No abstract

This work aims at estimating the ability of a sensor system to recognize the type and to determine the concentration of organic compounds emitted from solvents. The measuring system used in the experiment consisted of a sensor array and a unit for data analysis. Commercially available sensors (Figaro, TGS) were used as a basis for the construction of the first of the two elements. The second element utilized Discriminant Function Analysis (DFA) and Multiple Linear Regression (MLR). The investigated gas samples consisted of air mixed with variable amounts of two organic compounds. One of the two compounds was intended to have a dominant influence on the sensors' response. Depending on this substance, two groups of gaseous mixtures were considered. In the first one, hexane was the major compound. The choice of this particular group of chemical compositions was inspired by practical problems related to various applications of extraction solvents. In the case of the second group of gas mixtures, the prevailing compound was toluene. This type of mixture, on the other hand, is related to the use of paints. The data analysis unit was designed to achieve the following two goals: in the first place, to recognize binary gas mixtures and secondly, to determine the concentrations of VOCs in them. The data analysis was based on steady-state sensor responses. Our study demonstrates that the sensor system may be employed for classifying and quantitatively determining volatile organic compounds released into air as a result of the use of solvents. The presented solution proves the concept in the case of binary mixtures of VOCs.

No abstract

No abstract