Classifying combustible gases using micro-gas sensor array

A micro-gas sensor array with the four porous tin oxide thin films with noble metal catalysts on a micro-hotplate, dangling in air by Pt bonding wires and controlling the thickness by chemical mechanical process (CMP) has been fabricated. The sensing properties of sensors to some combustible gases, i.e. propane, butane, LPG, and carbon monoxide (CO), were evaluated. And we employed a principal component analysis (PCA) and a multi-layer neural network for classification of the gas species. The simulation and experimental results showed that the multi-layer neural network employing back-propagation learning algorithm with the micro-sensor array is very effective in identifying the kinds and some concentration levels of the combustible gases within the range of threshold limit values (TLVs). (C) 2003 Elsevier B.V. All rights reserved.