Following the kinetics of converting wood into kraft pulps by NIR spectroscopy

Kraft pulp is produced from wood by lignin removal (deslignification) from cell wall and middle lamella which allows individualising the fibres for papermaking applications. Concomitant to deslignification, polysaccharides removal (especially hemicelluloses) also takes place. The kinetics of conversion of wood into Kraft pulp fibres involves thus the quantification of lignin and polysaccharides (hemicelluloses and cellulose) content. A systematic study on the kinetics of kraft pulping of 6 hardwood species (beech, birch, chestnut, eucalyptus, oak, poplar) and 6 softwoods hardwoods (aleppo pine, black pine, Douglas-fir, maritime pine, scots pine, spruce), pure and in binary or ternary mixtures of the same botanic group was carried out in our laboratories. As around 1000 samples were produced, it was impossible to perform the traditional wet chemistry analysis on all samples for the determination of lignin and polysaccharides content due to time and financial constraints. The use of NIR spectroscopy appeared as an interesting alternative for tracking the changes in concentration of these biomacromolecules.

Kraft pulps were obtained from different softwood and hardwood species at three different temperatures (150°C, 160°C and 170°C for softwoods and 140°C, 150°C and 160°C for hardwoods) and cooking time. 36 pulp samples produced at three different temperatures and different cooking time were used for each wood species to create the NIRS calibrations. Lignin content was measured by Klason method (soluble + insoluble) and polysaccharides content measured after a two-stage acidic hydrolysis followed by the analysis of the monosugars. NIRS calibrations presented in this article were all obtained using a Bruker Bruker Tensor 37 NIR spectrometer equipped with a spinning cup module. 2 spectra (32 scans each spectrum) were obtained for each sample at 8 cm-1 spectral resolution. Both wood and pulp samples at different levels of deslignification were ground to 40/60 mesh fraction obtained by sieving. The spectra were acquired in reflectance mode using an integration sphere. Partial least-squares regression (PLSR) was used to correlate spectroscopic data to the wet chemistry measurements. .

As the number of samples per wood species was too small, pulp produced using different pure and mixtures of wood species were pooled into single calibrations for the two main botanic groups (softwoods and hardwoods). Globally, very good performance of the NIRS calibrations using multi-wood species were obtained for lignin and cellulose content (R2 higher than 0.98 for the calibrations and 0.97 for the validation and RMSEP lower than 0.75 for lignin and 1.8 % for cellulose. The performances of the calibrations for the prediction of hemicelluloses content were less impressive (R2 around 0.93 for calibrations and 0.81 for the validation, RMSEP below 1.7 %) but precise enough to be used in routine applications. NIRS calibrations were then used for predicting the lignin and polysaccharides content of all samples. Then, the kinetics and activation energies of the deslignification were measured for all the wood species and mixtures. Important variations in the kinetics constants and energy activation between the wood species were observed due to different factors such as the chemical composition and the anatomy of wood species.