Xiang Chen

The electrolyte, like “the blood for a body," is one of the most important parts of a battery and largely determines its practical performance. However, conventional trial-and-error approaches are facing tremendous challenges for electrolyte innovation due to the almost infinite molecular space and strong correlation effects between electrolyte species.

Xiang Chen focuses on the artificial intelligence design of advanced lithium battery electrolytes to overcome the disadvantages of conventional trial-and-error experimental and theoretical approaches with low efficiency, high expense, and long period. With data-driven approaches, especially machine learning, he is reconstructing the electrolyte design paradigm and promoting the upgradation of next-generation lithium batteries.

First, based on big data methods, Chen for the first time established the ion–solvent model to probe such interactions including the cation−solvent, cation−anion, and anion−solvent interactions, which are essential to the rational design of advanced electrolyte design. He has developed high-throughput electrolyte computation software, constructed the global largest electrolyte database, and developed deep learning potential for predicting electrolyte properties.

In addition, a machine learning model for electrolyte property prediction was developed, and new electrolyte molecules with great potentials were predicted through molecule generation models and high-throughput screening methods, which can significantly decrease the research period of electrolyte design and afford new protocols for high-energy-density and fast-charging lithium batteries.

Now, Chen’s team is cooperating with a number of lithium battery, electrolyte, and technology companies to promote the industrial applications of new lithium battery electrolyte.