Machine Learning-Assisted Active Center Exploration in Atomically Thin MoS_xTe_2-x Electrocatalysts for Efficient...

Area of Science:

• Materials Science
• Catalysis
• Artificial Intelligence

Background:

• Optimizing 2D molybdenum disulfide (MoS2) catalysts for the hydrogen evolution reaction (HER) involves modulating local atomic configurations.
• Transmission electron microscopy (TEM) provides atomic resolution but lacks automated methods for identifying active sites within complex micrographs.

Purpose of the Study:

• To develop a rapid and accurate machine learning (ML) framework for automated active site discovery in catalytic materials.
• To identify novel defect configurations in MoS2-based catalysts and understand their impact on HER performance.

Main Methods:

• Fabrication of a defective MoSxTe2-x alloy catalyst (S-MoTe2) via low-temperature sulfurization of 1T'-MoTe2.
• Application of an unsupervised ML framework utilizing Zernike features and uniform manifold approximation and projection (UMAP) for atomic structure exploration and clustering.
• Validation of ML-identified defects using Density Functional Theory (DFT) calculations and electrochemical experiments (HER overpotential and Tafel slope measurements).

Main Results:

• Discovery of a novel antisite Te adatom (Te_ads-Mo) defect configuration in the S-MoTe2 alloy.
• DFT calculations confirmed synergistic enhancement of hydrogen adsorption and electronic conductivity at the antisite defects.
• Experimental verification showed S-MoTe2 with Te_ads-Mo defects exhibited a halved overpotential and improved Tafel slope for HER compared to controls.

Conclusions:

• The developed ML framework offers an intelligent approach for efficient active center exploration in catalyst micrographs.
• The study demonstrates a closed-loop verification process for ML-assisted defect discovery, integrating theoretical calculations and experimental validation.
• This work highlights the seamless cooperation between ML tools and researchers in advancing catalyst development for HER.