Silicon is considered as one of the most promising lithium anode materials due to its high lithium storage capacity. However, the volume expansion effect and semiconductor properties of silicon hinder its practical application. Two effective strategies exist for mitigating this issue, i.e., reducing the size of silicon to nanoscale; and introducing non-active matrices to improve stability and conductivity. In recent decades, the incorporation of iron silicide (FeSi₂) with silicon (Si) has attracted much attention due to its high conductivity, mechanical hardness and environmental friendliness. Unfortunately, the Si/FeSi₂ nanocomposites are mainly prepared by high-energy ball milling, which has some drawbacks such as low purity, large particle size, and uncontrollable morphology, making it difficult to achieve the commercial application. Porous silica derived from clay minerals is extensively investigated for the preparation of silicon nanoparticles by magnesiothermic reduction. In addition, porous silica has a large specific surface area, which facilitates the effective dispersion and loading of iron oxides. The reduction of iron oxide and silica dioxide to Fe and Si can be achieved by magnesium. In this paper, a natural montmorillonite (Mnt) with a high content of silica (65%, in mass) was utilized as a raw material. Silicon/iron silicide (Si/FeSi₂) nanocomposites were synthesized via modification of Mnt (including acid etching and iron loading) and molten salt-assisted magnesiothermic reduction.

A porous silica (SiO₂) was prepared via leaching Mnt at a high concentration of HCl. Subsequently, a iron oxide was loaded on SiO₂ via precipitation in different Fe/Si mass ratios (i.e., 0, 10%, and 20%). Modified SiO₂ (Fe-SiO₂) was mixed with Mg and NaCl in a mass ratio of 1.00∶0.85∶3.00. Afterwards, the mixture was heated in an inert atmosphere at 650 ℃ for 5 h in a heating rate of 5 ℃/min. After cooling naturally, the product was washed with 1 mol/L HCl and 1% (in mass) HF in sequence to remove the by-product. Finally, it was rinsed with ultra-pure water until neutral and dried at 80 ℃. The sample were named as Si, Si/FeSi₂-1, and Si-FeSi₂-2, corresponding to the Fe/Si mass ratio of 0, 10%, and 20%, respectively.

The XRD patterns show that Si/FeSi₂ nanocomposites are consisted of Si and ɑ-FeSi₂. An increase in Fe content in precursors leads to a higher α-FeSi₂ content in the product. Based on the results of RIR value method and elemental composition analysis, the mass ratio of FeSi₂ to Si in Si/FeSi₂-1 and Si/FeSi₂-2 is 5% and 10%, respectively. The average grain sizes of Si, Si/FeSi₂-1, and Si/FeSi₂-2 are 20, 40 nm, and 48 nm, respectively, according to Scherrer's equation,. The grain size of silicon increases with the augmentation of Fe content in the precursor. This phenomenon analyzed by both transmission electron microscopy and scanning electron microscopy images is since the melting temperature of silicon nanoparticle is lower than that of bulk silicon (~1 410 ℃), the addition of iron greatly increases the local temperature (ΔH_Fe2O3= -309 kJ/mol vs. ΔH_SiO2= -130 kJ/mol), and the addition of Fe can effectively reduces the melting point of nanocrystalline grains in Fe/Si eutectic alloys. Moreover, the specific surface areas of Si, Si/FeSi₂-1, and Si/FeSi₂-2 are 89, 42 m²/g, and 33 m²/g, respectively. The gradual decrease in specific surface area coincides with the concurrent increase in grain size.

As anode electrodes, Si/FeSi₂ nanocomposites exhibit a higher initial coulombic efficiency, compared with Si due to the smaller specific surface area. The rate performance show that Si exhibits higher capacity than Si/FeSi₂-1 and Si/FeSi₂-2 at a low current density of 0.1 A/g. However, when the current density increases gradually, the capacity of Si is lower than that of Si/FeSi₂-1 and Si/FeSi₂-2. This phenomenon indicates that the incorporation of ɑ-FeSi₂ enhances both electron conduction and ion diffusion rates. After undergoing 500 cycles at a current of 1 A/g, the capacities of Si, Si/FeSi₂-1, and Si/FeSi₂-2 are 417, 781 mA·h/g, and 561 mA·h/g, respectively. The corresponding capacity retention rates are 35%, 77%, and 50%, respectively. The cycling performance demonstrates that Si/FeSi₂ nanocomposites have a more robust structure, compared to the Si electrode. In addition, the cyclic stability of Si/FeSi₂-1 is inferior to that of Si/FeSi₂-2, potentially due to the reduced structural stability resulting from a larger particle size.

The Si/FeSi₂ nanocomposites were prepared via modification of montmorillonite and molten salt-assisted magnesiothermic reduction. This method utilized magnesium powder to reduce Fe₂O₃ and SiO₂ for the in-situ formation of Si/FeSi₂ nanocomposites. The composition of Si and FeSi₂ nanocomposites could be adjusted via altering the ratio of Si/Fe in the precursor. Moreover, the reduction of Fe₂O₃ resulted in an increase in grain size and a decrease in specific surface area of silicon nanoparticles. As anode materials, Si/FeSi₂ composites demonstrated the superior rate performance and cycle stability, compared to Si. This phenomenon could be due to the incorporation of ɑ-FeSi₂, effectively enhancing the kinetics of lithium ions and electrons on the electrode. In addition, as an inert substance, FeSi₂ could reduce the volume expansion effect of silicon, thereby maintaining the stability of electrode structural. The structural stability of the Si/FeSi₂-2 electrode was relatively weaker than that of the Si/FeSi₂-1 electrode. Therefore, Fe content in the precursor could be of paramount importance for achieving high-performance silicon-based nanocomposite anodes.