The present study has aimed at comparing three methods for analyzing near-infrared (NIR) spectra: conventional spectral analysis methods, chemometrics, and generalized two-dimensional (2D) correlation spectroscopy. In a comparison of these approaches, NIR spectra were measured for aqueous solutions of human serum albumin (HSA) with the concentration range of 0.5-5.0 wt %. Synchronous and asynchronous 2D correlation spectra were generated from the concentration-dependent NIR spectral variations of HSA in distilled water. The first and second loadings plots were calculated for principal component analysis (PCA) models based upon the above NIR spectral data. It was found that slice spectra calculated from the synchronous spectra are very close to the first loadings plots and that slice spectra from the asynchronous spectra have a close resemblance to the second loadings plots. The reasons why the synchronous and asynchronous spectra bear close resemblance to the first and second loadings plots for the PCA model, respectively, are discussed in the paper.
Koichi Murayama, Bogsława Czarnik-Matusewicz, Yuqing Wu, Roumiana Tsenkova, and Yukihiro Ozaki, "Comparison between Conventional Spectral Analysis Methods, Chemometrics, and Two-Dimensional Correlation Spectroscopy in the Analysis of Near-Infrared Spectra of Protein," Appl. Spectrosc. 54, 978-985 (2000)