Use of time-dependent chemical sensor signals for selective identification

Abstract

A sensor is described for identifying and quantifying different components in a gas mixture by decoding information from a modulated output signal. The system operates by passing the unknown gas into a reaction chamber containing a heated catalytic filament. The output products of the chamber are detected by an electrochemical sensor. The concentration of the reaction products is modulated by varying the temperature of the catalytic filament periodically. A theory is presented for analyzing the modulated output signal. The validity of the analysis is confirmed by experiments conducted with benzene, CO and HCN, and by a computer simulations. It appears to be possible to identify different species in the inlet gas stream provided the species undergo chemical reactions in the reactor that have different activation energies. The signal magnitude at peaks or valleys in the output is directly proportional to concentration, even though a steady-state condition is not reached. The approach may be generalized to other detector systems.