoporous Si sponge with diameter over 20 µm
showed an area-specific-capacity of ≈1.5 mA h cm−2 and 92%
capacity retention over 300 cycles.[71] The highly porous struc-
ture with thin crystalline Si walls surrounding large pores can
help to accommodate volume change (Figure 8b,c,e).
Apart from the silicon/carbon nanocomposite, alloying sil-
icon with other inactive metals to form micron-sized particles
is another route to increase tap density. Lithium-inactive metal
silicides have good electrical conductivity and can buffer the
volume expansion from lithiation of silicon, providing stable
electrochemical performance. The microscale alloy particle
size leads to improved tap density compared to silicon nano-
particles. FeCuSi ternary composite with an average dia­
meter of 6.5 µm demonstrates the tap density of 0.8 g cm−3.[72]
Besides, the tap density of SiNiSi2Al2O3@C composite was
measured to be 1.34 g cm−3 due to the higher density of NiSi2
(4.83 g cm−3) compared to that of Si (2.33 g cm−3).[73]
Tap density of various Si-based anode materials can be found
in Figure 9. Compared with the low tap density of silicon nano-
particles (0.16 g cm−3), the tap density of SiC composite with
granule hierarchical bottom-up assembly,[74] SiC composite
from SiO disproportionation,[75] nano/microstructured Si/C,[46]
and pomegranate Si microparticle[42] are 0.49, 0.78, 0.81, and
0.53 g cm−3, respectively. Si-containing graphite composites
also have shown great potential for practical application. For
example, Si-nanolayer-embedded graphite/carbon hybrids show
the tap density of 1.0 g cm−3.[47] Therefore, it is confirmed that
secondary microstructure composites could improve tap den-
sity by reducing interparticle space.
4.2. Conductive Binder
Polymer binder is necessary for making slurry and main-
taining the integrity of the structure, and therefore critical for
stable electrochemical performance. Traditionally, the following
categories of binders have been shown effective for Si anode:
carboxymethyl cellulose (CMC) and alginate, conventional syn-
thetic polymer such as poly(acrylic acid) and polyimide. Recent
works have demonstrated that the polymeric binders with high
electronic conductivity and binding affinity with Si can help
achieve high areal capacity for silicon anodes.[76–79]
Self-healing binders (SHP) have both electrical and mechan-
ical healing capabilities, which heal the damages caused by
fracture of silicon microparticles during battery cycling.[80]
Figure 10a presents schematic illustration and cycling behavior
of a conventional microparticle silicon electrode and a self-
healing electrode. In a traditional silicon-based anode, the sil-
icon particles are surrounded by polymer binder to maintain
electrical connections between active materials and the current
collector. During cycling, large volume change results in frac-
ture of polymer binder and silicon particles and the loss of elec-
trical contact. However, self-healing binder is stretchable and

Adv. Energy Mater. 2017, 7, 1700715 1700715  (10 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 8.  a) A schematic of the mechanical-based micrometer-sized secondary clusters fabrication process. b) A schematic of the mesoporous silicon
sponge particle. c) SEM and TEM images of mesoporous silicon sponge. d) Cycling performance of micrometer-sized secondary clusters with different
areal mass loading. e) Cycling performance of mesoporous silicon sponge particle. a,d) Reproduced with permission.[70] Copyright 2015, The Royal
Society of Chemistry. b,c,e) Reproduced with permission.[71] Copyright 2014, Nature Publishing Group.

can repair spontaneously the mechanical destruction of anodes.
As shown in Figure 10b, dynamic reassociation of hydrogen
bonds due to amorphous structure and low glass-transition
temperature lead to spontaneous self-healing, which results
in more stable electrical and mechanical connections among
the anodes. Besides, good mechanical property of conductive
polymer binder is essential to accomplish the calendering or
winding during manufacturing process. The good mechanical
stretchability of crosslinked network of SHP enables structural
integrity when it is stretched to three times of its initial length.
Besides, the resistance of SHP/CB did not change significantly
even though the composite was subjected to a 100% strain, in
order to retain electrical conductivity.
Another functional conductive polymer binder, poly(9,9-dioc-
tylfluorene-co-fluorenone-co-methylbenzoic ester) (PFM) has
also been developed to provide electronic connections between
the conductive polymer matrix and active materials at mole­
cular-level, because of electrical conduction and tightly adhe-
sion (Figure 10c).[55] Chemical structure of the PFM functional
conductive polymer binder is presented in Figure 10d, in
which the blue shadow highlights the ester group for adhe-
sion bonds formation with the SiOH surface. Figure 10e shows
the trans-esterification reaction between the ester functional
group of the PFM binder and the SiOH surface group during
the electrode drying process, which provides strong adhe-
sion. In addition, poly (1-pyrenemethyl methacrylate-co-meth-
acrylic acid) (PPyMAA) is used for a high tap density nanoSi
anode.[77] PPyMAA can bind to nanoSi with strong affinity
and polymer is electronically conductive at the presence of
pyrene. High electronic conductivity together with mechanical
binding of polymer can help to significantly improve electrode
areal capacity. Detailed mechanical property of these polymers
needs to be further studied to ensure the requirement of cal-
endering or winding during manu­facturing process. We sum-
marize the specific areal capacity and mass loading of several
representative works in Table 1, which is divided into three
main categories. First, all of the selected designs with micro-
scale secondary structures present areal capacities above

Adv. Energy Mater. 2017, 7, 1700715 1700715  (11 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

and mechanical binding of polymer improve

Figure 9.  Tap density of various Si-based anodes.
2 mA h cm−2 after different numbers of cycles.[81,82] Among
electrode areal capacity significantly. Silicon-
self healing poly­ mer/carbon black com-
posite (Si-SHP/CB) have exhibited high areal
capacity (above 3 mA h cm−2) and stable
cycling for more than 140 cycles using low-
cost large Si particles.[83] Although areal
capacity and mass loading is drastically
improved, weight proportion of conductive
polymer is still high. Weight ratio between
active materials and polymer component
is 2:1 for PPy and Si-SHP/CB, 1:1 for self-
healing SiMP. Future study should focus on
reducing conductive polymer binder amount
and study volumetric or gravimetric energy
density considering all components.
Based on the areal capacity and electrode
thickness, volumetric-based information is
calculated and summarized, which is more
important than the gravimetric capacity for
certain applications. Self-healing SEI-based
silicon[84] and silicon@graphitic carbon
nanowire array present volumetric capacity of 1100 mA h cm−3
[85]

all, the areal capacity of pomegranate Si microparticle is above
3 mA h cm−2 after 100 cycles, with a high mass loading of
3.12 mg cm−2 (Si plus carbon weight). Second, composite with
Li-inactive matrix composite can buffer volume change, leading
to improved areal capacity. For example, Fe–Cu–Si ternary
composite with an average diameter of 6.5 µm demonstrated
the areal capacity of 1.968 mA h cm−2 after 300 cycles at a cur-
rent density of 3 mA cm−2. Finally, high electronic conductivity
(100th cycle) and 1500 mA h cm−3 (after 200 cycles), which
is higher than graphite of 550 mA h cm−3. Therefore, higher
tap density with close packing of anode materials demonstrate
high volumetric capacity. Taking the self-healing SEI-based Si@
TiO2 electrode as an example, the electrode thickness involved
in volumetric capacity is the one containing electrode mate-
rials before cycling. The electrode thickness of Si@TiO2 is
14.7 µm before cycling, with 25.5% increment after 100 cycles

Figure 10.  Polymer binder for higher areal capacity: a) Schematic illustration of the design and behavior of a conventional silicon electrode and stretch-
able self-healing electrode. b) Chemical structure of the self-healing polymer. c) Schematic of an electrode design with low concentration of functional
conductive polymer binder. d) Chemical structure of the PFM functional conductive polymer binder. e) Transesterification reaction between the ester
functional group of the PFM binder and the SiOH surface group. a, b) Reproduced with permission.[80] Copyright 2013, Nature Publishing Group.
c–e) Reproduced with permission.[55] Copyright 2014, American Chemical Society.

Adv. Energy Mater. 2017, 7, 1700715 1700715  (12 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Table 1.  Areal capacity and mass loading of different kinds of structures.

Active materials Mass loading Initial areal capacity Areal capacity No. cycles Applied current
[mg cm−2] [mA h cm−2] [mA h cm−2] density
Nano/microstructured Si/C[46] 3.1 (Si/C) 2.25 >2 200 0.2 A g−1
Mesoporous silicon sponge [71]
3.5–4 (electrode) 10 4.2 35 0.5 mA cm−2 for lithiation,
0.75 mA cm−2 for delithiation
Nonfilling carbon coating[82] 2.01 (Si/C) 3.22 2.84 100 0.05 mA cm−2 for first 3 cycles,
0.25 mA cm−2 for later cycles
Secondary Si particle by 2.02 (electrode) 3.5 2.3 100 0.1 mA cm−2 for first 3 cycles,
mechanical pressing[70] 0.5 mA cm−2 for later cycles
Pomegranate Si 3.12 (Si/C) 3.67 >3 100 0.03 mA cm−2 for first 3 cycles,
microparticle[42] 0.7 mA cm−2 for later cycles
FeCuSi[72] – 3.08 1.968 300 1C (1C = 3 mA h cm−2)
SiNiSi2Al2O3@C[73] 1.65 (active) 1.2 0.96 200 200 mA g−1
PPyMAA binder [77]
– 2.35 2 120 420 mA g−1
PPy[78] 1.34 (Si) 2.7 2.5 40 0.226 mA cm−2
Self-healing SiMP[80] 0.704 (Si) 2.06 2.02 8 1/30C (1C = 4200 mA h g−1)
Si-SHP/CB [83] 1.6 (Si) 4.16 >3 140 0.1 mA cm−2 for first 3 cycles,
0.3 mA cm−2 for later cycles

at delithiated state. Hence, the electrode swelling should also
be taken into account when considering the volumetric-based
information. Setting testing criteria to characterize volumetric
capacity is quite important to compare the performances of
different electrodes directly. Because of volumetric expansion,
it is important to clarify the state of electrode (lithiated or del-
ithiated) and cycling number when calculating the volumetric
capacity. At the aspect of energy density, SGC[49] demonstrate
energy density of 1043 W h L−1 (after 100 cycles) in a full coin
cell configuration with LiCoO2 cathode. In addition, monodis-
perse porous silicon nanospheres[86] and graphene coated sil-
icon particle with SiC-free graphene[87] growth can reach energy
density of 850 and 700 W h L−1 (200th cycle).
5. Low Raw Material Cost
While nanostructured silicon anodes show great potential in
lithium-ion battery, the commercial application can be limited
by high cost related to the complex fabrication and expansive
raw materials. Traditionally, silicon nanostructures are fabri-
cated either from top-down (chemical etching) or bottom-up
(CVD) method, which typically involve either toxic silane pre-
cursors or expensive high purity silicon sources. Recently, it
is found that various low grade silicon[7,48,88,89,90] and natural
sources[91–96] can serve as cost-effective sources to produce
nanostructured silicon for lithium-ion battery.
5.1. Low Grade Silicon
nanoparticles with controlled size can be produced at large-
scale through a simple and scalable high energy mechanical
milling process.[89] The size of obtained nanoparticles is
≈150 nm, which provides stable electrochemical performance
(Figure 11a).[89] It is found that inactive FeSi2 is formed in
nanoparticles produced from ferrosilicon sources, which can
serve as a buffer layer to alleviate the volume change during
cycling. Whereas, nanoparticles from metallurgical Si sources
have higher capacity and better kinetic properties due to higher
purity source materials and better electronic transport prop-
erties. The purity of these obtained nanoparticles can also be
improved by acid etching processes up to 99.999% (wt%).[7]
During high energy ball-milling processes, the impurity-rich
regions containing Fe, Al, and Ca are mechanically weak, can
be easily exposed on the surface, and be effectively removed
through acid etching. Therefore, it can be easily understood
that the purity goes up with the smaller size of Si particles in
this nanopurification process (Figure 11b).
Morphology control in porous network is also important to
buffer the volume expansion. A unique nanoperforation pro-
cess is developed to produce porous Si particles, combining
modified ball milling, annealing, and acid treating, with a
precise and continuous porosity control (from 17 to 70%)
(Figure 11c).[48] Metal-assisted chemical etching has also been
useful for porous silicon networks with low grade silicon
source (Figure 11d).[90] The similar structure is demonstrated
by Zhou and co-workers and Wehrspohn and co-workers.[88,97]
Above all, these results show great potentials of recycling low
grade silicon for commercial silicon anodes.

Low-grade silicon is an attractive material choice because of its
abundance and cheap price. Several low-grade silicon sources,
such as metallurgical-grade silicon (M-Si, ≈98 wt% Si, $1 kg−1)
and ferrosilicon (F-Si, ≈83.4 wt% Si, $0.6 kg−1) provide annual
global production over six million tones.[89] It is found that Si
5.2. Natural Source
Silicon is the second most abundant element on earth, widely
distributed in rocks, sands, and soils as various forms of sili-
cates or Si dioxide. Recently, there is progress in low-cost

Adv. Energy Mater. 2017, 7, 1700715 1700715  (13 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 11.  Low grade Si as the source for low cost Si anodes. a) Schematics of fabrications and electrochemical cycling of low grade silicon nano­
particles. b) Schematic of the nanopurification. c) Schematics of nanoperforation processes with three steps: high energy mechanical milling with water,
annealing, and acid etching. d) Schematic diagram of simultaneous perforation and purification processes through metal-assisted chemical etching.
e) Schematic diagram of the synthesis and morphology of porous silicon particles. a) Reproduced with permission.[89] Copyright 2015, American
Chemical Society. b) Reproduced with permission.[7] Copyright 2015, National Academy of Sciences. c) Reproduced with permission.[48] Copyright
2016, American Chemical Society. d) Reproduced with permission.[90] Copyright 2015, American Chemical Society. e) Reproduced with permission.[88]
Copyright 2014, American Chemical Society.

synthesis of porous Si nanomaterials from magnesiothermic
reduction with abundant and low-cost SiO2 sources, such as
rice husk and sand.
As a widespread food crops for human sustenance, rice is
widespread with massive amount of annual production near
108 tons worldwide. As demonstrated in Figure 12a, rice husk
is built up with interconnected nanoporous structure to venti-
late air and moisture and contains ≈15–20 wt% silica.[91] Sand
is a naturally abundant raw material with annual worldwide
production over 115.0 million metric tons and cost lower than
$10 per ton. The detailed fabrication process of turning low cost
silicon contended sources into functional silicon anodes is pre-
sented in Figure 12b.[96] The obtained 3D porous silicon shows
stable electrochemical performance, which prove that the infi-
nite natural source can easily satisfy the high and growing
demand for energy-storage in the next decades.
6. Challenges of Full Cells
Most works still focus on half-cells with Li metal as the counter
electrode for studying the electrochemical behavior of latest
synthesized materials. In half-cell configuration, lithium
amount in Li foil counter electrode is exhaustless, while lithium
amount is limited by cathode and electrolyte in full cell. The
drain of recyclable lithium is the main reason for the perfor-
mance degradation of the full cell. To some extent, long-term
cycling performance of silicon anode half-cell cannot provide
enough information since fading of lithium metal counter elec-
trode had effect on the whole performance. Therefore, testing
silicon anodes in full cell configuration is important for serious
evaluation and trouble-shooting toward practical application.
Energy and power density of full cells have a higher commer-
cial relevance.
In a full coin cell configuration with LiCoO2 cathode,
self-healing SEI-based silicon electrode demonstrates two
times higher gravimetric capacity (762 mA h g−1) and volu-
metric capacity (1100 mA h cm−3), compared with graphite of
340 mA h g−1 and 550 mA h cm−3.[84] Besides, binder compo-
nent and electrolyte additive are systematically studied to fur-
ther understand silicon-based full cell system.[98,99] At aspect
of energy density, Si-nanolayer-embedded graphite/carbon
hybrids[47] and monodisperse porous silicon nanospheres[86] can
reach 1043 and 850 W h L−1, respectively. Much effort should

Adv. Energy Mater. 2017, 7, 1700715 1700715  (14 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advenergymat.de

Figure 12.  Natural source for low cost Si. a) Generation of 3D nanoporous Si from rice husks. b) Schematic illustrations for the high-yield synthesis
of hierarchical Si nanostructures from sand (top) and detailed step-by-step reaction processes during the low-pressure magnesiothermic reduction
(bottom). a) Reproduced with permission.[91] Copyright 2013, National Academy of Sciences. b) Reproduced with permission.[96] Copyright 2014,
Wiley VCH.

be conducted to achieve high energy density and long cycle life
required for practical LIBs, through matching suitable high
capacity cathodes with silicon-based anodes. With full cell con-
figuration, parameters such as the average lithiation/delithia-
tion voltage and the Coulombic efficiency are more appropriate
to provide detailed electrochemical performance. Therefore, full
cell performance provides a more reliable and relevant prospect
of the practical applications of a specific material.
7. Conclusion and Outlook
Silicon has been intensively pursued as one of the most pro­
mising anode materials for the next-generation lithium-ion bat-
tery primarily because of high specific capacity. In past decades,
various nanostructures have been demonstrated to address
major challenges for stable electrochemical performance. How-
ever, the widespread commercial application is still limited due
to some new issues resulted by nanomaterials. In this review,
we focus on the issues and recent progress toward commer-
cialization of Si anodes for lithium-ion battery, including ini-
tial Coulombic efficiency, areal capacity, and material cost. The
principles for structure designs in addressing these challenges
were discussed in detail and typical examples were provided to
illustrate these principles.
Despite great progresses, there are still problems remained
to be addressed. As discussed earlier, none of the structures
can solve all the three problems at the same time. Future work
on designing a new structure or combining the above possible
solutions is expected, which can simultaneously provide high

Adv. Energy Mater. 2017, 7, 1700715 1700715  (15 of 17) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
initial Coulombic efficiency, high areal capacity, and low cost.
All-solid state lithium-ion microbatteries with silicon anodes
show great potential in systems-on-a-chip.[100,101] Since the
metal current collector on the anode side is usually a 10 µm
thick copper foil with an areal density ≈10 mg cm−2, which is
much larger in weight than silicon active materials, there are
many research progresses to avoid the use of heavy current col-
lector to enhance the overall gravimetric energy density.[102,103]
On the other hand, reports on freestanding or binder-free elec-
trode structures should also mention the electrode thickness
and areal capacity, for comprehensive evaluation of the pros and
cons. Even though there is still much room to further improve
the areal capacity; freestanding or binder-free electrode struc-
tures also show great potential application in flexible devices.
It is expected that, with extensive efforts, high energy density
and power density silicon-based lithium-ion battery can provide
better service in practical application.
Acknowledgements
The authors acknowledge the microfabrication center of the National
Laboratory of Solid State Microstructures (NLSSM) for technique
support. This work was jointly supported by the National Key Research
and Development Program of China (Grant No. 2017YFA0205700), State
Key Program for Basic Research of China (Grant No. 2015CB659300),
National Natural Science Foundation of China (Grant Nos. 11621091
and 11574143), Natural Science Foundation of Jiangsu Province (Grant
No. BK20150056), the Project Funded by the Priority Academic Program
Development of Jiangsu Higher Education Institutions (PAPD), and the
Fundamental Research Funds for the Central Universities (Grant Nos.
021314380068, 021314380089, and 021314380091).
Conflict of Interest
The authors declare no conflict of interest.
Keywords
areal capacity, cost, Coulombic efficiency, lithium-ion batteries,
silicon anodes
Received: March 16, 2017
Revised: July 18, 2017
Published online: September 21, 2017
[1] B. Dunn, H. Kamath, J. M. Tarascon, Science 2011, 334, 928.
[2] M. Armand, J. M. Tarascon, Nature 2008, 451, 652.
[3] J. M. Tarascon, M. Armand, Nature 2001, 414, 359.
[4] A. S. Arico, P. Bruce, B. Scrosati, J. M. Tarascon, W. Van Schalkwijk,
Nat. Mater. 2005, 4, 366.
[5] J. B. Goodenough, Y. Kim, Chem. Mater. 2010, 22, 587.
[6] C. K. Chan, R. Ruffo, S. S. Hong, Y. Cui, J. Power Sources 2009, 189,
1132.
[7] L. Zong, B. Zhu, Z. Lu, Y. Tan, Y. Jin, N. Liu, Y. Hu, S. Gu, J. Zhu,
Y. Cui, Proc. Natl. Acad. Sci. USA 2015, 112, 13473.
[8] H. Wu, Y. Cui, Nano Today 2012, 7, 414.
[9] M. N. Obrovac, V. L. Chevrier, Chem. Rev. 2014, 114, 11444.
[10] B. Liang, Y. Liu, Y. Xu, J. Power Sources 2014, 267, 469.
Adv. Energy Mater. 2017, 7, 1700715 1700715  (16 of 17)
www.advenergymat.de
[11] J. Li, J. R. Dahn, J. Electrochem. Soc. 2007, 154, A156.
[12] X. H. Liu, J. Y. Huang, Energy Environ. Sci. 2011, 4, 3844.
[13] J. R. Szczech, S. Jin, Energy Environ. Sci. 2011, 4, 56.
[14] X. Su, Q. Wu, J. Li, X. Xiao, A. Lott, W. Lu, B. W. Sheldon, J. Wu,
Adv. Energy. Mater. 2014, 4, 1300882.
[15] Y. Yao, M. T. McDowell, I. Ryu, H. Wu, N. Liu, L. Hu, W. D. Nix,
Y. Cui, Nano Lett. 2011, 11, 2949.
[16] C. K. Chan, H. Peng, G. Liu, K. McIlwrath, X. F. Zhang,
R. A. Huggins, Y. Cui, Nat. Nanotechnol. 2008, 3, 31.
[17] H. Wu, G. Chan, J. W. Choi, I. Ryu, Y. Yao, M. T. McDowell,
S. W. Lee, A. Jackson, Y. Yang, L. Hu, Y. Cui, Nat. Nanotechnol.
2012, 7, 310.
[18] Y. Park, N.-S. Choi, S. Park, S. H. Woo, S. Sim, B. Y. Jang, S. M. Oh,
S. Park, J. Cho, K. T. Lee, Adv. Energy Mater. 2013, 3, 206.
[19] B. Wang, X. Li, X. Zhang, B. Luo, Y. Zhang, L. Zhi, Adv. Mater.
2013, 25, 3560.
[20] J. Y. Li, Q. Xu, G. Li, Y. X. Yin, L. J. Wan, Y. G. Guo, Mater. Chem.
Front. 2017, https://doi.org/10.1039/c6qm00302h.
[21] M. Ashuri, Q. R. He, L. L. Shaw, Nanoscale 2016, 8, 74.
[22] X. X. Zuo, J. Zhu, P. Muller-Buschbaum, Y. J. Cheng, Nano Energy
2017, 31, 113.
[23] X. H. Liu, H. Zheng, L. Zhong, S. Huang, K. Karki, L. Q. Zhang,
Y. Liu, A. Kushima, W. T. Liang, J. W. Wang, J. H. Cho, E. Epstein,
S. A. Dayeh, S. T. Picraux, T. Zhu, J. Li, J. P. Sullivan, J. Cumings,
C. Wang, S. X. Mao, Z. Z. Ye, S. Zhang, J. Y. Huang, Nano Lett.
2011, 11, 3312.
[24] S. W. Lee, H. W. Lee, I. Ryu, W. D. Nix, H. Gao, Y. Cui, Nat.
Commun. 2015, 6, 7533.
[25] S. W. Lee, M. T. McDowell, L. A. Berla, W. D. Nix, Y. Cui, Proc. Natl.
Acad. Sci. USA 2012, 109, 4080.
[26] S. W. Lee, M. T. McDowell, J. W. Choi, Y. Cui, Nano Lett. 2011, 11,
3034.
[27] I. Ryu, J. W. Choi, Y. Cui, W. D. Nix, J. Mech. Phys. Solids 2011, 59,
1717.
[28] X. H. Liu, L. Zhong, S. Huang, S. X. Mao, T. Zhu, J. Y. Huang, ACS
Nano 2012, 6, 1522.
[29] N. Liu, H. Wu, M. T. McDowell, Y. Yao, C. Wang, Y. Cui, Nano Lett.
2012, 12, 3315.
[30] H. Wu, G. Zheng, N. Liu, T. J. Carney, Y. Yang, Y. Cui, Nano Lett.
2012, 12, 904.
[31] H. P. Jia, P. F. Gao, J. Yang, J. L. Wang, Y. N. Nuli, Z. Yang, Adv.
Energy Mater. 2011, 1, 1036.
[32] H. Ma, F. Y. Cheng, J. Chen, J. Z. Zhao, C. S. Li, Z. L. Tao, J. Liang,
Adv. Mater. 2007, 19, 4067.
[33] Z. L. Zhang, Y. H. Wang, W. F. Ren, Q. Q. Tan, Y. F. Chen, H. Li,
Z. Y. Zhong, F. B. Su, Angew. Chem., Int. Ed. 2014, 53, 5165.
[34] J. Luo, X. Zhao, J. Wu, H. D. Jang, H. H. Kung, J. Huang, J. Phys.
Chem. Lett. 2012, 3, 1824.
[35] R. Kumar, A. Tokranov, B. W. Sheldon, X. C. Xiao, Z. Q. Huang,
C. Z. Li, T. Mueller, ACS Energy Lett. 2016, 1, 689.
[36] C. Cao, H. G. Steinruck, B. Shyam, K. H. Stone, M. F. Toney, Nano
Lett. 2016, 16, 7394.
[37] G. Ferraresi, L. Czornomaz, C. Villevieille, P. Novak, M. El Kazzi,
ACS Appl. Mater. Interfaces 2016, 8, 29791.
[38] C. C. Nguyen, S.-W. Woo, S.-W. Song, J. Phys. Chem. C 2012, 116,
14764.
[39] E. Radvanyi, W. Porcher, E. De Vito, A. Montani, S. Franger,
S. Jouanneau Si Larbi, Phys. Chem. Chem. Phys. 2014, 16, 17142.
[40] E. Radvanyi, K. Van Havenbergh, W. Porcher, S. Jouanneau,
J.-S. Bridel, S. Put, S. Franger, Electrochim. Acta 2014, 137, 751.
[41] B. Philippe, R. Dedryvère, J. Allouche, F. Lindgren, M. Gorgoi,
H. Rensmo, D. Gonbeau, K. Edström, Chem. Mater. 2012, 24,
1107.
[42] N. Liu, Z. Lu, J. Zhao, M. T. McDowell, H. W. Lee, W. Zhao, Y. Cui,
Nat. Nanotechnol. 2014, 9, 187.
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
[43] S. Xun, X. Song, L. Wang, M. E. Grass, Z. Liu, V. S. Battaglia,
G. Liu, J. Electrochem. Soc. 2011, 158, A1260.
[44] C. C. Nguyen, H. Choi, S. W. Song, J. Electrochem. Soc. 2013, 160,
A906.
[45] S.-H. Park, D. Ahn, Y.-M. Choi, K. C. Roh, K.-B. Kim, J. Mater.
Chem. A 2015, 3, 20935.
[46] Q. Xu, J. Y. Li, Y. X. Yin, Y. M. Kong, Y. G. Guo, L. J. Wan,
Chem. - Asian J. 2016, 11, 1205.
[47] M. Ko, S. Chae, J. Ma, N. Kim, H.-W. Lee, Y. Cui, J. Cho, Nat.
Energy 2016, 1, 16113.
[48] L. Zong, Y. Jin, C. Liu, B. Zhu, X. Hu, Z. Lu, J. Zhu, Nano Lett.
2016, 16, 7210.
[49] Y. Li, K. Yan, H.-W. Lee, Z. Lu, N. Liu, Y. Cui, Nat. Energy 2016, 1,
15029.
[50] H. J. Kim, S. Choi, S. J. Lee, M. W. Seo, J. G. Lee, E. Deniz, Y. J. Lee,
E. K. Kim, J. W. Choi, Nano Lett. 2016, 16, 282.
[51] N. A. Liu, L. B. Hu, M. T. McDowell, A. Jackson, Y. Cui, ACS Nano
2011, 5, 6487.
[52] M. W. Forney, M. J. Ganter, J. W. Staub, R. D. Ridgley, B. J. Landi,
Nano Lett. 2013, 13, 4158.
[53] Z. Wang, Y. Fu, Z. Zhang, S. Yuan, K. Amine, V. Battaglia, G. Liu,
J. Power Sources 2014, 260, 57.
[54] J. Zhao, Z. Lu, N. Liu, H. W. Lee, M. T. McDowell, Y. Cui, Nat.
Commun. 2014, 5, 5088.
[55] H. Zhao, Z. Wang, P. Lu, M. Jiang, F. Shi, X. Song, Z. Zheng,
X. Zhou, Y. Fu, G. Abdelbast, X. Xiao, Z. Liu, V. S. Battaglia,
K. Zaghib, G. Liu, Nano Lett. 2014, 14, 6704.
[56] Z. Cao, P. Xu, H. Zhai, S. Du, J. Mandal, M. Dontigny, K. Zaghib,
Y. Yang, Nano Lett. 2016, 16, 7235.
[57] J. Zhao, Z. Lu, H. Wang, W. Liu, H. W. Lee, K. Yan, D. Zhuo, D. Lin,
N. Liu, Y. Cui, J. Am. Chem. Soc. 2015, 137, 8372.
[58] J. Zhao, H. W. Lee, J. Sun, K. Yan, Y. Liu, W. Liu, Z. Lu, D. Lin,
G. Zhou, Y. Cui, Proc. Natl. Acad. Sci. USA 2016, 113, 7408.
[59] J. Zhao, G. Zhou, K. Yan, J. Xie, Y. Zhang, L. Liao, Y. Jin, K. Liu,
P. Hsu, J. Wang, H. Cheng, Y. Cui, Nat. Nanotechnol. 2017;
https://doi.org/10.1038/NNANO.2017.129.
[60] X. Chen, X. Li, D. Mei, J. Feng, M. Y. Hu, J. Hu, M. Engelhard,
J. Zheng, W. Xu, J. Xiao, J. Liu, J.-G. Zhang, ChemSusChem 2014,
7, 549.
[61] C. Xu, F. Lindgren, B. Philippe, M. Gorgoi, F. Björefors,
K. Edström, T. Gustafsson, Chem. Mater. 2015, 27, 2591.
[62] Y. Okuno, K. Ushirogata, K. Sodeyama, Y. Tateyama, Phys. Chem.
Chem. Phys. 2016, 18, 8643.
[63] K. Kim, I. Park, S. Y. Ha, Y. Kim, M. H. Woo, M. H. Jeong,
W. C. Shin, M. Ue, S. Y. Hong, N. S. Choi, Electrochim. Acta 2017,
225, 358.
[64] H. Shin, J. Park, A. M. Sastry, W. Lu, J. Electrochem. Soc. 2015, 162,
A1683.
[65] S. Chen, M. L. Gordin, R. Yi, G. Howlett, H. Sohn, D. Wang,
Phys. Chem. Chem. Phys. 2012, 14, 12741.
[66] K. Schroder, J. Alvarado, T. A. Yersak, J. Li, N. Dudney, L. J. Webb,
Y. S. Meng, K. J. Stevenson, Chem. Mater. 2015, 27, 5531.
[67] J. Ryu, D. Hong, M. Shin, S. Park, ACS Nano 2016, 10, 10589.
[68] S. D. Beattie, D. Larcher, M. Morcrette, B. Simon, J. M. Tarascon,
J. Electrochem. Soc. 2008, 155, A158.
[69] T. Wada, T. Ichitsubo, K. Yubuta, H. Segawa, H. Yoshida, H. Kato,
Nano Lett. 2014, 14, 4505.
[70] D. Lin, Z. Lu, P.-C. Hsu, H. R. Lee, N. Liu, J. Zhao, H. Wang,
C. Liu, Y. Cui, Energy Environ. Sci. 2015, 8, 2371.
[71] X. Li, M. Gu, S. Hu, R. Kennard, P. Yan, X. Chen, C. Wang,
M. J. Sailor, J. G. Zhang, J. Liu, Nat. Commun. 2014, 5, 4105.
[72] S. Chae, M. Ko, S. Park, N. Kim, J. Ma, J. Cho, Energy Environ. Sci.
2016, 9, 1251.
[73] S.-O. Kim, A. Manthiram, J. Mater. Chem. A 2015, 3, 2399.
Adv. Energy Mater. 2017, 7, 1700715 1700715  (17 of 17)
www.advenergymat.de
[74] A. Magasinski, P. Dixon, B. Hertzberg, A. Kvit, J. Ayala, G. Yushin,
Nat. Mater. 2010, 9, 353.
[75] R. Yi, F. Dai, M. L. Gordin, S. Chen, D. Wang, Adv. Energy Mater.
2013, 3, 295.
[76] H. Zhao, A. Du, M. Ling, V. Battaglia, G. Liu, Electrochim. Acta
2016, 209, 159.
[77] H. Zhao, Y. Wei, R. Qiao, C. Zhu, Z. Zheng, M. Ling, Z. Jia, Y. Bai,
Y. Fu, J. Lei, X. Song, V. S. Battaglia, W. Yang, P. B. Messersmith,
G. Liu, Nano Lett. 2015, 15, 7927.
[78] S. J. Park, H. Zhao, G. Ai, C. Wang, X. Song, N. Yuca,
V. S. Battaglia, W. Yang, G. Liu, J. Am. Chem. Soc. 2015, 137,
2565.
[79] S. Xun, B. Xiang, A. Minor, V. Battaglia, G. Liu, J. Electrochem. Soc.
2013, 160, A1380.
[80] C. Wang, H. Wu, Z. Chen, M. T. McDowell, Y. Cui, Z. A. Bao, Nat.
Chem. 2013, 5, 1042.
[81] D. S. Jung, T. H. Hwang, S. B. Park, J. W. Choi, Nano Lett. 2013,
13, 2092.
[82] Z. D. Lu, N. Liu, H. W. Lee, J. Zhao, W. Y. Li, Y. Z. Li, Y. Cui, ACS
Nano 2015, 9, 2540.
[83] Z. Chen, C. Wang, J. Lopez, Z. Lu, Y. Cui, Z. Bao, Adv. Energy
Mater. 2015, 5, 1401826.
[84] Y. Jin, S. Li, A. Kushima, X. Q. Zheng, Y. M. Sun, J. Xie, J. Sun,
W. J. Xue, G. M. Zhou, J. Wu, F. F. Shi, R. F. Zhang, Z. Zhu,
K. P. So, Y. Cui, J. Li, Energy Environ. Sci. 2017, 10, 580.
[85] B. Wang, X. L. Li, T. F. Qiu, B. Luo, J. Ning, J. Li, X. F. Zhang,
M. H. Liang, L. J. Zhi, Nano Lett. 2013, 13, 5578.
[86] W. Wang, Z. Favors, C. Li, C. Liu, R. Ye, C. Fu, K. Bozhilov, J. Guo,
M. Ozkan, C. S. Ozkan, Sci. Rep. 2017, 7, 44838.
[87] I. H. Son, J. H. Park, S. Kwon, S. Park, M. H. Rummeli,
A. Bachmatiuk, H. J. Song, J. Ku, J. W. Choi, J. M. Choi, S. G. Doo,
H. Chang, Nat. Commun. 2015, 6, 7393.
[88] M. Ge, Y. Lu, P. Ercius, J. Rong, X. Fang, M. Mecklenburg, C. Zhou,
Nano Lett. 2014, 14, 261.
[89] B. Zhu, Y. Jin, Y. L. Tan, L. Q. Zong, Y. Hu, L. Chen, Y. B. Chen,
Q. Zhang, J. Zhu, Nano Lett. 2015, 15, 5750.
[90] Y. Jin, S. Zhang, B. Zhu, Y. Tan, X. Hu, L. Zong, J. Zhu, Nano Lett.
2015, 15, 7742.
[91] D. S. Jung, M. H. Ryou, Y. J. Sung, S. B. Park, J. W. Choi, Proc. Natl.
Acad. Sci. USA 2013, 110, 12229.
[92] Y.-C. Zhang, Y. You, S. Xin, Y.-X. Yin, J. Zhang, P. Wang,
X.-S. Zheng, F.-F. Cao, Y.-G. Guo, Nano Energy 2016, 25, 120.
[93] H. Xu, H. Zhang, Y. Mu, Y. Feng, Y. Wang, ACS Sustainable Chem.
Eng. 2015, 3, 1650.
[94] J. Liu, P. Kopold, P. A. van Aken, J. Maier, Y. Yu, Angew. Chem.
2015, 54, 9632.
[95] Z. Favors, W. Wang, H. H. Bay, Z. Mutlu, K. Ahmed, C. Liu,
M. Ozkan, C. S. Ozkan, Sci. Rep. 2014, 4, 5623.
[96] J.-K. Yoo, J. Kim, M.-J. Choi, Y.-U. Park, J. Hong, K. M. Baek,
K. Kang, Y. S. Jung, Adv. Energy Mater. 2014, 4, 1400622.
[97] X. Li, C. Yan, J. Wang, A. Graff, S. L. Schweizer, A. Sprafke,
O. G. Schmidt, R. B. Wehrspohn, Adv. Energy Mater. 2015, 5,
1401556.
[98] A. Bordes, K. Eom, T. F. Fuller, J. Power Sources 2014, 257, 163.
[99] J. G. Xu, L. Zhang, Y. K. Wang, T. Chen, M. Al-Shroofy, Y. T. Cheng,
ACS Appl. Mater. Interfaces 2017, 9, 3562.
[100] F. Le Cras, B. Pecquenard, V. Dubois, V. P. Phan, D. Guy-Bouyssou,
Adv. Energy Mater. 2015, 5, 1501061.
[101] V. P. Phan, B. Pecquenard, F. Le Cras, Adv. Funct. Mater. 2012, 22,
2580.
[102] Y. L. Chen, Y. Hu, Z. Shen, R. Z. Chen, X. He, X. W. Zhang, Y. Q. Li,
K. S. Wu, J. Power Sources 2017, 342, 467.
[103] C. Botas, D. Carriazo, W. Zhang, T. Rojo, G. Singh, ACS Appl.
Mater. Interfaces 2016, 8, 28800.
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim