All-solid-state batteries (ASSBs) hold promise as a safer and higher-energy-density alternative to state-of-the-art lithium-ion batteries. One of the key challenges in the development of ASSBs is the understanding and control of the interfaces between the solid electrolyte and the electrodes (the right figure^[1]). (Electro)chemical reactions take place at unstable battery interfaces and produce chemically-distinct interlayers known as interphases. The interphases' microstructural, thermodynamic, electrochemical, and ion-transport properties are crucial to the device performance such as lifetime, reliability, and (dis)charge rate of ASSBs. The group aims to understand and optimize the electrolyte-electrode interface in ASSBs using first-principles calculations aided with data-driven methods, and to realistically address the remaining interface-related key challenges for full-scale commercialization of ASSBs.

Example 1. First-principles investigations of solid-electrolyte surfaces and interfaces.^[1][2]

Relevant references on first-principles mechanism investigation

1. Y. Li, P. Canepa, P. Gorai, PRX Energy (2022)
2. Y. Li, A. Prabhu, T. Choksi, P. Canepa, J. Mater. Chem. A (2022)
3. Y. Lei#, Y. Li# et al., Nature (2022)
4. Y. Chen#, Y. Lei#, Y. Li et al., Nature (2020)

The urge to advance sustainable energy applications calls for insights into the microscopic mechanisms of phonomena related to materials interfaces, defects, and lattice vibrations. The theoretical investigation of these complex properties often requires long time and high computational cost for accurate first-principles calculations. Specialized machine learning models integrated with expert domain knowledge could speed up and scale up the investigation. Our group builds accurate databases from high-throughput (HT) computation in a fully automated fashion. Based on the in-house databases, we develop graph neural networks (GNNs) to achieve interpretable predictions of technologically relevant properties and to explore the design principles of energy storage and conversion materials (e.g., solid electrolytes, halide perovskites).

Example 1. High-throughput computational data as the foundation for materials design - perovskite derivatives.^[1]

Example 2. A database of computed Raman spectra of inorganic compounds with accurate hybrid functionals.^[2]

Relevant references

1. Y. Li and K. Yang, Energy Environ. Sci. (2019)
2. Y. Li, D. Lee, P. Cai, Z. Zhang, P. Gorai, P. Canepa, Sci. Data (2024)
3. Y. Li and K. Yang, WIREs Comput. Mol. Sci. (2021)

Aqueous zinc-ion batteries are competitive for grid-scale stationary energy storage applications due to their intrinsic safety and low cost. However, the narrow electrochemical stability window of the water-based electrolyte raises concerns about the stability and reactivity of the liquid-solid interfaces. Specifically, hydrogen evolution reactions result in corrosion of the zinc anode and reduce the Faradaic efficiency of the battery; non-uniform electro-deposition and dissolution of zinc may induce dendrite growth. The group aims to understand the kinetic processes at the aqueous battery interfaces using ab initio molecular dynamics and to improve the battery stability and performance.

Relevant references

1. S. Liu#, Y. Li#, D. Wang# et al., Nat. Commun. (2024)