The effects of microstructure on the gas sensing properties of chromium titanium oxide

• Main Author: Naisbitt, Simon Carl
• Published: University College London (University of London) 2005
• Subjects: 620.18930429
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.420304

The general principles relating to construction and operation of semiconducting metal oxide materials as gas sensitive resistors are discussed. The adsorbed oxygen model of sensor response is presented as the origin of the gas sensitivity of the conductivity of these materials. Microstructure is also noted to affect sensor response and various studies, simulations and models from the literature are presented and discussed. Based on a combination of the model developed by Williams et al. and the linear form of the empirical equation for sensor response the model that will be used in the present work is developed. Two batches of sensors made using two chromia raw material powders of differing particle size were obtained from City Technology Ltd. Groups of these sensors were sintered at different temperatures to generate sensors having different microstructures. The different sintering temperatures produced clear gradations in the microstructures of the sensors, observed using scanning electron microscopy. These sensors were then systematically tested to carbon monoxide (CO) and propane at a single operating temperature of 400 C. A further group of sensors created from the batch of sensors made from fine chromia raw material were sintered at a single temperature of 775 C but were operated at temperatures of 350, 400 and 450 C and tested to CO with exposures of 1800 second duration. The results of all the gas tests were successfully processed so as to obtain solutions to the microstructural model developed for the steady state and transient cases. The trends of the microstructural, sensitivity and time constant parameters with sintering temperature, microstructure, operating temperature and exposure length are successfully discussed with reference to how the model accounts for these trends. The unexpected or unusual results are explained in terms of potential inaccuracies in the model or how it has been applied in this work.