international journal of refrigeration 73 (2017) 125133

Available online at www.sciencedirect.com

ScienceDirect

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g

Surface tension measurement of oil/refrigerant

mixture by maximum bubble pressure method

Mitsuhiro Fukuta a,*, Junki Sumiyama b, Masaaki Motozawa a,

Tadashi Yanagisawa a
a
Department of Mechanical Engineering, Shizuoka University, Hamamatsu, Shizuoka, 432-8561, Japan
b
Graduate School of Engineering, Shizuoka University, Hamamatsu, Shizuoka, 432-8561, Japan

A R T I C L E I N F O A B S T R A C T

Article history: Lubrication oil used in refrigerant compressors forms oil mist in the compressor shell. Some
Received 22 May 2016 of the oil mist is discharged into a refrigeration cycle with refrigerant and causes degra-
Received in revised form 24 August dation of heat transfer in heat exchangers. Since the generation of the oil mist is related
2016 to the Weber number, it is necessary to measure the surface tension of the oil/refrigerant
Accepted 5 September 2016 mixture before discussing the oil mist generation in the compressor. In this study, the
Available online 9 September 2016 maximum bubble pressure method was adapted to measure the surface tension of PAG
(polyalkylene glycol) oil/CO2 mixture. The density of the mixture needed for the measure-
Keywords: ment was also carried out. It was found that the surface tension of PAG/CO2 mixture sharply
Oil mist decreases with an increase in the concentration of the refrigerant in the mixture. The surface
Weber number tension of the mixture under 10 MPa and 100 C is estimated to be 14.6 mN m1 by an
Surface tension extrapolation.
Oil/refrigerant mixture 2016 Elsevier Ltd and IIR. All rights reserved.
Maximum bubble pressure method

Mesure de la tension de surface dun mlange huile/

frigorigne par la mthode de la pression maximale des bulles
Mots cls : Vapeur dhuile ; Nombre de Weber ; Tension en surface ; Mlange huile/frigorigne ; Mthode de la pression maximale des
bulles

mist is generated in the compressor shell, some of the mist

1. Introduction
In refrigerant compressors, refrigeration oil is used to lubri-
cate sliding parts, to seal leakages through clearances in
compression chambers, to cool the compressor body, and to
clean the inside of the compression chamber. When fine oil
flows out from the compressor into the refrigeration cycle with
the refrigerant. Since the oil discharged from the compressor
causes a reduction of quantity of the lubrication oil stored in
the compressor, degradation of heat transfer efficiency, and in-
creased pressure loss in the heat exchangers (Dang et al., 2012),
it is important to reduce the amount of oil mist discharged into

* Corresponding author. Department of Mechanical Engineering, Shizuoka University, Hamamatsu, Shizuoka, 432-8561, Japan. Fax: +81
53 478 1054.
E-mail address: fukuta.mitsuhiro@shizuoka.ac.jp (M. Fukuta).
http://dx.doi.org/10.1016/j.ijrefrig.2016.09.004
0140-7007/ 2016 Elsevier Ltd and IIR. All rights reserved.
126 international journal of refrigeration 73 (2017) 125133
Nomenclature
a capillary constant
a sound velocity [m s1]
ai coefficients in Eq. (3)
C mass fraction of CO2
f correction factor
g gravitational acceleration [m s2]
h immersed depth of capillary tube [m]
L distance [m]
n index factor in Eq. (7)
Pmax maximum pressure [Pa]
r inner radius of capillary tube [m]
t1, t2 time lag [s]
surface tension [N m1]
density [kg m3]
density difference [kg m3]
Subscripts
mix mixture
oil oil
ref refrigerant
the cycle. Many researchers have studied the oil separation from
the refrigerant flow in the compressor shell (Ogata et al., 2006;
Song et al., 2015; Toyama et al., 2006) or in an oil separator in
a discharge line (Murakami et al., 2006; Shikazono et al., 2010;
Wiencke, 2011) to reduce the oil circulation ratio and to ensure
oil return to the compressor. However, the oil separator is limited
in its separation efficiency and it removes hardly any of the
fine oil mist from the refrigerant flow.
The motivation of this study is to find how to suppress the
oil mist generation in the compressor to improve the cycle per-
formance. In order to discuss the mechanism for generation
of the oil mist in the compressor, it is necessary to clarify the
phenomena involved in the breakup of the oil droplet. The
breakup of liquid is closely related to the Weber number
(Majithia et al., 2008; Sakurai et al., 2012), which is inversely
proportional to its surface tension. Since the refrigerant dis-
solves into the oil in the compressor, the surface tension of
the oil/refrigerant mixture existing in the oil sump of the com-
pressor should be measured. There are several studies
measuring the surface tension of refrigerant (Geller et al., 1996;
Kondou et al., 2015), but there are few researches measuring
that of the oil/refrigerant mixture, mainly because it is hard
to measure the surface tension in the pressurized condition
of the refrigerant atmosphere.
There are several techniques to measure the surface tension.
The Wilhelmy method (Cai et al., 2015) and the ring method
(Jensen and Jackman, 1984) are often used to measure the
surface tension of liquid mainly under atmospheric pressure
condition. The pulling force exerted by the surface tension on
a plate or a ring of known perimeter is measured in these
methods. In the pendant drop method (Khosharay and
Varaminian, 2014), the surface tension is calculated from the
droplet geometry suspended at the end of a capillary. It is ap-
plicable to a pressurized condition. The capillary rise method
(Geller et al., 1996; Kondou et al., 2015) is a method in which
the height of a liquid rising or sometimes falling inside the cap-
illary tube due to the surface tension is measured. Since the
pressure inside a bubble is proportional to the surface tension,
the surface tension is measured by measuring the bubble pres-
sure in the maximum bubble pressure method (MBPM) (Feenstra
et al., 2001; Seeton and Hrnjak, 2006). It also can be adapted
to the pressurized condition.
Jensen and Jackman (1984) investigated the surface tension
of an R113-oil mixture with several viscous mineral oils by using
the ring method. They showed that the surface tension of the
mixture decreases with an increase of dissolved refrigerant,
and they proposed the correlation of surface tension of the
mixture with the oil concentration as a parameter.
Feenstra et al. (2001) carried out surface tension measure-
ments of volatile fluids. They adapted the MBPM to measure
the surface tension of pure R134a and R134a of unknown purity
which contained dissolved contaminants such as lubrication
oil. The device they developed is applicable to the surface
tension measurement of volatile fluids, and the uncertainty
of the measurement is less than 2%. They concluded that the
surface tension of the R134a samples of unknown purity was
not significantly influenced by contaminants such as grease
or lubrication oil.
Seeton and Hrnjak (2006) measured the surface tension of
a POE32/CO2 mixture by MBPM, changing the temperature and
concentration of dissolved CO2. In their measurement by MBPM,
two capillaries having different radii are used so that the mea-
surement of a depth of a capillary tip in the mixture is
unnecessary. They showed that the surface tension of the
mixture decreases with an increase of its temperature and the
mass fraction of CO2.
In this study, an experimental setup applicable for the
surface tension measurement of an oil/refrigerant mixture,
using the MBPM under pressurized conditions is developed. CO2,
which is used as a refrigerant in heat pump water heaters, and
poly-alkylene glycol (PAG) oil are used in this study. The surface
tension of the PAG/CO2 mixture is measured at various tem-
peratures and with different concentrations of CO2 in the
mixture. The surface tension of the mixture in conditions ex-
pected in a compressor for water heater is estimated from a
tendency of surface tension change obtained by this study. The
density of the mixture needed to measure the surface tension
in MBPM is also measured using an ultrasonic sensor.
2. Maximum bubble pressure method
2.1. Principle of measurement
The inside of the high pressure shell of actual working com-
pressors for CO2 refrigerant has a high pressure and high
temperature. In this study, the MBPM is chosen to measure the
surface tension of the oil/refrigerant mixture under such pres-
surized conditions. Fig. 1 shows the principle of the MBPM. A
bubble is formed at the tip of a capillary tube immersed in
liquid, with gas supplied through the capillary tube. The bubble
generation cycle, i.e. appearance of the bubble at the tip of the
capillary, bubble growth and then its separation from the tip,
is repeated when the gas is supplied continuously to the cap-
illary. As the bubble size increases at the tip of the capillary,
 international journal of refrigeration 73 (2017) 125133 127

Fig. 2 Experimental setup for surface tension

measurement.
Fig. 1 Principle of maximum bubble pressure method.

pressure and temperature in the sub vessel, the mixture is sent

the radius of the bubble decreases and the pressure inside the
to the main vessel via a gear type oil pump, and the liquid level
bubble increases according to the bubble radius. The pres-
in the main vessel is adjusted so that the tip of the capillary
sure attains its maximum when the radius of the bubble
tube is immersed in the oil/refrigerant mixture. The im-
becomes equal to the inner radius of the capillary tube. Based
mersed depth of the capillary in the liquid is set to be about
on the YoungLaplace equation, the surface tension of the liquid,
5 mm. The immersed depth of the capillary tube is measured
, is calculated by Eq. (1) under static condition.
by a CCD camera through a sight glass set in the side of the
main vessel. The accuracy of measurement of the immersed
= f
r
2{
2
3 }
Pmax h + r g (1) depth is 0.03 mm. The upper part of the main vessel is con-
nected to a suction port of a vapor pump, and the discharge
where, r is the inner radius of the capillary tube, Pmax is the line of the pump is connected to the capillary tube through a
maximum pressure difference between inside the capillary tube stop valve and a needle valve. Fig. 3 shows the structure of the
and the atmosphere of the liquid phase, h is the immersed vapor pump made by us in this experiment. The vapor pump
depth of the capillary tube in the liquid, is the density dif- vessel is pressure proofed for 10 MPa. A rotary compressor
ference between the liquid and the gas, and g is the gravitational element is integrated into the vapor pump. No lubricant is used
acceleration. f is a correction factor for non-spherical bubble in the rotary compression element. This is to prevent con-
shape and is given by Eq. (2) (Fainerman and Miller, 2004). tamination by the lubricant. The rotary element is driven by
an external gear motor through a magnet coupling. The pres-
i sure inside the vapor pump vessel slightly increases after
f = ai
r
(2) applying a few rotations to the vapor pump. When the stop
a
valve installed in the connection line between the pump vessel
where a is the capillary constant defined by and the capillary tube opens, the slightly pressurized gas by
the pump is sent to the capillary tube due to the pressure dif-
2 ference between the pump vessel and the main vessel, and the
a= (3) bubble is generated at the tip of the capillary tube. The flow
g
rate from the vapor pump to the capillary is regulated by the
and the coefficients in Eq. (3) are: a0 = 0.99951, a1 = 0.01359, needle valve. A system volume is installed before the capillary
a2=0.69498, a3=0.11133, a4 = 0.56447, a5=0.20156. The range of
the factor f is from 0.989 to 0.994 in this study. As described
later, since the oil/refrigerant mixture is under saturated equi-
librium condition and it takes about 5 seconds to generate one
bubble, any dynamic effects such as dead time effects, inertia
effects and gas absorption can be eliminated and Eq. (1) is ap-
plicable in this study.

2.2. Experimental setup

Fig. 2 shows the experimental setup used in this study. A sub

vessel is used to prepare the PAG and CO2 mixture. After storing
a given amount of PAG oil in the sub vessel, CO2 gas is sup-
plied from a gas cylinder to the sub vessel. CO2 is dissolved
into the PAG using a magnet stirrer. A capillary tube of 0.5 mm
in inner diameter and 15 mm in length is set in the main vessel.
After the PAG/CO2 mixture reaches equilibrium under the given Fig. 3 Vapor pump.
128 international journal of refrigeration 73 (2017) 125133

Table 1 Uncertainty of measurements

Measurements Uncertainty
Immersed depth 0.03 mm
Pressure 0.25% f.s.
Sampling method 2%
Surface tension 0.2 mN m-1
Temperature 1.0 C

tube to stabilize the flow and the pressure (Fainerman et al.,

1994). The pressure difference between the main vessel and
the capillary tube is measured by a pressure differential sensor.
The accuracy of the pressure differential sensor is within 0.25%
of full scale. Since the density of super-critical CO2 is high and
varies according to temperature, the head of the super-critical
CO2 should be taken into account in the measurement of the
Fig. 4 Pressure change inside capillary tube.
pressure difference by the sensor when there is large tem-
perature difference between the main vessel and the line
connected to the sensor. In this study, zero point adjustment
creases as the pressure inside the pump shell decreases. After
of the pressure differential sensor is carried out in every ex-
20 seconds, the pressure level becomes almost constant and
periment under the equilibrium condition without the flow to
this means that the pressure loss through the capillary tube
the capillary tube to counteract the head difference caused by
becomes insignificant since the flow rate in this period becomes
the density change of super-critical CO2. Although CO2 does not
very small. The bubble generation during this period, shown
condense at the pressure differential sensor in the super-
as A in Fig. 4, is given in Fig. 5. The maximum pressure is ob-
critical condition in this study, please note that the lines outside
tained when the radius of the bubble becomes equal to that
the main vessel and the sensor should be heated to prevent
of the capillary tube. The maximum pressure at the end of each
the condensation of the refrigerant at the sensor when common
experiment is used as Pmax in Eq. (1) to calculate the surface
refrigerants are used.The maximum pressure occurs when the
tension. The average of five times measurement of ethanol at
radius of the bubble becomes equal to the inner radius of the
30 C under atmospheric pressure is 21.93 mN m1, where the
capillary during the bubble generation cycle. By measuring
relative error compared with the known value (Gonalves et al.,
the maximum pressure, the surface tension of the mixture is
2010) is 1.4%.
calculated from Eq. (1). A tape heater is wrapped around the
Secondly, the influence of pressure on the surface tension
main vessel, and the temperature of the test section is kept
measurement is confirmed. The gauge pressure in the test
constant. The temperature is measured by a T-type thermo-
vessel is increased up to 0.93 MPa by supplying nitrogen gas,
couple and the measurement error of the thermocouple is within
keeping the temperature at 30 C. Fig. 6 shows the result re-
1.0 C.The pressure is monitored by a semi-conductor type pres-
garding the surface tension of ethanol in a pressurized
sure gauge with an accuracy of 0.25% of full scale. The
condition. It is found that the surface tension of ethanol is not
concentration of refrigerant in the mixture is measured by a
significantly influenced by the pressure. The experimental ap-
sampling method. The accuracy of the sampling method is es-
paratus is applicable to the surface tension measurement of
timated to be less than 2%. In the experiment, the measurements
oil/refrigerant mixture in a pressurized condition.
of the surface tension and the refrigerant concentration of the
mixture are carried out by changing the pressure and the tem-
perature of the mixture.The purity of CO2 used in the experiment
is 99.95%. Viscosity grade of PAG used in this study is 100 mm2
s1 at 40 C. The experiment is carried out in the pressure range
up to 6 MPa and temperature range from 30 to 70 C. Uncer-
tainties of the measurements are summarized in Table 1.

2.3. Surface tension measurement of ethanol

Firstly, the surface tension of ethanol is measured under at-

mospheric pressure at 30 C to confirm the accuracy of the
measurement. Fig. 4 shows an example of the pressure change
inside the capillary tube just after starting the bubble genera-
tion, taking the pressure inside the main vessel as a baseline.
The pressure shown in Fig. 4 fluctuates due to the intermit-
tent bubble release from the tip of the capillary tube. The
pressure just after starting the experiment shows a higher value
due to pressure loss through the capillary tube and de- Fig. 5 Growth of bubble at tip of capillary tube.
 international journal of refrigeration 73 (2017) 125133
Fig. 6 Surface tension of ethanol versus gauge pressure.
3. Density of PAG/CO2 mixture
3.1. Experimental apparatus for density measurement
The density of the oil/refrigerant mixture is used in Eq. (1) to
calculate the surface tension of the mixture. Therefore, the
density of the PAG/CO2 mixture is measured before the surface
tension measurement. Fig. 7 shows the experimental setup for
the density measurement. The mixture is stored in a test vessel
and the density of the mixture is calculated by measuring its
volume and mass. A pre-determined amount of oil is firstly
stored in the vessel and then CO2 is charged and mixed with
the oil. The mass of charged CO2 is found by measuring the
increase of whole weight of the vessel. The amount of CO2 dis-
solved into the oil is calculated by subtracting the gas mass
calculated from the gas volume and density under given tem-
perature and pressure conditions from the charged amount of
CO2. The separate volumes of the gas and the mixture are ob-
tained by measuring the liquid level in the vessel. The liquid
surface is detected by using an ultrasonic sensor of a single
probe. The ultrasonic sensor consists of a transducer, a damper
and a plastic heat resistance part as shown in Fig. 8. A mesh
plate is set in the vessel over the sensor as a reflector. The sound
speed of the mixture can be measured from the distance
between the sensor and the reflector and the propagation time
of the ultrasonic wave from the reflector. The liquid level is cal-
culated by the sound speed and the propagation time from the
interface of the mixture. Fig. 9 shows a signal received from
Fig. 7 Experimental setup for density measurement.
129
Fig. 8 Ultrasonic sensor.
the ultrasonic sensor in the experiment. The first signal is a
pulsar signal, and echoes from the sensor surface, the reflec-
tor and the interface are detected in series. The time lag t1
shown in Fig. 9 is the propagation time of the ultrasonic wave
between the sensor surface and the reflector, and the time lag
t2 is that between the sensor surface and the interface of the
mixture. The sound velocity and the liquid level are ex-
pressed by Eqs. (4) and (5) respectively.
2L
a= (4)
t 1
a t 2
l= (5)
2
where, a is the sound velocity in the mixture and L is the dis-
tance between the sensor surface and the reflector. The data
of sound velocity of PAG/CO2 mixture were published in a ref-
erence paper (Fukuta et al., 2008).
3.2. Result of density measurement
Fig. 10 shows the density of the PAG/CO2 mixture measured
at each temperature with calculated one by Eq. (6) based on
an assumption of an ideal mixture.
Fig. 9 Received signal by ultra sonic sensor.
130 international journal of refrigeration 73 (2017) 125133

Fig. 12 Surface tension of PAG.

Fig. 10 Density of PAG/CO2 mixture.
tension of PAG is shown in Fig. 12 measured by the MBPM,
changing its temperature. The surface tension of PAG linearly
decreases with temperature. Fig. 13 shows the surface tension
of CO2 obtained by REFPROP ver. 9.1 (Lemmon et al., 2013). The
surface tension of CO2 is much smaller than that of PAG and
decreases with temperature, becoming 0 mN m1 at 31 C of
the critical temperature. When the temperature of CO2 exceeds
31 C, CO2 enters a super-critical condition and the surface
tension cannot be defined. Fig. 14 shows the surface tension
of the PAG/CO2 mixture at 30 C and 70 C. Since the surface
tension of CO2 is very small (1.2 mN m1 @ 20 C) compared

Fig. 11 Averaged density of PAG/CO2 mixture.

1
mix = (6)
1C C
+
oil ref

where, mix, oil, ref are densities of the PAG/CO2 mixture, PAG,
and CO2, and C is the mass fraction of the CO2 in the mixture.
The calculated density of the mixture decreases with an in-
crease in the CO2 mass fraction because the density of CO2 is
much less than that of PAG. On the other hand, the mea-
sured density of the mixture is almost constant at each
Fig. 13 Surface tension of CO2.
temperature, even though the concentration of CO2 increases.
In this study, the density of the PAG/CO2 mixture is assumed
to be constant against the concentration of CO2 and is only the
function of temperature. Fig. 11 shows an approximation of the
change in the density of the PAG/CO2 mixture against the tem-
perature. The density used in Eq. (1) is calculated from the
approximation curve shown in Fig. 11. It is found from Eq. (1)
that the error of surface tension caused by the error in the ap-
proximation of the density is 0.3% when the approximation
error of the density is 2% under the given condition.

4. Surface tension of the PAG/CO2 mixture

Prior to the surface tension measurement of the mixture, the

surface tension of PAG and CO2 is investigated. The surface Fig. 14 Surface tension of PAG/CO2 mixture.
 international journal of refrigeration 73 (2017) 125133 131

Fig. 15 Surface tension of PAG/CO2 mixture with correlation.

with that of PAG, the surface tension of the PAG/CO2 mixture
steeply decreases with the increase of CO2 dissolved in the
mixture. The surface tension of the mixture becomes about half
of that of pure PAG when the concentration of CO2 is 20%. The
surface tension of the mixture at 70 C is lower than that of
30 C, and is also lower than the surface tension of pure PAG.
In this study, the correlation between the surface tension
of the PAG/CO2 mixture and the concentration of refrigerant
in the mixture is shown by Eq. (7).
= r + ( o r ) (1 C)n (7) tains dissolved refrigerant and its change is expressed by Eq.
where, C is the concentration of refrigerant in the mixture, n
is the index factor, and ref and oil are the surface tensions of
refrigerant and oil respectively. Note that the surface tension
of CO2 at the temperature above the critical point (31 C) cannot
be defined and is regarded as 0 N m1 in this study.
Fig. 15 shows the surface tension measured by MBPM at each
temperature. The surface tension calculated by Eq. (7) is also
shown in Fig. 15, using the index factors of 2, 3 or 4 in Eq. (7).
Although the experimental results are scattered mainly due to
insufficient accuracy of the sampling method, it is shown that
relatively good agreement of the calculated surface tension with
the experimental one is obtained at all temperatures when n = 3
in Eq. (7). Jensen and Jackman (1984) proposed a correlation equa-
tion similar to Eq. (7) and the index factor n of 0.5 is used for
combinations of R113 and several kinds of oil. It may happen
that the decrease of the surface tension of oil by dissolving re-
frigerant largely depends on the combination of oil and refrigerant.
Since the surface tension of oil decreases steeply when it con-
(7), the change of the surface tension can be used to detect the
refrigerant concentration, and the maximum bubble pressure
method is utilized to develop a refrigerant concentration sensor.
The surface tension measurement of the mixture under high
temperature and pressure condition as occur inside an actual
working compressor with a high pressure shell is very diffi-
cult. The surface tension of the PAG/CO 2 mixture in a
compressor used in a heat pump water heater, therefore, is
estimated from the surface tension obtained in this study. The
132 international journal of refrigeration 73 (2017) 125133
Fig. 16 Solubility of CO2 and PAG.
Fig. 17 Correlation of surface tension at each temperature.
representative condition inside the compressor for the heat
pump water heater is supposed to be 100 C and 10 MPa. Fig. 16
shows solubility data for PAG used in this study and CO2. As
shown in this figure, the concentration of CO2 in PAG under
100 C and 10 MPa is found to be 15%. Therefore, the surface
tension of the mixture under such conditions is estimated from
the surface tension with the same CO2 concentration, i.e. 15%,
at 30 C ~ 70 C by an extrapolation. The surface tension of the
mixture at each temperature expressed by Eq. (7) with n = 3 is
shown in Fig. 17. The surface tension of the mixture with a CO2
concentration of 15% is obtained at each temperature from Eq.
(7). Fig. 18 shows the surface tension of the mixture having a
CO2 concentration of 15% obtained by the above method against
the temperature from 30 C to 70 C. The surface tension of the
mixture at the 15% CO2 concentration decreases lineally with
the temperature. By the extrapolation of this tendency to 100 C,
the surface tension of the mixture at 100 C and 10 MPa is es-
timated to be 14.6 mN m1.
5. Conclusion
In this study, the surface tension measurement of a PAG/CO2
mixture is carried out by the maximum bubble pressure method
Fig. 18 Prediction of surface tension at 10 MPa and 100C.
as the first step to examine the oil mist generation in refrig-
erant compressors. The following conclusions are obtained:
1. The experimental apparatus developed in this study is ap-
plicable for measuring the surface tension under high pressure
and temperature conditions. It has sufficient accuracy to
measure the surface tension of an oil/refrigerant mixture.
2. The density of the PAG/CO2 mixture is measured by using
the ultrasonic sensor. The density of the mixture shows little
change with CO2 concentration. It can be assumed that the
density of the PAG/CO2 mixture is only the function of
temperature.
3. The surface tension of the PAG/CO2 mixture sharply de-
creases as the concentration of CO2 in the mixture increases.
The surface tension of the mixture decreases with
temperature.
4. The correlation between the surface tension of the PAG/
CO2 mixture and the CO2 concentration in the mixture is
obtained by Eq. (7) in which the index n is proposed as 3.
5. From the surface tension of the mixture whose refriger-
ant concentration is 15% at each temperature, the surface
tension of the mixture at 100 C and 10 MPa, which is sup-
posed to be the representative condition in the high pressure
shell compressor for a heat pump water heater, is esti-
mated to be 14.6 mN m1 by the extrapolation.
Acknowledgement
The density measurement was done with help of Mr. Tatsuhiro
Suzuki, working at DENSO Corp., when he was a student at
Graduate School of Engineering, Shizuoka University, and the
authors appreciate his effort. The authors would also like to
thank Prof. Predrag Hrnjak of The University of Illinois at
Urbana-Champaign for giving us information about the pres-
sure differential sensor.
REFERENCES
Cai, W., Kong, W., Wang, Y., Zhu, M., Wang, X., 2015. Surface
tension of lithium bromide aqueous solution/ammonia with
additives and nano-particles. J. Cent. South Univ. 22, 19791985.
 international journal of refrigeration 73 (2017) 125133
Dang, C., Hoshika, K., Hihara, E., 2012. Effect of lubricating oil on
the flow and heat transfer characteristics of supercritical
carbon dioxide. Int. J. Refrigeration 35, 14101417.
Fainerman, V.B., Miller, R., 2004. Maximum bubble pressure
tensiometry an analysis of experimental constraints. Adv.
Colloid Interface Sci. 108109, 287301.
Fainerman, V.B., Miller, R., Joos, P., 1994. The measurement of
dynamic surface tension by the maximum bubble pressure
method. Colloid Polym. Sci. 272, 731739.
Feenstra, P.A., Judd, R.L., Weaver, D.S., 2001. A practical devise for
surface tension measurement in volatile fluids. HVAC&R Res
7 (1), 314.
Fukuta, M., Suzuki, T., Yanagisawa, T., 2008. Concentration
measurement of refrigerant/refrigeration oil mixture by
sound speed. Trans. Jpn. Soc. Refrigerating Air Cond. Eng. 25
(4), 391400, (in Japanese).
Geller, V.Z., Zhelezny, V.P., Bivens, D.B., Yokozeki, A., 1996.
Transport properties and surface tension of R23/116 Azeotropic
mixture. In: Proceedings of the International Refrigeration
and Air Conditioning Conference at Purdue. pp. 396402.
Gonalves, F.A.M.M., Trindade, A.R., Costa, C.S.M.F., Bernardo,
J.C.S., Johnson, I., Fonseca, I.M.A., et al., 2010. PVT, viscosity,
and surface tension of ethanol: new measurements and
literature data evaluation. J. Chem. Thermodyn. 42 (8), 1039
1049.
Jensen, M.K., Jackman, D.L., 1984. Prediction of nucleate pool
boiling heat transfer coefficients of refrigerantoil mixtures.
J. Heat Transf. 106, 184190.
Khosharay, S., Varaminian, F., 2014. Experimental and modeling
investigation on surface tension and surface properties of
(CH4+H2O), (C2H6+H2O), (CO2+H2O) and (C3H8+H2O) from
284.15 K to 312.15 K and pressures up to 60 bar. Int. J. Refrig.
47, 2635.
Kondou, C., Nagata, R., Nii, N., Koyama, S., Higashi, Y., 2015.
Surface tension of low GWP refrigerants R1243zf, R1234ze(Z),
and R1233zd(E). Int. J. Refrig. 53, 8089.
Lemmon, E.W., Huber, M.L., McLinden, M.O., 2013. REFPROP
reference fluid thermodynamic and transport properties.
NIST Standard Reference Database 23, Ver.9.1.
133
Majithia, A.K., Hall, S., Bowen, P.J., 2008. Droplet breakup
quantification and processes in constant and pulsed airflows.
In: Proceedings of the 22nd European Conference on Liquid
Atomization and Spray System, ILASS08-4-4.
Murakami, H., Wakamoto, S., Morimoto, O., 2006. Performance
prediction of cyclone oil separator. In: Proceedings of the
International Refrigeration and Air Conditioning Conference
at Purdue. p. R028.
Ogata, T., Hasegawa, H., Okaichi, A., Nishiwaki, F., 2006. Reduction
of oil discharge for rolling piston compressor using CO2
refrigerant. In: Proceedings of the International Compressor
Engineering Conference at Purdue C095.
Sakurai, Y., Kobayashi, K., Fujikawa, T., Sanada, T., Watanabe, M.,
2012. Observation of deformation of single droplet impact on
solid surface (effects of liquid properties, surface roughness
and surrounding gas pressure). Jpn. J. Multiphase Flow 26 (2),
164171, (in Japanese).
Seeton, C.J., Hrnjak, P., 2006. Thermophysical properties of CO2-
lubricant mixtures and their affect on 2-phase flow in small
channels (less than 1 mm). In: Proceedings of the
International Refrigeration and Air Conditioning Conference
at Purdue. p. R170.
Shikazono, N., Azuma, R., Sameshima, T., Iwata, H., 2010.
Development of compact gasliquid separator using surface
tension. In: Proceedings of the 2010 International Symposium
on Next-generation Air Conditioning and Refrigeration
Technology. p. NS11.
Song, S.J., Noh, K.Y., Min, B.C., Yang, J.S., Choi, G.M., Kim, D.J.,
2015. Observation of oil flow characteristics in rolling piston
rotary compressor for reducing oil circulation rate. In:
Proceedings of the 9th International Conference on
Compressors and their Systems 012071.
Toyama, T., Matsuura, H., Yoshida, Y., 2006. Visual techniques to
quantify behavior of oil droplets in a scroll compressor. In:
Proceedings of the International Compressor Engineering
Conference at Purdue C026.
Wiencke, B., 2011. Fundamental principles for sizing and design
of gravity separators for industrial refrigeration. Int. J.
Refrigeration 34, 20922108.