Assessment and process optimization for novel gas materials through the evaluation of mixed response surface models

In this paper an innovative methodology aimed at improving the development of novel gas sensors through a process optimization is carried out by applying mixed response surface models. High accuracy measurements of new conductometric metal oxide gas sensors, obtained by an efficient control of the working conditions, are gathered. The response of Metal Oxide Semiconductors (MOX) gas sensors changes significantly when the sensors operate at different temperatures and target gas concentrations. In order to consider all the sources of variability there involved, the response surface methodology was applied, including random effects, to improve and optimize the performance of these new gas sensors. More precisely, the optimization is performed exploiting a limited number of observations, systematically collected with an ad-hoc measurement system, and it takes into account external sources of variability, satisfying at the same time stringent requirements. Furthermore, the statistical results and the relative assessment of novel gas materials are obtained by considering fixed as well as random effects, where random variables are taken into account for better controlling the optimization step.