A review on the critical challenges and progress of SiO_x-based anodes for lithium-ion batteries

1.   Introduction

Non-renewable resources, such as coal, oil, and natural gas, are still used and consumed worldwide in large quantities, causing a rapid drop in their reserves and leading to greenhouse effect intensification, widespread water and air pollution, and other negative phenomena. To alleviate these issues, electric vehicles have been vigorously promoted worldwide in recent years. Lithium-ion batteries (LIBs) have dominated the power source market of electric vehicles because of their several advantages, such as high energy density, high working voltage, long cycling life, and low self-discharge [1]. However, the energy density of current LIBs cannot meet the ever-increasing mileage need of electric transportation, which limits their wide deployment [2]. Thus, exploring new electrode materials with high energy density has become the top priority in the R&D of LIBs.

Anodes exert a decisive influence on the electrochemical performance of LIBs. At present, graphite-based anodes are widely utilized in commercial LIBs, and the specific capacity of such anodes has already reached 340–365 mAh·g⁻¹, which is close to the theoretical value [3]. With the growing demand for high-energy-density LIBs, the development of new anodes with high specific capacities is imperative. Instead of intercalation-class materials, either conversion reaction or alloying-type compounds have been proposed and widely studied [4–5]. Silicon is a promising anode because of its high theoretical specific capacity (4200 mAh·g⁻¹), suitable operating voltage, abundant natural resources, and associated low costs [6]. However, the mechanical instability of Si-based electrodes due to large volume changes during lithiation/delithiation, typical for the alloying-type reaction [7–8], causes particle cracking and pulverization, resulting in poor cycling performance of Si-based anodes [9–10]. Many approaches, including the design of nanostructure-scale particles [11–12], formation of Si-containing composites [13–15], and optimization of electrode film structure [16–17], have been proposed to solve this problem. Despite numerous reports on the improvement of the electrochemical properties of Si anodes, less commercial breakthrough in high-Si-content electrodes has been achieved. In modifying silicon anodes, researchers found that SiO_x-type (0 < x < 2) nonstoichiometric oxides exhibit much better cycling stability than pure Si because of the reduced volume expansion in reaction with lithium. Therefore, SiO_x has become an excellent candidate anode for high-energy-density LIBs in recent years.

The fundamental difference regarding the charge/discharge mechanism of SiO_x with Si anodes stems from the irreversible formation of inert components (Li₂O, various lithium silicates) during the first lithiation, whose presence can dilute the volume changes of active particles and thus reduce the mechanical stresses of electrodes, ensuring a relatively stable cycling performance of SiO_x-based electrodes [18–20]. Nevertheless, the cycling stability of SiO_x electrodes is far from the requirement in practical applications. Irreversibly formed inert components contribute a large part of irreversible capacity, causing a much lower initial coulombic efficiency (ICE). In addition, the poor ionic and electronic conductivity of SiO_x limit the electrode reaction kinetics and thus the rate performance.

Facing the challenges of SiO_x application, various strategies, including particle morphology design [21–22], combination with conductive second phase [23–24], usage of proper binders [25– 26], and advanced additives [27], have been proposed to address the issues mentioned above. In this review, we first introduce the structural characteristics of SiO_x and the electrochemical reaction mechanism with lithium. Then, we summarize the progress in the fabrication of SiO_x materials in recent years. Finally, we discuss the improvement and advancement of the electrochemical performance of SiO_x anodes, with an outlook given regarding the possible further development of such electrodes.

2.   Oxygen content and atomic-scale structure of SiO_x

Despite many reports, the structural model of SiO_x remains controversial. Initially, in the 1970s, two main structural models were reported for the amorphous SiO_x, namely, Random-bonding (RB) [28–29] and random-mixture (RM) [30]. The RB model infers that the structure of SiO_x consists of a continuous random distribution formed by Si–Si and Si–O bonds, which run through the entire structural network. By contrast, the RM model proposes that SiO_x is generally a dual-phase material composed of regions comprising amorphous Si and crystalline SiO₂, with a particle size of both fractions being less than 1 nm. This corresponds to the composite-type material at nanoscale, but with respective regions being extremely small, i.e., technically not forming separate phases. Both models have been reported with supporting results [31–32].

Based on the two models mentioned above, Hohl et al. [33] proposed the interface cluster mixture (ICM) model (Fig. 1(a)), which assumes that SiO_x is composed of SiO₂ clusters (light gray) and Si clusters (black), as well as the ultrathin suboxide (in terms of local oxygen content) interface (gray), forming a matrix-like structure. Compared with previous models, the ICM model is more consistent with the reported electrochemical performance of SiO_x in elucidating the electrochemical reaction mechanism with lithium [34]. Using angstrom-beam electron diffraction combined with synchrotron X-ray scattering and computer simulations, Hirata et al. [35] studied the SiO_x structure. Their experimental results uncovered the non-uniform distribution of Si and O at the atomic scale and verified the heterostructural model of Si clusters mixing with amorphous SiO₂ (Fig. 1(b)). The two regions are connected by various suboxide-type interfacial regions with generally tetrahedral coordination.

Fig. 1.  Schematic of a-SiO structure: (a) the interface clusters mixture model (the gray, light gray, and black regions represent sub-oxidic interface, SiO₂ clusters, and Si clusters, respectively); (b) the reconstructed heterostructure model (the blue, red, and green spheres represent Si and O in amorphous SiO₂ and Si in Si clusters, respectively). (a) Reprinted from J. Non Cryst. Solids, 320, A. Hohl, T. Wieder, P.A. van Aken, T.E. Weirich, G. Denninger, M. Vidal, S. Oswald, C. Deneke, J. Mayer, and H. Fuess, An interface clusters mixture model for the structure of amorphous silicon monoxide (SiO), 255-280, Copyright 2003, with permission from Elsevier; (b) reprinted by permission from Springer Nature: Nat. Commun., Atomic-scale disproportionation in amorphous silicon monoxide, A. Hirata, S. Kohara, T. Asada, M. Arao, C. Yogi, H. Imai, Y.W. Tan, T. Fujita, and M.W. Chen, Copyright 2016.

DownLoad: Full-Size Img PowerPoint

With the intrinsic limitations of X-ray diffraction and other methods used for structural characterization, a complete picture showing the actual atomic-scale structure of SiO_x remains challenging to provide.

3.   Lithiation/delithiation mechanism of SiO_x

Many efforts have been exerted to elucidate the lithiation process of SiO_x. The reactivity of SiO_x depends on stoichiometry and crystallinity. Crystalline SiO₂ cannot react with lithium, whereas nanosized SiO₂ [36–38] and amorphous SiO₂ domains presenting in the SiO [39–42] can electrochemically react with lithium to form lithium silicates and Li₂O. Yamamura et al. [43] reported that even amorphous SiO₂ can react with lithium because of the presence of distorted SiO₄ tetrahedrons. Nevertheless, the stoichiometric SiO₂ is generally considered to be less active, and the nonstoichiometric SiO_x with x<2 is active toward lithium storage.

The lithiation/delithiation mechanism of SiO_x material is still not very clear, especially regarding the nature and reversibility of the electrochemical processes for the formed Li_xSi alloy and lithium silicates during the first lithiation. Nevertheless, Si, lithium silicates, and Li₂O are generally believed to form first during the first lithiation of SiO_x, and then Si is further lithiated (alloying-type of reaction) to form Li_xSi alloy, typically up to x = 3.75 [20,34,39,44–52]. Formation of Li_xSi alloy is reversible, occurs at advantageous low potentials vs. Li electrode, and provides the main part of the reversible capacity. By contrast, the formed lithium silicates (e.g., Li₄SiO₄, Li₆Si₂O₇, Li₂SiO₃) during the first lithiation are generally described as irreversible phases. The exception is Li₂SiO₅, in which the lithium is considered deliverable but does not contribute a substantial reversible capacity. In addition, the formation of Li₂O is most possibly irreversible [18,39,44–45]. Interestingly, numerous reports have proven that the electrochemically inactive lithium silicates and Li₂O can provide crucial advantages for the electrode reaction, not only relieving the volume change of the SiO_x-based anodes on lithiation/delithiation but also enhancing the diffusion of Li⁺ in the electrode, which contribute much to the improved cycling stability and rate performance [18,20,39,47–48]. In addition, the amount of the formed electrochemically inactive products is strongly dependent on the electrode charge/discharge current density. More irreversible products will be generated at low C-rates with sufficient Li⁺ provided, whereas the amount of the electrochemically inactive phases decreases faster at high C-rates [53].

For the initial lithiation products of SiO_x, Kim et al. [34] proposed the formation of Li₁₅Si₄ for Si clusters, Li₂O/Li₁₅Si₄ for SiO interphase, and Li₄SiO₄/Li₁₅Si₄ for silicon oxide part based on the RM structure model. As a result, 4.2 mol of Li are supposed to be stored by 1 mol SiO. However, subsequent investigation shows that various lithium silicates can form in the initial lithiation, and some of them can be reversible. The lithiation/delithiation mechanism of SiO_x powders prepared through high-energy mechanical milling (HEMM) was investigated using ex situ high-resolution transmission electron microscopy (HRTEM) [54]. In the first lithiation, unreacted Si and Li₂Si₂O₅ phases appeared first when the potential was lowered to 0.27 V (Fig. 2(a)). Then, Li₄SiO₄ formed subsequently at 0.24 V (Fig. 2(b)). Finally, the Li₁₅Si₄ phase was obtained when the potential reached 0 V (Fig. 2(c)). When Li⁺ was extracted, decomposition of Li₁₅Si₄ occurred first at 0.34 V (Fig. 2(d)). Then, Li₂Si₂O₅ reacted with Si to generate SiO₂ (Li₂Si₂O₅ + (1/2)Si → (5/2)SiO₂ + 2Li⁺ + 2e^–) at 0.39 V (Fig. 2(e)). This product could also be detected after 10 cycles, suggesting that the formation of Li₂Si₂O₅ is reversible.

Fig. 2.  Ex situ HRTEM images when SiO was discharged to (a) 0.27 V, (b) 0.24 V, and (c) 0.0 V, and charged to (d) 0.34 V and (e) 0.39 V; (f) transition between formed phases during Li insertion for SiO negative electrode at 298 K, data shown in the ternary Li-Si-O phase diagram; changes in the equilibrium SiO electrode potential at 298 K during (g) Li insertion and (h) Li removal after the formation of Li₁₃Si₄. (a–e) Reprinted with permission of Royal Society of Chemistry, from Quartz (SiO₂): A new energy storage anode material for Li-ion batteries, W.S. Chang, C.M. Park, J.H. Kim, Y.U. Kim, G. Jeong, and H.J. Sohn, 5, 5, Copyright 2012, permission conveyed through Copyright Clearance Center, Inc.; (f–h) reprinted from J. Power Sources, 329, K. Yasuda, Y. Kashitani, S. Kizaki, K. Takeshita, T. Fujita, and S. Shimosaki, Thermodynamic analysis and effect of crystallinity for silicon monoxide negative electrode for lithium ion batteries, 462-472, Copyright 2016, with permission from Elsevier.

DownLoad: Full-Size Img PowerPoint

The behavior of SiO_x suboxides with highly differing x values (x = 0.17, 0.51, 1.02, and 1.34) in the charge and discharge processes was investigated through X-ray photoelectron spectroscopy [19]. Fully lithiated Li_4.4Si, Li₂O, Li₄SiO₄, and Li₂SiO₃ form after the first discharge. Aside from the Li₂O and lithium silicates, a portion of Li_xSi (x < 2) could not be fully delithiated and remained after the first charge process, which also contributes to the irreversible capacity loss of SiO_x on cycling. In addition, Li₂SiO₃ might be a more stable irreversible phase than Li₄SiO₄ when SiO_1.34 was utilized, whereas Li₄SiO₄ was more stable when lower oxygen contents SiO_0.17, SiO_0.51, and SiO_1.02 were first lithiated.

Yasuda et al. [55] explored the lithiation mechanism of SiO by thermodynamic analysis and electrochemical experiments. The results are summarized in Fig. 2(f)–(h). (1) Lithium silicates (Li₂Si₂O₅, Li₂SiO₃, and Li₄SiO₄) should appear together with the formation of Si (from point 1 to point 7 in the graphs) up to the discharge capacity of 608 mAh·g⁻¹, which is consistent with 600 mAh·g⁻¹ capacity in the 0.35–0.5 V range from the lithiation curve of the amorphous SiO in practical experiments; (2) the formed Si reacts with Li to form Li-Si alloys (Li₁₂Si₇, Li₇Si₃, and Li₁₃Si₄) that are in equilibrium with Li₄SiO₄ (from point 7 to 13), corresponding to a total discharge capacity of 2090 mAh·g⁻¹, which does not match the measured capacity of 2600 mAh·g⁻¹ in the 0.05–0.35 V range, even if the capacities of the conductive additive and binder are deducted. This result suggests the partial decomposition of already formed lithium silicates into Li₂O, which occurs with the concomitant release of Si. The latter can contribute extra capacity with respect to the theoretical prediction; (3) Upon further lithiation, the Li₄SiO₄ phase is decomposed to form Li₂O and Li₁₃Si₄; (4) Further Li⁺ insertion produces Li₂₂Si₅ from Li₁₃Si₄ (from point 15 to 17), with a total discharge capacity of 2891 mAh·g⁻¹; (5) finally, metallic Li is deposited at the lowest voltages, close to 0.0 V (after point 18) [55].

Molecular dynamic simulations were adopted to investigate the lithiation mechanism of SiO. Results indicated that Li_xSiO can theoretically store up to x = 5.22 of lithium, corresponding to a higher capacity of 3172 mAh·g⁻¹, when Li₂O becomes dominant over Li-silicates. However, owing to large activation barriers that should be overcome to decompose the SiO₄ tetrahedral units in Li-silicates, the evolution of Li₂Si₂O₅, Li₆Si₂O₇, and Li₄SiO₄ into Li₂O is kinetically impossible under electrochemical conditions. Only Li_4.39SiO composition (i.e., average x = 4.39) can be achieved in practical experiments. They suggested that Li-silicates are dominant over Li₂O as the formed irreversible subphases. The kinetically controlled transformation of Li-silicates into Li₂O in the matrix should lead to great improvements in the reversible capacity and rate capability of SiO [20].

Although extensive investigations have been carried out on the lithiation/delithiation mechanism of SiO_x, some divergence still exists, especially on the issues of the final lithiation products and the reversibility of some of these lithium silicates. Nonetheless, the viewpoints that Li₁₅Si₄ is the final lithium-silicon alloying product and that Li₄SiO₄ is an irreversible lithiation product are widely accepted [34,43,45,55]. Apparently, the lithiation/delithiation mechanisms of SiO_x must be understood to improve the ICE and enhance the electrochemical performance effectively.

4.   Synthesis methods of electrochemically-active SiO_x

Many efforts have been devoted to the synthesis of SiO_x anodes because it is not a naturally stable phase structure. The successful methods reported can be generally grouped into three approaches: (1) thermal evaporation [18–19,56–57], (2) wet-chemistry [58–61], and (3) HEMM [48,52,62], which are illustrated in Fig. 3.

Fig. 3.  Schematic of major preparation methods of SiO_x: (a) thermal evaporation [63]; (b) wet-chemistry; (c) HEMM. (a) Reprinted from Electrochim. Acta, 148, S.S. Suh, W.Y. Yoon, D.H. Kim, S.U. Kwon, J.H. Kim, Y.U. Kim, C.U. Jeong, Y.Y. Chan, S.H. Kang, and J.K. Lee, Electrochemical behavior of SiO_x anodes with variation of oxygen ratio for Li-ion batteries, 111-117, Copyright 2014, with permission from Elsevier.

DownLoad: Full-Size Img PowerPoint

4.1   Thermal evaporation methods

The widely adopted method for SiO_x synthesis on an industrial scale is thermal evaporation, in which SiO_x powder is produced by sublimating and condensing SiO₂ and Si under high temperatures [56,64–66]. Moreover, SiO_x materials with different silicon/oxygen ratios can be prepared by adjusting the ratio of raw materials and process conditions [67]. Shin-Etsu Chemical Co., Ltd. of Japan, Osaka Titanium Technologies Co., Ltd. of Japan, and Samsung SDI Co., Ltd. of South Korea used this method to produce SiO_x powders as an anode component for LIBs on a commercial scale.

In addition, other physical and chemical techniques are explored to fabricate SiO_x powders or films include vacuum electron beam evaporation [19], plasma spray physical vapor deposition [57,63], and infrared ray nanosecond laser ablation [68]. SiO_x with different silicon/oxygen ratios (x value) can be obtained by controlling the processing parameters. As an example, Takezawa et al. [19] used vacuum electron beam evaporation to deposit amorphous SiO_x films on copper substrates through a reaction between evaporative Si and oxygen, and the x value could be controlled by oxygen flow rate. Therefore, SiO_x films with different silicon/oxygen ratios (x = 0.17, 0.51, 1.02, and 1.34) could be prepared, which enabled the study of the influence of total oxygen content on the electrochemical properties of SiO_x.

Considering industrial applications, thermal evaporation methods exhibit important advantages related to the control of the total oxygen content in the material. However, their usage for the commercial-scale synthesis of LIBs remains to have substantial issues, including large equipment investment, high energy consumption, and low productivity, resulting in the high cost of SiO_x products.

4.2   Wet-chemistry routes

Wet-chemistry synthesis routes are widely used in the preparation of various type powders because of the advantages of the homogenous, atomic-scale mixing of the elements ensured at the initial stages of the procedure, the controllable particle size and morphology of the product, as well as low-energy consumption and associated low costs [58,69–70].

The sol–gel methods are reported for the preparation of SiO_x materials [58–61]. The reaction conditions, including starting chemicals, surfactants, catalysts, hydrolysis time, drying, and reagent ratio, must be thoughtfully adjusted to obtain the desired phase composition, structure, and morphology of the product. Liu et al. [58] used vinyltriethoxysilane (VTES) as the silicon source for the preparation of SiO_x powders. Under Stöber reaction conditions, VTES was hydrolyzed and condensed to produce organosilicate, which was subsequently processed to generate SiO_x microspheres. In addition, tetraethoxysilane (TEOS) [21,60,71–75] and tetramethoxysilane [76–77] are frequently used as silicon sources. Various material structures were prepared by combining the template with a hydrothermal method [78–82]. With CTAB-assisted resorcinol self-assembly as a template, Ren and Li [81] prepared rod-like SiO_x@C composite with individual rod containing numerous interconnected nanospheres by using a hydrothermal method with a post pyrolysis process. The network-structured SiO_x/MWCNT/N-doped C composite was fabricated as LIB anode through precipitation and pyrolysis by using triethoxyethylsilane ((C₂H₅O)₃SiC₂H₅) as the silicon source [83].

Our group worked in this direction for more than ten years. Amorphous SiO_x powders with different particle sizes were synthesized as anode materials for LIBs from TEOS via sol–gel and hydrothermal methods [71–73,84]. With CTAB and PVP self-assembled template, the mesoporous SiO_x/C composite with an arrayed molecular-sieve architecture was prepared from TEOS and sucrose via a hydrothermal route [82]. SiO_x–C dual-phase glass was fabricated via a sol–gel method combined with calcination to increase the tapping density [71]. The phase composition and the oxygen nonstoichiometry of SiO_x can be controlled by optimizing the processing parameters. High Si content with a low x value of SiO_x can deliver a high specific capacity, which can be achieved by preparing a highly porous SiO_x-based material with a high specific surface area to realize sufficient reduction via a wet-chemistry route [74,85].

The wet-chemistry route for SiO_x synthesis has advantages in the chemical homogeneity and controllable morphology of target materials, as well as in capital investment reduction, which ensures a good demonstration of electrochemical performance and is conducive to the deployment of SiO_x anodes in the practical battery industry.

4.3   High energy mechanical milling

Compared with the traditional energy-consuming preparation processes, HEMM is simple in operation. The SiO_x powders can be synthesized using HEMM from Si and SiO₂, and the particle size of SiO_x can be controlled by adjusting the time of ball milling [86]. As reported, Si/O₂ can also be used as raw materials to synthesize SiO_x, and the x value in SiO_x can be tailored by tuning the exposure time of pure silicon powder to air [62,86–87]. For example, Cao et al. [62] prepared an amorphous SiO_x anode with oxygen content between 0 ≤ x ≤ 0.37 through HEMM of Si powder in the air. The amorphous SiO_x has a particle structure of the nano-Si core surrounding with amorphous SiO_x.

HEMM can reduce the activation energy of a reaction between the starting chemicals and realize the synthesis of materials at low temperatures. In general, the particle size decreases with ball milling time extension, but may cause secondary agglomeration and produce impurities originating from the milling chamber and balls. Thus, appropriate ball-milling conditions must be controlled accurately to obtain the desired oxygen content and particle size of the synthesized SiO_x.

5.   Fundamental challenges and solutions to overcome limitations of SiO_x anodes

As mentioned above, the inert buffer matrix composed of lithium silicates and Li₂O formed in the first lithiation of SiO_x anodes helps mitigate volume changes in reversible electrochemical processes and therefore improves the cycle stability of silicon suboxides as compared with pure Si. However, SiO_x anodes still have considerable volume changes (~200%) upon lithiation/delithiation processes, which has the potential to destroy the electrode structure and cause a rapid capacity decay. Moreover, the electrochemically inactive lithium silicates and Li₂O formed in the first lithiation result in a low ICE, and the low intrinsic electrical conductivity limits the rate performance of SiO_x electrodes [50]. These disadvantageous effects are associated with the following major challenges: (1) material pulverization effect on cycling, (2) unstable solid-electrolyte interphase (SEI) film, (3) low ICE, and (4) electrode failure by cracking.

5.1   Morphology design for mitigating the pulverization of SiO_x on cycling

The capacity of pure silicon suboxide electrodes usually degrades rapidly with cycling because of the particle pulverization caused by the mechanical stresses related to the volume changes. One of the effective approaches to address this issue is to build special particle structures, such as core-shell, yolk-shell, porous, and hollow structures, constructing buffer layers or providing sufficient space to accommodate the volume change.

For the core-shell structure, a layer of conductive material, such as carbon or metal oxide, can be coated on the surface of SiO_x [72,88–94]. This type of structure can effectively limit the volume expansion and improve the electronic conduction of SiO_x. For example, our group synthesized a core-shell structure composed of SiO_x nanoparticles coated with a thin carbon layer through a simple and efficient wet-chemical synthesis route, which provides a stable reversible capacity of 820 mAh·g⁻¹ after 100 cycles [72]. In addition, the dual-shell coating structural composites can further inhibit the structural damage caused by internal stress. Xiao et al. [89] synthesized SiO_x@TiO₂@C composite, and the introduced TiO₂ layer with a high-crystallinity anatase phase can improve the overall interface stability of the electrode. The volume change of the electrode is effectively buffered, the discharge capacity retention is 89.5% after 800 cycles at a current density of 1.0 A·g⁻¹.

Different morphology designs also aim to create a free space that can buffer the volume changes of SiO_x material [70,74,82,85,95–98]. As shown in Fig. 4(a)–(f), porous, hollow, and yolk-shell structures can provide abundant free space. Lee and Park [96] introduced a simple method to synthesize a 3D porous silicon monoxide anode through wet-chemical etching using bulk silica powder as the raw material (Fig. 4(a) and (b)). The porous structure can effectively buffer the volume expansion of SiO_x and promote a stable cycling performance. However, its high specific surface area causes excessive SEI generation, resulting in a low ICE of 60%. To address this issue, the author subsequently coated the surface with a carbon layer and increased the ICE to 76%. The yolk-shell and hollow structures were prepared using chemical etching, respectively (Fig. 4(c)–(f)) [98–99]. These structures enable SiO_x material to expand freely into the interior on lithiation, whereas the outer carbon-shell coating maintains the integrity of the whole particle structure, which is beneficial to the interface stability, thus promoting the steady growth of the SEI film. The yolk-shell structured SiO_x/C material shows a good cycle performance of 972 mAh·g⁻¹ after 500 cycles at 500 mA·g⁻¹ [98]. Our team prepared a dandelion-like highly porous SiO_x particle coated with a conformal carbon layer, in which the abundant mesopores effectively accommodate the volume variation of the particles upon lithiation, thereby endowing the material with an excellent cycle stability of 88.6% capacity retention at 2 A·g⁻¹ after 1000 cycles [74].

Fig. 4.  (a) Schematic showing the synthesis route of porous SiO using Ag catalytic etching; (b) SEM micrograph of the obtained porous SiO_x; (c) preparation of yolk-shell SiO_x/C and SiO_x/C microspheres; (d) TEM image of yolk-shell SiO_x/C; (e) schematic of synthesis process for void@SiO_x@C; (f) SEM image of hollow void@SiO_x@C. (a, b) Reprinted from Nano Energy, 2, J.I. Lee and S. Park, High-performance porous silicon monoxide anodes synthesized via metal-assisted chemical etching, 146-152, Copyright 2013, with permission from Elsevier; (c, d) reprinted from Energy Storage Mater., 19, Z.H. Liu, Y.L. Zhao, R.H. He, W. Luo, J.S. Meng, Q. Yu, D.Y. Zhao, L. Zhou, and L.Q. Mai, Yolk@shell SiO_x/C microspheres with semi-graphitic carbon coating on the exterior and interior surfaces for durable lithium storage, 299-305, Copyright 2019, with permission from Elsevier; (e, f) D.L. He, P. Li, W. Wang, Q. Wan, J. Zhang, K. Xi, X.M. Ma, Z.W. Liu, L. Zhang, and X.H. Qu, Small, 16, art. No. 1905736 (2020) [99]. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

DownLoad: Full-Size Img PowerPoint

5.2   SiO_x-based composite anode materials

Another effective way to resolve the particle pulverization is to combine SiO_x with a second phase to form multiphase architecture (carbon, metal, metal oxide, etc.). The second phase can buffer the volume change caused by SiO_x lithiation, prevent the aggregation of SiO_x particles, and/or construct a conducting network between the particles to improve the cycling performance of the electrode.

5.2.1   SiO_x/C composites

Combining SiO_x with carbon materials is a common, simple, economical, and overall promising approach to improve the performance of SiO_x anodes. Carbon materials can efficiently improve the conductivity and thereby promote the reaction kinetics occurring at the electrodes. Furthermore, carbon materials can properly buffer the volume changes of SiO_x during cycling and thus enhance electrode stability [51]. Several methods can be employed to synthesize SiO_x/C composites, but the electrochemical properties of the prepared composites differ greatly. A general overview of the recently published results is shown in Table 1.

Table  1.  Comparison of electrochemical performance of SiO_x–C composite electrodes.

Anode material	Carbon source	Carbon content / wt%	Initial discharge capacity (mAh·g−1)/ Charge capacity (mAh·g−1)/Coulombic efficiency (%)/Current density (mA·g−1)	Stable capacity (mAh·g−1) / Number of cycles	Year	Ref.
SiOx/C	Citric acid	1	1995/1297/65/100	844/200	2017	[100]
Si−SiOx−C	Citric acid	11	1948/1562/80/100	1373/100	2017	[101]
Si/SiOx@C	Sucrose	63	1619/1214/75/100	1095/100	2018	[102]
SiOx/C	RF	40	1460/965/66/100	854/150	2018	[58]
SiOx/rGO	rGO	—	3765/2564/68/100	580/200	2018	[78]
SiOx/G/C	Graphite, Pitch	74%	775/653/84/100	524/350	2019	[103]
Si/SiOx/N−C	Konjac flour	48	1757/1147/65/100	952/100	2019	[104]
SiOx/C	Sucrose	12	1284/830/65/100	755/300	2019	[69]
SiOx/ND−C	Citric acid	37	1793/1209/67/100	1182/100	2019	[105]
SiOx@graphite−Fe	Graphite	—	2331/1139/49/200	757/100	2019	[106]
SiOx@C	C2H4 (CVD)	7	1986/1491/75/1000	1290/500	2019	[107]
SiOx/C	Hexane	8	1666/1081/65/100	1029/100	2019	[108]
DC−HSiOx	Graphene, phenolic resin	22	1858/1113/60/100	982/200	2019	[70]
SiOx/graphene	Graphene	—	2091/1326/63/100	1270/120	2019	[109]
rGO@SiOx@C	rGO, PVP	6	1785/707/40/100	410/200	2020	[110]
N@C/SiOx	Rice husks	—	2225/1661/72/100	1019/100	2020	[111]
VG@SiOx/N−C	CH4 (CVD), graphene	33	1797/1324/74/100	1230/100	2020	[112]
Void@SiOx@C	Pitch	33	1226/751/61/100	696/500	2020	[99]
SiOx/C	Bamboo shoot hulls	77	1332/1258/94/200	1289/400	2021	[113]
Si−SiOx@C	Sugarcane leaves	60	1451/1277/88/200	1562/400	2021	[114]
SiOx/C	Citric acid	43	1067/684/64/100	721/202	2021	[115]
Note: RF—resorcinol/formaldehyde; MWCNT—multiwall carbon nanotube; PAN—polyacrylonitrile; G—graphite; rGO—reduced graphene oxide; ND—nano-diamond; DC—dual carbon conductive network; VG—vertical graphene; PVP—polyvinylpyrrolidone.

 | Show Table

DownLoad: CSV

As observed from Table 1, the carbon materials applied to form composites with SiO_x mainly include graphite, amorphous carbon, graphene, and carbon-based nanomaterials. Graphite is known for its good electrical conductivity, which can enhance the electrode reaction kinetics of SiO_x. SiO_x/graphite composites can be prepared by direct mixing or by coating the surface of SiO_x with graphitized carbon [103,106,116].

An amorphous carbon layer can be formed by mixing SiO_x with organic materials and calcining at high temperatures. The formed carbon shell layer effectively improves the conductivity of SiO_x-based composites and buffers the volume changes of SiO_x during lithiation/delithiation. Compared with a direct mixing method, adding carbon source in a sol–gel process allows to realize a uniform dispersion of carbon source with silicon source in controlled hydrolysis [69,71,84,104,117]. Uniformly dispersed SiO_x/C composites can also be synthesized using a hydrothermal method [72,108,118]. As a representative sample, SiO_x/C composite spheres prepared from the mixture of polydimethylsiloxane and hexane show a high reversible capacity of 1603 mAh·g⁻¹ with a capacity retention of 96.1% at 0.2 A·g⁻¹ after 400 cycles [108].

Graphene has a high surface area, superior electronic conductivity, and excellent mechanical flexibility. The incorporation of graphene into SiO_x can increase electronic conductivity, limit the volume change, and improve electrode reaction kinetics and cycling performance. Various SiO_x/graphene composites as anodes for LIBs have been reported. SiO_x/graphene composites with different ratios were prepared from multilayer graphene and SiO_x powders through HEMM [109]. The prepared composite electrode exhibits a reversible capacity of 1326 mAh·g⁻¹ with a capacity retention of 95.8% after 120 cycles, which is much higher than that of SiO_x electrodes (capacity retention of 62.6% after 120 cycles). With respect to the SiO_x/graphene composite, the introduction of other carbon components and the design of special structures were also adopted to improve the electrochemical performance of SiO_x anodes. Multicomponent nanosheets rGO-SiO_x@C were prepared (Fig. 5(a)) in a configuration of rGO nanosheets as the substrate, porous SiO_x as the intermediate layer, and nitrogen-doped nanoporous carbon as the shell. The highly porous structure of the composite ensures good cycling stability of the electrode (Fig. 5(a)–(c)) [110]. Similar multicomponent microspheres were prepared by Han et al. [112], in which SiOCN microspheres were obtained by pyrolysis of C₁₀H₂₈N₂OSi_2, and then graphene (VG) was vertically grown on their surface. The outer porous graphene and inside conductive N-doped carbon enable the multicomponent SiO_x-based microspheres to exhibit an attractive electrochemical performance with a high specific capacity and long-term cyclability (Fig. 5(d)–(f)).

Fig. 5.  (a) Schematic of the proposed fabrication route of rGO@SiO_x@C nanosheets; (b) TEM image of rGO@SiO_x@C; (c) CE and cycling stability at a current density of 1 A·g⁻¹ of rGO@SiO₂@C and rGO@SiO_x@C; (d) diagram for the synthesis of VG@SiO_x/NC microspheres; (e) SEM image of VG@SiO_x/NC; (f) Comparison of cycling performance of the prepared samples at 0.1 A·g⁻¹. (a–c) Reprinted from L.Y. Chen, J. Zheng, S.Y. Lin, S. Khan, J.L. Huang, S.H. Liu, Z.R. Chen, D.C. Wu, and R.W. Fu, Synthesis of SiO_x/C composite nanosheets as high-rate and stable anode materials for lithium-ion batteries, ACS Appl. Energy Mater., 3(2020), No. 4, p. 3562. Copyright 2020 American Chemical Society; (d–f) reprinted with permission of Royal Society of Chemistry, from Vertical graphene growth on uniformly dispersed sub-nanoscale SiO_x/N-doped carbon composite microspheres with a 3D conductive network and an ultra-low volume deformation for fast and stable lithium-ion storage, M.S. Han, Y.B. Mu, F. Yuan, J.B. Liang, T. Jiang, X.D. Bai, and J. Yu, 8, Copyright 2020, permission conveyed through Copyright Clearance Center, Inc.

DownLoad: Full-Size Img PowerPoint

Carbon nanotubes have attracted much attention because of their numerous advantages, including good conductivity, excellent flexibility, high strength, and good stability. Their usage to form composites with SiO_x is expected to strengthen electrical conductivity, homogenize the electrode reaction on the whole electrode surface, enhance the resistance against stress, and therefore improve the cycling performance of the electrode. For example, Guo et al. [119] prepared a flexible continuous thin film electrode composed of carbon nanotubes and SiO_x, which show a relatively stable and high specific capacity of ~1240 mAh·g⁻¹ after 100 cycles at 100 mA·g⁻¹.

5.2.2   SiO_x/metal and SiO_x/metal oxide composites

In addition to combining SiO_x with various carbon materials, compositing SiO_x with metals and metal oxides is another effective way to improve the electrochemical performance of SiO_x electrodes. The metal oxides formed by the reaction of SiO_x and metals can alleviate the large volume changes of SiO_x electrodes during cycling. Moreover, the added metals can enhance the electronic conductivity, which can further improve the charge transfer kinetics of SiO_x anodes.

Wen and other researchers [48,120] prepared a series of nanosized Si-based composites using SiO and metals, such as Li or Ni [121], by using HEMM followed by heat treatment. During milling, nanosized composites consisting of an inactive matrix (Li₂O, Li₄SiO₄, or Ni₂SiO₄) and uniformly distributed active particles (Si or Si–Ni) form (Fig. 6(a)). The inert phases can effectively buffer the volume expansion and enhance the cycling stability of electrodes. The ICE of this type of electrode can be considerably increased at the expense of specific capacity because the oxygen in SiO has been fixed by the addition of metals to form an inactive matrix.

Fig. 6.  (a) HRTEM image of the composite powder after milling SiO and lithium metal for 10 h, and the inset of the figure shows clear lattice fringes of Li₄SiO₄; (b) schematic particle structure; (c) HRTEM image; (d) cycle performance; (e) rate performance of SiO_x–TiO₂@C ; (f) cycling performance of cells with UHEM anode (current C/9 and C/3) and SPEX anode (current C/3) at room temperature. (a) Reprinted from J. Power Sources, 164, X.L. Yang, Z.Y. Wen, X.X. Xu, B. Lin, and S.H. Huang, Nanosized silicon-based composite derived by in situ mechanochemical reduction for lithium ion batteries, 880-884, Copyright 2007, with permission from Elsevier; (b–e) Z.L. Li, H.L. Zhao, P.P. Lü, Z.J. Zhang, Y. Zhang, Z.H. Du, Y.Q. Teng, L.N. Zhao, and Z.M. Zhu, Adv. Funct. Mater., 28, 1605711 (2018) [73]. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission; (f) reprinted from B. Liu, A. Abouimrane, Y. Ren, M. Balasubramanian, D.P. Wang, Z.Z. Fang, and K. Amine, New anode material based on SiO–Sn_xCo_yC_z for lithium batteries, Chem. Mater., 24(2012), No. 24, p. 4653. Copyright 2012 by American Chemical Society.

DownLoad: Full-Size Img PowerPoint

Metals coated on the SiO_x surface can effectively mitigate the pulverization of the electrodes. Apart from enhancing the conductivity, the integrity of the frame structure of the material can be improved with metal nanoparticle coating. This can be attributed to the homogenized electrode reaction on the particle surface enabled by the metal nanoparticles, which ensures a uniform particle volume change and so less structural stress inside the particles. Numerous SiO_x-based composite materials coated with W [122], Cr [123], or Cu [124] have been developed by physical vapor deposition or wet-chemistry methods.

In addition to improving the cycling performance of SiO_x, some additional benefits can be obtained by using metal oxides to form composites. For example, TiO₂ has various merits, including high thermal stability (which helps suppress reactions between SiO_x and the electrolyte), good mechanical properties to endure the volume changes of SiO_x, and low costs. Several SiO_x-based composites have been reported with the addition of TiO₂ [91,125–126]. Li et al. [73] designed a watermelon-like structured SiO_x-TiO₂@C nanocomposite by using the sol–gel method (Fig. 6(b)), TiO₂ nanocrystals were evenly embedded inside the SiO_x matrix (Fig. 6(c)). The addition of TiO₂ components effectively improves the electron and lithium-ion conductivity inside SiO_x particles and releases the structural stress occurring in the charge and discharge of SiO_x. The SiO_x–TiO₂@C electrode exhibits stable cycle performance (~910 mAh·g⁻¹ after 200 cycles at 0.1 A·g⁻¹) and excellent rate performance (Fig. 6(d) and (e)) [73].

Moreover, some other metal oxides, such as Fe₂O₃ [127], ZrO₂ [128], and SnO₂ [129], are employed to improve the cycle performance of SiO_x by increasing the conductivity of the composite and buffering the volume changes during lithiation/delithiation. Liu et al. [130] prepared a family of composite anode materials, SiO/Sn₃₀Co₃₀C₄₀, consisting of 50wt% SiO and 50wt% Sn₃₀Co₃₀C₄₀, by a custom-made ultrahigh-energy ball milling (UHEM) setup. The formed composite materials combine the high capacity of SiO_x and the good cycle performance of Sn₃₀Co₃₀C₄₀. Experimental results (Fig. 6(f)) showed that the UHEM SiO/Sn₃₀Co₃₀C₄₀ composite electrode delivered reversible capacities of 1040 and 900 mAh·g⁻¹ at currents of 100 and 300 mA·g⁻¹ after 100 cycles, respectively. Subsequently, Liu et al. [131] used a cheaper and environmentally friendly iron instead of expensive and toxic cobalt to prepare SiO/Sn_xFe_yC_z composite anodes through the same method. It was shown that the 50wt% SiO–50wt% Sn₃₀Fe₃₀C₄₀ composite electrode delivered reversible capacities of 900 and 700 mAh·g⁻¹ at current densities of 300 and 800 mA·g⁻¹, respectively. Table 2 summarizes the representative electrochemical performance of SiO_x/metal and SiO_x/metal oxide composite anodes.

Table  2.  Electrochemical performance of SiO_x/metal and SiO_x/metal oxide composite electrodes

Anode material	Initial discharge capacity (mAh·g−1)/ Charge capacity (mAh·g−1)/Coulombic efficiency (%)/ Current density	Stable capacity (mAh·g−1)/ Number of cycles	Year	Ref.
SiO/Li	951/770/81/0.1 mA·cm−2	762/50	2007	[132]
SiO/SnO/Li	1338/901/67/0.1 mA·cm−2	574/200	2009	[90]
SiO/Sn30Co30C40	1030/1480/70%/300 mA·g−1	900/300	2012	[130]
SiO/Sn30Fe30C40	1343/900/67/300 mA·g−1	900/50	2013	[131]
SiO/Fe2O3	2773/1893/68/160 mA·g−1	1335/50	2013	[127]
C/Cr/SiOx	780/591/76%/0.5 C	490/50	2014	[123]
SiO/W/graphite	964/566/59/0.5 C	358/100	2015	[122]
SiO/ZrO2/C	1737/1157/67/80 mA·g−1	721/100	2016	[128]
SiO/Cu/graphite	1177/2473/48/50 mA·g−1	836/100	2017	[124]
SiO/TiO2	1070/137/78/100 mA·g−1	730/100	2019	[125]
SiOx/SnO2/C	2186/1158/53/100 mA·g−1	1127/100	2020	[129]

 | Show Table

DownLoad: CSV

5.3   Electrolyte additives for stabilization of SEI film on SiO_x electrodes

The discussed above serious problem related to the instability of SEI formed on SiO_x materials can be addressed by proper usage of the liquid electrolyte additives. In fact, the degradation of a battery involves the active material and the electrolyte. Regarding typically used liquid electrolytes, Lin et al. [133] found that LiPF₆ salt can be hydrolyzed even in the presence of a small number of water molecules, which generates harmful HF [Eqs. (1)–(3)] [133–134]:

The strong tendency of Si to form Si–F bonds promotes the reaction between SiO and HF, which produces more H₂O and thus causes further hydrolysis of LiPF₆ [134]. This process can destroy the carbon layer on the surface of SiO and prevents the formation of the stable SEI film, which ultimately leads to the ongoing destruction of the electrode.

Electrolyte additives are the most simple, economical, effective, and feasible way to help construct a stable SEI film on anode materials, preventing the continuous decomposition of the electrolyte [135]. Mitigating the damage of HF on SiO_x anodes is also crucial for the cycling performance of SiO_x-based electrodes.

Vinylene carbonate (VC) and fluoroethylene carbonate (FEC) are the most widely used additives in SiO_x-based LIBs [27,136–138]. VC was first reported as the additive for Si-based LIBs by Chen et al. [139] in 2006. Such addition facilitates the formation of a uniform, smooth, and flexible SEI film with polycarbonate as the main component to withstand the volume changes of SiO_x [134]. FEC can also form a stable SEI film on the surface of Si, which prevents the penetration of the electrolyte breaking the –Si–Si– network bonds [140] and restricts the decomposition of the electrolyte, the oxidation of Si [141], and even the diffusion of HF to some extent [142]. The main components in the SEI film formed by FEC are LiF and polyene compounds. LiF nanocrystals not only have well Li⁺ conductivity but also can enhance Li⁺ conductivity by the creation of interface defects between LiF and the organic matrix in the SEI film, which improves the rate performance of SiO_x electrodes [134,141]. However, VC and FEC have their own limitations. The VC additive usually causes a dense SEI film, which blocks the migration of Li⁺ and so limits the power density of the battery. The film formed by FEC is less flexible and cannot fit well with the volume changes of SiO_x, thus reducing the cycle life [134]. Our group employed LiNO₃ and VC as synergistic additives for electrodes and electrolytes, respectively, to improve the electrode/electrolyte interface properties. The reduction products of Li₃N and LiN_xO_y from LiNO₃ and polycarbonate from VC endow the SEI film with high ionic conductivity and enhanced elasticity, which enable a fast and uniform electrode reaction. The synergistic effect of LiNO₃ and VC enables the bare SiO electrode to deliver a high reversible capacity of 1062.3 mAh·g⁻¹ and an excellent cycling performance with 94.5% capacity retention and 99.80% coulombic efficiency for 160 cycles [143].

Many functional electrolyte or electrode additives have been developed to promote the formation of stable and firm SEI film. For example, LiF was introduced into the electrode to compensate for the Li⁺ consumption during SEI film formation [116]. 4,5-Difluoro-1,3-dioxolan-2-one shows strong electronegativity because of the two fluorine elements and thus can be easily reduced to form a solid SEI film on the electrode surface, thus improving the cycling performance of SiO_x electrodes [144]. LiBOB can decompose to form a more resistive SEI film, which can control the alloying depth of SiO and prevent the formation of crystalline Li₁₅Si₄ phase, thus improving the cycling performance of SiO electrodes [145]. Moreover, LiTFSI/Py13-TFSI can passivate the SiO_x electrode surface to form a stable, dense, and flat SEI film composed of pyrrolidinium, TFSI ions, and their decomposition products, effectively improving the cycling life of SiO_x-based anodes [146].

5.4   Methods for overcoming the low ICE of SiO_x-based anodes

In the initial lithiation of SiO_x, the ICE of the SiO_x anode can be much lower than that of pure Si or graphite because of the formation of the SEI and irreversible phases, such as Li₂O, Li₄SiO₄, and other lithium silicates. A large part of Li ions from the electrolyte and cathode is immobilized in the first lithiation, decreasing the availability of lithium and significantly restricting the practical applications of SiO_x. Many effective methods have been proposed to mitigate this fundamental problem. At present, the two most important strategies are prelithiation and prereduction.

5.4.1   Prelithiation methods

Prelithiation is a viable method to offset the irreversible Li consumption during cycling by introducing extra Li into electrodes in advance. The effective approaches reported include direct contact with metal lithium, chemical and electrochemical prelithiation, and cathode-side prelithiation.

(I) Direct addition of metal lithium powders.

Direct addition of metal lithium powders into the SiO_x electrode is commonly adopted for Si-based anodes to compensate for the large irreversible lithium consumption. The direct contact of metal lithium with active SiO_x particles makes SiO_x partially prelithiated, which allows the generation of irreversible products before the lithium comes from the cathode during the initial charge. This phenomenon avoids the irreversible consumption of cathode lithium. Stable lithium metal powder (SLMP) is an efficient prelithiation agent composed of lithium carbonate and lithium metal powder [48]. Li metal particles are uniformly coated by lithium carbonate, which forms a protective layer to handle the material in dry air safely. As such, pressure must be applied to break the lithium carbonate layer in the prelithiation procedure to activate the lithium metal powder. The complete process of the SLMP route is shown in Fig. 7(a) [147]. In 2005, Jarvis et al. [148] first reported the application of SLMP in the graphite anode for LIBs. By introducing 1.93wt% SLMP, the ICE of the graphite anode could increase from 77.9% to 95.4%, demonstrating the high effectiveness of this approach. Subsequently, SLMP prelithiation was applied to several LIB anodes, including Si [149] and SiO_x [150–151]. With 15wt% SLMP added, the ICE of a micro-sized SiO-based anode was improved from 68.1% to 98.5% [150].

Fig. 7.  Schematic of different prelithiation procedures: (a) SLMP; (b) chemical one; (c) electrochemical prelithiation and its scaled-up roll-to-roll process scheme; (d) schematic for the electrochemical reaction of prelithiation reagent Li₆CoO₄. (a) Reprinted from M.W. Forney, M.J. Ganter, J.W. Staub, R.D. Ridgley, and B.J. Landi, Prelithiation of silicon-carbon nanotube anodes for lithium ion batteries by stabilized lithium metal powder (SLMP), Nano Lett., 13(2013), No. 9, p. 4158. Copyright 2013 by American Chemical Society; (b) reprinted from G.W. Wang, F.F. Li, D. Liu, D. Zheng, Y. Luo, D.Y. Qu, T.Y. Ding, and D.Y. Qu, Chemical prelithiation of negative electrodes in ambient air for advanced lithium-ion batteries, ACS Appl. Mater. Interfaces, 11(2019), No. 9, p. 8699. Copyright 2019 by American Chemical Society; (c) reprinted from H.J. Kim, S. Choi, S.J. Lee, M.W. Seo, J.G. Lee, E. Deniz, Y.J. Lee, E.K. Kim, and J.W. Choi, Controlled prelithiation of silicon monoxide for high performance lithium-ion rechargeable full cells, Nano Lett., 16(2016), No. 1, p. 282. Copyright 2016 by American Chemical Society; (d) reprinted from M. Noh and J. Cho, J. Electrochem. Soc., 159(2012) [160]. ©IOP Publishing. Reproduced with permission. All rights reserved.

DownLoad: Full-Size Img PowerPoint

Due to the high specific capacity of metallic lithium (3860 mAh·g⁻¹), only a small amount of SLMP is sufficient to realize the prelithiation of SiO_x electrodes. Furthermore, SLMP has important advantages related to the facile and controllable prelithiation route. However, the high cost, environmental pollution, and potential safety problems limit its wide application.

(II) Chemical prelithiation.

As early as 2005, Tabuchi et al. [152] developed a Li-doped SiO active material (Li_xSiO) with the reduced irreversible capacity by a process involving the dissolution of naphthalene and metallic Li into butyl methyl ether solvent to form a Li-organic complex solution and then immersing the SiO electrodes in the abovementioned solution. It is a facile method to compensate for the irreversible lithium-ion loss that the anode materials react with reductive prelithiation reagents directly. The degree of prelithiation can be tuned by the immersion time. The process of chemical prelithiation is shown schematically in Fig. 7(b) [153].

In recent years, many prelithiation reagents have been exploited to improve the ICE of negative electrodes [152–154], and considerable progress has been made. Yan et al. [155] reported an effective strategy to achieve the uniform pre-lithiation of SiO_x/C microparticles by mixing with a lithium-biphenyl complex solution so that the ICE of half-cell was as high as 90%. However, such reagents usually show high reactivity in ambient air, which is challenging for practical applications.

(III) Electrochemical prelithiation.

The electrochemical approach is a relatively simple method to achieve the prelithiation of SiO_x electrodes. It can be divided into direct and indirect contact prelithiation routes [156–159].

When the lithium foil is in direct contact with SiO_x electrodes with also a presence of the electrolyte, electrons transfer from Li metal to SiO_x, with concurrent insertion of Li⁺, because of the potential difference between the lithium foil and SiO_x electrode, achieving the required prelithiation [156–157]. The method is simple and operable, but the prelithiation depth and uniformity are not satisfactory. In 2019, Meng et al. [158] optimized the prelithiation of the direct contact of the anode and Li foil, proposing to add a resistance buffer layer (RBL) between the Li foil and SiO_x electrode to achieve uniform prelithiation. The rate and depth of lithiation can be adjusted by controlling the resistance of RBL and the time of the process. When assembled in a full cell with NCM622 cathode, the ICE of the prelithiated NCM622-SiO_x full cell can increase from 68.9% to 87.3% with respect to the one without prelithiation.

As for the indirect contact prelithiation, a temporary battery must be constructed using SiO_x and Li metal as the positive and negative electrodes, respectively. The SiO_x electrode is discharged with a specific current to achieve the insertion of Li⁺ [159]. The degree of prelithiation can be controlled by circuit parameters, such as voltage and current density. However, constructing a temporary battery is time-consuming and brings additional costs, which is unrealistic for commercial applications. To overcome this issue, Kim et al. [159] proposed a modified electrochemical prelithiation where pressure on a temporary constructed cell was applied, and a roll-to-roll manufacturing line was employed, as shown in Fig. 7(c). The degree of prelithiation can be accurately tuned by controlling short circuit time and monitoring voltage. When the resistance reached 100 Ω and the short circuit time was 30 min, the carbon-coated SiO_x (denoted as c-SiO_x) electrode showed an optimized ICE of 94.9%. When assembled with NCA cathode, the energy density of the prelithiated c-SiO_x/NCA full cell was 155% higher than that of the pristine c-SiO_x/NCA battery.

(IV) Cathode-side prelithiation.

The ICE of SiO_x can be improved by introducing Li-containing cathode additives. As shown in Fig. 7(d) [160], the introduction of such additives can effectively compensate for the irreversible capacity loss of SiO_x and reduce the Li⁺ loss of the cathode. Cathode additives, such as Li₆CoO₄ [160], Li₃N [161], and Li₂O₂ [162], could supply extra lithium to compensate for the irreversible capacity loss of the anode. Noh and Cho [160] assembled a LiCoO₂/SiO_x full cell with 15wt% Li₆CoO₄ added to the LiCoO₂ cathode, which exhibits a reversible capacity ratio from 43% to 66%.

In general, cathode-side prelithiation provides more options for improving the ICE of SiO_x. Different additives should be matched with the potential of the cathodes so that the extra-lithium can be released in the charge window. Moreover, the additive amount should be well-controlled.

5.4.2   Prereduction method

As discussed above, in the initial lithiation process, the irreversible capacity loss of SiO_x mainly comes from the side reactions between Li and SiO_x and the formation of SEI. The former contributes a large part of irreversible capacity because of the formation of various electrochemically inactive lithium silicates and lithium oxide. As such, the initial oxygen amount x in the material strongly influences such capacity loss [120]. Therefore, SiO_x reduction by reacting with reductive metals, such as Al [163], Li [48,120], and Ni [121], has been proposed to form metal oxides and silicates. Apart from decreasing the x value of SiO_x and avoiding the lithium consumption from oxygen, the formed metal oxides and silicates can buffer the volume changes of SiO_x to improve its cycle performance.

Jeong et al. [163] prepared a nanostructured composite material with chemical composition SiAl_0.2O through HEMM using SiO powder and Al metal powder as the raw materials. The SiO₂ in the disproportionated SiO can be reduced to Si, and the electrochemically inactive Al₂O₃ is generated in HEMM, which can mitigate the cracking and fracture of the active material in the cycling process. It was shown that such formed SiAl_0.2O composite was composed of silicon nanocrystallites and an amorphous matrix consisting of SiO₂ and Al₂O₃. The ICE was increased by 10% when compared with unmodified SiO, and the SiAl_0.2O material exhibited a stable cycling performance with a reversible capacity of 800 mAh·g⁻¹ over 100 cycles at 120 mA·g^–1.

5.5   Alleviation of the macro-cracking effect of SiO_x electrodes

It is of great importance to ensure tight bonding of SiO_x particles and the conductive agent between each other, and with the current collector, to prevent loss of electrical contact. However, providing sufficient adhesion to maintain electrode integrity under large volume changes upon cycling is difficult because of the weak Van der Waals forces between SiO_x and typically used polyvinylidene fluoride (PVDF) binders. In addition, the usage of PVDF results in strong swelling and the toxicity of concomitant solvent N-methyl-2-pyrrolidone. Various binders have been employed in SiO_x-based electrodes and can be classified into two categories based on their functions: (1) improvement of adhesion through hydrogen bonding between the binder and SiO_x and (2) formation of a 3D cross-linked network to improve the mechanical properties of the electrode.

5.5.1   Hydrogen bonding for improving adhesion

The Van der Waals force between the binder and the SiO_x particle is a weak intermolecular force. Thus, it can be easily destroyed by large volume changes. By contrast, the hydrogen bond is relatively strong and shows some self-healing function in volume change environments [164–165]. Therefore, the design of binders with functional groups that can form hydrogen bonds with SiO_x particles has become an important aspect for the deployment of SiO_x in LIBs.

Sodium carboxymethyl cellulose (CMC) and sodium alginate (Alg) have been adopted as binders in SiO_x electrodes; these binders, which benefit from the presence of polar groups on their molecular chains, can provide stronger adhesion than PVDF by forming abundant hydrogen bonds [156,166–167]. This phenomenon is also widely observed in other biomass-originating binders, such as xanthan gum [168], guar gum [169], chitosan [170], β-cyclodextrin [171], and konjac glucomannan (KGM) [26]. For example, compared with Alg, KGM can form more hydrogen bonds with Si–OH on the surface of Si@SiO₂. Molecular mechanics simulation revealed that the interfacial adhesion energy between KGM and Si@SiO₂ is 0.205 J·m⁻², which is 0.018 J·m⁻² higher than that of Alg (Fig. 8(a) and (b)). The strong adhesion of binder with SiO_x active particle and the self-healing ability of hydrogen bond make the electrode tolerant to the mechanical stress caused by the large volume change in SiO_x upon lithiation/delithiation. Some other non-biomass adhesives, including poly(acrylic acid) (PAA) [172–173], polyvinyl alcohol (PVA) [174–175], and polyacrylamide [176], exhibit similar functions and can be used to improve electrode integrity.

Fig. 8.  (a) Schematic of lithiation and delithiation for KGM/Si@SiO₂: the surface oxide layer of SiO₂ on Si nanoparticles enhances the interfacial adhesion with the KGM binder, and it protects the Si cores; (b) structures of KGM/SiO₂, KGM/Si(111), and SA/Si(111) model systems from molecular mechanics simulation; (c) schematic of surface modification of c-SiO using PyOH and its interaction with PAA; (d) Ring-sliding motion in PRPAA and its chemical structure; (e) cycling performance of Py-c-SiO-PRPAA electrode at 1.78 mA·cm⁻². (a, b) S.T. Guo, H. Li, Y.Q. Li, Y. Han, K.B. Chen, G.Z. Xu, Y.J. Zhu, and X.L. Hu, SiO₂-enhanced structural stability and strong adhesion with a new binder of konjac glucomannan enables stable cycling of silicon anodes for lithium-ion batteries, Adv. Energy Mater., 8, 1800434 (2018) [26]. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission; (c, d) Y. Cho, J. Kim, A. Elabd, S. Choi, K. Park, T.W. Kwon, J. Lee, K. Char, A. Coskun, and J.W. Choi, A pyrene–poly(acrylic acid)–polyrotaxane supramolecular binder network for high-performance silicon negative electrodes, Adv. Mater., 31, 1905048 (2019) [178]. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

DownLoad: Full-Size Img PowerPoint

5.5.2   Formation of 3D cross-linked network

Morphological analysis suggests that the contact between binders and active materials has three main types: point–point, point-line, and point–net. The point–point shows the smallest contact area and the weakest adhesion. This contact type mainly involves styrene-butadiene rubber and poly(tetrafluoroethylene) binders. Most of the single-polymer adhesives form the point-line contact, which provides a large combination area and considerable adhesion strength. Among them, PVDF, CMC, Alg, PAA, and PVA are typical examples. Although the adhesion area of the point-net contact type is not much larger than that of the point-line contact type, it can effectively accommodate the volume changes in SiO_x through the deformation between the branches to form a 3D bonding network, showing the best adhesion performance [177]. Therefore, forming a cross-linked 3D adhesive network is another important area regarding the development of high-performance binders.

PAA is a widely used material for 3D cross-linked network construction. It can not only form hydrogen bonds with SiO_x particles but also partly bridge and cross-link itself by carboxylic groups in PAA molecules, forming a 3D network that could accommodate the large volume changes of SiO_x [172]. However, PAA is apt to self-polymerize to form stiff bundles through carboxyl groups, which would reduce the contact sites of the binder with active SiO_x particles. He et al. [99] in situ cross-linked polyacrylic acid and d-sorbitol binder to form a 3D network structure, which enables the SiO electrode to have improved cycling stability than that with PAA by providing more contact sites for active particles. In another research, Cho et al. [178] prepared polyrotaxane cross-linked PAA (PRPAA) as a binder. PRPAA is composed of cyclodextrin rings and poly(ethylene glycol) chains. Cyclodextrin rings covalently cross-link to PAA through the ester linkage and thread along the poly(ethylene glycol) chain, providing a sliding motion effect (Fig. 8(c) and (d)), which ensures a good elasticity of this binder. Based on this design, 1-pyrenemethanol (PyOH)-modified carbon-coated SiO (c-SiO) electrode with mass loading between 2.3 and 2.5 mg preserves a capacity of approximately 2.54 mAh·cm^–2 with 97.6% capacity retention after 100 cycles at 1.78 mA·cm⁻². In addition, 92.6% capacity retention was achieved after 250 cycles (Fig. 8(e))

6.   Conclusions and outlook

Without a doubt, SiO_x is the most potential anode material for high-energy-density LIBs, owing to the high specific capacity and manageable volume changes in lithiation/delithiation processes. However, several challenges remain toward the practical deployment of SiO_x in LIBs, including low ICE and poor cycling performance. The latter is associated with particle pulverization, unstable SEI film, and electrode cracking failure. To overcome these issues, great efforts have been exerted in recent years, which resulted in significant advancements in the electrochemical performance of SiO_x-based anodes.

A series of novel structures and composites was designed to overcome the material pulverization caused by volume changes. Introducing free space and combining SiO_x with the conducting phase (different forms of carbon, metallic elements, MO_x, and so on) are effective ways to improve electrochemical performance. The synthetic technical routes and specific processing parameters have a strong influence on particle morphology, microstructure, and chemical compositions (x value), which decisively affect the electrochemical performance demonstration of SiO_x-based electrodes.

The development of additives and novel binders is an important direction to improve the performance of SiO_x-based anodes. Additives for electrolytes or electrodes can enhance the stability of the SEI film on SiO_x and reduce the continuous consumption of the electrolyte. Meanwhile, the undesirable thickening of the SEI film and the increase in the internal resistance of the battery can be avoided. Apart from exploring novel additives, a combination of different electrolyte additives could be effective and useful. Binders maintain electrode integrity and relieve the performance degradation due to loss of the electrical contact between the active material and the conductive agent or the current collector. Future works could focus on the exploration of new binders with specific groups that have high adhesion energy with SiO_x/Si, self-healing function, and/or the ability to form a 3D cross-linked network. Functional conductive binders are also meaningful.

Prelithiation and prereduction are efficient ways to overcome the low ICE of SiO_x electrodes. Prelithiation technologies include the direct addition of metal lithium powders or prelithiation agents, chemical reaction, electrochemical prelithiation, and cathode-side prelithiation. Prereduction of SiO_x with metal is another route to increase the ICE and improve the cycle performance of SiO_x electrodes. In this regard, the cost, safety, and scalability are crucial factors affecting the practical application of these techniques in SiO_x-involved batteries.

Although SiO_x anodes are still facing critical challenges in route to commercialization, significant progress has been achieved in electrochemical performance in recent years. Future breakthroughs most likely related to the development of simple and low-cost synthesis routes and the overall improvement of the performance of SiO_x anodes shall ensure their wide usage in next-generation high-energy-density LIBs.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Nos. U1637202 and 52074023), the National Key R&D Program of China (No. 2018YFB0905600), and the Beijing Municipal Education Commission-Natural Science Foundation Joint Key Project (No. KZ201910005003).

Conflict of Interest

The authors declared that they have no conflicts of interest.