(89a) Development and Modeling of Energy-Optimum Aromatic Separation Processes Using Ionic Liquids

Ionic liquid (IL) solvents are attractive solvents for selective separation of aromatics and aliphatics in hydrocarbon mixtures. There is incentive to extract aromatics from ethylene cracker naphtha feeds, where the current aromatic extraction technologies are not employed. ILs are considered attractive because of their non-volatility allowing easy separation from extract along with the vast number of synthesizable ILs that offer different chemical characteristics.

One of our goals to quantify the potential of IL solvent in aromatic extraction by understanding the best way to simulate and model an IL to capture interactions between ILs and hydrocarbons. We test this by evaluating how process simulation using a priori thermodynamic method COSMO-SAC compares to an experimental data supported UNIQUAC model. We also evaluate how feed simplifications can change design conclusions. Another goal is to develop an energy-optimum process that can meet high purity requirements for aromatic and aliphatic products and take full advantage of the non-volatile IL solvent. Our final goal is to use our novel aromatic extraction process design to evaluate 16 well-studied ILs to understand how they compare to an equivalent sulfolane process and identify key IL solvent properties that have the greatest impact on process energy.

We find that COSMO-SAC prediction is unsuitable for process simulation because of consistent and significant underprediction of activity coefficients for aliphatics which impacts the conclusion of separation feasibility. Regression of UNIQAC binary coefficients can much better reflect experimental equilibrium data. We also find simplifying complex feeds to binary mixtures significantly alters process results and design conclusions. Feed simplifications should not significantly alter the volatility or contribution of chemical groups in the feed.

We introduce a novel energy-optimum design that is advantageous for low aromatic content feeds due to independent control of the maximum solvent operating temperature and improved heat integration. Eleven of the 16 ILs we test achieve a lower process steam duty compared to sulfolane with the lowest energy process requiring 57% of the sulfolane steam duty. Finally, we introduce a new science-guided solvent variable related to the solvent heat load which by itself can be linearly correlated to process steam duty. This variable can be used as an objective function to minimize in IL solvent screening.