Spectrophotometric methods for studying the composition of mixtures are considered. The need to increase the stability and accuracy of the results of spectrophotometric measurements when analyzing substances with low absorption coefficients or a low concentration of the determined substance is shown. A spectrophotometric complex with a thermostated cell is described, which enables to study gaseous and liquid mixtures. Results of study at this complex of absorption spectra of isomers of liquid xylene and their mixtures are given. Methods of increasing the stability of spectrophotometric measurement results, hardware and data processing methods necessary for this are considered. It has been shown that thermostating of not only vaporous, but also liquid samples is one of the necessary conditions for ensuring the stability of measurement results. The procedure for performing experiments and processing the results of spectrophotometric measurements is presented. A differential method has been developed to reduce the impact of spectrophotometer signal instabilities on the accuracy of determining the composition of mixtures. The method consists in using as input data for multiple linear regression not the absorption spectrum, but its wavelength derivative. An analysis of the composition of mixtures of two xylene isomers showed that the use of this method can reduce RMS of determining the composition of mixtures by about 1.6 times. It is noted that additional filtering (smoothing) of the measured absorption spectrum derivative may be required to reduce errors in determining the composition of mixtures. The developed method of reducing the influence of the spectrophotometer signal instabilities on the results of the mixture composition analysis is of the greatest interest for the analysis of substances with low absorption coefficients or in the case of low concentrations of the studied substances and can be used to solve environmental monitoring problems, for example, when determining hydrocarbon contaminants in the atmosphere.
Read full abstract