Quantitative optical analysis of component content in multivariate mixture solution

Abstract

The partial least square method (PLS) was introduced to quantitatively analyze component content of multivariate mixture by using ultraviolet-visible spectroscopy. Two models were built to establish component content of multivariate mixture based on the transmittance spectra of multivariate mixture, which is the full band PLS model and the feature band iPLS model, respectively. The comparative study of the models was carried out with ethanol solution as the research object and the samples were divided into two groups, including the calibration sets and the prediction sets. This study indicates that there is a certain relationship between the spectra information of the multivariate mixture and its component content, and the above two models are feasible for solving the component content. Compared with the full band PLS model, the feature band iPLS model has higher accuracy and the model is simpler. Therefore, the feature band iPLS model, which is an efficient method, can be used to analyze the component content in multivariate mixture solution based on the ultraviolet-visible spectrum.