Systematic Investigation of Prelithiated SiO2 Particles for High-Performance Anodes in Lithium-Ion Battery

Prelithiation is an important strategy used to compensate for lithium loss during the formation of a solid electrolyte interface (SEI) layer and the other irreversible reactions at the first stage of electrochemical cycling. In this paper, we report a systematic study of thermal prelithiation of SiO2 particles with different sizes (6 nm, 20 nm, 300 nm and 3 μm). All four lithiated anodes (LixSi/Li2O composites) show improved performance over pristine SiO2. More interestingly, lithiated product from micron-sized SiO2 particle demonstrates optimum performance with a charge capacity of 1859 mAhg−1 initially and maintains above 1300 mAhg−1 for over 50 cycles.


Introduction
Next generation high-capacity electrode materials are needed to meet the demands of the explosion of electricity-fueled forms of transportation [1][2][3]. Graphite, the commercial anode material, cannot completely satisfy the requirements of electric vehicles due to its relatively low theoretical specific capacity. To overcome this problem, numerous researchers have focused on silicon-based anode materials because of its high theoretical specific capacity and relatively low charge/discharge voltage. Among them, silica (SiO 2 ), the main institutions of sand and quartz, is generally considered to be a promising candidate [4][5][6][7]. Silica is easier to prepare and is much more economical than silicon, while also stores a large quantity of lithium and has a relatively low potential platform. However, bulk-SiO 2 is generally conformed to be lithium inactive due to instinctive insulation [8][9][10][11]. Therefore, considerable efforts have been devoted to various nanostructures, such as hollow nano-spheres [12], nano-films [8,11], nano-cubes [13], nanotubes [14] and nano-belts [15].
There is no denying that nano-structures can facilitate the diffusion of electron and lithium ion effectively. However, nano-structuring also has its limitations. Its higher specific surface area consumes the more electrolyte to form solid electrolyte interface (SEI), which results in lower Coulombic efficiency (CE) and capacity fading [16]. In addition, nano-structured materials usually have a low tap density, leading to a low volumetric capacity and a thick electrode at high mass loading [16,17]. As a result, maintaining the electrical and ionic pathways during cycling is difficult. Furthermore, nano-structuring usually requires a multi-step and advanced preparation process, leading to a higher cost. More seriously, silica can form a great deal of irreversible lithium silicates and Li 2 O during the first lithiation cycle. This process consumes an excess amount of cathode materials [6][7][8]13,18,19], which significantly reduces the total energy density of full-cells, thereby preventing their practical applications. In order to tackle this issue, researchers have proposed prelithiation of the electrode material, which directly compensates for the irreversible loss of lithium during the first cycle [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37]. For example, Tarascon's

Synthesis of Li x Si/Li 2 O Composites
SiO 2 powders (about 6 nm, 20 nm, 300 nm, and 3 µm) were first dried under vacuum for 24 h to remove trapped water. Typically, the variously sized SiO 2 particles (200 mg) were mixed with Li metal (196 mg) in a tantalum crucible. Then, the mixtures were heated at 200 • C while they were mechanically stirred inside the tantalum crucible at 400 rpm for at least 6 h in a glove box (Ar atmosphere, O 2 level < 1.2 ppm, and H 2 O level < 0.1 ppm, respectively). The powder turned to black during the reaction. The products were denoted as Li x Si/Li 2 O-1, Li x Si/Li 2 O-2, Li x Si/Li 2 O-3 and Li x Si/Li 2 O-4, corresponding to the different pristine SiO 2 particle sizes (6 nm, 20 nm, 300 nm, and 3 µm).

Morphological and Structural Characterizations
The crystal structures of the products were identified by X-ray diffractometer (XRD, Ultima III, Rigaku, Tokyo, Japan) using Kα radiation (40 Kv, 40 mA). The structure and morphology details of the products were observed by field-emission scanning electron microscopy (FESEM, Ultra 55, Zeiss, Germany) and transmission electron microscopy (TEM, Tecnai G 2 F20 X-TWIN, FEI). The particle size and size distributions were examined by Dynamic Light Scattering (DLS, Zetasizer Nano-ZS90, Malvern, UK).

Electrochemical Measurements
The electrochemical properties were carried out using LIR 2032-type coin cells. The active material was mixed with binder acetylene black and carboxymethyl cellulose (CMC) (weight ratio of 65:20:15) in tetrahydrofuran (THF) to form a slurry. Then, the slurry was uniformly spin-coated on copper foils. The mass loading of the slurry on each copper foil (12 mm) is about 0.3-3.5 mg. The electrodes were assembled in an Ar-filled glove box using lithium foil as the counter electrode and a Celgard 2300 film as the separator. An electrolyte solution consisting of 1 M LiPF6 dissolved in ethylene (EC)/dimethyl carbonate (DMC) (1:1 by volume) was used to assemble the coin cells. Galvanostatic charge-discharge measurements were tested between 0.01 and 1.0 V at a rate of 0.05 C on the cell tester (LAND CT2001A, Wuhan, China). Cyclic voltammetry (CV) was conducted on an electrochemical workstation (Bio-logic VMP3) at a scan rate of 0.1 mVs −1 in a potential range from 0.01 to 3.0 V. Electrochemical impedance spectroscopy (EIS) measurements were carried out on the same electrochemical workstation with frequencies ranging from 100 kHz to 10 MHz, with an alternating voltage of 5 mV. All electrochemical measurements were carried out at room temperature.

Results and Discussion
To systematically study the effect of particle size, we selected different sizes of SiO 2 particles. As shown in Figure 1, the morphology and size distributions of particles were characterized by TEM, SEM and DLS measurement. Figure 1a-d show the morphology of pristine SiO 2 particles with different sizes from 6 nm to 3 µm. The TEM (Figure 1a,b) and SEM (Figure 1c,d) images clearly demonstrate all the silica particles have the spherical structures and uniform size. The size statistics by DLS further confirm that SiO 2 particles have a quite narrow size distribution (Figure 1e,h).

Results and Discussion
To systematically study the effect of particle size, we selected different sizes of SiO2 particles. As shown in Figure 1, the morphology and size distributions of particles were characterized by TEM, SEM and DLS measurement. Figure 1a-d show the morphology of pristine SiO2 particles with different sizes from 6 nm to 3 μm. The TEM (Figure 1a  The SiO2 particles with different sizes were alloyed with molten Li metal under Ar atmosphere to form LixSi/Li2O NPs, as schemed in Figure 2a. The powder color changes from white to black within ten minutes, indicating the formation of LixSi alloy. SEM and TEM images (Figure 2b-e) demonstrate the morphology of as-prepared LixSi/Li2O composites. After prelithiation, the surface of SiO2 particles becomes rougher due to the coverage of oxide layer and volume expansion. Moreover, the images show that they have similar morphology despite being derived from variously sized SiO2 particles. It is found that although the pristine SiO2 particles varied from 6 nm to 3 μm, the over-all sizes of aslithiated samples are all around several micrometers, which can be clearly seen from the enlarged SEM images in the insets. This can be attributed to the smaller particles undergo more sever aggregation during SiO2 lithiation. As a result, Nanoparticles trend to aggregate into microparticles, while micro-or submicro-ones prefer to retain their size in the lithiation process. Figure 3a shows the XRD patterns of prelithiated SiO2 particles. A home-made sealed sample holder was used to avoid exposing the products to the air. The diffraction peaks of Li2O and Li21Si5 are shown in the Figure 3a, demonstrating the formation of Li2O (PDF# 04-001-893) and Li21Si5 (PDF# 00-018-747) phases. It has been reported that Li21Si5 is the most thermally stable phase among the crystalline lithium silicates [21]. The XRD patterns of the four LixSi/Li2O composites are similar, indicating that pristine particle size does not have a big effect on the formation of crystalline Li21Si5. As calculated by the Scherrer After prelithiation, the surface of SiO 2 particles becomes rougher due to the coverage of oxide layer and volume expansion. Moreover, the images show that they have similar morphology despite being derived from variously sized SiO 2 particles. It is found that although the pristine SiO 2 particles varied from 6 nm to 3 µm, the over-all sizes of as-lithiated samples are all around several micrometers, which can be clearly seen from the enlarged SEM images in the insets. This can be attributed to the smaller particles undergo more sever aggregation during SiO 2 lithiation. As a result, Nanoparticles trend to aggregate into microparticles, while micro-or submicro-ones prefer to retain their size in the lithiation process. Figure 3a shows the XRD patterns of prelithiated SiO 2 particles. A home-made sealed sample holder was used to avoid exposing the products to the air. The diffraction peaks of Li 2 O and Li 21 Si 5 are shown in the Figure 3a, demonstrating the formation of Li 2 O (PDF# 04-001-893) and Li 21 Si 5 (PDF# 00-018-747) phases. It has been reported that Li 21 Si 5 is the most thermally stable phase among the crystalline lithium silicates [21]. The XRD patterns of the four Li x Si/Li 2 O composites are similar, indicating that pristine particle size does not have a big effect on the formation of crystalline Li 21 Si 5 . As calculated by the Scherrer equation, the crystalline sizes of the four composites are 28.2, 29.3, 31.9 and 37.5 nm, indicating that the SiO 2 microparticles produce relatively large Li 21 Si 5 domains.   To invest the electrochemical behavior of the prelithiated products, half cells were fabricated with Li metal as a counter electrode. Figure 3b shows the charging/discharging profiles of LixSi/Li2O   To invest the electrochemical behavior of the prelithiated products, half cells were fabricated with Li metal as a counter electrode. Figure 3b shows the charging/discharging profiles of LixSi/Li2O To invest the electrochemical behavior of the prelithiated products, half cells were fabricated with Li metal as a counter electrode. Figure 3b shows the charging/discharging profiles of Li x Si/Li 2 O electrodes. The open-circuit voltage (OCV) of the half cells is around 0.35 V, which is significantly lower than~2.0 V of the SiO 2 /Li cell (SI, Figure S1). This proves that most SiO 2 particles were well lithiated. All of the lithiated SiO 2 electrodes exhibit the similar potential plateaus around 0.45 V, indicating the same active materials (Li x Si) of the electrodes and their high crystallinity, consistent with the XRD results. The prelithiation capacities of the SiO 2 particles are 1972, 2267, 1776, and 1412 mAhg −1 , which are listed in Table 1. Here, the prelithiation capacity is determined by subtracting the lithiation capacity from the delithiation capacity at the first cycle. The thermal lithiation can effectively compensate for the huge irreversible lithium loss during the first cycles via the formation of lithium silicates and Li 2 O. The cyclic voltammetry (CV) tests are conducted for both the pristine and prelithiated SiO 2 half cells for 3 consecutive cycles between 3.0 V and 0.01 V at a scan rate of 0.05 mVs −1 (Figure 4). The pristine SiO 2 electrode exhibits cathodic peaks above 1.0 V corresponding to the irreversible electrochemical reactions between the SiO 2 and Li, which disappears in the subsequent cycles. The cathodic peak below 0.2 V can be ascribed to the lithiation process of silicon. The electrochemical equations can be expressed as followed: 5SiO 2 + 4Li + + 4e − lithiated. All of the lithiated SiO2 electrodes exhibit the similar potential plateaus around 0.45 V, indicating the same active materials (LixSi) of the electrodes and their high crystallinity, consistent with the XRD results. The prelithiation capacities of the SiO2 particles are 1972, 2267, 1776, and 1412 mAhg −1 , which are listed in Table 1. Here, the prelithiation capacity is determined by subtracting the lithiation capacity from the delithiation capacity at the first cycle. The thermal lithiation can effectively compensate for the huge irreversible lithium loss during the first cycles via the formation of lithium silicates and Li2O. The cyclic voltammetry (CV) tests are conducted for both the pristine and prelithiated SiO2 half cells for 3 consecutive cycles between 3.0 V and 0.01 V at a scan rate of 0.05 mVs −1 (Figure 4). The pristine SiO2 electrode exhibits cathodic peaks above 1.0 V corresponding to the irreversible electrochemical reactions between the SiO2 and Li, which disappears in the subsequent cycles. The cathodic peak below 0.2 V can be ascribed to the lithiation process of silicon. The electrochemical equations can be expressed as followed: For the LixSi/Li2O-4 anode, the corresponding peak of silicon lithiation is very pronounced. The cathodic peak at 0.5 V at the first cycle confirms the formation of highly crystalline LixSi. Two distinct anodic peaks at about 0.34 and 0.51 V during the second and third cycle agree with the characteristic peaks of delithiation process from amorphous LixSi to Si. The electrochemical equation can be expressed as followed: The corresponding Nyquist plots measured at the first cycle are illustrated in supplement information (SI, Figure S2). LixSi/Li2O-4 anode demonstrated a smaller semicircle than the corresponding SiO2 anode, indicating a lower charge transfer resistance (Rct). It is clear that the prelithiated sample provides a fast electron transport pathway between the LixSi domains in the microparticles.  electrodes. The open-circuit voltage (OCV) of the half cells is around 0.35 V, which is significantly lower than ~2.0 V of the SiO2/Li cell (SI, Figure S1). This proves that most SiO2 particles were well lithiated. All of the lithiated SiO2 electrodes exhibit the similar potential plateaus around 0.45 V, indicating the same active materials (LixSi) of the electrodes and their high crystallinity, consistent with the XRD results. The prelithiation capacities of the SiO2 particles are 1972, 2267, 1776, and 1412 mAhg −1 , which are listed in Table 1. Here, the prelithiation capacity is determined by subtracting the lithiation capacity from the delithiation capacity at the first cycle. The thermal lithiation can effectively compensate for the huge irreversible lithium loss during the first cycles via the formation of lithium silicates and Li2O. The cyclic voltammetry (CV) tests are conducted for both the pristine and prelithiated SiO2 half cells for 3 consecutive cycles between 3.0 V and 0.01 V at a scan rate of 0.05 mVs −1 (Figure 4). The pristine SiO2 electrode exhibits cathodic peaks above 1.0 V corresponding to the irreversible electrochemical reactions between the SiO2 and Li, which disappears in the subsequent cycles. The cathodic peak below 0.2 V can be ascribed to the lithiation process of silicon. The electrochemical equations can be expressed as followed: For the LixSi/Li2O-4 anode, the corresponding peak of silicon lithiation is very pronounced. The cathodic peak at 0.5 V at the first cycle confirms the formation of highly crystalline LixSi. Two distinct anodic peaks at about 0.34 and 0.51 V during the second and third cycle agree with the characteristic peaks of delithiation process from amorphous LixSi to Si. The electrochemical equation can be expressed as followed: The corresponding Nyquist plots measured at the first cycle are illustrated in supplement information (SI, Figure S2). LixSi/Li2O-4 anode demonstrated a smaller semicircle than the corresponding SiO2 anode, indicating a lower charge transfer resistance (Rct). It is clear that the prelithiated sample provides a fast electron transport pathway between the LixSi domains in the microparticles.   electrodes. The open-circuit voltage (OCV) of the half cells is around 0.35 V, which is significantly lower than ~2.0 V of the SiO2/Li cell (SI, Figure S1). This proves that most SiO2 particles were well lithiated. All of the lithiated SiO2 electrodes exhibit the similar potential plateaus around 0.45 V, indicating the same active materials (LixSi) of the electrodes and their high crystallinity, consistent with the XRD results. The prelithiation capacities of the SiO2 particles are 1972, 2267, 1776, and 1412 mAhg −1 , which are listed in Table 1. Here, the prelithiation capacity is determined by subtracting the lithiation capacity from the delithiation capacity at the first cycle. The thermal lithiation can effectively compensate for the huge irreversible lithium loss during the first cycles via the formation of lithium silicates and Li2O. The cyclic voltammetry (CV) tests are conducted for both the pristine and prelithiated SiO2 half cells for 3 consecutive cycles between 3.0 V and 0.01 V at a scan rate of 0.05 mVs −1 (Figure 4). The pristine SiO2 electrode exhibits cathodic peaks above 1.0 V corresponding to the irreversible electrochemical reactions between the SiO2 and Li, which disappears in the subsequent cycles. The cathodic peak below 0.2 V can be ascribed to the lithiation process of silicon. The electrochemical equations can be expressed as followed: For the LixSi/Li2O-4 anode, the corresponding peak of silicon lithiation is very pronounced. The cathodic peak at 0.5 V at the first cycle confirms the formation of highly crystalline LixSi. Two distinct anodic peaks at about 0.34 and 0.51 V during the second and third cycle agree with the characteristic peaks of delithiation process from amorphous LixSi to Si. The electrochemical equation can be expressed as followed: The corresponding Nyquist plots measured at the first cycle are illustrated in supplement information (SI, Figure S2). LixSi/Li2O-4 anode demonstrated a smaller semicircle than the corresponding SiO2 anode, indicating a lower charge transfer resistance (Rct). It is clear that the prelithiated sample provides a fast electron transport pathway between the LixSi domains in the microparticles.
The corresponding Nyquist plots measured at the first cycle are illustrated in supplement information (SI, Figure S2). Li x Si/Li 2 O-4 anode demonstrated a smaller semicircle than the corresponding SiO 2 anode, indicating a lower charge transfer resistance (Rct). It is clear that the prelithiated sample provides a fast electron transport pathway between the Li x Si domains in the microparticles.
The cycling performance of prelithiated SiO 2 composites is shown in Figure 5a and listed in Table 1. The Li x Si/Li 2 O composites deliver a first delithiation capacity of around 2000 mAhg −1 at C/20, with a corresponding Coulombic efficiency (CE) that is above 90% at initial cycles (Figure 5a and Figure S3). Very interestingly, the Li x Si/Li 2 O-4 anode, prepared from micron-sized silica, has the best cycling stability among the four samples. It is able to maintain a stabilized reversible capacity of around 1300 mAhg −1 with acceptably small decay (~71% capacity retention) after prolonged 50 cycles. The longer cycling performance is shown in supporting information, remaining around 800 mAhg −1 after 500 cycles. (Figure S4). The rate performance of Li x Si/Li 2 O-4 anode was shown in Figure 5b. It yields a reversible capacity of over 1500 mAhg −1 at 0.05 C, around 1200 mAhg −1 at 0.1 C and 800 mAhg −1 at 0.2 C. At a high rate of as high as 1 C, the capacity is as low as 100 mAhg −1 . However, when the current rate reversed to 0.05 C, a high specific capacity have been recovered.
The cycling performance of prelithiated SiO2 composites is shown in Figure 5a and listed in Table  1. The LixSi/Li2O composites deliver a first delithiation capacity of around 2000 mAhg −1 at C/20, with a corresponding Coulombic efficiency (CE) that is above 90% at initial cycles (Figures 5a and S3). Very interestingly, the LixSi/Li2O-4 anode, prepared from micron-sized silica, has the best cycling stability among the four samples. It is able to maintain a stabilized reversible capacity of around 1300 mAhg −1 with acceptably small decay (~71% capacity retention) after prolonged 50 cycles. The longer cycling performance is shown in supporting information, remaining around 800 mAhg −1 after 500 cycles. (Figure S4). The rate performance of LixSi/Li2O-4 anode was shown in Figure 5b. It yields a reversible capacity of over 1500 mAhg −1 at 0.05 C, around 1200 mAhg −1 at 0.1 C and 800 mAhg −1 at 0.2 C. At a high rate of as high as 1 C, the capacity is as low as 100 mAhg −1 . However, when the current rate reversed to 0.05 C, a high specific capacity have been recovered.
As schemed in Figure 6, LixSi nano-domains are homogeneously dispersed in robust Li2O matrix after the thermal prelithiation. For nano-sized SiO2 particles, the LixSi domain is small and each domain is covered by a Li2O layer. This Li2O layer, formed by trace air oxidization of surface LixSi, is electrical insulating, which hinders the electrical conductivity between the LixSi/Si nano-domains. After many electrochemical cycles, the electrical contact becomes even worse due to the SEI formation on these small domains. While for micron-sized particles, the LixSi domains are relatively large. In addition, oxidized Li2O surface layer is mainly wrapped out of the whole microparticle, leaving LixSi domains contacting with each other inside the particle. The Li2O outer layer can spatially limit the direct SEI formation on LixSi/Si cores during the cycling. As a result, the LixSi/Li2O from SiO2 microparticles can achieve enhanced cycling stability in half cells. Therefore, SiO2 microparticle should be a very promising anode material, considering its low cost and the high cycling stability.   As schemed in Figure 6, Li x Si nano-domains are homogeneously dispersed in robust Li 2 O matrix after the thermal prelithiation. For nano-sized SiO 2 particles, the Li x Si domain is small and each domain is covered by a Li 2 O layer. This Li 2 O layer, formed by trace air oxidization of surface Li x Si, is electrical insulating, which hinders the electrical conductivity between the Li x Si/Si nano-domains. After many electrochemical cycles, the electrical contact becomes even worse due to the SEI formation on these small domains. While for micron-sized particles, the Li x Si domains are relatively large. In addition, oxidized Li 2 O surface layer is mainly wrapped out of the whole microparticle, leaving Li x Si domains contacting with each other inside the particle. The Li 2 O outer layer can spatially limit the direct SEI formation on Li x Si/Si cores during the cycling. As a result, the Li x Si/Li 2 O from SiO 2 microparticles can achieve enhanced cycling stability in half cells. Therefore, SiO 2 microparticle should be a very promising anode material, considering its low cost and the high cycling stability.

Conclusions
In conclusion, this study explores the potential of using prelithiated SiO 2 particles as anode materials for high-capacity batteries. We systematically invested SiO 2 particles with different sizes varying from 6 nm to 3 µm. All prelithiated composites exhibit better performance than that of untreated SiO 2 . The nano-particle, which is often thought to be more competitive in battery performance, however, lost its superiority after prelithiation. Contrary to our expectations, the micron-sized SiO 2 particle demonstrates more stable cycling performance, delivering a prelithiation capacity of 1412 mAhg −1 and a reversible specific capacity over 1300 mAhg −1 after 50 cycles. Considering the encouraging battery performance displayed, together with its abundance and low-cost, SiO 2 microparticle demonstrates a promising future as a novel anode material with high capacity in lithium-ion batteries.
Author Contributions: Y.H., X.L. and Z.L. did the experiment and wrote this work together.