In this paper, we describe a technique for the qualitative and quantitative analysis of such gas mixtures as “hydrogen sulfide in air” and “methane in air” using temperature modulation of a single metal oxide sensor. Using regression analysis in the principal components plane (PC1, PC2), we performed a selective determination of analytes on the minimum set of their concentrations in the training set, which is essential for solving practical problems. An important feature of this work is the difference in test gas concentrations from their concentrations in the training set. For the qualitative analysis of gas mixtures in a wide range of concentrations, we have developed an improved method for processing arrays of multidimensional data. For this improvement, we form a Mahalanobis neighborhood for polynomial regression lines constructed from the projection of training samples for each analyte on the (PC1, PC2) plane. Using the temperature modulation mode for the metal oxide sensor allowed us to increase its response when determining hydrogen sulfide by two to four orders of magnitude compared with the constant temperature mode.

The application of a non-stationary regime of temperature modulation in metal oxide semiconductor sensor based on SnO₂–Ag leads not only to a strongly increased sensor response, but also to a considerably improved sensor selectivity toward hydrogen sulfide. Selectivity with respect to other reducing gases (CO, NH₃, H₂) is about five orders of magnitude, enabling a correct selective determination of H₂S in the presence of interfering gas components.

Using the successively inverted projection method, we studied the dynamics of ultralean hydrogen–air flames propagating freely in a horizontal cylindrical Hele–Shaw cell. To quantify the two revealed characteristics of the flame dynamics—the dependence of the average flame velocities on time and the dependence of the initial flame velocity on the stoichiometry of the initial hydrogen–air mixture—we proposed time and stoichiometric scaling relations. The first relation approximates the dependence of the path of the flame front in hydrogen–air mixtures with an initial hydrogen concentration exceeding a certain critical value. The second relation approximates the dependencies of the initial flame front velocities on the hydrogen concentration. The general relationships for topologically different types of ultralean hydrogen–air flames can be interpreted as additional evidence of the presence of a general mechanism for the transition from discrete fronts of isolated drifting ball flames to a quasi-continuous deflagration flame front through a cascade of bifurcations.

In this work, the kinetics and mechanisms of the interaction of carbon monoxide and hydrogen with the surface of a nanosized SnO₂–PdO_x metal oxide material in air is studied. Non-stationary temperature regimes make it possible to better identify the individual characteristics of target gases and increase the selectivity of the analysis. Recently, chemometric methods (PCA, PLS, ANN, etc.) are often used to interpret multidimensional data obtained in non-stationary temperature regimes, but the analytical solution of kinetic equations can be no less effective. In this regard, we studied the kinetics of the interaction of carbon monoxide and hydrogen with atmospheric oxygen on the surface of SnO₂–PdO_x using semiconductor metal oxide sensors under conditions as close as possible to classical gas analysis. An analysis of the influence of catalytic surface temperature on the mechanisms of chemisorption processes allowed us to correctly interpret and mathematically describe the electrophysical characteristics of the sensor in the selective determination of carbon monoxide and hydrogen under nonstationary temperature conditions. The reaction mechanism is applied as well to the analysis of the operation scheme of the CO sensor TGS 2442 of Figaro Inc.