Prediction of important sulphite pulp properties from NIR spectra.

In the pulp and paper industry many important process parameters require time consuming wet chemical analyses. These parameters often describe the pulp quality and are used in process control. The need to replace these with fast, accurate measurements is therefor significant. NIR spectroscopy is one method that has been investigated to predict such parameters in kraft pulp research in recent years. Borregaard Industries Ltd. produces sulphite pulp in a wide range of qualities from Norway Spruce. These pulps are mainly used as raw materials in the manufacture of cellulose derivatives such as cellulose ethers, esters, nitrocellulose and viscose, as well as for microcrystalline cellulose. In the process of transforming wood chips to the desired products, the first steps take place in batch reactors. While the wood chips are being processed in the reactors, there is at present no possibility of measuring the properties of the solid phase contents. Such information would be very useful of when the cook should be ended, and thereby minimize deviation from specifications of unbleached pulp. In the present work NIR spectroscopy has been used to predict kappa number and viscosity in unbleached sulphite pulp. Viscosity is an indirect measure of the chain length of the carbohydrates in the pulp and the kappa number is an indirect method for measuring the remaining lignin content in the pulp. NIR spectra have been measured on both pulp and the accompanying cooking liquor and calibrated against wet chemical analyses of the pulp. The main focus was to use the spectra of the cooking liquor to predict chosen pulp properties. The cooking liquor is like a fingerprint of the pulp contents, and should therefore be a god descriptor of these properties. The best calibration method for this purpose was partial least square regression with one response (PLS1) on the 1st derivatives of the NIR spectra. The prediction models will be further developed and eventually applied in the optimization of the pulping process at Borregaard.