Application of a Genetic Algorithm to Optimize Purification in the Zone Refining Process

Abstract

Zone refining has been extensively used as a process of final purification of materials. Extensive theoretical and experimental efforts have been made to improve the purification efficiency of zone refining by optimizing parameters such as the length and the number of passes of the liquid zone. In the present study, a genetic algorithm (GA) integrated to a mass transfer numerical model has been used in the search of optimized operational parameters of the zone refining process. Experiments were carried out with naphthalene as the reference material having rhodamine as the impurity. Transparent organic substances, such as naphthalene, have been widely used as metals analogues materials because they allow in situ observations of microstructure evolution. Numerical predictions of the impurity profile along the refined naphthalene ingot is also ascertained by the present experimental data. Zone refining experiments were carried out by varying the zone length in each zone pass in order to permit the optimized data to be compared. It was found that the use of a GA integrated to the numerical mass transfer model has permitted the planning of operational conditions which are able to increase the purification efficiency up to 40% if compared with that attained when using conventional simulation approaches.