Integration of machine learning approaches for accelerated discovery of transition-metal dichalcogenides as Hg0 sensing materials

Abstract

The detrimental impact of urban airborne Hg0 from fossil fuel utilization has necessitated the discovery and development of Hg0 sensing materials for effective Hg0 detection and mitigation of the pollutant. Earlier studies have hypothetically and experimentally supported 2-dimensional transition-metal dichalcogenides (2D TMDCs), particularly MoS2 to have excellent performance for Hg0 removal. However, the potential of other TMDCs is yet to be investigated for Hg0 sensor application. In this study, a total of 28 transition metals within periods 4–6 of the periodic table, excluding the lanthanides series, were examined. To ensure proper data management flow, a high-throughput data mining approach with integrated machine learning and cheminformatics simulation approaches is developed. The systemic approach integrates the Pymatgen, Factsage, Aflow and density functional theory simulation tools for accelerated discovery of suitable TMDCs from raw data via the chemical vapour reaction route. Predicted results showed that TiS2, NiS2, ZrS2, MoS2, PdS2 and WS2 exhibited TMDCs characteristics. Furthermore, first-principles calculation shows Hg-uptake capacity is in the order NiS2 > PdS2 > TiS2 > ZrS2 > WS2 > MoS2, while Hg sensing response is in the order PdS2 > MoS2 > WS2 > ZrS2 > NiS2 > TiS2. Accordingly, PdS2 depicted to be the most suitable TMDCs for airborne Hg0 sensor application. The proposed systemic approach is an initial platform for materials discovery using integrated machine learning approaches and is well-suited for the screening and the discovery of new materials based on component-oriented structures.