Optimization of Gas Sensors Based on Advanced Nanomaterials through Split-Plot Designs and GLMMs

Abstract

This paper deals with the planning and modeling of a split-plot experiment to improve novel gas sensing materials based on Perovskite, a nano-structured, semi-conductor material that is sensitive to changes in the concentration of hazardous gas in the ambient air. The study addresses both applied and theoretical issues. More precisely, it focuses on (i) the detection of harmful gases, e.g., NO 2 and CO, which have a great impact on industrial applications as well as a significantly harmful impact on human health; (ii) the planning and modeling of a split-plot design for the two target gases by applying a dual-response modeling approach in which two models, e.g., location and dispersion models, are estimated; and (iii) a robust process optimization conducted in the final modeling step for each target gas and for each gas sensing material, conditioned to the minimization of the working temperature. The dual-response modeling allows us to achieve satisfactory estimates for the process variables and, at the same time, good diagnostic valuations. Optimal solutions are obtained for each gas sensing material while also improving the results achieved from previous studies.

Keywords

• experimental design
• Generalized Linear Mixed Models
• metal oxide semiconductors
• resistive gas sensors
• robust process optimization