Abstract.
Near infrared spectrometry and multivariate data analysis were applied to predict the chemical composition and physico-chemical
characteristics of eucalypt unbleached kraft pulps obtained at different laboratory pulping conditions. Viscosity, degree of
polymerization (DP), kappa, brightness and contents of glucan, xylan, uronic acids, and lignin were the modeled variables using
diffuse reflectance near infrared spectra obtained on pulp handsheets and the partial least squares (PLS) method. Models with two
to four PLS components and good predictive ability were established after first derivative spectra pre-processing and application of
cross-validation methodology. The predictive models can reduce the time consuming traditional analyses in the pulping industry
laboratories, and also lead to a better process monitoring for suitable applications.

Keywords.
Kraft Pulp; Carbohydrates; Automation; Multivariate Calibration; NIR; Spectrometry.

Keywords Plus.
Near-Infrared Spectroscopy; Orthogonal Signal Correction; Least-Squares Regression; Scatter Correction; Spectra; Kinetics; Lignin; Wood; NMR.