How Can We Unravel Complicated near Infrared Spectra?—Recent Progress in Spectral Analysis Methods for Resolution Enhancement and Band Assignments in the near Infrared Region

Abstract

This review paper reports recent progress in spectral analysis methods for resolution enhancement and band assignments in the near infrared (NIR) region. Spectra in the NIR region are inherently rich with information on the physical and chemical properties of molecules. However, it is not always straightforward to analyse the spectra because an NIR spectrum consists of a number of overlapped bands due to overtones and combination modes. An NIR spectrum may be analysed by conventional spectral analysis methods, chemometrics or two-dimensional correlation spectroscopy. The following conventional methods are currently utilised to analyse NIR spectra: (a) derivatives, (b) difference spectroscopy, (c) Fourier self-deconvolution and (d) curve fitting. The derivative method is powerful in separating superimposed bands and correcting for a baseline slope. Conventional experimental methods for spectral analysis, such as isotope exchange and measurement of polarisation spectra, are also valid in the NIR region. Chemometrics is very useful for extracting information from NIR spectra. Among a variety of chemometrics methods, multiple linear regression, principal component analysis, principal component regression and partial least squares regression are most often used for qualitative and quantitative analysis. Recently, chemometrics has been used for resolution enhancement of NIR spectra. Particularly, loadings plots or regression coefficients are useful for separating overlapped bands and for making band assignments. Notable recent advances in the analysis of NIR spectroscopy are the development or introduction of new spectral analysis methods such as two-dimensional (2D) correlation spectroscopy and self-modelling curve resolution methods (SMCR). 2D correlation analysis enables enhancement of apparent spectral resolution by spreading spectral peaks over a second dimension. SMCR allows one to resolve the experimental matrix into concentration profiles and pure spectra of the involved species without prior knowledge of any of these features.