Resolving Near-Infrared Spectra by Generalized Window Factor Analysis for Understanding Interactions in Aqueous Solution

Near-infrared (NIR) spectroscopy has been demonstrated as a sensitive technique to characterize the structure of water for understanding its interactions in aqueous solutions. Chemometrics, however, is needed to enhance the resolution or to obtain the spectrum of water and analytes. In this work, generalized window factor analysis (GWFA) was developed by defining a virtual window using a high order derivative spectrum. Continuous wavelet transform (CWT) was used for calculating the high order derivative of the determined component. The virtual window was defined as the spectral data that are not zero in the derivative spectrum. The method was tested by separating the water spectrum from aqueous solution. In comparison with the measured spectrum, the calculated water spectrum was more accurate at describing the water structures in the mixture. Therefore, by analyzing the difference spectrum of water between the measured and the calculated one, the structural changes and the interactions may be observed. The spectra of water in the solutions were obtained from which the structure, interactions and the function of water in the systems were determined.