Si attracts significant attention as an alternative anode material to replace the conventional graphite anode in lithium-ion batteries (LIBs). Si offers extremely high theoretical energy-storage capacity of 4200 mAh/g for Li_4.4Si, while the theoretical energy-storage capacity of graphite is only 372 mAh/g for Li/C₆. In addition, Si is abundant, eco-friendly, and non-toxic, and it also has a safe thermodynamic potential with an average voltage of about 0.4 V vs. Li⁺/Li, making them attractive candidates for LIB anodes. However, the capacity of the Si anode exhibits a relatively high initial charge capacity and then rapidly decreases as the charge/discharge cycle proceeds, hampering its extensive application. This rapid performance degradation can be attributed to the poor connection between the active material and the current collector resulting from the severe volume expansion and contraction of the Si particles during the repeated charge and discharge cycles, respectively.

Controlled addition of O to Si is often taken as a viable approach to utilize Si as an anode material. Silicon suboxides (SiO_x, 0 < x < 2), compared with Si, offer a low energy capacity but a high stability against volume expansion. As a result, silicon monoxide (SiO) anodes are indeed commercialized by being blended with graphite. However, only low amounts (about 5 wt%) of SiO are blended with graphite because of the poor first cycle efficiency of SiO. During the charge process, SiO_x reacts with Li and produces in-situ byproducts such as Li₂O and Li-silicates, which are irreversible in following discharge cycles. This irreversible Li consumption is known to decrease the initial Coulombic efficiency (ICE) of SiO, typically below 75%. The low ICE of SiO_x along with its inherently poor electronic conductivity leads to poor rate performance and increased capacity decay, hindering its substantial application in LIBs. To realize more extensive application of SiO_x for LIB anodes, physical properties of SiO_x should be improved in a direction to mitigate the low ICE issue.

In this study, we investigate the motion and dynamics of a Li atom in SiO_x by performing a series of first principles-based atomistic simulations. To this end, we perform Monte Carlo simulations within a continuous random network model to generate realistic SiO_x structures with varying O-to-Si ratios. Then, we implement a density-functional theory calculation-based path sampling scheme to obtain detailed thermodynamic information when a Li atom penetrates a SiO matrix. Subsequent electronic structure analysis reveals that the thermodynamic stability of a Li atom is determined by local Si-O network environment surrounding Li. The identified thermodynamic information regarding the Li dynamics in SiO provides additional insight into the origin and solutions of the low ICE, highlighting the importance of controlling the SiO morphology during the synthesis of SiO. This fundamental understanding can be an important theoretical basis for developing practically applicable high-ICE silicon suboxide-based anode materials.