Selection of a minimum toxicity and high performance ionic liquid mixture for the separation of aromatic - aliphatic mixtures by extractive distillation

Abstract

Abstract The separation of aromatic compounds from naphtha (mainly aliphatics) has a great interest for the petrochemical industry. This separation is commonly carried out by liquid- liquid extraction using different solvents which entails very high energy cost during the solvent recovery steps. Recently, ionic liquids (ILs) have been proposed as entrainers for the separation of this mixtures. In this paper we present a multiobjective optimization to select of ILs. One objective is minimize the toxicity of the mixture of ionic liquids used as entrainer in the separation of aromatics and aliphatics mixtures. The second target is to improve the separation performance in the extractive distillation process. The estimation of the toxicity has been carried out by training an artificial neural network (ANN) from structure information and toxicity values of ionic liquids. Results show a high correlation between the presence of heteroatoms and toxicity. The separation of aromatic and aliphatic hydrocarbons was evaluated through rigorous simulation of an extractive distillation process as detailed in Diaz et al. (2016). The evaluation of objectives was carried out in MATLAB connecting Aspen Plus simulations of the single stage extractive distillation process (vapor-liquid-liquid equilibria)through COM objects and evaluating the trained ANN for the ILs mixtures evaluated in each iteration. The multiobjective optimization of the problem was performed using a derivative free algorithm (genetic algorithm). Results show that 1-ethyl-3-methylimidazolium dicyanamide ([EMIM][DCA]) is a promising solvent in terms of separation performance and low toxicity.