(275b) Integrated Solvent Design for the Separation of Aromatics from Aliphatics with Ionic Liquids

Current processes for the separation of aromatics from aliphatics are plagued with challenges including: feed composition limitations, low aromatics/aliphatics selectivity, low aromatics capacity, and high energy consumption. Collectively, these limitations call for the development of new technologies to improve separation. Solvent design of ionic liquids (ILs) for the separation process is one of the most promising research directions in this field, because ILs offer several advantages including nonvolatility, thermal stability, and tunability [1,2,3]. In this study, we integrate conceptual process design and predictive thermodynamic models into the molecular design of an IL for this separation process, using toluene/heptane as a model system for aromatic/aliphatic.

Process design and predictive thermodynamic models are integrated into the IL design using an iterative process. First, we apply computational chemistry methods to characterize the candidate IL molecule. This involves using Gaussian to do molecular geometry optimization and to do Conductor-like Polarizable Continuum Model (CPCM) calculations to characterize solvation effects. These results are then used to do Conductor-like Screening Model (COSMO) calculations to determine a screening charge density profile for the IL, which is used by the COSMOtherm software to obtain molecular parameters and physical properties. Second, we evaluate the candidate IL using a simple process simulation (Aspen Plus) model of the IL-based separation process with a simple flash drum as the solvent recovery unit. We use COSMO - Segment Activity Coefficient (COSMO-SAC) as the activity coefficient model, with the molecular parameters determined in the first step. From the process simulation, we can obtain separation process results such as toluene purity, recovery, energy duty, capital cost, etc. After the evaluation of the second step, the IL molecule can be modified, and the two-step procedure repeated. For comparison of the IL process modeling results with the current commercial sulfolane process we have developed an Aspen Plus model of this process using the non-random two liquid (NRTL) approach as the activity coefficient model. The goal of this strategy is to design the best IL that not only meets the separation product criteria but also maintains a lower energy and capital cost compared to the current commercial process.

At this time, we have tested multiple ILs based on the design strategy described above. Some are previously tested ILs in the aromatic/aliphatic separation with modified functional groups and alkyl chain length, such as bis(trifluoromethylsulfonyl)imide ([Tf₂N]^−) anion paired with “modified” pyridinium, imidazolium, thiolanium cations. Others are newly designed ILs that can potentially interact with aromatics via electromagnetic interactions such as hydrogen bonding, σ-π interactions, π-π interactions, etc. The process modeling results of tested ILs are compared with other published IL processes using the same solvent recovery method [4,5], as well as to the commercial sulfolane process, to show their applicability.

[1] Lin, W.-C., Tsai, T.-H., Lin, T.-Y., Yang, C.-H. J. Chem. Eng. Data2008; 53: 760-764.

[2] Cháfer, A., de la Torre, J., Font, A., Lladosa, E. J. Chem. Eng. Data2015; 60: 2426-2433.

[3] Brennecke. J. F., Maginn. E. J. AIChE J. 2001; 47: 2384-2389.

[4] Meindersma G. and Haan A. de. Chemical Engineering Research and Design 2008; 86(7): 745–752.

[5] Ferro, V. R., de Riva, J., Sanchez, D., Ruiz, E., & Palomar, J. Chemical Engineering Research and Design 2015; 94(1):632–647.