Nano Energy 38 (2017) 477–485

Contents lists available at ScienceDirect

Nano Energy
journal homepage: www.elsevier.com/locate/nanoen

Full paper

A high-performance Li-ion anode from direct deposition of Si nanoparticles MARK

a b a c a a,d
Yaolin Xu , Ellie Swaans , Sibo Chen , Shibabrata Basak , Peter Paul R. M. L. Harks , Bo Peng ,
⁎
Henny W. Zandbergenc, Dana M. Borsab, Fokko M. Muldera,
a
Materials for Energy Conversion and Storage (MECS), Department of Chemical Engineering, Faculty of Applied Science, Delft University of Technology,
Van der Maasweg 9, 2629 HZ Delft, The Netherlands
b
Meyer Burger (Netherlands) B.V., Luchthavenweg 10, 5657 EB Eindhoven, The Netherlands
c
Kavli Institute of Nanoscience Delft, Department of Quantum Nanoscience, Delft University of Technology, Lorentzweg 1, 2628 CJ Delft, The Netherlands
d
Department of Physics, Renmin University of China, No. 59 Zhongguancun Street, Haidian District, 100872 Beijing, China

A R T I C L E I N F O A BS T RAC T

Keywords: Nanostructured silicon has been intensively investigated as a high capacity Li-ion battery anode. However, the
Li ion battery commercial introduction still requires advances in the scalable synthesis of sophisticated Si nanomaterials and
Si nanoparticles electrodes. Moreover, the electrode degradation due to volume changes upon de-/lithiation, low areal electrode
Direct deposition capacity, and application of large amounts of advanced conductive additives are some of the challenging aspects.
One-off SEI formation
Here we report a Si electrode, prepared from direct deposition of Si nanoparticles on a current collector without
any binder or conducting additives, that addresses all of the above issues. It exhibits an excellent cycling stability
and a high capacity retention taking advantages of what appears to be a locally protective, yolk-shell
reminiscent, solid electrolyte interphase (SEI) formation. Cycling an electrode with a Si nanoparticle loading
of 2.2 mg cm−2 achieved an unrivalled areal capacity retention, specifically, up to 4.2 mAh cm−2 and ~
1.5 mAh cm−2 at 0.8 mA cm−2 and 1.6 mA cm−2, respectively.

1. Introduction methods that do use potentially more scalable methods involve large
amounts of additional relatively costly nanoscale carbonaceous con-
Energy storage is one of the key issues of modern society. The ducting networks (graphene [21], carbon nanotube [22], carbon
development of Li-ion battery based energy storage devices is flourish- nanofiber [23], etc.) and advanced binder materials [24,25], which
ing, in an effort to meet the increasing power demand of a diverse lowers the cost-effectiveness and overall capacity that is realized.
range of applications, such as electrical vehicles and widely used Another major challenge for Si based anodes is to enhance the areal
portable electronics [1–5]. Si based anode materials have been capacity, which is generally poor owing to the low mass loading of
extensively studied for Li ion batteries mostly due to the unparalleled active materials [14–17,26–28]. Progress has been made to increase
theoretical capacity of Si (4200 mAh g−1 for Li4.4Si). However, the huge the mass loading of Si and/or electrode packing density to improve the
volume change and its resulting electrode pulverization upon Li uptake performance [29–36], but these electrode manufacturing approaches
leads to electrical conduction loss and rapid capacity degradation, are still hardly industrially viable as they introduce either high cost
which has severely reduced its promise for practical applications [6–9]. materials or complicated production processes.
Research has been focusing on various kinds of nanostructured Si Thus the main challenge for the practical implementation of Si
materials to resolve the different aspects leading to capacity loss, and to anodes is to develop cost-effective and commercial viable techniques to
achieve stable cycling performance [10–12]. Current studies mostly prepare Si anodes. The technique should be capable of producing Si
synthesize sophisticated nanostructured Si, including nanowires particles that are nanosized to accommodate volume changes, and that
[13,14], nanotubes [15,16], nanospheres in a flexible matrix or 3D are electronically connected throughout the structure. In addition the
architecture [17,18], etc., or apply protective coatings [19,20] to technique should have a high production rate and be scalable. When
prevent electrode pulverization and to provide continuous electronic possible advanced carbons and binders in the process should be
and ionic conduction pathways during cycling. However, these syn- minimized and further processing for electrode manufacturing should
thetic approaches on a lab scale appear to have poor scalability and also be eliminated. Progress has been made to realize the scalable
therefore until now limit large scale commercial viability. Moreover, synthesis of nanostructured Si anodes [37–40] as well as to produce

⁎
Corresponding author.
E-mail address: F.M.Mulder@tudelft.nl (F.M. Mulder).

http://dx.doi.org/10.1016/j.nanoen.2017.06.011
Received 15 March 2017; Received in revised form 5 June 2017; Accepted 6 June 2017
Available online 13 June 2017
2211-2855/ © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Y. Xu et al. Nano Energy 38 (2017) 477–485

Fig. 1. A schematic of electrode preparation process via direction deposition of Si nanoparticles.

binder-free Si electrodes [41–46]. information (SI). The as-synthesized Si NP-C sample was directly
In this study, a novel but simple and industrially scalable process applied as the electrode after drying in a vacuum oven for 12 h. For
was developed to synthesize Si anodes. (Fig. 1) In this process silicon comparison, Si NP deposition was also carried out on a non-porous
nanoparticles (Si NP) are directly deposited on a porous carbon current planar Cu foil (Goodfellow, 12.7 μm) and its morphology before and
collector by plasma enhanced chemical vapour deposition (PECVD). after de-/lithiation is shown in SI.
The obtained additive- and binder-free electrode can be used as an
anode in a Li-ion battery without further processing. 2.2. Sample characterization
Previous studies on the Si deposition based Li ion anodes consider
Si nanowires, nanoscale Si films on 3D substrates, or thin structured Si For the pristine Si NP electrode, Scanning Electron Microscopy
films [13–16,28,42,47–50]. The novel approach in this work intro- (SEM) images were taken using a JEOL JSM 6010 F scanning electron
duces a facile one-step synthesis by direct deposition of Si nanoparti- microscope at an accelerating voltage of 5 kV; The high resolution SEM
cles with at least an order of magnitude higher mass loading of Si images were acquired using Hitachi S4800 working at 2 kV. For
compared to the previous studies (max. 5 µm Si layer in previous Transmission Electron Microscopy (TEM) measurement Si particles
reports), and without the need for catalysts or other additional 3D were carefully scraped off from the electrode. Bright field TEM images
nanostructures. The Si electrode in this work exhibits several advan- were acquired using FEI-Tecnai operating at 200 kV. X-ray diffraction
tages: (1), CVD based Si deposition is industrially well-developed and (XRD) patterns were measured with a PANalytical X′Pert Pro
scalable due to its present widespread applications in solar cells and PW3040/60 diffractometer with Cu Kα radiation operating at 45 kV
electronics. The thickness and porosity of Si NP deposition layers can and 40 mA. Raman spectra on Si NP particles were obtained from a
be monitored and controlled. (2), The plasma aided self-assembly of Renishaw InVia Raman spectrophotometer. The wavelength of the
nanoparticle layers produces hierarchical nano and micro structured Si laser applied is 785 nm and the laser power is set at 0.2 mW. Thermal
NP branched structures, which results into sufficient porosity to Gravimetric Analysis (TGA) was carried out with a Perkin Elmer TGA 7
accommodate the volume expansion during Li ion uptake. (3), The Si thermogravimetric analyser. The Si NP sample were scraped from the
nanoparticles appear naturally interconnected, which eliminates the carbon substrate and heated in a 20% O2/Ar mixture from room
need for binders and conductive additives and thus improves the temperature to 1000 °C followed by a full oxidation at 1000 °C for
overall capacity of the electrode and reduces the processing cost. (4), 200 min. X-ray photoelectron spectroscopy (XPS) was performed with
Controlled oxidation on the surface of Si NP is achieved through low a K-Alpha XPS System.
pressure O2 exposure between the plasma deposition cycles. The All cycled electrodes were washed with diethyl carbonate (DEC) for
inactive and solid SiOx scaffold appears to reinforce the structural 3 times to remove soluble electrolyte residuals in the electrode, and
integrity of the electrode during cycling. (5), The inter-particle poros- were dried in a glove box before the measurement. The SEM images
ities are connected in nanoscale, which provides an express way for were acquired using Hitachi S4800 with an accelerating voltage of
electrolyte infiltration and thus good Li ion transport throughout the 15 kV. STEM-EDX (Energy Dispersive X-ray spectroscopy) mapping
electrode. Hence high performance can be reached for electrodes with was done at FEI-Titan operating at 300 kV with Oxford instrument
increased thicknesses, enabling high areal capacities with increased Si EDX system. TEM, XRD and XPS measurement were performed using
mass loading. (6), A templating effect [51,52] is introduced by the open the same equipment and working conditions as were utilized for
porosity of the used carbon substrate. This results in interconnected pristine samples.
micron scale porosities allowing electrolyte to enter the Si NP layer,
which shortens the diffusion pathway of Li ion towards active materials 2.3. Electrochemistry
and thus enables fast Li ion transport within the electrode. (7), The
density of this Si NP electrode (~ 0.5 g cm−3) may not appear high but The half-cell employed for the electrochemical measurement was a
this is about 20% of the density of bulk silicon, significantly higher than lab-designed prototype (Fig. S1) which consists of two stainless steel
in other reports where a much smaller fraction of the bulk Si density is flanges, an O-ring in between the flanges, and a plastic vacuum clamp
reached [15,16,28,47]. The density is however sufficiently low to allow to hold them together. There is a separate stainless steel plate
space for the large volume expansion of Si; up to 400% expansion supported by a metal spring to provide the necessary pressure and
would just fill the total available volume, let alone the space required mechanical compaction inside the battery. Half-cell Li ion batteries for
for the SEI growth. test were assembled inside an Ar environment glovebox with the O2
and H2O levels < 0.1 ppm. A Li metal foil (Aldrich) worked as the
2. Methods counter electrode. A borosilicate glass micro fibre (Whatman) was
applied as the separator, and 1 M LiPF6 dissolved in DEC, ethylene
2.1. Electrode preparation carbonate (EC) and fluoroethylene carbonate (FEC) (1:1:1 in volume)
was utilized as the electrolyte.
The Si NP electrodes were prepared via PECVD from silane using The galvanostatic electrochemical performance was measured with
an Expanding Thermal Plasma (ETP) source [53,54]. Si NP depositions a MACCOR 4600 battery cycler at room temperature. The cut-off
were performed on a porous carbon sheet (Caplinq) layer by layer and voltages for discharge and charge were 0.005 V and 2 V versus Li/Li+,
each layer includes 30s′ deposition of Si NP. The specific process and respectively. Cyclic voltammetry was perform with a PGSTAT302N
conditions for the Si NP deposition are stated in the Supporting Autolab potentiostat within the same voltage range. Electrochemical

478
Y. Xu et al. Nano Energy 38 (2017) 477–485

impedance spectroscopy (EIS) measurements were carried out with a carbon sheet. The thickness of the Si NP layer is around 40 µm (Si NP:
PGSTAT302N Autolab within the frequency range of 1 MHz and ~ 2 mg cm−2). SEM (Fig. 2a–g) shows that small silicon nanoparticles (
0.01 Hz. < 20 nm) form clusters with typical sizes of approximately100 nm
scale, and a hierarchically multi-branched porous nanostructure ap-
2.4. Subtraction of the carbon capacity pears throughout the Si NP layers. TEM (Fig. 2h–j) demonstrates that
the particles reside as clusters of ~ 100 nm scale; while individual
Specific capacities of the Si NP reported in this paper were particles (~ 20 nm) are spherical with a very thin native SiOx layer on
calculated based on the mass of the Si NP layer (excl. the carbon the surface. The oxidation layers are amorphous as only XRD peaks of
substrate). The capacity of the carbon substrate was determined by crystalline Si are observed (Fig. 2k). While the presence of both
measuring the capacity retentions of the carbon paper at a series of amorphous and crystalline Si phases is evidenced by Raman spectro-
current rates and then fitting the capacity data with a logistic function; scopy (Fig. 2l). EDX elemental mapping performed on the Si NP layer
(Fig. S3) using this function, the capacities of the carbon substrate (Fig. S5) shows that the oxygen distribution is uniform throughout the
could be calculated regarding the corresponding current rates applied Si NP layers. Individual Si nanoparticles are thus homogeneously
on carbon. The specific capacities of the Si NP were determined by oxidized on the surface, which is also consistent with the TEM results.
subtracting the capacity of carbon from that of the Si NP-C electrode. To determine the amount of oxygen present in the sample, TGA (Fig.
(More details about the subtraction of the capacity contribution from S6) was applied by heating the Si NP under 20% O2/Ar gas and fully
the carbon substrate are described in SI.) oxidizing Si into SiO2. The result indicates the fraction of Si inside the
sample amounts to 46.3 wt % (i.e. Si: SiO2 = 0.65: 0.35 in mole). Such
3. Results and discussions an amount of SiO2 is consistent with the observation of a very thin layer
(1.6 nm) on a 15 nm diameter Si particle in the high resolution TEM
3.1. Electrode manufacturing and characterization images.

Si NP anodes were synthesized via expanding thermal plasma 3.2. Electrochemical performance
chemical vapour deposition from silane. Si NP depositions were
performed on a planar but porous carbon substrate layer by layer; The electrochemical performance of the Si NP-C electrode was
cooling and air exposure were applied in between. In this study 1–4 verified by cycling the electrode against a Li metal counter electrode.
layers of Si NP depositions were carried out on the carbon substrates. The electrochemical performance of Si NP-C electrodes has contribu-
The as-prepared Si NP-C electrodes are tested in a half cell with a Li tions from both Si NP and the carbon substrate. The capacity
metal counter electrode, separator and liquid electrolyte. contribution from the carbon substrate was determined and subtracted
Fig. 2 shows the morphology of a 4-layer Si NP deposition on the from that of the Si NP-C electrode. Details about the capacity

Fig. 2. Morphology of as-synthesized Si NP on a porous C substrate. (a), A SEM image of the Si NP electrode. inset of (a): SEM micrograph of the porous C fibre sheet applied as a
substrate for Si deposition. (b) - (d), cross-sectional and (e) - (g), top view SEM micrographs of the nanostructured Si layer imaged at different magnifications. (h) - (j), TEM images of a
Si nanoparticle cluster and individual particles as-synthesized from PECVD. (k), XRD patterns of Si nanoparticles as synthesized (the specimen for X-ray diffraction was loose Si
nanoparticles scraped from the substrate). (l), Raman spectra of the Si nanoparticles indicating the co-existence of crystalline and amorphous Si.

479
Y. Xu et al. Nano Energy 38 (2017) 477–485

Fig. 3. Electrochemical performance of the Si NP-C electrode for de-/lithiation at room temperature. (a), Capacity retentions and Coulombic efficiencies of Si NP (0.5 mg cm−2) at
0.1 A g−1 and the capacity retentions of the carbon substrate cycling with the same measurement program. (b), Electrochemical performance of the Si NP-C electrodes with different
mass loadings cycling at 0.8 A g−1. (c), Areal capacity retentions and Coulombic efficiencies of the Si NP-C (2.2 mg cm−2) electrodes cycling at 0.8 and 1.6 mA cm−2. (d), Rate capabilities
of Si NP electrodes with different mass loadings. (In Fig. 3a – d, solid symbols: lithiation capacities; open symbols: delithiation capacities; cross-centered open symbols: Coulombic
efficiencies.) (e), Voltage profiles of Si NP (0.5 mg cm−2) at 0.1 A g−1. Inset of Fig. (e): voltage profiles of the carbon substrate. (f), Voltage responses of the Si NP-C electrodes with
different mass loadings cycling at 0.8 A g−1. (g), Voltage profiles of the Si NP-C (2.2 mg cm−2) electrodes cycling at 0.8 and 1.6 mA cm−2. (h), GITT de-/lithiation of the Si NP-C electrode.
GITT was carried out on the thickest Si NP-C electrode (2.4 mg cm−2) following the rate performance test in Fig. (g). Current pulses: 0.1 A g−1 for 1 h during charge/discharge;
Relaxation for 4 h. (i), Cyclic voltammograms of the Si NP-C electrode (scan rate: 0.5 mV s−1).

subtraction are described in Section 2 and SI. In this paper, all the (Fig. 3a). A delithiation capacity retention of 1825 mAh g−1 is obtained
reported specific capacities are calculated based on the mass of the Si after 30 cycles; Given that the amount of SiOx accounts for about half of
NP layer, including the mass of SiO2. the sample, the specific capacity of Si (excl. SiOx) would be doubled and
Cycling a Si NP (mass loading of 0.5 mg cm−2) electrode on the therefore be remarkably close to the theoretical capacity assuming SiOx
carbon substrate at 0.1 A g−1 shows a highly stable reversible capacity does not contribute to the reversible capacity.

480
Y. Xu et al. Nano Energy 38 (2017) 477–485

Fig. 4. Impedance measurements. (a), The electronic DC resistance of the Si NP layer during chemical lithiation; inset: experimental set-up of in-situ impedance measurement during
chemical lithiation of Si NP. The influence of the Cu substrate, Li foil and cell tester has been subtracted by carrying out a parallel experiment with the same set-up and test conditions
but replacing the Si NP-Cu electrode with a blank Cu foil without Si NP. (b), Nyquist plots of the Si NP anode in a half cell with Li metal counter electrode and liquid electrolyte: before
cycling and after 1, and after 100 cycles.

Fig. 5. Morphology of the Si NP electrode at different stages: pristine, 1st cycle and 100th cycle cycling at 0.8 A g−1. (a) - (f), SEM images showing the morphology of Si NP clusters at
different stages: (a) - (b), pristine; (c) - (d), SEI shell after 1st cycle; (e) - (f), SEI shell after 100th cycle; (g), A schematic of the morphology change of the hierarchically nanostructured Si
NP during de-/lithiation process. (h) - (k), STEM-EDX result: layered image and element mapping of Si, O and F, respectively. The measurements were performed on electrode after 100
de-/lithiation cycles at 0.8 A g−1.

481
Y. Xu et al. Nano Energy 38 (2017) 477–485

Fig. 6. Characterization on the Si NP electrode at different stages: pristine, 1st cycle and 100th cycle cycling at 0.8 A g−1. (a), XRD patterns of the Si NP electrode. (b), XPS spectra of the
Si NP electrode. The electrodes were washed with DEC for 3 times to remove any chemical residuals (mainly LiPF6) on the electrode and were then dried in glove box to evaporate any
organic volatile compounds (mostly DEC) before XPS and XRD measurement.

Fig. 3b demonstrates a remarkable cycling stability over 100 cycles Measuring the electrochemical performance of the electrode at
for Si NP-C electrodes with various mass loadings cycling at 0.8 A g−1. different constant temperatures between 10 ℃ and 40 ℃ (Fig. S7), it is
The small and periodic capacity fluctuations stem from ambient observed that a higher temperature leads to a higher initial capacity
temperature variations. At this charge rate the achieved capacity of Si and a relatively faster degradation. A lower temperature results into a
NP-C declines with growing Si NP loadings; the capacity reaches lower capacity retention but a much more stable cycling stability.
around 1600 mAh g−1, 1200 mAh g−1 and 800 mAh g−1 for electrodes Moreover, regardless of the capacity retention disparity at different
with 0.7 mg cm−2, 1.3 mg cm−2 and 2.0 mg cm−2 Si NP, respectively. temperatures, no obvious capacity swings can be observed. This reveals
The slight capacity fading along cycling originates from the gradual that the phenomenon of apparent capacity fluctuations observed in
structural deformation along repetitive dis-/charges. Fig. 3 results from the temperature variations between day and night in
Cycling at 0.8 mA cm−2 the areal capacity for the Si NP-C electrode the laboratory. It is more pronounced for the electrode with a higher
(Fig. 3c) drops gradually, regardless of the slight swings, from the mass loading because the absolute value of the applied current
initial reversible capacity of 4.2 mAh cm−2 to 2.5 mAh cm−2 after 100 (0.8 A g−1) is higher for a larger mass.
cycles. A greatly promoted cycling stability is observed when the The main de-/lithiation of Si NP occurs at a lower voltage range ( <
current density is increased to 1.6 mA cm−2 which demonstrates an 0.8 V), which is evidenced by the voltage profiles (Fig. 3e–h) and the
areal capacity retention of about 1.5 mAh cm−2 and no apparent cyclic voltammograms (Fig. 3i). A solid electrolyte interphase (SEI)
capacity degradation after 100 cycles. The improvement of cycleability, layer is formed, due to the decomposition of the organic electrolyte
when the current rate increases from 0.8 mA cm−2 to 1.6 mA cm−2, solvent and reaction with the electrode materials, at a higher voltage
results from the rapid and incomplete lithiation of Si NP at high plateau (0.7 V, Fig. 3e) during the initial Li ion uptake. Nevertheless,
current rate and therefore less structural deformation along cycling. this phenomenon is not observed in the following cycles; instead, later
The achieved areal capacity is high compared to the recent literature cycles show highly reversible capacities and rather symmetrical voltage
reports on advanced Si based electrodes (Table S2). profiles for charge and discharge indicating a low overpotential
Galvanostatic dis-/charge of the Si NP-C electrodes also shows an between charge and discharge. Fig. 3f & g demonstrate that the
excellent rate capability. (Fig. 3d) A remarkable rate performance is overpotential grows when the mass loading of Si NP is higher and at
observed for the Si NP-C electrode with a relative lower mass loading elevated current rates, which is due to the limited Li ion diffusivity in
(1.0 mg cm−2). The capacity achieved at 0.1 A g−1 reaches > the Si NP electrode and intrinsic electronic conductivity. However,
1600 mAh g−1; and no apparent capacity degradation occurs when regardless of the increasing overpotentials, considerable reversible
the current rate is increased to higher currents up to 1 A g−1. capacity retentions are still achieved.
Increasing the mass loading of Si on carbon sheet slightly reduces To study the equilibrium potential the galvanostatic intermittent
capacities. ~ 1400 mAh g−1 for 1.7 mg cm−2 and above 1300 mAh g−1 titration technique (GITT) was applied by pulsing a current of 0.1 A g−1
for 2.4 mg cm−2, respectively, have been realized for deep lithiation (i.e. for 1 h and a relaxation period of 4 h after each pulse. Fig. 3h reports
low current densities); The capacity fades with either rising current that overpotentials during both Li uptake and release decrease with
rates or increases in mass loadings. higher lithiation levels.

482
Y. Xu et al. Nano Energy 38 (2017) 477–485

3.3. Electronic / ionic conductivity individual particles and their SEI. In this way the contact throughout
the electrode is not destroyed during cycling, as illustrated in Fig. 5g.
The electronic conductivity of Si can be expected to be greatly This one-off SEI formation with the remaining coherence throughout
enhanced upon trace amounts of Li insertion, leading to even metalli- the nanoparticle strands is of remarkable advantage. Specifically the
cally conducting Si [55–57]. This could be an explanation of why an local SEI surrounding the particles apparently has outstanding elasti-
initially poorly conducting electrode that merely consists of Si nano- city and mechanical strength, which protects the electrode. It therefore
particles is capable of fast Li ion uptake and release. For the as- prevents further SEI formation while it also provides ionic conduction.
prepared nanostructured electrode, with its packed structure of in- Furthermore, the SEI layer growth is also investigated by EDX
dividual insulating SiO2 coated particles, the electrical conduction element mapping of F and O (mostly from LiF, Li2CO3, and Li-Si-O
enhancement induced by Li insertion/doping will be of paramount (see below) in the SEI layer). Fig. 5j & k shows that F and O appear
importance. A simple in-situ resistance measurement was performed more concentrated in the shell of the Si NP cluster, which may be
during chemical lithiation by pressing a Li foil against the Si NP layers; indicative of its yolk-shell configuration on the scale of the clusters (~
purely in the solid state, without any electrolyte addition. Fig. 4a 100 nm). The SEI thus surrounds the clusters of NP, rather than the
demonstrates that the DC electrical resistance of the pristine Si NP individual NP. On the larger scale of the entire electrode (10 µm scale)
layer (1.5 mg cm−2) is as high as ~ 500 kΩ. Most remarkably the Fig. S9 shows that F and O distribute evenly throughout the electrode
electrical resistance reduces by one order of magnitude within as little indicating a uniform SEI growth throughout the Si NP layer. The
as 2 min of all solid chemical lithiation; it further decreases rapidly to resulting morphology of a spontaneous SEI layer around the Si-NP
< 50 Ω in 2.5 h. The electronic conductivity of the Si NP layer thus clusters, extending throughout the electrode without breaking the
increases by about 4 orders of magnitude during solid state lithiation at electronic contacts between the clusters apparently exhibits high
room temperature, which is comparable to the published results on resistance against stresses caused by the Si volume expansion and
electrochemical lithiation of a single Si nanowire / thin solid film contraction of individual particles, therefore it ensures the structural
(100 nm) [56,57]. This finding may illustrate how the Si NP electrode, integrity of the Si NP layer and hence supports the cycling stability. In
even without addition of any conducting additives, achieves sufficient addition it may be deduced from the electrochemical performance that
electronic conductivity for Li ion insertion and extraction. The presence the entire SEI formation takes place only in the first cycles and thus
of SiO2 apparently has no large detrimental effect on the DC con- provides a stabilised electrode morphology afterwards.
ductivity. XRD patterns (Fig. 6a) indicate that crystalline Si is observed in
Furthermore, EIS was performed to investigate the ionic conduc- pristine Si NP, and that it has been totally amorphized after the first
tivity of the Si NP anode in a half cell with a Li counter electrode and cycle. LixSiOy compounds can be identified from the XRD spectra after
liquid electrolyte. Nyquist plots (Fig. 4b) of the Si NP anode demon- dis-/charge for 100 cycles. Irreversible formation of LixSiOy (mostly
strate semicircles in the high frequency region corresponding to the Li4SiO4, reaction (1)) is commonly recognized for Si/SiOx based
interface resistance of the Si NP anode related to the SEI layer and materials; whereas reversible Li2Si2O5 formation (reaction (2)) may
charge transfer resistance of Li ions. Linear tails in the low frequency also be present [64–66].
range are related to the long range Li-ion diffusion in the electrode
irreversible reaction: 2 SiO2 + 4 Li+ + 4 e- → Li4SiO4 + Si (1)
[58,59]. It is observed that before cycling, the charge transfer resis-
tance is large ( > 300 Ω) as no electrochemical reaction has happened
and a large barrier for Li ion diffusion exists. After the first cycle it
reversible reaction: 5 SiO2 + 4 Li+ + 4 e- ↔2 Li2Si2O5 + Si (2)
decreased significantly to around 70 Ω. After dis-/charge for 100 cycles
at 0.8 A g−1, the interface resistance increases only slightly, indicating The Si 2p XPS spectra (Fig. 6b) of the electrode demonstrate that
that there is no apparent change in the transfer resistance of Li-ions pristine Si NP consists of Si (99.4 eV) and SiO2 (103.3 eV). The peak
from and into the Si NP electrode. This indicates that the composition from SiO2 dominates since XPS is surface sensitive (~ few nm). Upon
as well as structure of the electrode has been retained after the first initial dis-/charge SiO2 has been partially reduced to SiOx (~ 102.4 eV);
cycle and essentially stays intact during cycling. the production of LixSiOy (~ 100.5 eV) is also evident according to
reactions (1) and (2); The peak of Si is almost invisible under XPS due
3.4. One-off in-situ uniform SEI formation to the coverage of the SEI. After 100 cycles, the Si 2p spectrum is
dominated by LixSiOy due to the further consumption of SiOx and
SEI formation is commonly believed to be a crucial disadvantage for formation of LixSiOy (mainly in SEI).
nanostructured Si anode materials since repetitive cracking of SEI The XPS spectra after 1st cycle indicate that the SEI layer is
layers due to the large volume expansion and contraction of the composed of a series of substances, such as LiF, Li2CO3, LixSiOy, and
electrode and continuous formation of new SEI ceaselessly reduces ROCOOLi, etc., which are produced upon the decomposition of organic
available active material, leading to a rapid capacity degradation electrolyte and the following reactions with Si NP. The commonly
[4,6,11,60]. However, in this article the SEI layer turns out to behave recognized LiF is evidenced from the dominating peak at 684.8 eV in
differently and is even beneficial for the cycling stability. the F 1 s spectrum and the one at 55.5 eV in the Li 1 s spectrum.
FEC addition into the electrolyte has been proved to be advanta- Meanwhile, LixPFy and LixPOyFz are also observed, which originate
geous as it creates a compact and stable LiF-dominant SEI layer that from the decomposition of LiPF6 in the electrolyte.
limits the further formation of SEI [61–63]. Moreover, it is anticipated Comparing the XPS spectra after 100 cycles and after the 1st cycle,
that a SEI layer will form on the exposed surface of the nano-porous it shows that the composition of the SEI varies slightly. The amounts of
electrode layer of individual particles during the 1st Li ion uptake. LixSiOy, ROCOOLi, ionic C-F and LixPOyFz appear to increase, and LiF
Electron microscopy in Fig. 5a–f shows that, regardless of the initial and LixPFy turn out to be less. This phenomenon can be ascribed to a
growth of cluster size, the SEI covered Si NP clusters after 100 cycles couple of reasons: (1) Different reactions may take place over time due
appear to remain the same typical size as after the 1st cycle. The same to the presence of the initial SEI stratum; (2) The SEI layer thickens
is true for the micro-structure of the electrode. Apparently, after the slightly with fresh layers along cycling. Since XPS is surface sensitive,
rigid protective SEI is formed, it largely stays intact during later cycles the SEI formed in later cycles is more visible.
without apparent loss of coherence. There may be a little additional
growth, which is consistent with the EIS analysis in Fig. 4b. The 4. Conclusions
rationale for the cycling stability may be that now the volume change of
individual particles is accommodated by the space available between In conclusion, the binder/carbon-free Si NP anode synthesized via

483
Y. Xu et al. Nano Energy 38 (2017) 477–485

direct PECVD promises a particularly rapid and scalable synthesis [29] Z. Chen, C. Wang, J. Lopez, Z. Lu, Y. Cui, Z. Bao, Adv. Energy Mater. 5 (2015)
1614–6840.
method for Si anodes of Li ion batteries. The electrodes show high [30] I.H. Son, J. Hwan Park, S. Kwon, S. Park, M.H. Rummeli, A. Bachmatiuk,
specific and areal capacity as well as cycling stability. The Si NP anode H.J. Song, J. Ku, J.W. Choi, J.-m. Choi, S.-G. Doo, H. Chang, Nat. Commun. 6
shows sufficient electronic conductivity facilitated by the lithiation of Si (2015).
[31] D. Lin, Z. Lu, P.-C. Hsu, H.R. Lee, N. Liu, J. Zhao, H. Wang, C. Liu, Y. Cui, Energy
throughout the solid. Excellent cycling stability results from the in-situ Environ. Sci. 8 (2015) 2371–2376.
formation of locally protective SEI on the clusters of the individual [32] X. Li, M. Gu, S. Hu, R. Kennard, P. Yan, X. Chen, C. Wang, M.J. Sailor, J.-G. Zhang,
nanoparticles throughout the electrode; it apparently reinforces the J. Liu, Nat. Commun. 5 (2014).
[33] N. Liu, Z. Lu, J. Zhao, M.T. McDowell, H.-W. Lee, W. Zhao, Y. Cui, Nat.
cohesion of the clusters, the single particles as well as the entire Nanotechnol. 9 (2014) 187–192.
electrode, while the electronic conduction throughout the electrode is [34] Y. Li, K. Yan, H.-W. Lee, Z. Lu, N. Liu, Y. Cui, Nat. Energy 1 (2016) 15029.
maintained. Overall, this novel and binder/C-free Si anode exhibits [35] Q. Xu, J.Y. Li, J.K. Sun, Y.X. Yin, L.J. Wan, Y.G. Guo, Adv. Energy Mater. 6 (2016)
1601481.
unprecedented high areal capacity at high current rates and surprising
[36] M. Ko, S. Chae, J. Ma, N. Kim, H.-W. Lee, Y. Cui, J. Cho, Nat. Energy 1 (2016)
resistance to electrode degradation. Using modified but large through- 16113.
put solar photovoltaic production equipment the method also may [37] Y. Jin, S. Zhang, B. Zhu, Y. Tan, X. Hu, L. Zong, J. Zhu, Nano Lett. 15 (2015)
provide a significant promise for the commercial large scale production 7742–7747.
[38] L. Zong, B. Zhu, Z. Lu, Y. Tan, Y. Jin, N. Liu, Y. Hu, S. Gu, J. Zhu, Y. Cui, Proc. Natl.
of Si anodes for Li ion batteries. Acad. Sci. USA 112 (2015) 13473–13477.
[39] L. Zong, Y. Jin, C. Liu, B. Zhu, X. Hu, Z. Lu, J. Zhu, Nano Lett. 16 (2016)
Acknowledgement 7210–7215.
[40] G. Shoorideh, Y.S. Kim, Y.L. Joo, Electrochim. Acta 222 (2016) 946–955.
[41] X. Hu, Y. Jin, B. Zhu, Y. Tan, S. Zhang, L. Zong, Z. Lu, J. Zhu, ChemNanoMat 2
This research is financially supported by A green Deal in Energy (2016) 671–674.
Materials (ADEM) grant funded by Dutch Ministry of Economic Affairs [42] L.-F. Cui, L. Hu, J.W. Choi, Y. Cui, ACS Nano 4 (2010) 3671–3678.
[43] D.P. Wong, R. Suriyaprabha, R. Yuvakumar, V. Rajendran, Y.-T. Chen, B.-
and ADEM industrial partners. H. W. Zandbergen and S. Basak would J. Hwang, L.-C. Chen, K.-H. Chen, J. Mater. Chem. A 2 (2014) 13437–13441.
like to thank the European Research Council (ERC) project No. 267922 [44] D. Nan, Z.-H. Huang, R. Lv, Y. Lin, L. Yang, X. Yu, L. Ye, W. Shen, H. Sun, F. Kang,
for the financial support. J. Nanomater. 2014 (2014) 10.
[45] C. Li, C. Liu, W. Wang, J. Bell, Z. Mutlu, K. Ahmed, R. Ye, M. Ozkan, C.S. Ozkan,
Chem. Commun. 52 (2016) 11398–11401.
Appendix A. Supporting information [46] L. Fei, S.H. Yoo, R.A.R. Villamayor, B.P. Williams, S.Y. Gong, S. Park, K. Shin,
Y.L. Joo, ACS Appl. Mater. Interfaces 9 (2017) 9738–9746.
[47] J. Graetz, C.C. Ahn, R. Yazami, B. Fultz, Electrochem. Solid-State Lett. 6 (2003)
Supplementary data associated with this article can be found in the
A194–A197.
online version at doi:10.1016/j.nanoen.2017.06.011. [48] H. Wolf, Z. Pajkic, T. Gerdes, M. Willert-Porada, J. Power Sources 190 (2009)
157–161.
References [49] H. Liu, L. Hu, Y.S. Meng, Q. Li, Nanoscale 5 (2013) 10376–10383.
[50] E. Biserni, A. Scarpellini, A.L. Bassi, P. Bruno, Y. Zhou, M. Xie, Nanotechnology 27
(2016) 245401.
[1] S. Chu, A. Majumdar, Nature 488 (2012) 294–303. [51] F. Cheng, Z. Tao, J. Liang, J. Chen, Chem. Mater. 20 (2008) 667–245681.
[2] B. Dunn, H. Kamath, J.-M. Tarascon, Science 334 (2011) 928–935. [52] D.P. Singh, F.M. Mulder, A.M. Abdelkader, M. Wagemaker, Adv. Energy Mater. 3
[3] M. Armand, J.M. Tarascon, Nature 451 (2008) 652–657. (2013) 572–578.
[4] Y. Sun, N. Liu, Y. Cui, Nat. Energy (2016) 16071. [53] B. Hoex, A.J.M. van Erven, R.C.M. Bosch, W.T.M. Stals, M.D. Bijker, P.J. van den
[5] J.M. Tarascon, M. Armand, Nature 414 (2001) 359–367. Oever, W.M.M. Kessels, M.C.M. van de Sanden, Prog. Photovolt.: Res. Appl. 13
[6] M.N. Obrovac, L. Christensen, Electrochem. Solid-State Lett. 7 (2004) A93–A96. (2005) 705–712.
[7] D. Ma, Z. Cao, A. Hu, Nano-Micro Lett. 6 (2014) 347–358. [54] I. Dogan, N.J. Kramer, R.H.J. Westermann, K. Dohnalova, A.H.M. Smets,
[8] M. Ashuri, Q. He, L.L. Shaw, Nanoscale 8 (2016) 74–103. M.A. Verheijen, T. Gregorkiewicz, M.C.M. v.d. Sanden, J. Appl. Phys. 113 (2013)
[9] X. Zuo, J. Zhu, P. Müller-Buschbaum, Y.-J. Cheng, Nano Energy 31 (2017) 113–143. 134306.
[10] A.S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon, W. van Schalkwijk, Nat. Mater. 4 [55] W. Wan, Q. Zhang, Y. Cui, E. Wang, J. Phys. Condens. Matter 22 (2010) 415501.
(2005) 366–377. [56] M.T. McDowell, Y. Cui, Adv. Energy Mater. 1 (2011) 894–900.
[11] H. Wu, Y. Cui, Nano Today 7 (2012) 414–429. [57] E. Pollak, G. Salitra, V. Baranchugov, D. Aurbach, J. Phys. Chem. C. 111 (2007)
[12] X. Su, Q. Wu, J. Li, X. Xiao, A. Lott, W. Lu, B.W. Sheldon, J. Wu, Adv. Energy 11437–11444.
Mater. 4 (2014) 1300882. [58] R. Ruffo, S.S. Hong, C.K. Chan, R.A. Huggins, Y. Cui, J. Phys. Chem. C 113 (2009)
[13] C.K. Chan, H. Peng, G. Liu, K. McIlwrath, X.F. Zhang, R.A. Huggins, Y. Cui, Nat. 11390–11398.
Nanotechnol. 3 (2008) 31–35. [59] J. Guo, A. Sun, X. Chen, C. Wang, A. Manivannan, Electrochim. Acta 56 (2011)
[14] L.-F. Cui, R. Ruffo, C.K. Chan, H. Peng, Y. Cui, Nano Lett. 9 (2009) 491–495. 3981–3987.
[15] T. Song, J. Xia, J.-H. Lee, D.H. Lee, M.-S. Kwon, J.-M. Choi, J. Wu, S.K. Doo, [60] M.B. Pinson, M.Z. Bazant, J. Electrochem. Soc. 160 (2013) A243–A250.
H. Chang, W.I. Park, D.S. Zang, H. Kim, Y. Huang, K.-C. Hwang, J.A. Rogers, [61] V. Etacheri, O. Haik, Y. Goffer, G.A. Roberts, I.C. Stefan, R. Fasching, D. Aurbach,
U. Paik, Nano Lett. 10 (2010) 1710–1716. Langmuir 28 (2012) 965–976.
[16] H. Wu, G. Chan, J.W. Choi, I. Ryu, Y. Yao, M.T. McDowell, S.W. Lee, A. Jackson, [62] Y.-M. Lin, K.C. Klavetter, P.R. Abel, N.C. Davy, J.L. Snider, A. Heller, C.B. Mullins,
Y. Yang, L. Hu, Y. Cui, Nat. Nanotechnol. 7 (2012) 310–315. Chem. Commun. 48 (2012) 7268–7270.
[17] H. Ma, F. Cheng, J.Y. Chen, J.Z. Zhao, C.S. Li, Z.L. Tao, J. Liang, Adv. Mater. 19 [63] K. Schroder, J. Alvarado, T.A. Yersak, J. Li, N. Dudney, L.J. Webb, Y.S. Meng,
(2007) 4067–4070. K.J. Stevenson, Chem. Mater. 27 (2015) 5531–5542.
[18] B. Liu, P. Soares, C. Checkles, Y. Zhao, G. Yu, Nano Lett. 13 (2013) 3414–3419. [64] W.-S. Chang, C.-M. Park, J.-H. Kim, Y.-U. Kim, G. Jeong, H.-J. Sohn, Energy
[19] B. Zhu, N. Liu, M. McDowell, Y. Jin, Y. Cui, J. Zhu, Nano Energy 13 (2015) Environ. Sci. 5 (2012) 6895–6899.
620–625. [65] N. Yan, F. Wang, H. Zhong, Y. Li, Y. Wang, L. Hu, Q. Chen, Sci. Rep. 3 (2013) 1568.
[20] X. Zhou, Y.-X. Yin, L.-J. Wan, Y.-G. Guo, Adv. Energy Mater. 2 (2012) 1086–1090. [66] X. Hu, K. Zhang, L. Cong, F. Cheng, J. Chen, Chem. Commun. 51 (2015) 15827.
[21] K. Evanoff, A. Magasinski, J. Yang, G. Yushin, Adv. Energy Mater. 1 (2011)
495–498.
[22] L. Xue, G. Xu, Y. Li, S. Li, K. Fu, Q. Shi, X. Zhang, ACS Appl. Mater. Interfaces 5 Yaolin Xu received his B.Sc. in physics in Renmin
(2013) 21–25. University of China in 2011, and obtained his M.Sc. in
[23] S.-M. Jang, J. Miyawaki, M. Tsuji, I. Mochida, S.-H. Yoon, Carbon 47 (2009) Sustainable Energy of Technology at Delft University of
3383–3391. Technology in 2013. He is currently a Ph.D. student in
[24] G. Liu, S. Xun, N. Vukmirovic, X. Song, P. Olalde-Velasco, H. Zheng, V.S. Battaglia, Department of Chemical Engineering at Delft University of
L. Wang, W. Yang, Adv. Mater. 23 (2011) 4679–4683. Technology and his research is mainly focused on the
[25] C. Wang, H. Wu, Z. Chen, M.T. McDowell, Y. Cui, Z. Bao, Nat. Chem. 5 (2013) materials for Li- and Na-ion batteries.
1042–1048.
[26] Q. Xiao, Y. Fan, X. Wang, R.A. Susantyoko, Q. Zhang, Energy Environ. Sci. 7 (2014)
655–661.
[27] H. Wu, G. Yu, L. Pan, N. Liu, M.T. McDowell, Z. Bao, Y. Cui, Nat. Commun. 4
(2013).
[28] Y. Yao, M.T. McDowell, I. Ryu, H. Wu, N. Liu, L. Hu, W.D. Nix, Y. Cui, Nano Lett.
11 (2011) 2949–2954.

484
Y. Xu et al. Nano Energy 38 (2017) 477–485

Ellie Swaans received her M.Sc. degree in chemical Bo Peng received his B.Sc. degree in physics from Renmin
engineering from Technical University Eindhoven in University of China in 2011. He is currently a Ph.D. student
1998. She is currently senior process development engineer in Department of Physics at Renmin University of China.
at Meyer Burger Netherlands, where she is working on the His research interest mainly focuses on the anode materials
development of production processes regarding Silicon for Li- and Na-ion batteries.
based anodes for Li-ion batteries.

Sibo Chen received her Bachelor of Engineering (B.E.) in Prof. Dr. Henny W. Zandbergen is a professor in Kavli
environmental engineering in Nanjing Normal University Institute of Nanoscience Delft at Delft University of
in 2013. She accomplished her M.Sc. study in Sustainable Technology. He received his Ph.D. degree in solid-state
Energy Technology in Delft University of Technology in chemistry from Leiden University and had started working
2015 and is now working as a process engineering in Meyer in the field of TEM from 1984. His research interests are
Burger. Her research interests include plasma enhanced mainly directed towards the development of quantitative
chemical vapour deposition (PECVD) and Li-ion batteries. electron diffraction and in-situ TEM and his interest in
developing in-situ TEM has led to the fabrication of
dedicated sample holders.

Shibabrata Basak received his M.Tech. in solid state Dr. Dana M. Borsa is the Research Manager of the Thin
physics from IIT Kharagpur in 2010. He is currently a Film Coating and Projects Business Unit at Meyer Burger
Ph.D. student in Kavli Institute of Nanoscience at Delft (Netherlands) B.V. responsible for all process and applica-
University of Technology. His research interest mainly tion developments. Her Ph.D. in physics from the Rijks
focuses on understanding the fundamental of Li-ion bat- Universiteit Groningen was focused on the growth and
teries using electron microscopy including in-situ TEM characterization of nitride-based thin films and multi-
study. layers. After some years of research on switchable mirrors
and hydrogen storage materials at Vrije Universiteit
Amsterdam she moved to ECN and worked on high
efficiency solar cell concepts before joining OTB Solar
(currently Meyer Burger (Netherlands) B.V.) in 2009.

Peter-Paul R.M.L. Harks received his M.Sc. degree in Prof. Dr. Fokko M. Mulder is professor of materials for
physics from Utrecht University in the Netherlands. He is integrated energy systems in Faculty of Applied Sciences at
currently working as a Ph.D. student in Department of Delft University of Technology. During his Ph.D. in physics
Chemical Engineering at Delft University of Technology. In at Leiden University he published on electronic properties
his research he mainly focuses on high-energy density of metal nanoparticles and on permanent magnetic materi-
electrode materials for Li- and Na-ion batteries. als. Currently his research is focused on energy storage and
conversion materials including Li- and Na-ion batteries
and Hydrogen storage and production.

485