A novel approach, utilizing a two-dimensional (2D) statistical correlation of mid- and near-infrared spectra, is presented as a means to assist with qualitative spectral interpretation. The method utilizes cross-correlation by least-squares to assess changes in both regions that result from changes in sample composition. The technique has been applied to complex agricultural samples that differ in wax (cuticle), carbohydrate, protein, and lignin content. Dispersive near-infrared (NIR) and interferometric mid-infrared (FT-IR) diffuse reflectance spectra were obtained on each of the samples, and point-for-point 2D cross-correlation was obtained. The technique permits the correlation of the combination and overtone region of the NIR to the fundamental vibrations in the mid-infrared (MIR) region. This allows the determination of the most probable source of NIR signals and verification of the "real" information content of the purely statistically derived signals whose intensities currently are used for quantitative analysis in this spectral region.
F. E. Barton, D. S. Himmelsbach, J. H. Duckworth, and M. J. Smith, "Two-Dimensional Vibration Spectroscopy: Correlation of Mid- and Near-Infrared Regions," Appl. Spectrosc. 46, 420-429 (1992)