Accelerated Screening of Highly Sensitive Gas Sensor Materials for Greenhouse Gases Based on DFT and Machine Learning...

Area of Science:

• Materials Science
• Environmental Science
• Computational Chemistry

Background:

• Greenhouse gases (GHGs) pose significant environmental risks, necessitating effective detection methods.
• Developing efficient and cost-effective gas sensor materials is crucial for monitoring GHGs like CO2, CH4, N2O, and SF6.
• Traditional screening methods for gas sensor materials can be time-consuming and expensive.

Purpose of the Study:

• To propose a high-throughput, cost-effective strategy for screening gas sensor materials for key GHGs.
• To leverage machine learning (ML) combined with density functional theory (DFT) for rapid material identification.
• To identify optimal transition metal-doped WSe2 (TM-WSe2) monolayers as potential gas sensor candidates.

Main Methods:

• Utilized DFT to construct adsorption structures of four target GHGs (CO2, CH4, N2O, SF6) on 28 TM-WSe2 monolayers.
• Implemented and evaluated ten fine-tuned ML models to predict adsorption energy (Eads) and distance (D).
• Validated promising materials by analyzing band structure, work function, and predicted recovery time.

Main Results:

• Successfully screened a large number of candidate TM-WSe2 materials using ML-DFT approach.
• Identified optimal ML models for predicting gas adsorption properties with high accuracy.
• Selected promising TM-WSe2 materials exhibiting favorable gas-sensing characteristics.

Conclusions:

• The ML-DFT strategy offers a viable, low-cost, and efficient method for rapid screening of ideal gas sensor materials.
• The study demonstrates the effectiveness of artificial intelligence in accelerating materials discovery for environmental monitoring.
• Identified TM-WSe2-based materials show potential for practical application in GHG detection.