i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2

www. i ifi i r .org

ava i lab le at www.sc iencedi rec t .com

journa l homepage : www. e lsev ier . com/ loca te / i j re f r ig

The fluorinated olefin R-1234ze(Z) as a high-temperatureheat pumping refrigerant

J. Steven Browna,*, Claudio Ziliob, Alberto Cavallinib

aDepartment of Mechanical Engineering, Catholic University of America, 620 Michigan Ave., NE, Washington, DC 20064, USAbDipartimento di Fisica Tecnica, Universita di Padova, Padova, 35131, Italy

a r t i c l e i n f o

Article history:

Received 12 November 2008

Received in revised form

13 February 2009

Accepted 8 March 2009

Published online 21 March 2009

Keywords:

Heat pump

Refrigerant

Review

R-1234ze(Z)

Thermodynamic property

* Corresponding author. Tel.: þ1 202 319 517E-mail address: brownjs@cua.edu (J.S. Bro

0140-7007/$ – see front matter ª 2009 Elsevidoi:10.1016/j.ijrefrig.2009.03.002

a b s t r a c t

Estimates are provided for R-1234ze(Z) of its: (1) critical temperature, pressure, and density,

acentric factor, and ideal gas specific heat at constant pressure, and (2) various thermo-

dynamic and transport properties, which are used to predict the performance potential of

R-1234ze(Z) in high-temperature heat pumping applications. In particular, for an idealized

cycle, the coefficient of performance and volumetric heating capacity for R-114 are 3.24 and

1667 kW m�3, respectively, and for R-1234ze(Z) are 3.40 and 1645 kW m�3, respectively. The

attractiveness of R-1234ze(Z) is confirmed further through heat exchanger simulations.

This paper demonstrates that R-1234ze(Z) deserves further consideration as a possible

R-114 replacement.

ª 2009 Elsevier Ltd and IIR. All rights reserved.

Le R-1234ze(Z) : un olefine fluore utilise dans les applicationshaute temperature des pompes a chaleur

Mots cles : Pompe a chaleur ; Frigorigene ; Enquete ; R-1234ze(Z) ; Propriete thermodynamique

1. Introduction

R-114, a refrigerant that has been widely used in high-

temperature heat pumping applications, has an ozone

depletion potential (ODP) of 1 (EPA, 2009) and a global warming

potential (GWP) of 10,400 (EPA, 2009). Because of its ODP, it is

a regulated substance as per the Montreal Protocol (1987).

While R-114 has been replaced in some applications with

R-134a, R-227ea, R-236fa, or R-245fa, for example, there is no

0; fax: þ1 202 319 5173.wn).

er Ltd and IIR. All rights

clear and obvious substitute for R-114 since these refrigerants

all have relatively large GWPs and R-134a and R-227ea have

relatively low critical point temperatures, limiting their

applicability to lower-temperature applications. Thus,

a search for acceptable R-114 replacement refrigerants

continues. For example, Devotta and Rao Pendyala (1994)

investigated 30 potential substitutes for R-114 and identified

R-143 and R-E134 as the most promising candidates; Rakhesh

et al. (2003) investigated R-227ea as a substitute for R-114 in

reserved.

Nomenclature

COP Coefficient of performance

cp specific heat at constant pressure (kJ kg�1 K�1)

cop ideal gas specific heat at constant pressure

(kJ kg�1 K�1)

e % relative error, ðyp � ykÞ=yk � 100

ea mean absolute error,Pn

i¼1 jeij=ner mean relative error,

Pni¼1 ei=n

GWP global warming potential

h specific enthalpy (kJ kg�1or kJ kmol�1)

M molar mass (kg kmol�1)

NBP normal boiling point temperature (K)

ODP ozone depletion potentinal

P pressure (kPa)

PF penalty factor (a performance evaluation

criterion)

Pr Prandtl number

Q heat transfer rate (kW)

R specific gas constant (kJ kg�1 K�1)

s specific entropy (kJ kg�1 K�1 or kJ kmol�1 K�1)

T temperature (�C or K)

VCC volumetric cooling capacity (kJ m�3)

VHC volumetric heating capacity (kJ m�3)

y property value, e.g., P, T, l, .

Z compressibility factor

g surface tension (N m�1)

l thermal conductivity (W m�1 K�1)

m dynamic viscosity (mPa s)

u acentric factor

r density (kg m�3)

s standard deviation

Subscripts

c critical

cond condenser

evap evaporator

f saturated liquid

fg difference between saturated vapor and saturated

liquid property values

g saturated vapor

k known value

p predicted value

r reduced value

sub subcooling

sup superheat

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2 1413

heat pumps with evaporating temperatures above 30 �C and

condensing temperatures up to 85 �C; Toms et al. (2004) report

on U.S. Navy activities to replace R-114 with R-236fa; and

Brown (in press) investigated 53 potential R-114 replacements

and identified 15 of these as warranting further investigation.

Recently, research and development activity has been

initiated and is ongoing to investigate fluorinated propene

isomers as potential refrigerants possessing low GWPs. The

catalyst for much of this effort can be attributed to European

regulations regarding the use of R-134a in automotive appli-

cations. In particular, the European Union’s f-gas regulations

(Regulation (EC) No 842/2006 and Directive 2006/40/EC) specify

beginning on January 1, 2011 new models and on January 1,

2017 new vehicles fitted with air-conditioning cannot be

manufactured with fluorinated greenhouse gases having

global warming potentials (GWP) greater than 150. Ongoing

research and development efforts by chemical manufacturers

have identified that, from among the fluorinated propene

isomers, the most promising R-134a replacement fluid is R-

1234yf (e.g., Spatz and Minor, 2008). However, there are other

potential applications for fluorinated propene isomers beyond

the automotive industry. For example, the civilian aircraft

industry is moving towards the replacement of hydraulic

power and control systems with power electronic systems. To

realize this goal many issues remain to be resolved; however,

among the challenges, one of the most difficult is the

management of thermal loads. This is because as the number

of electrical and electronic systems increase, their physical

sizes decrease, and the spacing between electrical compo-

nents decrease, both the total amount of heat generated (and

thus which needs to be dissipated) and the power density (the

heat generated per unit volume) increase significantly. One

approach for solving this problem could be to use vapor

compression refrigeration cycles to manage the thermal

loads, requiring refrigerants appropriate for civilian aircraft

that have little to no toxicity, are essentially nonflammable,

and which have low GWP. Thus, the newly developed fluori-

nated propene isomers, among which is R-1234ze(Z), could be

potentially interesting for these civilian aircraft applications

where the sink temperatures can be as high as 70 �C.

This paper investigates one of the fluorinated propene

isomers, namely, R-1234ze(Z) as a potential refrigerant for use

in high-temperature heat pumping applications. (The reader

can consult ASHRAE (2008) for the naming convention for the

fluorinated propene series.) While little thermodynamic and

transport property data are available in the open literature for

R-1234yf (the most actively investigated fluorinated propene

isomer), an even scarcer amount of data are available in the

open literature regarding the other, less studied, fluorinated

propene isomers, including R-1234ze(Z). Thus, to begin, this

paper uses the methodologies illustrated in Brown (2007a–c,

2008, in press) to predict the thermodynamic property values

of R-1234ze(Z) knowing only its normal boiling point temper-

ature (NBP) and its structural formula, and provides in Section

2.2 several methodologies for estimating the transport prop-

erties of R-1234ze(Z).

Briefly, the paper proceeds as follows. Sections 2.1 and 2.2

present methodologies for predicting thermodynamic and

transport properties of not-so-well-defined refrigerants (ones

where little or no thermodynamic or transport property data

are known). Then, to demonstrate the capabilities of these

methodologies, Sections 2.3 and 2.4 provide comparisons

between predicted property values and known property

values for a group of 28 well-defined refrigerants. Sections 3.1

and 3.2 discuss approaches for evaluating the performance

potential of refrigerants first in an idealized vapor compres-

sion refrigeration cycle and secondly by considering non-ideal

behavior in the evaporator and in the condenser. Then,

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 21414

Section 4 proceeds to analyses for R-1234ze(Z). In particular, to

demonstrate the accuracy and usefulness of the estimation

techniques of Section 2.1, Section 4.1 uses these techniques to

estimate the thermodynamic properties of R-1234yf and

compares them with published values. Sections 4.2–4.4

compare the fundamental thermodynamic parameters, ther-

modynamic properties, and transport properties of R-114 and

R-1234ze(Z). Section 4.5 compares the performance of R-114

and R-1234ze(Z) in an idealized vapor compression refrigera-

tion cycle. Section 4.6 compares the performance of R-114 and

R-1234ze(Z) in a non-ideal evaporator and in a non-ideal

condenser. Section 4.7 discusses the environmental charac-

teristics, flammability, and toxicity of R-114 and R-1234ze(Z).

Finally, Section 5 presents conclusions.

2. Estimating properties

2.1. Thermodynamic properties

If little experimental data are available for a refrigerant,

a simple, fast, and accurate approach for estimating thermo-

dynamic properties is to use group contribution methods

coupled with a simple cubic equation of state. Brown (2007a)

demonstrated the ease of implementing such an approach by

using several group contribution techniques coupled with the

Peng–Robinson Equation of State (P–R EoS) contained in

REFPROP 8.0 (Lemmon et al., 2007). In order to use the P–R EoS,

one needs to know values for the thermodynamic parameters

critical temperature (Tc), critical pressure (Pc), critical density

(rc), acentric factor (u), and ideal gas specific heat at constant

pressure ðcopÞ. For not-so-well-defined fluids, most, if not all, of

these are unknown and thus one must resort to estimating

them using group contribution techniques based only on the

fluid’s NBP and its structural formula. To this end, one could:

(1) use the group contribution method of Ambrose (Reid et al.,

1987) to predict Tc, Pc, and rc, (2) use the technique given in

Poling et al. (2001) to calculate u, and (3) use the group

contribution method of Joback (Poling et al., 2001) to deter-

mine cop. Then, the P–R EoS implemented in REFPROP 8.0

(Lemmon et al., 2007) could be used to determine the neces-

sary thermodynamic properties, e.g., T, P, r, h, s, and cp,g. For

cp,f, one either could use the same method described above or

one could obtain somewhat better predictions by using the

modified Bondi method described in Poling et al. (2001).

An approach that provides somewhat more accurate

results than the one described above and thus the one rec-

ommended by the current authors is to follow the approach of

Brown (2007b). In particular, he showed that Tc and rc

predictions are generally more accurate than the Pc predic-

tions, with absolute errors (ea) being 0.7%, 2.4%, and 7.7%,

respectively, for the group of refrigerants of Saleh and

Wendland (2006). Because of this, Brown (2007b) used Tc and rc

group contribution estimates to predict Pc through the critical

compressibility factor:

Zc ¼Pc

rcRTc

(1)

where Zc is a function of Tc, Pc and M, and R is the specific gas

constant. Zc typically ranges in value from 0.22 to 0.30 for

a wide range of fluids (Wark, 1995). Brown (2007b) used a group

of 42 typical refrigerants and their known values of Tc, Pc and

rc to calculate Zc, resulting in a mean value of 0.27, a minimum

value of 0.24, a maximum value of 0.28 and a standard devi-

ation of 0.012. Therefore, the approach recommended by the

authors of this paper is to: (1) use the group contribution

method of Ambrose (Reid et al., 1987) to predict Tc and rc, (2)

use Eq. (1) to predict Pc, (3) use the technique given in Poling

et al. (2001) to calculate u, and (4) use the group contribution

method of Joback (Poling et al., 2001) to determine cop. Then, the

P–R EoS implemented in REFPROP 8.0 (Lemmon et al., 2007)

can be used to determine the necessary thermodynamic

properties.

2.2. Transport and other properties

For not-so-well-described fluids, it is unusual to find transport

property data published in the open literature. Thus, one

almost certainly will need to resort to estimation techniques,

which require little thermodynamic property data as input. In

fact, the objective of this subsection is to highlight some

appropriate models for the estimation of transport properties.

In particular, the liquid thermal conductivity (lf) can be esti-

mated using the method of Latini et al. (Poling et al., 2001), the

vapor thermal conductivity (lg) can be estimated using the

method of Chung et al. (Poling et al., 2001), the liquid dynamic

viscosity (mf) can be estimated using the group contribution

method of Sastri and Rao (2000) and Poling et al. (2001), the

vapor dynamic viscosity (mg) can be estimated using the

method of Chung et al. (Poling et al., 2001), and the liquid

surface tension (g) can be estimated using the method of

Sastri and Rao (Poling et al., 2001). Note: if one knows one

experimental value for lf, improved predictions over those of

Latini et al. (Poling et al., 2001) can be made using Latini (2008).

A further note: the thermodynamic properties needed to make

lf, lg, mf, mg, and g predictions all can be estimated following

the approach described in the previous subsection.

2.3. Errors in property predictions

As an example, Table 1a shows relative errors (er) for ther-

modynamic and transport properties for a group of 28 of the 36

refrigerants of Saleh and Wendland (2006) at a reduced

temperature (Tr) of 0.70. Table 1b then shows for the group of

28 refrigerants, ea, er, and standard deviation (s) values at

Tr¼ 0.70 and Tr¼ 0.90. Note: only 28 refrigerants of the original

36 are included since they are well-described and are con-

tained in REFPROP 8.0 (Lemmon et al., 2007); whereas the

remaining eight refrigerants are not as well described and do

not make up part of the database of REFPROP 8.0.

2.4. Discussion

Generally, errors for the predicted values of the saturation

thermodynamic properties P, r, hfg, sfg, and cp,f are within

a few percent of known values over the temperature range

from Tr¼ 0.60 to Tr¼ 0.90. The predictions for cp,g are some-

what larger, though this property plays a less significant role

in estimating heat transfer effects than does cp,f. The errors for

the thermodynamic saturation properties begin to increase

Table 1a – Relative errors for thermodynamic and transport properties for 28 refrigerants of Saleh and Wendland (2006) atTr [ 0.70.

P rl rv hfg sfg cp,l cp,v ll lv ml mv sl

R-11 �2.0 4.7 �2.4 1.2 1.2 0.9 �8.3 1.4 �7.4 14.4 10.6 0.7

R-12 �2.2 4.6 �2.6 0.6 0.6 0.6 �6.1 5.7 �5.7 15.4 10.5 2.3

R-22 �3.9 1.4 �5.0 0.5 0.5 4.3 �2.8 7.6 6.1 7.1 4.0 �2.3

R-32 �3.1 �1.4 �6.3 4.6 4.6 8.5 �18.5 �26.6 8.5 29.3 0.8 �4.0

R-41 0.6 �0.1 �2.9 7.0 7.0 7.3 �32.0 �29.8 �12.7 �11.9 8.5 5.8

R-114 �1.8 3.2 �1.6 �0.1 �0.1 �3.8 �7.6 11.9 �7.2 28.4 6.8 3.3

R-123 �2.5 1.9 �3.5 0.5 0.5 0.6 �3.1 5.7 �1.2 �0.8 �0.2 2.5

R-124 �3.1 0.9 �3.9 �0.7 �0.6 �1.3 �6.4 2.1 �8.2 �5.8 2.2 �0.5

R-125 �3.8 �1.5 �4.5 �2.3 �2.3 �3.6 �7.3 �6.3 �8.8 �11.7 1.6 �2.9

R-134a �4.1 �0.7 �5.5 �0.4 �0.4 �1.0 �7.8 �8.7 �3.7 �9.0 1.6 �2.7

R-141b �1.6 3.8 �2.3 2.3 2.3 1.5 �8.5 6.2 �11.6 7.1 6.8 0.7

R-142b �1.9 4.0 �3.0 2.3 2.3 0.7 �9.7 6.0 �1.5 �2.1 4.0 2.3

R-143a �2.3 3.5 �4.5 3.0 3.0 0.5 �12.6 2.4 �6.9 �11.2 9.6 17.2

R-152a �2.4 2.1 �4.6 3.6 3.6 5.0 �7.0 0.4 0.4 2.7 4.7 2.8

R-170 �3.2 0.9 �3.3 �2.1 �2.1 �6.3 �16.1 12.3 �13.1 21.5 �2.2 4.8

R-218 �1.2 �1.5 �1.6 �1.3 �1.3 �6.2 �14.2 6.1 �5.1 7.0 0.7 1.4

R-227ea �1.5 �1.9 �2.0 �1.8 �1.8 �3.3 �7.6 �2.5 �14.0 �11.5 1.2 �1.5

R-236ea �2.0 �3.4 �2.8 �1.2 �1.2 �3.0 �7.9 �11.5 �23.2 �15.6 �0.6 �7.7

R-236fa �2.2 �1.0 �2.8 �2.0 �2.0 �3.7 �4.8 �7.5 �8.5 �19.2 1.5 �4.5

R-245ca �2.0 �3.7 �3.2 1.2 1.2 �0.8 �8.5 �13.8 �19.6 �12.9 �1.5 �10.4

R-245fa �2.9 �2.3 �4.4 �1.0 �1.0 0.1 �7.9 �10.2 �11.8 �20.5 0.3 �8.9

R-290 0.1 15.1 �0.8 3.4 3.4 4.2 �3.9 18.2 �6.6 45.2 0.3 8.6

R-600 �2.2 1.3 �2.8 0.2 0.2 1.2 �4.4 14.0 �5.4 2.2 �0.3 4.8

R-600a �2.2 2.3 �2.6 �0.1 �0.1 1.2 �2.9 26.4 �5.1 �8.3 0.8 3.8

R-1270 �3.0 2.4 �3.5 �0.9 �0.9 �2.2 �7.4 1.3 �6.2 �9.0 2.8 3.0

R-C318 �1.7 1.2 �1.5 �3.7 �3.7 �15.7 �19.7 �11.0 �21.0 �0.6 3.5 �3.3

R-13I1 �2.0 5.0 �2.8 0.9 0.9 7.4 �2.7 9.8 0.4 �3.0 10.1 3.3

R-E170 �2.1 2.3 �4.4 3.8 3.8 4.6 �12.7 �1.6 �9.6 4.2 �1.3 7.0

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2 1415

somewhat from Tr¼ 0.90 to Tr¼ 0.95 and begin to increase

more rapidly above Tr¼ 0.95. Although not shown here, ther-

modynamic property predictions in the superheat region are

within a few percent of known values.

On the other hand, errors for the predicted values of the

transport properties l and m are somewhat higher than those

for thermodynamic properties. Generally, the errors are: (1)

within �15% but as high as 30% for lf, (2) 0 to �20% for lg, (3)

within �20% but as high as 40% for mf, and (4) 0 to þ10% for mg.

Finally, errors for the predicted values of g generally are�10%.

Note: as mentioned in Section 2.2, if one knows an experi-

mental value for lf, improved predictions can be made using

Latini (2008). It should be noted, however, that accurate

predictions of the liquid transport properties are more impor-

tant for making good heat transfer and pressure drop estimates

than are accurate predictions of the vapor transport properties.

Table 1b – Errors and standard deviations for thermodynamicTr [ 0.70 and Tr [ 0.90.

P rf rg hfg sfg cp,f

Tr¼ 0.70

ea 2.27 2.79 3.25 1.88 1.88 3.56

er �2.22 1.54 �3.25 0.64 0.63 �0.09

s 1.03 3.63 1.30 2.42 2.42 4.96

Tr¼ 0.90

ea 3.24 5.14 4.05 2.70 2.70 2.70

er �1.89 �4.62 �2.69 1.20 1.20 1.37

s 3.54 3.37 4.31 3.72 3.72 3.45

3. Predicting thermodynamic performance ofrefrigerants

3.1. Idealized performance

Once a refrigerant’s thermodynamic properties are known,

one can estimate its performance potential [coefficient of

performance (COP) and volumetric cooling (or heating)

capacity (VCC or VHC)] in an idealized vapor compression

refrigeration cycle. Brown (2007a) proposed three such meth-

odsdfrom a less accurate one to a more accurate onedwith

each method being based on the P–R EoS. The least accurate

method used group contribution techniques to estimate Tc, Pc,

rc, u, and cop from a refrigerant’s known NBP and its structural

formula. The other two methods used known critical state

and transport properties of the refrigerants of Table 1a at

cp,g lf lg mf mg g

9.23 9.54 8.54 12.05 3.53 4.39

�9.23 0.28 �7.45 1.12 3.09 0.91

6.35 12.45 7.19 15.83 3.90 5.66

16.07 10.86 9.39 14.54 8.38 5.00

�16.07 �7.94 �9.39 11.32 7.13 2.91

6.29 11.25 6.24 15.85 6.19 5.53

Table 2 – Errors and standard deviations for 26refrigerants.

COP VCC

Brown(2007a)

Currentpaper

Brown(2007a)

Currentpaper

ea 2.04 1.74 10.10 3.08

er 1.26 0.93 �6.76 0.17

s 2.74 2.49 10.88 4.17

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 21416

properties, with the distinction between the two methods

being that the less accurate of the two used a group contri-

bution method to estimate cop and the more accurate one used

a known cop model. The current paper presents an improved

version (by mainly improving VCC predictions) of the least

accurate of the three methods, which is particularly useful for

not-so-well-described refrigerants, such as R-1234ze(Z). Note:

the least accurate method of Brown (2007a) is the one

described in the first paragraph of Section 2.1 and the

improved version presented in this paper is described at the

end of the same section.

Table 2 summarizes the errors in COP and VCC predictions

for 26 of the 36 refrigerants of Brown (2007a). In particular, the

ea for COP and VCC are 2.04% and 10.10%, respectively, for the

Brown (2007a) predictions and are 1.74% and 3.08%, respec-

tively, using the improved version described at the end of

Section 2.1. Note: only 26 refrigerants are used since they form

part of and fall within the operating limits of the high-accu-

racy database of REFPROP 8.0 (Lemmon et al., 2007) and thus

easily and accurately can be verified. However, to further give

a sense of the possible improvements in VCC predictions

using the approach described above, the remaining nine fluids

(R-744 is excluded since its Tc is too close to the condensing

temperature specified below and thus biases the results) not

contained in REFPROP 8.0 (Lemmon et al., 2007) are compared

with the BACKONE predictions of Saleh and Wendland (2006).

In particular, the ea for COP and VCC are 1.39% and 10.31%,

respectively, for the Brown (2007a) predictions and are 1.35%

and 6.05%, respectively, using the improved version described

at the end of Section 2.1. In each case described above, the COP

and VCC were simulated in the idealized vapor compression

refrigeration cycle of Saleh and Wendland (2006), specified by

a constant condensation temperature of 30 �C, a constant

evaporation temperature of �40 �C, an evaporator superheat

Table 3a – EVAP–COND 2.3 evaporator simulation results using

Qevap (kW) DTsu

Use ofactual

propertyvalues

Use ofpredicted

property values

% Relativeerror in

Qcond

Use of actualpropertyvalues

R-22 3.55 3.58 0.9 17.6

R-32 5.73 6.21 8.3 15.2

R-134a 2.93 3.06 4.5 18.7

R-290 4.86 4.93 1.3 17.4

R-600a 2.54 2.55 0.2 20.2

of 5 �C, a condenser subcooling of 5 �C, a compressor isen-

tropic efficiency of unity, and with no pressure losses in the

heat exchangers.

3.2. Non-ideal performance: heat transfer and pressuredrop effects

An actual refrigeration cycle deviates from ‘‘ideal’’ due to

several exergy losses; for example, heat transfer and pressure

drop effects in the refrigerant lines, heat transfer effects in the

compressor, re-expansion losses in the compressor, pressure

losses across the compressor’s suction and discharge valves,

and heat transfer and pressure drop effects in the heat

exchangers, among others. The heat transfer and pressure

drop effects in the refrigerant lines will not be considered

further here, and the cumulative compressor effects can be

modeled using compressor efficiencies, leaving only heat

transfer and pressure drop effects in the heat exchangers to be

considered in this paper.

In order to assess the performance potentials in an evap-

orator and a condenser, we use the public domain simulation

model EVAP–COND 2.3 (Domanski, 2008), which is a tube-

finned, tube-by-tube simulation model for both the evapo-

rator and condenser. The refrigerant library of EVAP–COND

consists of seven single-component refrigerants (R-22, R-32,

R-134a, R-290, R-600a, R-717, R-744) and four blends (R-404A,

407C, R-410A, R-507A).

To assess the capabilities of the transport property

prediction methodologies described in Section 2.2, this section

compares simulation results for an evaporator and

a condenser (the heat exchangers used are the ones named

EVAP and COND in EVAP–COND’s library). Two sets of simu-

lations are made for five single-component refrigerants (R-22,

R-32, R-134a, R-290, R-600a), namely: (1) using EVAP–COND’s

default fluid properties and (2) using thermodynamic and

transport properties estimated using the methodologies

described at the end of Section 2.1 and in Section 2.2.

Table 3 presents the simulation results for the evaporator

and the condenser. The operating conditions used in the

simulations are not reported since they are of secondary

importance. The point of the simulation results in this

subsection is not the absolute results, but rather is to

demonstrate the ability of and the usefulness of the property

prediction methodologies to be used in a heat exchanger

actual property values and using predicted property values.

p (�C) DPevap (kPa)

Use ofpredicted

property values

Use of actualpropertyvalues

Use ofpredicted

property values

% Relativeerror inDPcond

17.8 26.6 27.6 3.7

14.6 26.2 36.5 39.3

20.3 28.9 38.8 34.3

17.4 28.2 29.0 2.8

20.1 25.1 24.7 �1.6

Table 3b – EVAP–COND 2.3 condenser simulation results using actual property values and using predicted property values.

Qcond (kW) DTsub (�C) DPcond (kPa)

Use of actualpropertyvalues

Use ofpredictedpropertyvalues

% relativeerror in

Qcond

Use of actualpropertyvalues

Use ofpredictedpropertyvalues

Use ofactual

propertyvalues

Use ofpredictedpropertyvalues

% Relativeerror inDPcond

R-22 8.67 8.73 0.7 9.2 9.1 32.3 35.4 9.6

R-32 11.98 12.62 5.3 9.1 6.1 32.8 47.5 44.8

R-134a 7.33 7.74 5.5 9.1 12.4 27.9 20.5 �26.5

R-290 9.76 9.88 1.2 8.9 9.0 32.4 32.5 0.3

R-600a 7.41 7.47 0.8 8.9 8.3 30.0 26.2 �12.7

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2 1417

model. To aid the reader, Table 4 presents er values for the five

fluids of Tables 3a and 3b.

4. R-1234ze(Z)

4.1. R-1234yf used as an example to illustrate propertyprediction capabilities of fluorinated olefins

In this section, comparisons are made for some thermody-

namic properties of R-1234yf between known values and ones

predicted by the methods described at the end of Section 2.1.

R-1234yf is chosen to illustrate the ability of the prediction

methods to accurately estimate thermodynamic properties

for a not-so-well-described refrigerant, and secondly because,

of the fluorinated propene isomers, R-1234yf is the only one

where more than a bare minimum of data have been pub-

lished in the open literature thus making a comparison

possible. Note: a similar illustration cannot be made for the

transport properties of R-1234yf since none have been pub-

lished in the open literature.

Figs. 1–3 compare the predictions of this paper with those of

Spatz and Minor (2008). In particular, Fig. 1 is a Mollier diagram,

Fig. 2 shows the saturation pressure curve, and Fig. 3 shows the

saturated vapor density curve. The largest er values for the

saturated liquid and saturated vapor enthalpies are approxi-

mately 3.5%. The largest er value for the saturation pressure

occurs at the lowest saturation temperature and is on the order

of 3%. The largest er value for the saturated vapor density also

occurs at the lowest saturation temperature and is on the order

of 4%. In summary, the predictive methods provide quite

reasonable estimates for the thermodynamic properties of

a not-so-well-described fluorinated propene isomer and thus

demonstrate their usefulness for predicting other even less-

well-described propene isomers, e.g., R-1234ze(Z).

Table 4 – Relative errors in predicted property values for the fivfrom REFPROP 8.0 (Lemmon et al., 2007).

P rf rg hfg sfg cp,

R-22 �3.9 1.4 �5.0 0.5 0.5 4.

R-32 �3.1 �1.4 �6.3 4.6 4.6 8.

R-134a �4.1 �0.7 �5.5 �0.4 �0.4 �1.

R-290 0.1 15.1 �0.8 3.4 3.4 4.

R-600a �2.2 2.3 �2.6 �0.1 �0.1 1.

4.2. Fundamental thermodynamic parameters of R-114and R-1234ze(Z)

Table 5 provides the fundamental thermodynamic parame-

ters Tc, Pc, rc, u, and cop for R-114 and R-1234ze(Z). The values

for R-114 are taken from REFPROP 8.0 (Lemmon et al., 2007)

and the values for R-1234ze(Z) are estimated by the method

described at the end of Section 2.1 knowing only its NBP and

its structural formula. The NBP for R-1234ze(Z) is taken from

Mukhopadhyay et al. (2008). The values provided in Table 5

demonstrate the possible attractiveness of R-1234ze(Z) as

a substitute for R-114.

4.3. Thermodynamic properties of R-114 and R-1234ze(Z)

Figs. 4–7 compare some of the thermodynamic properties of

R-114 (taken from REFPROP 8.0) with those predicted for

R-1234ze(Z). In particular, Fig. 4 shows the Mollier diagrams,

Fig. 5 shows the T–s state diagrams, Fig. 6 shows the saturation

pressure curves, and Fig. 7 shows the saturated vapor density

curves. The R-1234ze(Z) properties are estimated using the

fundamental thermodynamic parameters provided in Table 5

coupled with the P–R EoS implemented in REFPROP 8.0 (Lem-

mon et al., 2007). Note: the properties on the abscissas of Figs.

4 and 5 are plotted per kmol in order to make possible

a comparison between the two refrigerants, which possess

widely different molecular weights. Also, in each case the

reference state has been modified from the IIR convention

(h¼ 200 kJ kg�1 and s¼ 1 kJ kg�1 K�1 for saturated liquid at

T¼ 0 �C) in order to make the comparison easier. In particular,

the translated property values for h (the ones plotted in Fig. 4)

are simply the ones calculated using the IIR convention sub-

tracted by the quantity [(M ) (hrefer,IIR)], and the translated

property values for s (the ones plotted in Fig. 5) are simply the

e fluids of Tables 3a and 3b. The actual property values are

f cp,g lf lg mf mg g

3 �2.8 7.6 6.1 7.1 4.0 �2.3

5 �18.5 �26.6 8.5 29.3 0.8 �4.0

0 �7.8 �8.7 �3.7 �9.0 1.6 �2.7

2 �3.9 18.2 �6.6 45.2 0.3 8.6

2 �2.9 26.4 �5.1 �8.3 0.8 3.8

Fig. 1 – Mollier diagram for R-1234yf.Fig. 3 – Saturated vapor density curve for R-1234yf.

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 21418

ones calculated using the IIR convention subtracted by the

quantity [(M ) (srefer,IIR)].

A few interesting observations immediately can be made

regarding Figs. 4 and 5. First, the critical temperature and

critical pressure of R-1234ze(Z) are somewhat greater than the

R-114 values, which implies that the COP of R-1234ze(Z) in an

ideal cycle should be somewhat larger than the R-114 value in

the same cycle. Second, for a given saturation temperature,

the width of the two-phase saturation region in Fig. 4 is wider

for R-1234ze(Z) than it is for R-114, that is, hfg on a kmol-basis

at a given saturation temperature is greater for R-1234ze(Z)

than it is for R-114. Third, the saturated vapor line in Fig. 5 for

R-1234ze(Z) does not possess a positive slope as does the one

for R-114. This implies that smaller values of superheat could

be used for R-1234ze(Z) as compared to R-114 to avoid wet-

compression.

Fig. 6 shows that the saturation pressure curves for R-

1234ze(Z) and R-114 are similar. Fig. 7 shows that the satu-

rated vapor densities for R-1234ze(Z) are approximately 40–

50% less than are the R-114 values at the same temperatures.

However, since the M of R-1234ze(Z) is approximately 35%

smaller than that of R-114 and since hfg on a kmol-basis for R-

1234ze(Z) is somewhat greater than that of R-114, the VHC of

R-1234ze(Z) in an ideal cycle should not be too different from

the R-114 value in the same cycle.

Fig. 2 – Saturation pressure curve for R-1234yf.

4.4. Transport properties of R-114 and R-1234ze(Z)

Table 6 compares mf, lf, cp,f, and saturated liquid Prandtl (Pr)

numbers for R-114 and R-1234ze(Z) at saturation temperatures

of 25 �C and 85 �C. The R-114 values are from REFPROP 8.0

(Lemmon et al., 2007) and the R-1234ze(Z) values are predicted

by the methodologies of Section 2.2. The results show that

heat transfer performance should be somewhat better for R-

1234ze(Z) than for R-114 in the evaporator and in the

condenser. The pressure drop penalties in both the evaporator

and the condenser should be slightly greater for R-1234ze(Z)

than for R-114. Therefore, considering heat transfer and

pressure drop effects together, we expect similar perfor-

mances for the evaporator and the condenser for both R-114

and R-1234ze(Z).

4.5. Idealized thermodynamic performance of R-114 andR-1234ze(Z)

The thermodynamic properties presented in Section 4.3 were

used to simulate the COP and VHC of R-114 and R-1234ze(Z) in

an idealized cycle comprising a constant evaporation

temperature of 25 �C, a constant condensation temperature of

85 �C, a condenser subcooling of 10 �C, a compressor isen-

tropic efficiency of 85%, and the minimum amount of evapo-

rator superheat that ensures saturated or superheated vapor

at the compressor outlet.

Using this cycle, the COP and VHC for R-114 are 3.24 and

1667 kW m�3, respectively, and for R-1234ze(Z) are 3.40 and

1645 kW m�3, respectively, or expressed differently, the COP

and VHC of R-1234ze(Z) are 4.9% higher and 1.3% lower,

respectively, than the R-114 values. These values are consis-

tent with the observations made in Section 4.3, and once

again, show the possible attractiveness of R-1234ze(Z) as

a possible substitute for R-114.

4.6. Non-ideal performance: heat transfer and pressuredrop effects for R-114 and R-1234ze(Z)

EVAP–COND 2.3, described in Section 3.2, was used to simu-

late the performances of R-114 and R-1234ze(Z) in an

Table 5 – Fundamental thermodynamic property values for R-114 and R-1234ze(Z).

M (kg kmol�1) Tc (K) Pc (kPa) rc (kg m�3) u (–) cop at 273 K (kJ kg�1 K�1)

R-114a 170.9 418.8 3257 580.0 0.252 0.661

R-1234ze(Z)b 114.0 426.8 3970 472.6 0.333 0.758

a Values taken from REFPROP 8.0 (Lemmon et al., 2007).b Values estimated by the method described at the end of Section 2.1 knowing only its NBP and its structural formula. The NBP is taken from

Mukhopadhyay et al. (2008).

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2 1419

evaporator and in a condenser to confirm the assessment

made at the end of Section 4.4. The thermodynamic and

transport properties for R-114 are from REFPROP 8.0 (Lemmon

et al., 2007) and for R-1234ze(Z) are from Sections 4.3 and 4.4.

While detailed analyses were not performed and thus are not

reported here, the simulations conducted do confirm that the

R-1234ze(Z) heat transfer rate can be up to approximately 20%

greater than the R-114 values in both the evaporator and the

condenser for constant saturation temperature drops (Cav-

allini, 2002). Furthermore, the penalty factor (PF) concept of

Cavallini (2002) was used to compare the two fluids (Fig. 8),

with the resulting PF of R-1234ze(Z) being some 12–15% lower,

at the same heat transfer coefficient, than R-114 value for

condensation heat transfer. Note: the PF is a performance

evaluation criterion developed by Cavallini to capture exergy

losses due to heat transfer and frictional pressure drop in the

heat exchanger, with a lower PF, at the same heat transfer

coefficient, indicating better heat transfer performance

potential for the particular refrigerant.

4.7. Other characteristics of R-114 and R-1234ze(Z)

While there is no ‘‘ideal’’ refrigerant for a particular application,

asaminimumarefrigerantshouldpossessgoodthermodynamic

and transport properties. These are obviously not the only, or

even necessarily, the primary characteristics that a ‘‘good’’

refrigerant should possess. For example, other important

characteristics include: cost, stability, environmental impact,

availability, compatibility with common materials, toxicity,

flammability, among others. However, nowadays, environ-

mental impactdas measured by ODP and GWPdare becoming

increasingly more important characteristics for a refrigerant.

Fig. 4 – Mollier diagrams for R-114 and R-1234ze(Z).

Thus, these two characteristics, together with toxicity and

flammability, will now be considered for R-114 and R-1234ze(Z).

R-114 is a CFC with an ODP of 1 (EPA, 2009) a GWP of 10,400

(EPA, 2009), and is classified as an A1 (low toxicity, nonflam-

mable) refrigerant by ASHRAE (2005). On the other hand, R-

1234ze(Z) has no ODP and almost certainly has a GWP on the

order of probably no more than 10 or so based on values of

other fluorinated propene isomers. What is unknown (no data

has been published in the open literature) is its toxicity or

flammability. However, based on published data for other

fluorinated propene isomers, R-1234ze(Z) could very likely be

toxic and/or flammable to some greater or lesser degree, and

thus before R-1234ze(Z) could be used in a particular appli-

cation extensive toxicity and flammability testing would need

to be conducted. Note: for comparison purposes, R-1234ze(E)

has a GWP of 6, a LC-50 toxicity value of at least 100,000 ppm,

and an NOAEL value greater than 12% by volume (Luly and

Singh, 2008).

5. Conclusions

The first part of this paper summarizes methods to predict

thermodynamic and transport properties knowing only

a refrigerant’s normal boiling point temperature (NBP) and its

structural formula. The steps in the approach are to:

(1) Estimate the critical temperature (Tc) and the critical

density (rc) using the Ambrose (Reid et al., 1987) group

contribution method, the critical pressure (Pc) using Eq. (1),

the acentric factor (u) using the approach of Poling et al.

Fig. 5 – T–s state diagrams for R-114 and R-1234ze(Z).

Fig. 7 – Saturated vapor density curves for R-114 and

R-1234ze(Z).

Fig. 6 – Saturation pressure curves for R-114 and R-

1234ze(Z). Fig. 8 – Condensation heat transfer penalty factors for

R-114 and R-1234ze(Z).

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 21420

(2001), and the ideal gas specific heat at constant pressure

ðcopÞ using the Joback group contribution method (Poling

et al., 2001).

(2) Use the Peng–Robinson (P–R) Equation of State (EoS)

implemented in Lemmon et al. (2007) to estimate the

thermodynamic properties: temperature (T ), pressure (P),

density (r), specific enthalpy (h), specific entropy (s), and

specific heat at constant pressure for the saturated vapor

(cp,g). Note: if the refrigerant’s critical state properties are

known, they should be used in this step in lieu of the

critical state properties estimated in Step 1.

Table 6 – Transport properties for R-114 and R-1234ze(Z).

mf (mPa s) cp,f (kJ kg�1 K�1)

R-114a R-1234ze(Z)b R-114a R-1234ze(Z)b

25 �C 273.6 306.5 0.992 1.209

85 �C 141.4 188.6 1.133 1.399

a Values taken from REFPROP 8.0 (Lemmon et al., 2007).b Values estimated by the methods described in Section 2.2.

(3) Use the modified Bondi method described in Poling et al.

(2001) to estimate the specific heat at constant pressure for

the saturated liquid (cp,f).

(4) Estimate the liquid thermal conductivity (lf) using the

method of Latini et al. (Poling et al., 2001), the vapor

thermal conductivity (lg) using the method of Chung et al.

(Poling et al., 2001), the liquid dynamic viscosity (mf) using

the group contribution method of Sastri and Rao (2000) and

Poling et al. (2001), the vapor dynamic viscosity (mg) using

the method of Chung et al. (Poling et al., 2001), and the

surface tension (g) using the method of Sastri and Rao

(Poling et al., 2001).

Following the steps described above, errors for the ther-

modynamic properties are generally within a few percent over

the reduced temperature range from Tr¼ 0.6 to Tr¼ 0.9;

whereas, errors for transport properties are generally within

�20% but can be as high as 40%. In summary, at Tr¼ 0.7 for

a group of 26 refrigerants the mean absolute errors (ea) are

P¼ 2.3%, rf¼ 2.8%, rg¼ 3.3%, hfg¼ 1.9%, sfg¼ 1.9%, cp,f¼ 3.6%,

cp,g¼ 9.3%, lf¼ 9.5%, lg¼ 8.5%, mf¼ 12.0%, mg¼ 3.5%, and

g¼ 4.4%.

Once the thermodynamic and transport properties have

been determined, one can use the thermodynamic properties

determined in Step 2 to predict the thermodynamic perfor-

mance [coefficient of performance (COP) and volumetric

cooling (or heating) capacity (VCC or VHC)] of the refrigerant in

an idealized vapor compression refrigeration cycle. The mean

lf (W m�1 K�1) Pr

R-114a R-1234ze(Z)b R-114a R-1234ze(Z)b

0.061 0.086 4.45 4.33

0.049 0.065 3.30 4.04

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 2 1421

absolute errors (ea) for COP and VCC for a group of 26 refrig-

erants are 1.7% and 3.1%, respectively. Finally, one can use the

transport properties estimated in Step 3 to gauge the refrig-

erant’s possible deviation from ‘‘ideal’’ due to exergy losses

from heat transfer and pressure drop effects.

The second part of the paper provides estimates for R-

1234ze(Z) of its: (1) fundamental thermodynamic parameters,

(2) thermodynamic properties, and (3) transport properties.

These properties then are used to estimate the thermody-

namic performance of R-1234ze(Z) in an idealized high-

temperature heat pumping cycle, and are used to investigate

its performance in a non-ideal evaporator and in a non-ideal

condenser. In particular, Tc¼ 426.8 K, Pc¼ 3970 kPa,

rc¼ 472.6 kg m�3, u¼ 0.333, and cop ¼ 0.758 kJ kg�1 K�1 at

T¼ 273 K. The thermodynamic properties are presented in

a P–h state diagram, a T–s state diagram, a saturation pressure

curve, and a saturated vapor density curve, and the saturated

liquid transport properties m and l are presented at saturation

temperatures of 25 �C and 85 �C. The thermodynamic perfor-

mance in an idealized high-temperature heat pump cycle

provides a COP of 3.40 and a VHC of 1645 kW m�3, which are

4.9% higher and 1.3% lower, respectively, than the R-114

values in the same cycle. The penalty factor (PF) concept of

Cavallini (2002) shows that the PF of R-1234ze(Z) is some 12–

15% lower than the R-114 value. Furthermore, heat exchanger

simulations were performed which indicate that the R-

1234ze(Z) heat transfer rate can be up to approximately 20%

greater than the R-114 values in both the evaporator and the

condenser for constant saturation temperature drops.

Finally, R-1234ze(Z) has no ODP and almost certainly has

a GWP on the order of probably no more than 10 or so based on

values of other fluorinated propene isomers. What is

unknown (no data has been published in the open literature)

is its toxicity or flammability. However, based on published

data for other fluorinated propene isomers, R-1234ze(Z) could

very likely be toxic and/or flammable to some greater or lesser

degree, and thus before R- 1234ze(Z) could be used in

a particular application extensive toxicity and flammability

testing would need to be conducted. Despite these potential

drawbacks, R-1234ze(Z) deserves further consideration as

a possible R-114 replacement.

Acknowledgements

J. Steven Brown expresses his appreciation to the Diparti-

mento di Fisica Tecnica of the Universita di Padova for hosting

him for a sabbatical stay during the 2008–2009 academic year.

The authors thank MOET (More Open Electrical Technologies),

an FP6 European Integrated Project (http://www.moetproject.

eu), for their financial support.

r e f e r e n c e s

ASHRAE, 2005. 2005 ASHRAE Handbook–Fundamentals Chapter19. American Society of Heating, Refrigerating andAir-Conditioning Engineers, Inc., Atlanta, USA.

ASHRAE, December 15, 2008. Proposed addendum z to Standard34-2007. Designation and Safety Classification of Refrigerants.Available from: https://osr.ashrae.org/Public%20Review%20Draft%20Standards%20Lib/34z-2007%201st%20PPR%20Draft.pdf.

Brown, J.S., 2007a. Predicting performance of refrigerants usingthe Peng–Robinson Equation of State. International Journal ofRefrigeration 30, 1319–1328.

Brown, J.S., 2007b. Preliminary selection of R-114 replacementrefrigerants using fundamental thermodynamic parameters(RP-1308). HVAC&R Research 13, 697–709.

Brown, J.S., 2007c. Evaluation of potential R-22 substituterefrigerants using fundamental thermodynamic parameters.In: Proceedings of the 22nd IIR International Congress ofRefrigeration, August 21–26, Beijing, China.

Brown, J.S., 2008. Methodology for estimating thermodynamicparameters and performance of alternative refrigerants.ASHRAE Transactions 114 (1), 230–238.

Brown, J.S., Potential R-114 replacement refrigerants. ASHRAETransactions 114, 2.

Cavallini, A., 2002. Heat transfer and energy efficiency of workingfluids in mechanical refrigeration. Bulletin of the InternationalInstitute of Refrigeration 2002 (6), 4–21.

Devotta, S., Rao Pendyala, V., 1994. Thermodynamic screening ofsome HFCs and HFEs for high-temperature heat pumps asalternatives to CFC114. International Journal of Refrigeration17, 338–342.

Directive 2006/40/EC of The European Parliament and of the Councilof 17 May 2006 relating to emissions from air-conditioningsystems in motor vehicles and amending Council Directive 70/156/EC, 2006, September 18, 2008. Official Journal of theEuropean Union, Available from: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri¼OJ:L:2006:161:0012:0018:EN:PDF.

Domanski, P.A., August 8, 2008. NIST EVAP-COND 2.3: simulationmodels for finned-tube heat exchangers. Available from:http://www2.bfrl.nist.gov/software/evap-cond/.

EPA, January 16, 2009. Class I ozone-depleting substances.Available from: http://www.epa.gov/ozone/science/ods/classone.html.

Latini, G., July 14–17, 2008. An estimation method for the thermalconductivityof the refrigerants inthe liquid state. In: Proceedingsof the International Refrigeration and Air ConditioningConference at Purdue, West Lafayette, Indiana, USA.

Luly, M.H., Singh, R.R., 2008. Cover gas composition for moltennon-ferrous metals such as magnesium. WIPO patentapplication WO 2008/005920 (10.01.2008).

Lemmon, E.W., Huber, M.L., McLinden, M.O., 2007. NIST referencefluid thermodynamic and transport properties – REFPROP.NIST Standard Reference Database 23 – Version 8.0.

Montreal Protocol on Substances that Deplete the Ozone Layer,1987. United Nations, New York, NY, USA (1987 withsubsequent amendments).

Mukhopadhyay, S., Nair, H.K., Tung, H.S., Van Der Puy, M., 2008.Processes for synthesis of 1,3,3,3-tetrafluoropropene. U.S.patent 7,345,209 (18.03.2008).

Poling, B.E., Prausnitz, J.M., O’Connell, J.P., 2001. In: The Propertiesof Gases and Liquids, fifth ed. McGraw-Hill, New York, pp. 2.23–2.34, 3.6–3.7, 6.22, 9.41, 9.61–9.68, 9.75–9.76, 10.23–10.24, 10.44–10.45, 12.8–12.9, C.1–C.4.

Rakhesh, B., Venkatarathnam, G., Srivivasa Murthy, S., 2003.Performance comparison of HFC227 and CFC114 incompression heat pumps. Applied Thermal Engineering 23,1559–1566.

Regulation (EC) No 842/2006 of The European Parliament and of theCouncil of 17 May 2006 on certain fluorinated greenhouse gases,2006, September 18, 2008. Official Journal of the EuropeanUnion. Available from: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri¼OJ:L:2006:161:0001:0011:EN:PDF.

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 2 ( 2 0 0 9 ) 1 4 1 2 – 1 4 2 21422

Reid, R.C., Prausnitz, J.M., Poling, B.E., 1987. In: The properties ofgases and liquids, fourth ed. McGraw-Hill, New York, pp.12–14.

Saleh, B., Wendland, M., 2006. Screening of pure fluids asalternative refrigerants. International Journal of Refrigeration29, 260–269.

Sastri, S.R.S., Rao, K.K., 2000. New method for predictingsaturated liquid viscosity at temperatures above the normalboiling point. Fluid Phase Equilibria 175, 311–323.

Spatz, M., Minor, B., September 18, 2008. HFO-1234yf low GWPrefrigerant update. In: Proceedings of the InternationalRefrigeration and Air Conditioning Conference at Purdue, WestLafayette, Indiana, USA. Available from: http://refrigerants.dupont.com/Suva/en_US/pdf/MAC_Purdue_HFO_1234yf.pdf.

Toms, G.S., Frank, M.V., Bein, T.W., 2004. Navy shipboard CFC-114elimination program. Naval Engineers Journal 116, 55–68.

Wark, K., 1995. Advanced Thermodynamics for Engineers.McGraw-Hill, New York, p. 165.