The Viscosity of Liquid R134a: C. M. B. P. Oliveira 1 and W. A. W A K e H A M 1

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

International Journal of Thermophysics, Vol. 14, No.

1, 1993

The Viscosity of Liquid R134a


C. M. B. P. Oliveira 1 and W. A. W a k e h a m 1

Received September 23, 1992

The paper reports new measurements of the viscosity of liquid R134a over
the temperature range 235 to 343 K and pressures up to 50MPa. The
measurements have been carried out in a vibrating-wire viscometer calibrated
with respect to the viscosity of several liquid hydrocarbons. It is estimated that
the uncertainty in the viscosity data reported is +0.6%. The data therefore have
a lower uncertainty than that of earlier measurements of the viscosity of this
environmentally acceptable regrigerant. The viscosity data have been repre-
sented as a function of density by means of a formulation based upon the rigid,
hard-sphere theory of dense fluids with a maximum deviation of _+0.3%. This
representation allows the present body of experimental data to be extended to
regions of thermodynamic state not covered by the measurements.

KEY WORDS: high pressure; refrigerants; R134a; viscosity.

1. I N T R O D U C T I O N

R134a is now commercially available as an environmentally acceptable


replacement refrigerant for domestic refrigeration. Nevertheless, there
remains a considerable uncertainty with respect to its viscosity in the liquid
phase as a result of several discordant and independent measurements [-1 ].
More recent measurements [2] than those contained in the review of
Stephan and Krauss [1] do provide support for some of the original
results but have a claimed accuracy of only • Thus, it is clear that
further measurements of greater precision are necessary before unequivocal
statements about the viscosity of this fluid can be made. The present work
is concerned with the measurement of the viscosity of R134a in the liquid
phase from the saturation pressure to 5 0 M P a over the range of

1 Department of Chemical Engineering and Chemical Technology, Imperial College, Prince


Consort Road, London SW7 2BY, United Kingdom.

33
840/'14/1-3 0195-928X/93/0100-0033$07.00/0 9 1993 Plenum Publishing Corporation
34 Oliveira and Wakeham

temperature from 235 to 343 K. The measurements have been made


with a vibrating-wire viscometer described in detail elsewhere [3], which
has a precision of _+0.3% and an accuracy of _+0.6%.

2. E X P E R I M E N T A L
The vibrating-wire viseometer employed for the present measurements
was exactly that described earlier E3]. It comprises a 100-#m-diameter
straight tungsten wire, suspended from the top plate of a pressure vessel,
and carrying a tensioning weight on its lower end. The wire is set into
transverse vibration electromagnetically and the subsequent motion of the
wire within the test fluid observed by electromagnetic induction. The decay
of the transverse vibrations, which conform to a damped sinusoidal
oscillation, is related to the viscosity and the density of the test fluid by the
working equation

A - (P/Ps) k ' + 2Ao (1)


211 + (p/ps) k ]

Here, A is the logarithmic decrement of the oscillation in the fluid, Ao


the logarithmic decrement in vacuo, p the fluid density, and Ps the density
of the wire material. In addition, k and k' are quantities that depend [4]
upon the viscosity of the fluid through the equations

k = - 1 + 2 Im(A) (2)
k ' = 2 Re(A) + 2A Im(A) (3)

where

2Ko(s)]
A = (i-- Ao) 1 + s K ~ J (4)

and
s = E(i- A) f2] m (5)
f2 = p ~ R 2 / r l (6)

in which R is the radius of the vibrating wire and ~o the frequency of


oscillation. The symbols Ko,1 are modified Bessel functions.
It follows from these equations that measurements of the decrement
and frequency of the oscillation of the wire in vacuo and in the fluid serve
to determine the viscosity of the fluid given a knowledge of its density,
provided that the radius of the wire and its density are available. It has not
Viscosity of Liquid R134a 35

been possible to determine the latter quantities directly with sufficient


accuracy so that they have been determined by calibration of the instru-
ment with fluids of known viscosity as described in detail elsewhere [5]. As
in our earlier work we have employed viscosity data under saturation
conditions at 303.15 K for toluene, n-hexane, n-heptane, n-octane, and
n-decane reported by Gonqalves et al. [6] and Knapstad et al. [7]. to
perform the calibration. The results are identical to those described
earlier [5].
The samples of R134a were provided by ICI Chemicals and Polymers
Ltd. with a purity of better than 99.9%. During the measurements various
samples were withdrawn for analysis to confirm that there had been no
contamination. In the evaluation of the viscosity of R134a we have con-
sistently made use of the equation of state of Saltoh et al. [8], for which
the estimated uncertainty in the density is +0.1%. Thus, accounting for
the accuracy of the calibration data and the precision of individual
measurements, it is estimated that the overall accuracy of the reported
viscosity data is +0.6%. The precision of the measurements is +0.3%,
which is significantly better since the majority of the inaccuracy arises from
the calibration viscosity data.

Table I. Viscosity of R134a Along the Saturation Line

Expt. Expt. Saturation Viscosity at


pressure Density viscosity pressure saturation
Temperature P p( T, P) q( T, P) P~ rl( T, P~)
(K) (MPa) (kg. m-3) (mPa- s) (MPa) (mPa- s)

237.74 0.092 1400.4 0.4298 0.065 0.4296


244.14 0.136 1382.6 0.3955 0.089 0.3951
248.23 0.138 1370.5 0.3767 0.108 0.3765
251.06 0.253 1362.5 0.3632 0.122 0.3626
253.86 0.193 1353.9 0.3512 0.138 0.3510
261.39 0.209 1330.9 0.3183 0.188 0.3183
268.10 0.305 1310.0 0.2925 0.244 0.2923
270.96 0.624 1302.0 0.2812 0.271 0.2798
271.26 0.703 1301.3 0.2801 0.274 0.2785
279.07 0.759 1275.6 0.2532 0.362 0.2517
289.30 0.729 1239.6 0.2229 0.507 0.2221
293.35 0.664 1224.6 0.2116 0.576 0.2110
303.14 0.923 1187.8 0.1877 0.770 0.1871
313.15 1.159 1147.2 0.1654 1.016 0.1648
323.15 1.389 1102.3. 0.1451 1.317 0.1448
333.15 1.875 1054.9 0.1277 1.680 0.1267
343.15 2.303 999.6 0.1116 2.114 0.1104
36 Oliveira and Wakeham

3. RESULTS
The viscosity data for R134a are contained in Table I for
measurements near the saturation line and in Table II for those isotherms
for which measurements were performed as a function of pressure. The
tables list the temperature, pressure, density, and viscosity for the fluid. In
view of the fact that it may be possible to recalibrate the viscometer in the
future if more accurate viscosity data for the reference fluids become
available, the raw experimental data for calibration and the measurements
on R134a have been retained. However, they are omitted from the
tabulations in the interests of brevity.

3.1. Saturation Line

Not all of the measurements near the saturation vapor pressure were
performed exactly at saturation. Thus, in order to provide values exactly at
the saturation conditions given by Saltoh et al. I-8], the data have been
corrected to the saturation density for each temperature by the application
of a small (1.0%) correction. The correction has been deduced from the
representation of the viscosity as a function of density discussed later. The
magnitude of this correction is sufficiently small that the additional
uncertainty introduced by its application is no more than 0.1%. The data
at the saturation density given by Saltoh et al. 1-8] are included in Table I.
The temperature dependence of the viscosity of R134a along the
saturation line has been represented by the equation

t/= exp aiT -i (7)


i

where

ao = - 39.05765, al = 3.616708 • 10 4 K
a2 = - 1.372566 • 10 7 K 2, a 3 = 2.409684 • 10 9 K 3 (8)

a4 = -1.61014 • 1011 K 4

and q is measured in m P a . s. Figure 1 contains a plot of the deviations of


the present experimental data from this correlation; they do not exceed
+0.4%, while the standard deviation is +0.2%. The same figure contains
a comparison with the results of earlier work. The data of Ripple I-2], for
the lower temperature range, depart from the present representation by no
more than _ 4 % , which is broadly consistent with the estimated error in
his measurements of + 3 % . The results of Kumagai and Takahashi [9]
Viscosity of Liquid R134a 37

Table II. Viscosity of R134a as a Function of Pressure

Temperature Pressure Density Viscosity


T P p
(K) (MPa) (kg. m-3) (mPa. s)

293.35 0.664 1224.6 0.2116


2,222 1232.3 0.2172
4.976 1245.8 0.2271
6.808 1253.6 0.2337
9.997 1267.1 0.2445
17.05 1292.0 0.2677
20.06 1301.6 0.2772
27.94 1323.8 0.3017
29.76 1328,5 0.3075
40.17 1353.0 0.3387
41.06 1354.9 0.3409
51.01 1375.1 0.3705
303.14 0.923 1187.8 0.1877
0,929 1187.9 0,1873
4.806 1211.2 0.2019
6.207 1218.6 0.2069
10.20 1237.5 0.2204
16.85 1264.0 0.2416
20.30 1276.0 0.2519
30.45 1306.5 02814
39.90 1330.5 0.3083
50.47 1353.8 0.3377
313.15 1.160 1147.2 0.1654
6.844 1186.5 0.1860
10.30 1205.2 0.1981
25.92 1267.6 0.2441
39.79 1307.4 0,2815
50.97 1333.9 0.3106
323.15 1.389 1102.3 0.1451
3.159 1119.8 0.1529
7.324 1152.3 0.1684
333.15 1.875 1054.9 0.1276
3,893 1080.2 0.1367
5.552 1097.1 0.1435
343.15 2.303 999.6 0.1116
3,911 1027.9 0.1183
5,842 1053.7 0.1270
9.829 1093.1 0,1417
18.84 1152.9 0.1712
26.64 1190.1 0.1895
30.07 1204.0 0.2003
40.36 1240.0 0.2245
50.24 1269.0 0.2470
38 Oliveira and Wakeham

30

20

4.
"i ~0
0 0
0
o O0 ~) 0
o
o
§
§

§
I -10
+
v §
+
-20 §

+
i i
3O
23O 270 310 350

T, K

Fig. 1. Deviations of experimental viscosity data for


R134a along the saturation line from the representation
of them by Eq. (7) and (8). (HI) Present work; (O)
Ripple [2]; (O) Kumagai etal. [9]; ( + ) Diller etal.
[10]; ( ~ ) Shankland etal. [11].

show quite good agreement in the intermediate temperature range,


although the departture increases to _+6% at the highest temperatures. The
data of Diller et al. [-10] and Shankland et al. [11], on the other hand,
show a quite different temperature dependence, leading to very large dis-
crepancies at high and low temperatures. Figure 1 illustrates the wide

Table IiI. Coefficients of the Correlating Eq. (9) for the Viscosity of R134a

Temperature Pressure
T P'
(K) (MPa) bo bl b2

293.35 25.8 0.2061 0.0184 0.0704


303.14 25.7 0.1803 0.0241 0.0631
313.15 26.1 0.1560 0.0308 0.0572
323.15 4.4 0.1344 0.0133 0.0100
333.15 3.7 0.1139 0.0126 0.0095
343.15 27.1 0.0844 0.0773 0.0308
Viscosity of Liquid R134a 39

0.50

o
+A
z~
A +U
O + +
O 9 o
,II-
0.00 n~v A
A
A
u
I + +

I i
-0.50
0 20 40 60

P, M P a

Fig. 2. Deviations of the viscosity data for R134a


from the representation of it as a function of pressure
by Eq. (9): (+) 293.35 K; (•) 303;15 K; (O) 313.15 K;
(V) 323.15 K; (41,) 333.15 K; (ll) 343.15 K.

divergence of results for the viscosity of R134a that exist in the literature.
While the present data support the results of Ripple [ 2 ] and of Kumagai
and Takahashi [9], there are still discrepancies at elevated temperatures
greater than can be explained by the claimed uncertainties.

3.2. Pressure Dependence


The pressure dependence of the viscosity of R134a along each
isotherm has been represented, for the purposes of interpolation, by the
equation
2
l~ = ~ bi(P/Pt) i (9)
i:0

with the coefficients listed in Table III. Figure 2 contains the deviations of
the data for the various isotherms from this representation. The deviation
does not exceed +0.4%, the standard deviation being 0.2%. There are no
published data for comparison.
40 Oliveira and Wakeham

4. DENSITY DEPENDENCE OF THE VISCOSITY


Semiempirical representations of the viscosity of liquid hydrocarbons
and their mixtures based upon the rigid-sphere model of a fluid have
proved to be remarkably successful [-12]. Such methods not only allow the
interpolation of data, but also the extension of data to regions of
thermodynamic state beyond that covered by direct measurement. The
present experimental data provide an opportunity to examine whether a
similar procedure is applicable to refrigerants.
The most important result of the rigid-sphere theory of transport in
dense fluids is the conclusion that the dimensionless quantity, ~/*, defined
by equation [12]
t/* = 6.035 x 108 [ 1 / M R T ] i/2 tl V2/3/R,I (10)
is a function only of the ratio of the molar volume, V, to a characteristic
molar volume Vo. Here, R is the universal gas constant, M the molar mass
of the species, and R, a scaling factor that accounts for the roughness of
the rigid-sphere system. It is expected that R, is dependent on the
molecular species, but independent of temperature and density, while Vo

25

20

'=. 15

10

1.60 1.80 2.00 2.20 2.40

V/V o

Fig. 3. The reduced viscosity for R134a as a func-


tion of the reduced molar volume: ( + ) 293.35 K;
(A) 303.15K; (O) 313.15K; (V) 323.15K; (41,)
333.15 K; ( i ) 343.15 K.
Viscosity of Liquid R134a 41

Table IV. Coefficients of the Correlating


Eq. (11) for the Viscosity of R134a

Index
i ci

0 0.65031
1 2.41753
2 -0.70771
3 0.153854

may be weakly temperature dependent, reflecting the fact that the true
forces between the molecules in the fluid have a repulsive component which
is not infinitely steep. Accordingly, it is first appropriate to examine
whether, for R134a, the experimental data support the fact that rl* is a
function only of (V/Vo.) For this purpose we have evaluated t/* from our
experimental data and adopted an arbitrary reference value for Vo at a
temperature of 293.35 K. Then, by selecting appropriate values for Vo at
other temperatures, we have constructed the single curve of t/* against

0.40

0.20
O

A V?
+ Oo +I+
0 0,00 +R+~ n

A A 0 7
0
a

-0.20

i
0.40 ' =
1.60 1.80 2.00 2.20 2.40

V/Vo

Fig. 4. Deviations, [100(q~xp-qcalc] , of the present


viscosity data from the representation of them as a
function of volume by means of Eq. (11 ): ( + ) 293.35 K;
(zl) 303.15K; ( O ) 313.15K; ( V ) 323.15K; ( , )
333.15 K; ( i ) 343.15 K.
42 Oliveira and Wakeham

Table V. The Characteristic Molar Volume


Vo for R134a

Temperature Characteristic volume


T 106V0
(K) (m3 .mol 1)

293.35 44.46
303.14 44.26
313.15 44.00
323.15 43.89
333.15 43.74

(V/Vo.), which is shown in Fig. 3. This single function has been represented
by the equation
3
ln~*= Y ciEVo/(V- v0)] i (11)
i=O

The coefficients ci are listed in Table IV. Figure 4 contains plots of the
deviations of the experimental data from this common representation and
0.40

0.20

oo

0
d +-,+%
++ V
,0i 0.00 +~,~ +~- 7
%
4)

-0.20

0.40 i J =
1.60 1.80 2.00 2.20 2.40

V/V,,
Fig. 5. Deviations, [100(t/exp- qcalc)/?/calc], of the pre-
sent viscosity data for R134a from the representation of
them by means of the correlation of Assael et al. [12]
for n-alkanes: ( + ) 293.35K; ( A ) 303.15K; (O)
313.15 K; ( V ) 323.15 K; (~') 333.15 K; ( i ) 343.15 K.
Viscosity of Liquid R134a 43

it can be seen that they do not exceed 0.3 %, which provides substantial
evidence for the result that t/* depends only upon (V/Vo).
A second point of interest is to what extent the data for R134a con-
form to the representation of ~/* for the normal alkanes given by Assael
et al. [12]. We have therefore sought to secure the optimum representation
of the present experimental data for R134a by means of the function t/*
given by Assael et al. [12]. In this process we have allowed both Vo(T)
and R, to be determined in the fitting. The resulting deviations are shown
in Fig. 5 and the values of Vo listed in Table V; the optimum value of R,
was R, = 1.022175. It can be seen from Fig. 5 that the representation of the
experimental data for R134a by an equation based on results for nonpolar
systems is essentially the same as when the fluid is treated in isolation. This
is a remarkable result that is well within the combined uncertainty of the
viscosity and the density.

5. CONCLUSIONS
The viscosity of R134a has been measured along the saturation line
and over a range of pressures in the temperature range 235 to 343 K. The
results along the saturation line confirm the most recent measurements by
other workers but have a lower uncertainty. The results at higher pressures
have been used to demonstrate that methods of representing the viscosity
of this liquid refrigerant based upon the rigid-sphere model are as success-
ful as they have proved for liquid hydrocarbons. This makes it possible to
extend the range of data beyond that covered by direct measurement.

ACKNOWLEDGMENTS
The authors are grateful to ICI Chemicals and Polymers Ltd. for
the supply of high-purity samples of R134a and subsequent analysis of
samples used in the equipment. In particular, they wish to thank
Mr. R. Wheelhouse for his advice on several aspects of their work.

REFERENCES
1. K. Stephan and R. Krauss, Int. J. Thermophys. (in press).
2. D. Ripple, Personal communication (1992).
3. M. J. Assael, C. P. Oliveira, M. Papadaki, and W.A. Wakeham, Int. J. Thermophys.
13:593 (1992).
4. T. Retsina, S. M. Richardson, and W.A. Wakeham, Appl. Sci. Res. 43:325 (1987).
5. C. M. B. P. Oliveira and W. A. Wakeham, Int. J. Thermophys. 13:805 (1992).
6. F. A. Gon~alves, K. Hamano, J. V. Sengers, and J. Kestin, Int. J. Thermophys. 8:641
(1987).
44 Oliveira and Wakeham

7. B. Knapstad, P. A. Sjolsvik, and H. A. Oye, J. Chem. Eng. Data 34:37 (1989).


8. A. Saltoh, S. Nakagawa, H. Sato, and K. Watanabe, J. Chem. Eng. Data 35:107 (1990).
9. A. Kumagai and S. Takahashi, Int. J. Thermophys. 12:105 (1991).
10. D. E. Diller, A. S. Aragon, and A. Laesecke, Fluid Phase Equil. (in press).
11. I. R. Shankland, R. S. Basu, and D. P. Wilson, Proe. Meet. IIR, Comm. BI, B2, El, E2,
Purdue Univ. 2:305 (1988).
12. M. J. Assael, J. H. Dymond, M. Papadaki, and P.M. Patterson, Int. J. Thermophys.
13:659 (1992).

You might also like