Machine-Learning-Assisted Density Functional Theory Calculations: A New Approach to Screening Thermal Runaway Gas...

Area of Science:

• Materials Science
• Computational Chemistry
• Battery Technology

Background:

• Lithium-ion batteries pose safety risks, including fire and explosion, due to thermal runaway.
• Thermal runaway releases combustible gases like methane, carbon dioxide, hydrogen, and carbon monoxide.
• Early detection of these gases is crucial for preventing battery accidents.

Purpose of the Study:

• To develop a machine learning (ML) approach for efficiently identifying gas-sensitive materials.
• To screen candidate materials for detecting specific gases released during lithium-ion battery thermal runaway.
• To reduce the time and cost associated with discovering novel gas-sensing materials.

Main Methods:

• Utilized density functional theory (DFT) calculations to determine stable configurations of target gases on doped MoS2 substrates.
• Constructed and evaluated eight machine learning models to predict adsorption energy.
• Analyzed feature importance within the optimal ML model.
• Validated the sensing performance of selected materials.

Main Results:

• Identified optimal machine learning models for predicting material adsorption energies.
• Determined the importance of various features in the ML prediction process.
• Screened and validated the performance of doped MoS2 materials for gas sensing.
• Successfully combined DFT with ML for efficient material discovery.

Conclusions:

• Machine learning, combined with DFT, significantly accelerates the screening of gas-sensitive materials for lithium-ion battery safety.
• This integrated approach offers a novel and cost-effective method for rapid nanomaterial screening.
• The study provides a pathway for developing advanced gas detection systems to prevent battery thermal runaway incidents.