Electrochemical performance of thermally-grown SiO₂ as diffusion barrier layer for integrated lithium-ion batteries

Abstract

Direct integration of lithium-ion battery (LIB) with electronic devices on the same Si substrate can significantly miniaturize autonomous micro systems. For achieving direct integration, a barrier layer is essential to be inserted between LIB and the substrate for blocking Li⁺ diffusion. In this paper, the feasibility of thermal SiO₂ film as the barrier layer is investigated by electrochemical characterization and X-ray photoelectron spectroscopy (XPS). Due to the negligible side reactions of thermal SiO₂ with electrolyte, the solid electrolyte interphase (SEI) layer formed on the surface of the barrier layer is thin and the SEI content mainly consists of hydrocarbon together with slight polyethylene oxide (PEO), Li_xPO_yF_z, and Li₂CO₃. Although 8-nm thermal SiO₂ effectively prevents the substrate from alloying with Li⁺, the whole film changes to Li silicate after electrochemical cycling due to the irreversible chemical reactions of SiO₂ with electrolyte. This degrades the performance of the barrier layer against the electrolyte penetration, thus leading to the existence of Li⁺ (in the form of F-Si-Li) and solvent decompositions (with the products of hydrocarbon and PEO) near the barrier layer/substrate interface. Moreover, it is found that the reaction kinetics of thermal SiO₂ with electrolyte decrease significantly with increasing the SiO₂ thickness and no reactions are found in the bulk of the 30-nm SiO₂ film. Therefore, thermal SiO₂ with an appropriate thickness is a promising barrier layer for direct integration.

Keywords

autonomous micro system / direct integration / barrier layer / thermal SiO₂ film / X-ray photoelectron spectroscopy (XPS)

Introduction

Microelectronic industry is one of the most influential industries in the modern society. Recently, many efforts have been made to integrate non-electronic devices (e.g., sensors, transducers) with electronic devices for establishing a micro system with multi functionality as well as miniaturized dimensions [¹–⁶]. A typical micro system developed consists of a micro-electro-mechanical systems (MEMS) capacitive sensor and electronic circuits, both of which are fabricated on the same silicon substrate [⁷]. The whole volume of this micro system is only about 4 mm³. Such systems are quite suitable for many applications which have strict limit in geometric dimensions (e.g., biomedical implants, micro air vehicles, and wireless sensors). Moreover, integration of power sources with electronic devices on one substrate cannot only establish an autonomous system but also minimize the system size [³–⁶]. Among various power sources, micro lithium-ion battery (LIB) is a promising candidate as power supply mainly because of its low self-discharge rate and high energy densities in both gravimetric and volumetric terms. Therefore, it is worthwhile investigating LIB directly integrated with electronic devices.

On the other hand, it is known that Li⁺ is easy to electrochemically react with Si substrate (a common housing material in the microelectronic industry) and the large stress induced by alloying and dealloying reactions of Si with Li⁺ would even lead to cracks of the substrate [⁶,⁸], thus destroying electronic devices built on the substrate. Also, mobile ionic charges (e.g. Na⁺, Li⁺) have been regarded as killers of the performance of electronic devices. Therefore, in order to directly integrate LIB with electronic devices, it is essential to insert a barrier layer between LIB and the substrate for blocking Li⁺ diffusion. Many materials (e.g., Ti, Ta, TaN, and TiN) have been investigated as the barrier layer [³–⁶]. Among them, TiN deposited by atomic layer deposition (ALD) technique has been regarded as a promising candidate (0.02 Li involved per TiN formula unit) [⁴,⁵]. However, because even traces of Li⁺ would severely degrade the performance of electronic devices, the ability of the ALD TiN against Li⁺ diffusion needs to be further improved. In addition, to achieve direct integration of LIB with electronic devices on the same substrate, the substrate needs to electrically isolate from LIB; otherwise, the voltage of the substrate would change as the voltage of LIB changes during the charging/discharging process and the voltage fluctuation of the substrate would lead to serious interferences with the electronic devices on the substrate. Therefore, the barrier layer should be used not only for blocking Li⁺ diffusion but also for electrically isolating LIB from the substrate.

SiO₂ grown by thermal oxidation has a high bulk quality as well as a good interface with the Si substrate and thus is widely used in the microelectronics, e.g., as agate insulator in transistors and electrical isolation between electronic devices [⁹]. Thermal SiO₂ also has a very dense structure which is desirable to block ion diffusion. Furthermore, thermal SiO₂ can achieve conformal coverage even in high-aspect-ratio substrates [¹⁰], suggesting its great potential in three-dimension (3D) type batteries. Unfortunately, even though thermal SiO₂ has received increasing interests in LIBs (e.g., as surface coating to suppress the volume expansion of nano-structured Si anodes) [¹¹,¹²], little attention has been paid to thermal SiO₂ as the barrier layer. Moreover, it has been reported that SiO₂ in other morphologies (e.g., nano particles) or fabricated by other techniques (e.g., physical vapor deposition) can work as an active anode material for LIB [¹³–¹⁶]. Therefore, the feasibility of thermal SiO₂ as the barrier layer needs to be carefully investigated.

In this paper, the electrochemical behaviors of thermal SiO₂ film are carefully investigated for checking the feasibility of thermal SiO₂ as the barrier layer. The chemical states of the films before and after electrochemical characterization are also studied by X-ray photoelectron spectroscopy (XPS) to gain more insight into the electrochemical behaviors of the films. Considering that solid electrolyte interphase (SEI) layer formed on the film surface has a great influence on the electrochemical performance and the bonding states near the SiO₂ and substrate interface can provide direct evidence on ion diffusion, special attention will be paid to the bonding states near the SiO₂ surface and the SiO₂/substrate interface. The influences of the SiO₂ thickness on the electrochemical performance are investigated, too, for achieving optimum effectiveness of thermal SiO₂ as the barrier layer.

Experimental

Sample preparation

n-Type heavily doped Si wafers (11 × 11 mm²) with a resistivity of 0.001–0.01 W·cm were used as the substrates. After the standard RCA (Radio Corporation of America) cleaning, the SiO₂ with thicknesses from 8 nm to 30 nm was grown on the wafers by dry thermal oxidation at 900^oC by using a diffusion furnace (HDC-8000A) and the flow rate of pure O₂ gas with a purity of approximately 99.999% is adjusted to be 1000 sccm/min by using a flow controller. The films display an amorphous state as demonstrated by X-ray diffraction (XRD, not shown). Then, the SiO₂ at the backside of the substrate was removed by using HF (hydrofluoric acid) wet etching. After that, a current collector consisting of 10-nm Cr/50-nm Cu was deposited on the backside of the substrate by sputtering using meal Cr and Cu targets.

Characterization

To simulate the LIB environment, the Si wafers were assembled into CR2016-type coin cells with Celgard 2400 as the separator, 1 M LiPF₆ dissolved in ethylene carbonate/dimethyl carbonate (v/v= 1:1) as the electrolyte and Li foil as the counter electrode. Cyclic voltammetry (CV) was recorded at a scan rate of 0.1 mV/s between 0 V and 3 V versus Li/Li⁺ (CHI 660E). Galvanostatic cycling was conducted at a current density of 10 mA/cm² between 0.01 V and 3 V versus Li/Li⁺ (Land CT2001A). After 500 electrochemical cycles, the cells were discharged to 0.01 V. This corresponds to the process that Li⁺ intercalates toward the Si substrate, and thus is quite suitable for examining the ability of the thermal SiO₂ against the Li⁺ diffusion. After that, the wafers were took out from the cells and washed in acetonitrile to remove residual electrolyte and then dried under an Ar ambient for physical characterization.

The thickness of the SiO₂ films was determined by spectroscopic ellipsometry (HorribaUVISEL). During the ellipsometry characterization, the angle of incidence is fixed at 70^o and the wavelength of incidence increases from 210 nm to 880 nm with a step of 10 nm. The chemical states of the films were characterized by using XPS (PHI Quantera II) with Al K_a monochromatic X-ray source, and the XPS penetration depth is about 10 nm. XPS depth profiling was performed by using Ar⁺ beam sputtering and the etching rate of the barrier layer is about 3.3 nm/min. The XPS spectra were calibrated by using C1s (~285.3 eV) in hydrocarbon and Si 2p (~99.8 eV) in monocrystalline Si substrate. The surface morphology of the films was measured by using scanning electron microscopy (SEM, Zeiss Ultra Plus).

Results and discussion

Electrochemical results

Figure 1(a) shows the CV curves of the sample with and without SiO₂ growth. For the sample without SiO₂, two cathodic peaks at around 0.01 – 0.1 eV and two anodic peaks at around 0.3 – 0.5 eV are clearly presented. These peaks are consistent with the literatures and correspond to Li⁺ alloying and dealloying with Si, respectively [¹⁷–²⁰]. For comparison, these peaks are not found in the sample with SiO₂, suggesting that the 8-nm SiO₂ film is effective to prevent the underlying Si from participating in the reactions. In addition, for the sample with SiO₂, only one broad and weak cathodic peak appears at 1.8 V in the 1st CV scanning and this peak is ascribed to the SEI formation due to reactions of SiO₂ with electrolyte [¹⁴]. This peak becomes negligible in the following cycles, suggesting that the reactions are irreversible and mainly happen in the 1st scanning. The redox area for the CV loop is normally proportional to the energy-storage capacity; therefore, the much smaller area for the sample with SiO₂ than that without SiO₂ indicates much lower reactivity (and thus much higher stability) of the SiO₂film than the Si electrode. It is also found that the redox area decreases as the SiO₂ thickness increases from 8 nm to 30 nm, suggesting that the electrochemical reactivity can be further suppressed by increasing the SiO₂ thickness.

Figure 1(b) demonstrates the capacities of the samples with various SiO₂ thicknesses in the 1st galvanostatic discharging/charging cycle. The discharging/charging capacities in the 1st cycle are about 1.1 mAh/cm²/0.5 mAh/cm² for the sample with 8-nm SiO₂, 0.5 mAh/cm²/0.2 mAh/cm² for the sample with 15-nm SiO₂, and 0.1 mAh/cm²/0.0 mAh/cm² for the sample with 30-nm SiO₂, respectively. The high charge loss of each sample is caused by the irreversible reactions of SiO₂ with electrolyte [¹⁴]. The capacity decreases with increasing the SiO₂ thickness, indicating that the energy storage is not a bulk effect but is mainly determined by the electrochemical reactions near the film surface [⁴]. The resistance of the barrier layer increases with increasing the film thickness, thus reducing the reaction kinetics and the capacities. Figure 1(c) depicts the galvanostatic cycling capacities of the samples with various SiO₂ thicknesses. Both the charging and discharging capacities degrade rapidly with cycling, suggesting that the irreversible electrochemical reactions tend to end in the following cycles. This further confirms that the SiO₂ film can effectively protect the underlying Si substrate from alloying and dealloying with Li⁺ even under repeated cycling. It has also been reported that the SiO_x film [¹³] or nanostructured SiO_x (x≤ 2) [¹⁴–¹⁶] can be used as the anode electrodes for LIBs. Compared with those in Refs. [¹³–¹⁶], the much lower capacities in this paper should be mainly caused by the high quality of the SiO₂ film grown by dry thermal oxidation. Moreover, the morphology of the materials has a great effect on its electrochemical reactivity and the nanostructures have a higher surface area and a higher reactivity than the films [²⁰], thus also contributing to the much higher capacities in Refs. [¹³–¹⁶].

XPS results

Figure 2 shows the XPS Si 2p spectrum of the sample with 8-nm SiO₂ before and after the galvanostatic cycling. The Si 2p spectrum before cycling includes an intense peak at 103.1 eV (from the SiO₂ film) and a weak peak at 99.8 eV (from the Si substrate). After cycling, the main peak (~102.8 eV) and the weak peak (~100.2 eV) make a negative and positive shift, respectively. The negative shift of the main peak should be caused by the formation of silicate resulting from the irreversible reactions of SiO₂ with electrolyte [¹³,¹⁴]. The positive shift of the weak peak indicates that the Si-Si bonds near the surface of the substrate become oxidation states after cycling. This weak peak agrees well with Si-F-Li bonds [²¹] whose details will be discussed later. It is worth pointing out that the Si 2p peak from the substrate can be also observed even after cycling, suggesting slight formation of the SEI layer; therefore, the total thickness of the SEI layer and the barrier layer is smaller than the XPS penetration depth (~10 nm).

SEI layer on the surface of the barrier layer

To gain more insight into the SEI layer, the XPS P 2p, F 1s and C 1s spectra near the surface are investigated for the sample with 8-nm SiO₂. Figure 3(a) is the P 2p spectrum, which displays two peaks at 134.9 eV and 137.9 eV,corresponding to Li_xPO_yF_z and residual LiPF₆, respectively [²²]. Li_xPO_yF_z is a reduction product of the LiPF₆ electrolyte and the formation of Li_xPO_yF_z can be further confirmed by the F 1s spectrum shown in Fig. 3(b). For the F 1s spectrum, in addition to two weak Li_xPO_yF_z (~687.6eV) and LiPF₆ (~688.8 eV) peaks, a strong peak at 685.6 eV is also observed, suggesting F bonded to the silicate after cycling [²²]. The C 1s spectrum in Fig. 3(c) exhibits a strong peak at 285.3 eV together with a shoulder at 286.8 eV and a weak peak at 290.3 eV. These three species result from the decomposition of the carbonate-based solvent and can be assigned to hydrocarbon, polyethylene oxide (PEO), and Li₂CO₃, respectively [²²]. Therefore, according to the above analysis, the SEI layer mainly consists of hydrocarbon combined with small amounts of PEO, Li_xPO_yF_z and Li₂CO₃ species. Moreover, as shown in Figs. 3(a) – 3(c), the species in the SEI layer decrease rapidly upon sputtering, indicating that the SEI layer is thin, which is consistent with the observation from the Si spectrum in Fig. 2. Compared with the SEI layer formed by the Si film [²⁰,²¹], the compositions of the SEI layer in this paper are much simpler and the thickness (<3.3 nm) is much smaller, too. This suggests that the thermal SiO₂ is effective to suppress the side reactions of the Si substrate with electrolyte.

XPS depth profiling

The depth profiling of each element along the barrier layer is also characterized by XPS sputtering. As shown in Fig. 4, the F, O, Si, and Li contents decrease with increasing sputtering time, suggesting that these elements distribute in the whole barrier layer. For comparison, the C content decreases rapidly after the 1st sputtering (due to removal of the SEI layer) and then remain almost unchanged with further sputtering, indicating that the C content is negligible in the bulk of the barrier layer and the remaining C after the 1st sputtering is mainly located near the barrier layer/Si interface. The strength of the Si-O bonds for the thermal SiO₂ is quite strong, and thus only active elements (e.g., Li and F) can react with the SiO₂ film.

Bonding states near barrier layer/substrate interface

To further examine the feasibility of thermal SiO₂ film as the barrier layer, the chemical states near the SiO₂/Si interface are also investigated for the sample with 8-nm SiO₂ after cycling. Figure 5(a) shows the Si 2p spectrum near the interface, which consists of a weak peak (from the barrier layer) and a strong peak (from the substrate). The position of the peak corresponding to the substrate decreases from 100.0 eV to 99.8 eV with increasing sputtering time. The peak at 100.0 eV suggests the formation of F-Si-Li bonds near the interface after cycling [²¹], while the peak at 99.8 eV agrees well with the monocrystalline Si, indicating that the cycling has no effect on the Si-Si bonds in the bulk of the substrate. In addition, no Li_xSi species (~98.0 eV [¹³,²⁰]) are found in the Si 2p spectrum, indicating that 8-nm SiO₂ is effective to prevent the Si substrate from alloying with Li⁺. Moreover, the weak peak corresponding to the barrier layer can still be observed even with further increasing the sputtering time (as shown in the inset of Fig. 5(a)), and this should result from increased roughness of the interface after cycling. It is known that the SiO₂ film grown by thermal oxidation is very dense and has a high bulk quality and an abrupt interface with the Si substrate. However, the electrochemical reactions of SiO₂ with electrolyte under cycling make the bonds disorder and the SiO₂ film becomes less dense [¹⁵], thus facilitating the electrolyte and solvent penetration through the barrier layer. This can be evidenced by the C 1s spectrum near the interface in Fig. 5(b), which shows that two decomposition products (PEO and hydrocarbon) of the solvent concentrate near the interface. The Li silicate has a poor ion conductivity [¹²,¹⁵] which is helpful to prevent the Si substrate from alloying with Li⁺. Moreover, the formation of F-Si-Li and the coverage of hydrocarbon and PEO near the interface can further block the substrate from reacting with the electrolyte. Consequently, the 8-nm SiO₂ film is effective to protect the substrate against Li⁺ alloying. However, because electronic devices are quite sensitive to mobile ionic charges (e.g. Na⁺, Li⁺) [³], even traces of Li⁺(in the form of F-Si-Li in this paper) near the interface is harmful to the performance (e.g., deteriorating leakage and working speeds) of the electronic devices which are directly integrated with LIB. Besides, the rough interface induced by the degradation of the barrier layer would reduce the carrier mobility by enhancing carrier scattering [⁹], thus further degrading the performance of the electronic devices. Therefore, considering the above reasons, the 8-nm thermal SiO₂ is not competent to act as the barrier layer for direct integration of LIB with electronic devices.

Effects of SiO₂ thickness on its performance as the barrier layer

As discussed in Fig. 1, the electrochemical reaction kinetics (and thus the capacities) decrease rapidly with increasing the SiO₂ thickness; therefore, it should be effective to improve the feasibility of thermal SiO₂ as the barrier layer by increasing the thickness. For verifying this viewpoint, the XPS F 1s and C 1s spectra for the sample with 30-nm SiO₂ are also studied after cycling, as shown in Figs. 6(a) and 6(b). Compared with the spectra in Fig. 3, both the samples have similar spectrum profiles, indicating similar products formed on the surface after cycling. However, the peak intensities for the sample with 30-nm SiO₂ are much weaker than the sample with 8-nm SiO₂, suggesting suppressed side reactions with increasing the SiO₂ thickness. This is consistent with the conclusions made from Fig. 1. Moreover, after 3.6-min sputtering, only Si and O elements are found for the sample with 30-nm SiO₂, which is different from the sample with 8-nm SiO₂ (as seen in Fig. 4). It is further found that for the sample with 30-nm SiO₂, the Si 2p spectrum in the bulk film after cycling matches well with the spectrum before cycling (as seen in Fig. 6(c)).The above phenomena suggest that SiO₂ in the bulk is intact and has no reactions with electrolyte even after cycling, demonstrating that the 30-nm thermal SiO₂ film is effective to block Li⁺ diffusion. Moreover, the intact SiO₂ is a good insulator to electrically isolate LIB from the substrate. Therefore, the 30-nm thermal SiO₂ film is competent to act as a barrier layer for direct integration of LIB with electronic devices.

Film morphology

Figure 7(a) shows the SEM image of the sample without SiO₂ after the 1st discharging, where obvious cracks are observed because of large stress induced by Si alloying with Li⁺ [²⁰,²¹]. For comparison, as shown in Fig. 7(b), the surface of the sample with 8-nm SiO₂ is intact even after cycling, further suggesting that the SiO₂ film is effective to block the substrate from alloying with Li⁺. However, considerable white spots are observed on its surface (as seen in the inset of Fig. 7(b)). These white spots can be assigned to SEI formed on the surface and can be further confirmed by EDX (energy dispersive X-ray) spectrum shown in Fig. 7(c), where the presence of C, O, F, and P in the white-spot region agrees well with the contents of SEI as discussed above [¹⁵]. As shown in Fig. 7(d), the density of the white spots reduces significantly for the sample with 30-nm SiO₂ mainly due to suppressed side reactions with increasing the SiO₂ thickness. This is consistent with the conclusions made from the XPS analysis as mentioned earlier. The sample with 30-nm SiO₂ exhibits an intact and smooth surface, suggesting that the thermal SiO₂ can act as a good insulator to electrically isolate LIB from the substrate even after cycling. To investigate the morphology of the Si substrate for the sample with 30-nm SiO₂ after cycling, the SiO₂ on the surface is removed by using HF wet etching. As shown in Fig. 7(e), the Si substrate displays a smooth and uniform surface, further demonstrating the effectiveness of thermal SiO₂ as the barrier layer for blocking Li⁺ diffusion. This is consistent with the observations from the XPS analysis as seen in Fig. 6(c).

Conclusions

The electrochemical behaviors of thermal SiO₂ film as the barrier layer are investigated by electrochemical characterization combined with XPS and SEM analysis. As evidenced by thin thickness and simple compositions of the SEI layer, the thermal SiO₂ film is effective to suppress side reactions compared with the Si electrodes. 8-nm SiO₂ film can effectively block the Si substrate from alloying with Li⁺, but the whole film is transformed to Li silicate after galvanostatic cycling due to the irreversible chemical reactions. This makes the barrier layer less dense, thus facilitating penetration of the electrolyte and solvent through the barrier layer into the substrate. The F-Si-Li and solvent decompositions near the barrier layer/Si interface are undesirable for integration of electronic devices with LIB. Due to reduced reaction kinetics, the reactions of SiO₂ with electrolyte are significantly suppressed with increasing the SiO₂ thickness and no electrochemical reactions are found in the bulk of 30-nm SiO₂ film. Therefore, 30-nm thermal SiO₂ film is effective to act as the barrier layer for direct integration of LIB with electronic devices.

References

[1] Wang Y, Liu B, Li Q, Cartmell S, Ferrara S, Deng Z D, Xiao J. Lithium and lithium ion batteries for applications in microelectronic devices: a review. Journal of Power Sources, 2015, 286(14): 330–345

[2] Liu J, Banis M N, Li X, Lushington A, Cai M, Li R, Sham T K, Sun X. Atomic layer deposition of lithium tantalate solid-state electrolytes. Journal of Physical Chemistry C, 2013, 117(39): 20260–20267

[3] Janski R, Fugger M, Sternad M, Wilkening M. Lithium distribution in monocrystalline silicon-based lithium-ion batteries. ECS Transactions, 2014, 62(1): 247–253

[4] Baggetto L, Oudenhoven J F M, Van Dongen T, Klootwijk J H, Mulder M, Niessen R A H, de Croon M H J M, Notten P H L. On the electrochemistry of an anode stack for all-solid-state 3D-integrated batteries. Journal of Power Sources, 2009, 189(1): 402–410

[5] Knoops H C M, Baggetto L, Langereis E, van de Sanden M C M, Klootwijk J H, Roozeboom F, Niessen R A H, Notten P H L, Kessels W M M. Deposition of TiN and TaN by remote plasma ALD for Cu and Li diffusion barrier applications. Journal of the Electrochemical Society, 2008, 155(12): G287–G294

[6] Baggetto L, Niessen R A H, Roozeboom F, Notten P H L. High energy density all-solid-state batteries: a challenging concept towards 3D integration. Advanced Functional Materials, 2008, 18(7): 1057–1066

[7] Huang X D, Huang J Q, Qin M, Huang Q A. A fully integrated capacitive pressure sensor with high sensitivity. In: 6th IEEE Conference on SENSORS (IEEE SENSORS 2007), Atlanta, United States, 2007, 1052–1055

[8] Chan C K, Peng H, Liu G, McIlwrath K, Zhang X F, Huggins R A, Cui Y. High-performance lithium battery anodes using silicon nanowires. Nature Nanotechnology, 2008, 3(1): 31–35

[9] Streetman B G, Banerjee S K. Solid State Electronic Devices, 6th ed.New Jersey: Prentice Hall, 2007

[10] Roozeboom F, Elfrink R, Verhoeven J, Van den Meerakher J, Holthuysen F. High-value MOS capacitor arrays in ultradeep trenches in silicon. Microelectronic Engineering, 2000, 53(1–4): 581–584

[11] Wu H, Chan G, Choi J W, Ryu I, Yao Y, McDowell M T, Lee S W, Jackson A, Yang Y, Hu L, Cui Y. Stable cycling of double-walled silicon nanotube battery anodes through solid-electrolyte interphase control. Nature Nanotechnology, 2012, 7(5): 310–315

[12] Sim S, Oh P, Park S, Cho J. Critical thickness of SiO₂ coating layer on core@shell bulk@nanowire Si anode materials for Li-ion batteries. Advanced Materials, 2013, 25(32): 4498–4503

[13] Chan C K, Peng H, Liu G, McIlwrath K, Zhang X F, Huggins R A, Cui Y. High-performance lithium battery anodes using silicon nanowires. Nature Nanotechnology, 2008, 3(1): 31–35

[14] Liu B, Soares P, Checkles C, Zhao Y, Yu G. Three-dimensional hierarchical ternary nanostructures for high-performance Li-ion battery anodes. Nano Letters, 2013, 13(7): 3414–3419

[15] Luais E, Ghamouss F, Wolfman J, Desplobain S, Gautier G, Tran-Van F, Sakai J. Mesoporous silicon negative electrode for thin film lithium-ion microbatteries. Journal of Power Sources, 2015, 274: 693–700

[16] Takezawa H, Iwamoto K, Ito S, Yoshizawa H. Electrochemical behaviors of nonstoichiometric silicon suboxides (SiO_x) film prepared by reactive evaporation for lithium rechargeable batteries. Journal of Power Sources, 2013, 244(6): 149–157

[17] Tu J, Yuan Y, Zhan P, Jiao H, Wang X, Zhu H, Jiao S. Straightforward approach toward SiO₂ nanospheres and their superior lithium storage performance. Journal of Physical Chemistry C, 2014, 118(14): 7357–7362

[18] Zhang Y, Li Y, Wang Z, Zhao K. Lithiation of SiO₂ in Li-ion batteries: in situ transmission electron microscopy experiments and theoretical studies. Nano Letters, 2014, 14(12): 7161–7170

[19] Miyachi M, Yamamoto H, Kawai H, Ohta T, Shirakata M. Analysis of SiO anodes for lithium-ion batteries. Journal of the Electrochemical Society, 2005, 152(10): A2089–A2091

[20] Chan C K, Ruffo R, Hong S S, Cui Y. Surface chemistry and morphology of the solid electrolyte interphase on silicon nanowire lithium-ion battery anodes. Journal of Power Sources, 2009, 189(2): 1132–1140

[21] Arreaga-Salas D E, Sra A K, Roodenko K, Chabal Y J, Hinkle C L. Progression of solid electrolyte interphase formation on hydrogenated amorphous silicon anodes for lithium-ion batteries. Journal of Physical Chemistry C, 2012, 116(16): 9072–9077

[22] Ensling D, Stjerndahl M, Nytén A, Gustafsson T, Thomas J O. A comparative XPS surface study of Li₂FeSiO₄/C cycled with LiTFSI-and LiPF₆-based electrolytes. Journal of Materials Chemistry, 2009, 19(1): 82–88

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature