Optimization of Anthraquinone Pulping and Oxygen Delignification Through Statistical Design, 1996 Pulping Conference Proceedings

D. L Oldroyd and F. M. Wadley

ICI Forest Products

North American pulp mills continue to utilize modern technologies and strategies to optimize operations and improve economic and environmental performance. A good deal of this effort has focussed on extending delignification via modified cooking processes and oxygen delignification. This results in lower bleaching chemical charges and reduces environmental impact. Mills are also under pressure to maximize their utilization of wood, again from an economic and environmental standpoint. This has refocussed concerns with respect to pulp yield and yield losses. Technologies such as oxygen delignification and anthraquinone have been shown to increase selectivity and maximize pulp yield compared to conventional pulping. This has led to considerable commercial growth in these areas. There are a few examples reported in the literature investigating extended delignification with a view to maximizing one or more responses (e.g. pulp yield or strength). However, statistical optimization studies of a variety of pulping variables coupled with the measurement of numerous responses are quite rare. The present study addresses these issues by constructing a pulping model applicable to mills wishing to examine different strategies in optimizing their operations. Selected changes in operating parameters and mill objectives can also be accounted for. Using statistically designed methods, laboratory-scale pulping and oxygen delignification experiments were performed on a northern softwood furnish. Pulping parameters varied in the study were anthraquinone addition and effective alkali (EA) level. Oxygen delignification variables included caustic charge and temperature. Responses measured were kappa number (post-digester and post-O2 , pulp yield (post-digester and post-0,), viscosity and brightness. The observations were analyzed using experimental design software to generate a mathematical model. Equations for kappa number, pulp yield, viscosity and brightness were derived. With this model, a global maximum for each response (e.g. final pulp yield) within the range of all variables examined was determined. Alternatively, the value of any combination of variables which are “untouchable” can be fixed to determine a local maximum.