Conceptual design of unit operations to separate aromatic hydrocarbons from naphtha using ionic liquids. COSMO-based process simulations with multi-component “real” mixture feed

• Unit operations for removing aromatic from naphtha with ILs were modeled by Aspen Plus.
• COSMOSAC predictions were successfully validated for compound and mixture properties.
• Multicomponent “real” mixtures of hydrocarbons and ILs were used in the simulations.
• IL selection greatly affected the extraction and the solvent regeneration efficiencies.
• (IL–IL) mixtures presented enhanced performance as extracting agents in the process.

COSMO-based process simulations using Aspen Plus and Aspen HYSYS were systematically applied to the conceptual design of the two main unit operations commonly proposed to separate aromatic and aliphatic hydrocarbons with ionic liquids (ILs) as extracting solvents; the extraction itself and the vacuum distillation for regenerating the ionic liquid. By the first time, multi-component (up to 28 components) “real” mixture feeds were taken into account to model the naphtha in the process design. Binary model (n-hexane + benzene, n-heptane + toluene, n-octane + ethylbenzene, n-octane + m-xylene) mixture feeds were also considered to validate the computational procedure. Nine different ionic liquids and mixtures of them, (IL–IL) mixtures, were selected as extracting solvents. Ionic liquids were introduced in the process simulations as (pseudo)components and the COSMOSAC property model was used for estimating the activity coefficients of the individual components in the mixtures. The information needed to both create the non-data bank ionic liquid (pseudo)components and to specify the COSMOSAC property model was gathered from COSMO-RS calculations. COSMO-based models exhibited a reasonably good predictability of both the thermo-physical properties of the pure (hydrocarbons and ionic liquids) components and the LL and VL equilibria of their mixtures. The performances of extraction and regeneration individual operations were analyzed at different operating conditions, including the nature of the IL-based extracting agent, the solvent-to-feed ratio, and the hydrocarbon mixture composition. The present results suggest that COSMO-supported process simulations are capable of confidently dealing with complex multicomponent mixtures of hydrocarbons and ionic liquids. This opens new perspectives to improved developments of this process based on ionic liquids and their mixtures.