The Properties of Nanocrystalline Metal Oxide Semiconductor - based Gas Sensor Materials and Gases Identification by Neural networks and Multi-gas Sensors

• Main Authors: Shah-Jye Tzeng, 曾世杰
• Other Authors: Hong-Ming Lin
• Format: Others
• Language: en_US
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/71818965112719693120

博士 === 大同大學 === 材料工程研究所 === 89 === ABSTRACT Metal oxide gas sensing materials are known those have high sensitivity due to inflammable gases in air such as CH4, LPG. There are some succeeding examples with thin or thick film gas sensors in detecting these gases have been reported. In this study, Nanocrystalline metal oxide semiconductor gas sensor array had been constructed. There are three components in this gas sensor array: ZnO-based, WO3-based, and Pd doped WO3-based gas sensors. While completely finished nanocrystalline gas sensors, the mean particle size of ZnO-based gas sensors is about 80 nm; and the WO3-based is about 30 nm. When detecting NO2 gas, the sensitivity of our NC ZnO-based sensor approached to 128, this result is the highest value in the current reports. We constructed these three gas sensors into detected chamber and then detecting various gases. The selective problem is the defect of the most metal oxide semiconductor gas sensors. Added noble metal, such as Pt, Pd, might increase selective property. Artificial neural network technique to identify various gases has been examined in this study. The neural network can recognize CO, CO2, NO2 and CO + NO2 precisely.