Concentration Determination of Ternary Mixtures in Water using Ultraviolet Spectrophotometry and Artificial Neural Networks

摘要

Phenylethylamine (PEA) is an important pharmaceutical and dyestuff intermediate. During the production of PEA, NaOH is usually used as a reactant, and NaCl will be produced. Finally, the remains which usually conclude these three components are discharged into the water and bring serious impact on the water quality. Therefore, in order to protect the safety of water quality, it is very important to determine the concentration of pollutants in water. However, one of the main difficulties in quantification of NaCl, NaOH, and PEA in water is the fact that the water usually contains impurities rather than being pure. Here we report the development of a rapid and powerful method, artificial neural network (ANN), for spectral resolution of a highly overlapping ternary mixtures in the presence of interferences. To this end, NaCl, NaOH and PEA were selected as three calibration models whose UV-Vis absorption spectra highly overlapped each other. After calibration, the prediction models of NaCl, NaOH and PEA were validated through testing with an independent spectra-concentration dataset, in which high correlation coefficients (R2) of 0.9876, 0.9912 and 0.9864 were obtained by ANN for NaCl, NaOH and PEA, respectively. Having shown a relative error of prediction of less than 4% for the independent datatest, the ANN model was found to be highly accurate in simultaneous determination of NaCl, NaOH and PEA in pure water samples. Finally, the ANN model coupled with net-analyte signal concept were successfully applied to simultaneously determine eight ternary mixtures of NaCl, NaOH and PEA added into the natural water samples. The results showed that the average recovery of NaCl, NaOH and PEA were 95%-102%. Thus, UV spectrophotometry coupled with ANN model can be considered as a promising strategy to conduct the concentration measurements of mixtures in water.