Prediction of thermo mechanical pulp brightness using NIR spectroscopy on wood raw material

Partial Least Squares regression to latent structures was used to find a correlation between near-infrared reflectance spectra of samples of milled Norway Spruce (Picea abies) and ISO brightness of bleached thermo mechanical pulp. Logs of spruce were stored in climate chambers with different conditions for 14 weeks. A factorial design based on five variables was constructed to control the storage and to provide different brightness properties of the stored spruce logs. Wood samples were collected during the storage and they were analysed with near-infrared reflectance spectroscopy and later refined to thermo mechanical pulp. The pulp was bleached with both hydrogen peroxide and sodium dithionite. The ISO brightness of the bleached pulp was correlated to the near-infrared spectra of the untreated samples using Partial Least Squares regression to latent structures. The modelling showed that it is possible to predict the ISO brightness of bleached thermo mechanical pulp from near-infrared spectra of original wood raw material.