Abstract
A new approach is presented for the study of the variability of Portuguese reproduction cork using solidโstate ^13^CโNMR spectroscopy and photoacoustic (PAS) FTIR (FTIRโPAS) spectroscopy combined with chemometrics. Cork samples were collected from 12 different geographical sites, and their ^13^Cโcrossโpolarization with magic angle spinning (CP/MAS) and FTIR spectra were registered. A large spectral variability among the cork samples was detected by principal component analysis and found to relate to the suberin and carbohydrate contents. This variability was independent of the sample geographical origin but significantly dependent on the cork quality, thus enabling the distinction of cork samples according to the latter property. The suberin content of the cork samples was predicted using multivariate regression models based on the ^13^CโNMR and FTIR spectra of the samples as reported previously. Finally, the relationship between the variability of the ^13^CโCP/MAS spectra with that of the FTIRโPAS spectra was studied by outer product analysis. This type of multivariate analysis enabled a clear correlation to be established between the peaks assigned to suberin and carbohydrate in the FTIR spectrum and those appearing in the ^13^CโCP/MAS spectra. ยฉ 2001 John Wiley & Sons, Inc. Biopolymers (Biospectroscopy) 62: 268โ277, 2001