The Solubility of R-32/125 in Modified. Polyalkylene GI ycols. R.H.P. Thomas, R. P. Robinson, R. H. Chen and W-T. Wu. Allied-Signal Corporation Buffalo Research Laboratory Buffalo, New York 14210 Introduction: The air-conditioning and refrigeration industry is well on its way towards the design and commercialization of systems using R- 134a as the preferred environmentally acceptable replacement for CFC-12 in many applications. The industry is now turning some of its attention to replacements for HCFC-22. The US Clean Air act and its subsequent amendment calls for phaseout of CFCs by the year 2000 and a production freeze on HCFCs by 2015, a use limitation as refrigerants in new equipment by 2020 and a total production phaseout by 2030. In Germany, it is proposed that appliances and equipment using HCFCs be banned by January 1, 2000. Historically the commercialization of a new refrigerant product has taken up to 10 years. Given the current US and international time frames for phaseout of HCFC-22, it is important that research and development on the next generation of zero ozone depletion refrigerants for stationary air conditioning and commercialandustria1 refrigeration be undertaken now. This paper discusses some of the lubricant requirements for a potential zero ozone depletion replacement for HCFC-22. A March 19, 1982 EPA sponsored Symposium on R-32 and R-32 mixtures in refrigeration applications supported earlier indications that HFC- 32 blends and azeotropes could be considered as potential replacements for R-22 and R-502. At this Symposium, Allied-Signal announced a new refrigerant, R-32/125 (US Patent 4,978,467) as a potential future member of a family of environmentally safer refrigerants. Both HFC-32 and HFC-125 have zero ozone depletion potentials. However, the combination of these two HFCs into a non-segregating composition appears to provide an optimum balance between the physical and performance characteristics of each component, as will be discussed in other papers at this conference. One of the key decisions to made when developing a new refrigerant is the selection of the lubricant. R-12 is miscible with the mineral oils with which it is used. Much of the equipment used with R-12 takes this into account. To minimize the extent of redesign required R-l34a, the 37s
replacement for R-12, should be miscible with its lubricants. R-22 is not miscible with mineral oils, yet it is primarily used with mineral oils. Alkylbenzenes are used with R-22 when miscibility problems, Le. oil return, become severe. Much of the equipment that uses R-22 is designed to "handle" oil immiscibility. Should R-22 replacements therefore be required to be miscible with their lubricants? The only correct answer to this question is that a means of providing return of oil to the compressor must be provided for every air-conditioner and refrigeration system. It is clear that refrigerant oil combinations that are miscible will allow the oil the to return to the compressor. We will discuss briefly the miscibility of R3U125 with various lubricants. Having decided on a lubricant for use with the refrigerant, it is very useful to know the solubility of the refrigerant in the oil. We ( Thomas, Pham and Wu) have previously discussed the solubility of R-l34a with appropriatdlubricants. Here we present solubility data for R32/125 with lubricants. R-32/125 is most likely to be considered as a R-22 replacement for high temperature and medium temperature applications. The evaporator temperatures in this range of applications is approximately -1 0 to above 40' F. The lowest temperature chosen for this work was as -10' C. Other workers such Glova (1 965), Sandvordenkar (1 990), Spauschus et al(l987) and van Gaalen have discussed the measurement of solubility of refrigerants in lubricants. This is the first time that the solubility of a zero ODP mixed refrigerant in appropriate lubricants has been discussed. 2. MATERIALS. The R-32 and R-125 used in this work were 99.95% pure. The modified PAGs and PAG diol used were BRL-150 and APE0 respectively. The viscosity at 40% is 32 cst. The lubricants are proprietary materials now being tested for use with R-32/125. The exact composition of the refrigerant mixture used was 60.32 weight percent R-32 and 39.68 weight percent R- 125. 3. EXPERIMENTAL METHODS. The miscibility of the refrigerant lubricant mixtures was determined by sealing the mixtures in thermostatted glass tubes and visually observing the contents. The solubility of the refrigeranvoil mixtures was studied by determining the equilibrium vapor pressure of the systems at constant composition as a function of temperature. The apparatus consists of a 376
sample cell to which a transducer is attached. The lubricant is weighed into the cell. The entire system is assembled and evacuated. The desired amount of refrigerant was transferred to and degassed in a separate cylinde J. The refrigerant was then charged into the cell. The amount of charge was such that 80% to 90% of the cell was filled with liquid. The composition charged into the cell was therefore the composition of the liquid. The entire assembly was positioned on top of a stir plate and submerged in a thermostatted bath. The system was stirred continuously and allowed to equilibrate for an hour before taking measurements. At low temperature and high concentrations of lubricant, the system required longer time for equilibrium. The pressure transducer was calibrated using a dead weight tester and is accurate to 0.3% of the pressure. The temperature of the system was measured with a platinum resistance thermometer with an accuracy of.03% of the temperature. The vapor pressure of the refrigerant /oil system was determined at three concentrations of oil in refrigerant at -10, 20, 30, 40 and 50 C. 4. DATA AND DISCUSSION. Miscibility measurement were done using mineral oils, alkylbenzenes, PAG diols and a modified PAG. The refrigerant is not miscible with mineral oils or alkylbenzenes. This indicates that esters, PAGs and modified PAGs should be studied. We found that the refrigerant is miscible in AP150, and a modified PAG BRL150. The miscibility range found for both these lubricants is -60 to at least 55 C. This miscibility range is excellent for high and medium temperature applications. It is interesting that PAGs and modified PAGs are suitable tor use with R-i34a and R32il25. The preferred situation is that there should be one lubricant that can be used with as many HFC refrigerant candidates. This modified PAG and PAG diols meet that requirement. More miscibility work needs to be done in order to fully evaluate the lubricants that are available on the market today. The esters tested were immiscible with this refrigerant. It is possible some esters may be appropriate for use with the R32/125. In order to check the solubility equipment and our refrigerant mixture, we began by measuring its vapor pressure at three temperatures. These values were then compared with a vapor pressure equation developed by Singh et al (1991). The agreement was within 0.5%. All vapor pressure values were therefore generated using this equation. 377
Table 1. R32/125 Vapor Pressure at selected temperatures. - Temperature Vapor Pressure (psia) This work Singh et al. % Difference 0.00 1 17.2 1 17.8.36 30.00 277.0 276.0 -.28 11 40.00 I 352.7 1353.7.51 II Thk solubility data for R32/125 with BRLl50 is shown in Figure 1. In that figure the points are experimental. We have previously shown (Thomas, Wu and Pham, 1991) that solubility data for R134a mixtures with modified PAGs is well represented by Flory-Huggins (Flory, 1953) theory. The lines on this graph are calculated using Flory-Huggins theory as applied by Bawn et al, 1950. The theory is expressed by equations 1 and 2. Equation 2 describes the volume fraction of the lubricant in the refrigerant/oil mixture. Volume fraction is a way of representing the concentration of the refrigerant in the oil. Table 2 lists the values of the relative pressure P/P,( where P is the vapor pressure above the refrigerant oil mixture and Po is the vapor pressure of the pure refrigerant ) vs volume fraction of the lubricant. Equation 1. P In - - lnv, + vz + x vt PO Equation 2. 378
Table 2. Relative Pressure at fixed Oil Volume - Fraction at different Temperatures. The oil concentrations studied were 0.536, 0.755 and 0.907 volume fractions. The values of P/P, ratio should be constant for a given composition. For the first composition listed the relative pressure is constant with an average of 0.930. We observed a small decrease in this ratio as the temperature increased for the other two concentrations. This can perhaps be explained partly by the fact that at lower refrigerant concentrations and temperatures it is more difficult to reach equilibrium because the viscosity of the solution Is higher. More vigorous stirring might be required. It can also be partly caused by the fact the Flory-Huggins parameter is not truly temperature independent. When the line In P/P, -In v, -v, vs vt is plotted, its slope (Bawn et al, 1950) should give the value of the Flory-Huggins parameter, chi. The correlation coefficient for this line is 0.94. It seems that the line curves at higher concentrations for this refrigerant mixture. In Figure 2, we plot the average values of the relative pressure vs volume fraction of the lubricant. The points are experimental. The line is calculated using Flory-Huggins theory. It can be seen that at lower oil concentrations the predicted value Of the relative pressure is lower than the experimental value. Preliminary data on the solubility of 133211 25 in a PAG diol indicates the solubility is of the same order as that for the modified PAG. A decrease in the values of relative pressure with temperature at fixed composition also seems to occur. 379
5. CONCLUSIONS..L In order to better understand the systems that are being now developed, engineers need to have as much relevant data as is possible. For R22 replacements, this data is not yet available. This paper provides an initial step in providing such data for R32/125. We have shown that the solubility behavior of this mixture is not that different from that a pure refrigerant such as R-l34a. We have provided solubility data that should be useful to those are considering the design of systems for use with this potential R-22 replacement. 1 6. NOMENCLATURE P : Vapor pressure over the refrigerant oil mixture. Po : Vapor pressure of the pure refrigerant. V : Volume fraction of refrigerant or lubricant. W : Weight percent of refrigerant or lubricant in the mixture. M : Molecular weight 1 : Refrigerant 2 : Lubricant x : Flory-Huggins parameter. References:- REFERENCES Bawn, C. E. H ; Freeman, R. E. J. and Kalamaliddin (1950). Transactions of the Faraday Society 3, 50, 677 Flory, P. J. ( 1953 ), Principles of Polymer Chemistry, Cornell University Press. Glova, D. J.(1984) ASHRAE Transactions, 90,Pt IIB, KC-84-14, #4, 806. Sandvordenkar, K. S.( 1989)." CFCs: Time of Transition ", ASHRAE, 21 1. Singh, R. R., Lund, E.A. E. and Shankland, 1. R. (1991). International CFC and Halon Alternatives Conference.
Spauschus H. 0. and Henderson D. R. (1990). Proceedings of the USNC/IIR Purdue Refrigeration Conferen&. ASHRAE/ Purdue Conference, 173. Thomas, R. H. P., Pham, H. T. and Wu, W-T (1991). ASHRAE Journal, February, 37. Thomas, R. H. P., Wu W-T. and Pham, H.T. (1991) Paper #48, XVlll* International Congress of Refrigeration. Thomas, R. H. P. and Pham, H. T. (Accepted for ASHRAE 1992) van Gaalen, Nj A., Pate M. B.,Zoz, S.C. (1990). ASHRAE Transactions, 96, Pt. 2, 100. 381
. Figure 1: Pressure Vs. Volume Fraction of Oil at different temperatures for Solubility of R32/R125 in BRL 150. 500 400 n 4 Z 300 CL W E 7 v) E v, 200 L1 100 0 : 10 C : 20 C v : 30 C : 40 C a : 50 C. Calculaticr 0 0.0 0.2 0.4 0.6 0.8 Volume Fraction of Oil 1.0 382
. Figure 2 : Relative Pressure Vs. Volume Fraction of Oi 1 for R32/R125, BRL 150. L 7 cn cn aj.- > c 0 : Experimental. Calculation 7. Volume Fraction of Oil 383