|Home||Search | Sitemap||Technique | Links | News | Training | Employment | Forums | Base|
|EFPG DAYS 2003||cerig.efpg.inpg.fr|
|Last update : June 19, 2003|
|12 - Modeling full bleaching sequences. A tool for process modification|
G. Mortha (EFPG)
Single stages chlorine dioxide bleaching have been individually modeled and combined together to build a full DEDED sequence model. As a preliminary step to bleaching understanding, the main features of the reaction mechanism of ClO2 and the influence of process parameters during bleaching are reviewed. It is shown that one key aspect for an optimized action of ClO2 during bleaching is a good control of the pH-time profile, to avoid the presence of residual species such as chlorite and chlorate. If the latter species are stable, (in the case of chlorite, this depends of the end pH of the stage), they are responsible for an important loss in available oxidant. Their formation is due to the presence of intermediate molecular chlorine and hypochlorous acid, and attributed to the complex reactions involving ClO2, chlorite, and possibly radical species in solution or in the pulp matrix. Both the pH and the delignification level play a major role on the chlorate yield.
During Do stages, the pH decreases strongly and rapidly, and the best pH-time profile can be obtained with a correct control of the starting pH, that should not excess neutrality. The better delignification yield in acidic medium is attributed to the more efficient reaction of molecular chlorine with lignin than with hypochlorous acid.
During D1 and D2 stages, the end pH has a particular importance, and it should be maintained not lower than 3 to allow a progressive consumption of chlorite initially formed by the reaction of ClO2. The presence of chlorite also leads to no excess chlorine, which also decreases formation of chlorate.
Modeling a single D stage is based on a simplified reaction mechanism that takes into account the general reaction pattern, and by the use of power-law stoichiometric and kinetic equations that have been empirically identified to model kappa number and brightness evolution. A pH equation is also introduced, that relies on the buffering effect of alkaline groups in the pulp or in the liquor. In the multi-stage DEDED model, the lignin concentration is represented by the kappa number during prebleaching, and K457 (Kubelka-Munk parameter) during full bleaching, from which brightness is calculated. The transition between kappa number and K457 relies on the linear relationship between these two variables; the latter relation depends on the pulp studied and has to be empirically identified for a given pulp.
Some issues of the model are presented. The multi-stage model is appropriate to predict the optimal dosage of ClO2 to reach a target brightness after D2, as well as the best share of ClO2 between the different D stages. A generally observed tendency is that better bleaching can be achieved by decreasing the charge of ClO2 applied in Do below 50 % of the total charge of the sequence. Another challenging issue is predicting the optimal kappa number after the DoE stages to reach an optimal brightness after D2. A third issue is predicting the time for full brightness achievement during the D2 stage; additionally, it is shown the requirement for a good brightness control after the D1 stage, unless the final target brightness after D2 cannot be achieved.
As a conclusion, we demonstrate the interest for modeling ECF bleaching. Further progress would be required and could be accomplished by more deepened theoretical work and fitting a lot of laboratory and mill data. Since a large part of process control in bleaching mills remains under the direction of operators, modeling can help for a better understanding and anticipation of process actions in mill situations, but we consider that it may not, at present, represent an absolute guide for it.