(472b) Thermodynamic Modeling and Solubility Measurements for the Design of Anti-Solvent Precipitation Processes | AIChE

(472b) Thermodynamic Modeling and Solubility Measurements for the Design of Anti-Solvent Precipitation Processes

Authors 

Trevisan Melfi, D. - Presenter, The University of Kansas
To, A., The University of Kansas
Altman, K., The University of Kansas
Doolan, J., Honeywell KCNSC
Scurto, A., University of Kansas
Camarda, K., University of Kansas
This project seeks to develop new thermodynamic models for fine chemical and polymer intermediate molecules, and to verify those models using experimental solubility measurements. The models are then used in the design of more efficient separation processes, both in terms of energy use and material recycle. The synthesis of N-phenyl succinimide (NPS) involves the ring-closing reaction of succinanilic acid (SA) in a polar organic solvent such as dimethyl formamide (DMF). The product succinimide must be separated and the acid recycled. This separation is usually performed by anti-solvent precipitation using another organic solvent or water. The anti-solvent needs to be soluble in the DMF solvent yet precipitate the product and remaining reactant. The mixture of the solvent and anti-solvent then must be separated usually using conventional distillation. The choice of the anti-solvents has one of the primary influences on the energy use and operating costs. In an effort to develop a methodology to help screen potential anti-solvents, phase equilibrium experiments (solid-liquid equilibrium (SLE)), a thermodynamic correlation, and predictions were explored. Our group has investigated a range of thermodynamic models, including correlations using the Non-Random Two-Liquid (NRTL) GEX/activity coefficient model, and two predictive activity coefficient models: the group-contribution approach in UNIFAC, the COSMO-SAC quantum-based approach. These activity coefficient models have been verified using experimental approaches. Sampling techniques with UVvis spectroscopy have been used to measure the binary solubility (SLE) of each solid solute (NPS and SA) in the solvent DMF, and potential anti-solvents of varying polarity: water, ethanol, and ethyl acetate at 10°C and 25°C. Quantitative 1H NMR spectroscopy techniques were utilized to measure the ternary solubilities of both solute/solute/solvent and solute/DMF/anti-solvent. Several quaternary systems were also measured: solute/solute/DMF/anti-solvent. The activity coefficients computed and verified in this work have then been used within a set of process simulations to optimize the design of the overall process. Simulations using the software ASPEN have been performed, and a techno-economic analysis allows the consideration of a range of anti-solvents to find a system with low energy use.