3282 (RP-588) Measurement and Formulation of the Thermodynamic Properties of Refrigerants 134a (1,1,1,2-Tetrafluoroethane) and 123 (1,1-Dichloro-2, 2, 2-Trifluoroethane) M.O. Mclinden, Ph.D. J.S. Gallagher L.A. Weber, Ph.D. G. Morrison, Ph.D. Associate Member ASHRAE D. Ward A.R.H. Goodwin, Ph.D. M.R. Moldover, Ph.D. J.W. Schmidt, Ph.D. H.B. Chae, Ph.D. T.J. Bruno, Ph.D. J.F. Ely, Ph.D. M.L. Huber, Ph.D. ABSTRACT duction present a major challenge to the refrigeration The thermodynamicproperties of R134a and R123 industry. There are a number of very promising fluids that are formulated using a modified Benedict..JNebb-_Flubin may serve as substitutes, either as pure fluids. as constitu­ (MBWR) equation of state fit to experimental measure­ ents of mixtures, or both. In particular R134a (CF3CFH2} ments of thecritical point, vaporpressure, saturatedliquid and R123 (CCl2HCF3) have emerged as leading candi­ and vapor volumes, superheated pressure-volume­ dates to replace R12 and R11, respectively;They are under temperature(p-V-T) behavior, and secondvif1a/ coefficients active worldwidedevelopment by a number ofchemical derived from p-V-T and sound speed measurements. manufacturers working with the refrigeration and air­ The heat capacityof the ideal gas referencestate is de­ conditioning industry. termined from sound speed measurements on the low In order to evaluate the performance (energy effi­ densityvapor. Surfacetensions arealso presented. The ciency, capacity, etc.) of these or any other replacement experimentalmethods and resultsare summarized, com­ working fluids in heat pumping, air-conditioning, or re­ paredto the propertyformulation and, where'possible, frigerating applications, a complete set of thermodynamic compared to other sources in the literature. Tables and properties is required. Developmentefforts by equipment diagramsof the thermodynamicproperties of R134a a.nd manufacturers havebeen hampered by property data that R123, preparedusing the MBWR equation of state, are are sparse, conflicting, proprietary or, in some cases, presented. While the various measurements cover dif­ simply unavailable (Hickman 1988). The situation with ferent ranges of temperature and pressure, the MBWR R134a has been considerably improved by the recent formulation is applicable in both the liquid and vapor publication of measurements by Wilson and Basu (1988) phases at pressures up to 10,000 kPa (1500 psla); the and Kabata eta!. (1988). The data for·R123, however, re­ applicable temperaturerange Is 233 to 450 K (-400 to main very sparse. at least in the openliterature. Even for 350°F) for R134a and 255 to 450K (0° to 350°F)for R123. R134a, values listed on propertysummary sheets pre­ pared by various chemical manufacturers are often INTRODUCTION inconsistent. The advent of the Montreal Protocol restricting the The objectives of the present project are to measure future production of certain fully halogenated chlorofluoro­ for R134a and R123 the basic data necessary to define the carbon (CFC)refrigerants and more recentcalls frommany thermodynamic surface, to fit these data, along with other quarters to completely or nearly completely ban their pro- data available in the literature, to an equation of state, and M.O. Mclinden, T.J. Bruno, and M.L. Huber are chemical engineers and J.F. Ely is a group leader in the Thermophysics Divi­ sion, National Institute of Standards and Te chnology, Boulder, CO; J.S. Gallagher, L.A. Weber, and J.W. Schmidt are physicists, G. Morrison is a research chemist, A.R.H. Goodwin and H.B. Chae are guest workers, and M.A. Moldover is a group leader in the ThermophysicsDivision and D. Wardis an engineering technician in the Building Environment Division, NIST,Gaithersburg, MD. 263 to use this equation of state to generate tables and Once the Helmholtz free energy is evaluated. the other diagrams of the thermodynamic properties. This paper will thermodynamic functions follow Entropy 1s given by focus on the last two aspects-fitting the data and present­ = ing them in a convenient and usable form. The measure­ S - (5) ment techniques themselves and the rationale behind [!i] V them will be outlined, but the full details, including the actu­ · so that the difference between S and the reference state al experimental data, will be presented elsewhere (e.g., entropy is given by the derivative of Equation 4: Chae et al. 1989; Goodwin and Mo/dover 1989; Weber 1989; Morrison and Ward 1989; Weber and Levelt-Sengers s-S' = -[: (A-A')J (6) 1989). Any property formulation is merely a correlation of r , the actual thermodynamic behavior of a fluid and is The results of the integration in Equation 4 and the differen­ therefore subject to change as additional experimental or theoretical information becomes available. The property tiation in AEquation 6 represent only the volume depen­ dence of and S. The temperature dependence of these formulations to be presented here are felt to represent the quantities is contained in the ideal gas reference state; for actual thermodynamic properties of these fluids as well as entropy: possible given the present data; while fully adequate for T engineering calculations, they must. however, be con­ r q,/ So = S er + f. T dT (7) sidered provisional. A final formulation of properties must T ref await more extensive measurements, which should be under way by the time this report is published. where T,et is an arbitrary reference temperature at which REQUIREMENTS FOR THE DEFINITION entropy is set to an arbitrary reference= value, =S,e1; a com­ OF THE THERMODYNAMIC SURFACE mon °choice for refrigerants is S,et 0 at T,e1 233.15 K (-40 F). Thus, in addition to an equation of state (e.g., For calculations involving thermodynamic properties, Equation 1) a second fundamental requirement to define one rarely has a collection of experimental data that is suffi­ the thermodynamic behavior of a fluid is knowledge of cg, ciently large to permit interpolation at all points of interest. the heat capacity of the vapor in the limit of zero pressure. Instead one employs an equation of state. An equation of (Other reference states, which, in turn, would require calori­ state is a function that can correlate all the thermodynamic metric information other than cg, can be defined. The properties of a fluid in a consistent way. Furthermore, it is choice of the ideal gas reference state and the resulting a fundamental relationship from which all the thermo­ need for � information is the most convenient from both dynamic properties of a material can be derived by apply­ numerical ;nd experimental viewpoints.) ing the appropriate thermodynamic operations. An _ The other thermodynamic quantities arise by similar equation of state thus allo'NSthe calculation of quantities, operations. As examples: such as entropy, that may not have been, or cannot be, H -A+ TS+ pV directly measured. = (8) An equation of state typically expresses the pressure, p, O as a function of absolute temperature, T, molar volume, H...:.1f0 = (A-A )+ T(S-So)+(pV.:..RT) (9) 1/, and a set of adjustable parameters, a;: = p f(T, V, a1) Cv T [:� (10) = (1) V Other thermodynamic quantities are obtained by = manipulation of Equation 1. Integration of Equation 1 over Cv-� .T }_(S-5°) -· (11) . _ [ar . ] volume yields the Helmholtz free energy: . V A= - p dV f (2) -� Gg-R = (12) Care must be taken in choosing the limits of this integration; 2 the most convenient choice for the upper limit is V = oo, = T_-[!2.1 /[!2.l Cp C.,·� (13) where all gases follow the ideal gas law: · · arJ. avJt = p RT/V (3) The saturation pressure (vapor pressure) is obtained from an equation of state by_ the equality of pressures and V pdV In the limit of = oo,however, the integral of diverges; by use of t.he criteria for phase equilibria: this problem is avoided by integrating over the difference between the actual pressure as given by Equation 1 and (14) the pressure of the ideal gas reference state (Equation 3) where G, the Gibbs free energy, is evaluated forthe specific to obtain the difference between the actual Helmholtz0 free A. A : r v energy, and that of the reference state, volumes of the liquid and vapor phases. V and V . G is m given by · = (4) A(V,T)-A'(V,T) i (p-RT/V)dv = 0 + V G- (JO (A-A ) (pV-RT) (15) 264 An equation of state, along with information on the Usually U1e volume ratio is chosen to have a value between ideal gas heat capacity, thus ties together all the thermo­ 1.5 and 2 (1.78 in our case); it is a very weak function of dvnamic properties of a fluid. Which propertiesto measure pressure as indicated by the p's in Equation 16. The in. order to define a fluid is dictated more by the form of the· volume ratio, including its pressure dependence. was equation of state and the accuracy and ease with which determined in a separate experiment by a calibration vvith one can measure various quantities than by which quanti­ helium. ties are required for a particular application, such as the The specific volume of the sample in successive ex­ calculation of cycle efficiency. The measurements carried pansions follows from the definition of the volume ratio: out in this work consist of pressure-volume-temperature (p-V-T) measurements along lines of constant volume in the (17) vapor phase, saturated liquid density and vapor pressure over a range of temperature, and a determination of the The volume for any given expansion can then be ex­ critical point parameters. These data represent the mini­ pressed as: mum information necessary to define an equation of state covering the range of temperature and pressure of interest (18) for refrigeration applications.