Data-Driven Search for the Optimal Ag–Pd–Pt-Based Electrode Alloy Chemistry for ZnO-Based Methane Sensor

摘要

Abstract The present study aims to explore the favorable chemistry of catalytic metal contact in Ag–Pd–Pt alloy system for sensing low concentration of CH4 gas (0.5–1.5) at low temperature range (100–125?°C) by using data-driven computing techniques. Artificial neural network (ANN) model has been developed considering ZnO particle size, alloy chemistry, temperature and CH4 as the input parameters and response time, recovery time and response magnitude as output parameters. The ANN model is used as the objective function for optimization of alloy chemistry employing multi-objective genetic algorithm. The alloy chemistry suggested by the Pareto solution has been used as the contact electrode for experimental fabrication of sensing device. The ZnO–p-Si hetero-junction structure has been fabricated by sol–gel technique and explored as a potential methane sensor. The proposed approach suggests the scope of using data-driven techniques to explore the wider range of alloy chemistry of the catalytic metal contact keeping in view the time and cost constraints in experimental search and/or theoretical deduction of target alloy chemistry in multi-component systems.