Sensitive Impurity Detection and Recognition in Inert Gaseous Media with SnO2 Based Gas Sensor Microarray

Abstract

Contrary to the conventional use of semi-conductive metal oxides as gas detectors for components in air the analytical performance of segmented SnO₂-layers were investigated to detect and discriminate impurities in inert gases such as nitrogen or helium. A broad variety of technical processes require high purity inert gases which have to be kept under permanent surveillance to detect leaks to the atmosphere, a gas release of contaminated fittings or other contaminations. The unique gradient gas sensor microarray of the Karlsruhe micronose (KAMINA) was equipped with segmented layers of either nanoparticle or sputtered SnO2 layers and exposed to a high purity nitrogen flow of 1lit/min into which temporarily model impurities have been admitted with a variety of concentrations. Pulsed exposures were performed to nitrogen containing oxygen (0.2 to 10 vol %), water vapour (5 to 15 % relative humidity) and 2-propanol as well as tetrachloromethane in the ppm range. For all model contaminations a fast and sensitive response has been found. The measured concentration dependence revealed a detection limit of about 0.01 vol% for oxygen and 0.03% of relative humidity for an operating temperature of the KAMINA chip at 325..273 °C. The detection limits for the organic compounds were even found to be in the lower ppb range. The conductivity patterns obtained from the simple segmentation microstructure of the KAMINA SnO2 layers operated with a temperature gradient were evaluated by a Principal Component Analysis (PCA) which proved an appropriate gas discrimination power to distinguish pure nitrogen and the different kinds of contaminations for both types of SnO2 layers. However, the sputtered layer showed better detection limits while the nanoparticle layers offered a better sensitivity, i.e. a higher slope of the concentration dependence.