Separation of Azeotropic Hydrofluorocarbon Refrigerant Mixtures: Thermodynamic and Kinetic Modeling for Binary Adsorption of HFC-32 and HFC-125 on Zeolite 5A.
Andrew D YanceyDarren P BroomMark G RoperMichael J BenhamDavid R CorbinMark B ShiflettPublished in: Langmuir : the ACS journal of surfaces and colloids (2022)
Hydrofluorocarbons (HFCs) have been used extensively as refrigerants over the past four decades; however, HFCs are currently being phased out due to large global warming potentials. As the next generation of hydrofluoroolefin refrigerants are phased in, action must be taken to responsibly and sustainably deal with the HFCs currently in circulation. Ideally, unused HFCs can be reclaimed and recycled; however, many HFCs in circulation are azeotropic or near-azeotropic mixtures and must be separated before recycling. Previously, pure gas isotherm data were presented for both HFC-125 (pentafluoroethane) and HFC-32 (difluoromethane) with zeolite 5A, and it was concluded that this zeolite could separate refrigerant R-410A (50/50 wt % HFC-125/HFC-32). To further investigate the separation capabilities of zeolite 5A, binary adsorption was measured for the same system using the Integral Mass Balance method. Zeolite 5A showed a selectivity of 9.6-10.9 for HFC-32 over the composition range of 25-75 mol % HFC-125. Adsorbed phase activity coefficients were calculated from binary adsorption data. The Spreading Pressure Dependent, modified nonrandom two-liquid, and modified Wilson activity coefficient models were fit to experimental data, and the resulting activity coefficient models were used in Real Adsorbed Solution Theory (RAST). RAST binary adsorption model predictions were compared with Ideal Adsorbed Solution Theory (IAST) predictions made using the Dual-Site Langmuir, Tóth, and Jensen and Seaton pure gas isotherm models. Both IAST and RAST yielded qualitatively accurate predictions; however, quantitative accuracy was greatly improved using RAST models. Diffusion behavior of HFC-125 and HFC-32 was also investigated by fitting the isothermal Fickian diffusion model to kinetic data. Molecular-level phenomena were investigated to understand both thermodynamic and kinetic behaviors.