Abstract
High syringly/guaiacyl (S/G) ratios are advantageous for chemical pulp production due to higher delignification rates, higher pulp yields and lower chemical consumption. Near infrared-based partial least-squares regression (PLS-R) models were developed to assess the S/G ratio of Eucalyptus globulus wood using analytical pyrolysis as the reference method. The PLS-R models obtained using the wavenumber range from 6100cm−1 to 5450cm−1 (1639–1835nm) of the preprocessed spectra using first derivative (1stDer) and first derivative in combination with; vector normalisation (1stDerVN), multiplicative scatter correction (1stDerMSC) and straight-line-subtraction (1stDerSLS), and the second derivative (2ndDer), are well qualified for rapid screening the S/G ratio of Eucalyptus globulus wood. Overall, models using 1stDerVN and 1stDerMSC preprocess (78 samples) requiring only three PLS components have the best statistics with coefficient of determination (r2) = 0.97, root mean square error of cross-validation (RMSECV)=0.025 and residual prediction deviation (RPD) = 5.7
Get full access to this article
View all access options for this article.
