High-throughput prediction of physical and mechanical properties of paper from Raman chemometric analysis of pulp fibres.

Better pulp quality control, especially relating to product sheet strength, can offer an important means to improve the market superiority of a pulp. However, aside from lignin content, few pulp properties can be easily measured in a timely manner for process control. The present report proves a principle on a novel Raman system and a systematic chemometric approach, allowing for rapid spectral data acquisition and definitive spectrochemical analysis of wet pulp under harsh manufacturing conditions. By incorporating sophisticated chemometrics strategies that combine wavelet transform with a template-oriented genetic algorithm feature selection and partial least squares multivariate classification, this instrument system extracts maximum analytical information from raw Raman spectra to accurately predict the physicomechanical properties of sheet products. Our work has tested and refined these routines by drawing upon the systematic analysis of 26 bleached softwood kraft pulps formulated to yield challenges representative of those encountered in common pulp fibre and sheet analyses. Satisfactory calibration results suggest that a Raman gauge can be developed that has the capacity to continuously assay pulp to predict the physicomechanical properties of sheet paper in real time. [ABSTRACT FROM AUTHOR]