Restrictions for environmental reasons on the use of CFCs and HCFCs, the traditional working fluids in heat pumps, refrigeration plants and air conditioning, have led to the use of pure HFCs and blends of HFCs. However, since HFCs have a different influence on compressor oils, new oils that are compatible with HFCs need to be found. A likely alternative for compressor oils is polyol ester oils (POE), but POEs differ in solubility, which is one important factor when selecting a successful compressor oil. To minimize costly experimentation, new thermodynamic tools that can predict the solubility of HFCs in compressor oils will be very useful.
To develop a predictive model for solubility, based on a group-contribution concept, experimental values for pure substances with known structure are needed. Thus, solubility measurements were made for 13 binary systems of four HFCs in four long-chained hydrocarbons (lcHCs) and for 20 binary systems of five HFCs in four pentaerythritol tetra alkyl esters (PEs), where the latter is one type of POEs.
Based on these data a modified Flory-Huggins-based predictive model, termed MFH, was developed. This model is of the heterogeneous type; the vapour phase is calculated with an equation of state and the liquid phase with an activity factor model. A simple version of this model was used to describe data for 14 binary HFC-lcHC mixtures with 7 general parameters and a relative deviation between calculated and experimental data, DP, of 4.1%. A more extended version was used to describe data for the 20 binary HFC-PE mixtures with 15 general parameters and a DP of 3.4%.
While DP increases from 1.8 to 3.4% when the MFH model is used instead of correlation with the traditional Flory-Huggins model, the number of empirical parameters decreases from 60 to 15. Furthermore, the 15 parameters are general and the MFH model can thus also be used to predict solubility of HFCs in PEs that have no measured data available.
Available at Chalmers Library