Computational Evaluation of Mixtures of Hydrofluorocarbons and Deep Eutectic Solvents for Absorption Refrigeration Systems.

We used computational tools to evaluate three working fluid mixtures for single-effect absorption refrigeration systems, where the generator (desorber) is powered by waste or solar heat. The mixtures studied here resulted from combining a widely used hydrofluorocarbon (HFC) refrigerant, R134a, with three common deep eutectic solvents (DESs) formed by mixing choline chloride (hydrogen bond acceptor, HBA) with urea, glycerol, or ethylene glycol as the hydrogen bond donor (HBD) species. The COSMOtherm/TmoleX software package was used in combination with refrigerant data from NIST/REFPROP, to perform a thermodynamic evaluation of absorption refrigeration cycles using the proposed working fluid mixtures. Afterward, classical MD simulations of the three mixtures were performed to gain insight on these systems at the molecular level. Larger cycle efficiencies are obtained when R134a is combined with choline chloride and ethylene glycol, followed by the system where glycerol is the HBD, and finally that where the HBD is urea. MD simulations indicate that the local density profiles of all species exhibit very sharp variations in systems containing glycerol or urea; furthermore, the Henry's law constants of R134a in these two systems are larger than those observed for the HFC in choline chloride and ethylene glycol, indicating that R134a is more soluble in the latter DES. Interaction energies indicate that the R134a-R134a interactions are weaker in the system where ethylene glycol is the HBD, as compared to in the other DES. Radial distribution functions confirm that in all systems, the DES species do not form strong directional interactions (e.g., hydrogen bonds) with the R134a molecules. Relatively strong interactions are observed between the Cl anions and the hydrogen atoms in R134a; however, the atom-atom interactions between R134a and the cation and HBD species are weaker and do not play a significant role in the solvation of the refrigerant. In all systems, R134a has the largest diffusion coefficients, followed by the HBD, the anion and the cation; the diffusion coefficients are the largest in the systems containing ethylene glycol, followed by those having glycerol and urea. This work is our first step toward our long-term goal of designing and demonstrating optimal working fluid mixtures for use in absorption refrigeration systems. Our results suggest that COSMO-RS can be used to perform a rapid screening of a large number of working fluid mixtures, and select a few candidates for further exploration using molecular simulations and experiments. These latter approaches can be used to refine the accuracy of the COSMO-RS predictions, and to optimize the selection of optimal working fluid mixtures for demonstration in absorption refrigeration systems powered by solar or waste heat sources.