SIMULTANEOUS UV-SPECTROPHOTOMETRIC DETERMINATION OF BENZENE TOLUENE AND XYLENE WITH APPLICATION OF ARTIFICIAL NEURAL NETWORK

The problem of serious overlapping of the UV-spectra of benzene,toluene and xylene in their UV-spectrophotometric determination was solved satisfactorily by the application of the principle of three layers artificial neural network and the back-propagation algorithm,and simultaneous determination of the three homologous compounds was effectively performed.Sixteen absorbance values (A) at their characteristic wavelengths in the UV-spectral range of 230-280 nm were taken as network parameters,and the network training and calculation of the samples were carried out by the arrangement of uniform designing.Tests for recovery and precision were carried out by standard addition method at different concentration of the 3 homologues,giving values of average recovery of 987%,99.4% and 97.4%,and of RSD′s of 2.0%,2.8% and 2.7% for benzene,toluene and xylene respectively.