Applied Thermal Engineering 29 (2009) 509514

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

An experimental investigation of the optimum geometry for the cold end orice
of a vortex tube
Sachin U. Nimbalkar *, Michael R. Muller
Center for Advanced Energy Systems, Mechanical Engineering, Rutgers University, 98 Brett Road, Piscataway, New Jersey, USA

a r t i c l e

i n f o

Article history:
Received 14 June 2007
Accepted 9 March 2008
Available online 25 March 2008
Keywords:
RanqueHilsch vortex tube
Energy separation
Secondary ow

a b s t r a c t
A vortex tube is a simple mechanical device, which splits a compressed gas stream into a cold and
hot stream without any chemical reactions or external energy supply. This paper presents the results
of a series of experiments focusing on various geometries of the cold end side for different inlet
pressures and cold fractions. Specically, the tests were conducted using different cold end orice
diameters.
Energy separation and energy ux separation efciencies are dened and used to recover characteristic properties of the vortex tube. These are used to show an appropriate scale to non-dimensionalize
the energy separation effect. The experimental results indicate that there is an optimum diameter of
cold end orice for achieving maximum energy separation. The results also show that the maximum
value of energy separation was always reachable at a 60% cold fraction irrespective of the orice
diameter and the inlet pressure. The results are compared with the previous studies on internal ow
structure, and optimal operating parameters are shown to be consistent with a matching of orice
size with the secondary circulation being observed.
2008 Elsevier Ltd. All rights reserved.

1. Introduction
Energy separation is the re-distribution of the total energy in
a uid ow without external work or heat, so that some portion of
the uid has a higher total energy, and the other portion has a lower total energy than the inlet uid. The RanqueHilsch tube [1] is
an ideal example to this phenomenon, which separates the ow
of gas into two streams simultaneously, using the generation of a
strong vortex whose axis is aligned with the directions of ow. This
has resulted in a large number of engineering applications utilizing
this ow behavior, such as the separation of gases of different
molecular weight, refrigeration schemes, cyclone separation, gas
turbine cyclones, combustion chambers, and spray dryers.
The literature concerning the design, operation and performance of vortex tubes is extensive, with descriptions rst appearing in 1948 ([2]) with excellent review papers by Takahama [3],
Ahlborn and Groves [7], and Gao et al. [18]. The studies focusing
on the role of internal geometry include Takahama and Soga
[19], Hartnett and Eckert [20], Marshall [5], Saidi and Valipour
[6], and Eiamsa-ard and Promvonge [16].
During 1960s, Takahama carried out experimental studies on
pressure and temperatures inside the vortex tube and studied

* Corresponding author. Tel.: +1 7324455540; fax: +1 7324450730.

E-mail address: nsachin@caes.rutgers.edu (S.U. Nimbalkar).
1359-4311/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.applthermaleng.2008.03.032

the effect of various geometric factors on its energy separation

characteristics [3,4]. In 1977, Marshall [5] conrmed experimentally that separation is primarily dependent upon centrifugation.
His results appear substantially to indicate in comparing the standard and the large tubes that the gas separation performance is the
same if the effect of overall pressure drop is considered. In 2003,
Saidi and Valipour [6] found that the generator with more ow intakes causes the cold air temperature and the efciency of the vortex tube to decrease.
Finally, of importance to the work discussed here, several
researchers [11,17] have sought to characterize the internal ow
details including the existence of a secondary ow gyre. Specically, in 1997, Ahlborn and Groves [7] used a pitot tube to observe
a secondary ow within the vortex tube. From the measured velocity eld they determined that the return ow at the center of the
tube is much larger than the cold mass ow emerging out of the
cold end. Therefore, the vortex tube must have a secondary circulation imbedded into the primary vortex, which moves uid from
the back ow core to the outer regions.
The current study examines the role of the orice, which restricts the ow out of the cold end side of the tube seeking to
determine the conditions for maximum temperature and energy
ux separation (two different features) over a range of parameters.
The next section details the experimental setup and denes the
geometric terms. The following section presents the experimental
results along with an interpretation.

510

S.U. Nimbalkar, M.R. Muller / Applied Thermal Engineering 29 (2009) 509514

2. Experimental apparatus

Table 1
Geometric parameters and test sequence

The typical setup for the experiments is shown in Fig. 1. Clean,

dry air from a compressed air system is reduced through a pressure
reduction valve to the desired value and is then dumped into a plenum chamber. From there the ow accelerates into the inlet port of
the vortex tube. The ow is then forced into an energetic vortex
through multiple jets or a single jet oriented tangentially. This
spinning uid is turned 90 and can exit from either end of the
tube. The ow out of the hot end side is restricted by a tapered
valve, which creates back pressure and preferentially selects uid
near the outer wall. The ow out of the cold end side is restricted
by a disk with a circular hole in its center. This orice preferentially selects the ow from the center of the tube, which will be
colder due to the adiabatic expansion associated with turbulence
contained within the vortex.
The purpose of this study was to examine the effect of the cold
end orice diameter on the overall energy separation of the vortex
tube. Different generators with varying cold-end orices were
manufactured in the lab (see Table 1 and Fig. 2), and tests were
conducted using a commercially manufactured vortex tube (Exair
Corp., Model No. 3299, maximum mass ow rate 5 kg/min at
690 kPa inlet pressure) for different inlet mass ow rates
(0.45, 0.68 and 0.82 kg/min).
The temperature of the air entering into the vortex tube and
the air leaving at hot and cold-end sides were measured by
thermistors. The total mass ow entering the system, the mass
ow leaving the cold end and the mass ow leaving the hot
end were measured by the turbine type of ow meters. The turbine type of ow meter produces less pressure drop across it as
compared to variable area ow meters. It was observed that the
back pressures generated at cold and hot ends due to the variable area ow meters affect the ow characteristics of the vortex
tube, and may force a recirculating ow inside it. Inlet pressure
was measured with the help of a diaphragm type pressure transducer which was calibrated by the manufacturer and cross
checked with the help of a dead weight calibrator in the measurement lab.

Generator
number

Generator
Generator
Generator
Generator

1
2
3
4

Inner
diameter
D (cm)

Generator nozzle
Width
(cm)

Length
(cm)

Nozzles

1.905
1.905
1.905
1.905

0.3175
0.3175
0.3175
0.3175

0.4064
0.4064
0.4064
0.4064

4
4
4
4

Cold end orice

diameter dc (cm)

dc1
dc2
dc3
dc4

0:3454
0:6985
0:9576
1:2636

dr dDc

0.18
0.37
0.50
0.66

The vortex tube used while performing all the experiments had
an internal diameter D = 1.905 cm and a length L = 25.4 cm (L/D ratio equal to 13.33). The tube was made of stainless steel and thermally insulated from the atmosphere to avoid errors due to heat
conduction and to maintain repeatable steady-state.
3. Appropriate non-dimensionalization of energy separation
effect
To recover the characteristic properties of the vortex tube and
its energy separation effect, it is necessary to non-dimensionalize
various geometric and thermo-physical parameters involved.
These parameters are inlet pressure (Pi), orice diameter (dc), inlet
mass ow rate (mi), cold end mass ow rate (mc) and pressure drop
across the hot fraction control valve.
In the past, most of the researchers did not use the dimensionless parameters in their research work, and the remaining of them
used improper scaling parameters for non-dimensionalization. For
example, the use of inlet temperature to non-dimensionalize temperature splitting effect (Fig. 3A). Temperature splitting effect is
dened as the difference between hot and cold end side temperatures (Eq. (1)). Basically, it is the maximum possible temperature
gradient achieved by the vortex tube. If the compressed air is generated at some remote location, then the inlet temperature would
not change with the inlet pressure.
Hence, it would be inappropriate to use the inlet temperature to
scale the temperature gradients. But if compression is done onsite,

Fig. 1. Schematic of the experimental apparatus.

511

S.U. Nimbalkar, M.R. Muller / Applied Thermal Engineering 29 (2009) 509514

Fig. 2. Lab manufactured generators with varying orice diameter.

where k is specic heat ratio for air = 1.4, R is gas constant = 287 J/
kg K.
Hence, non-dimensionalized version of the energy splitting effect is

2.5

(Th-Tc)/Ti

90% cf

77% cf

60% cf

50% cf

37% cf

15% cf

Energy splitting effect

1.5

1.50

2.00

2.50

3.00

3.50

Inlet Presure Ratio (Pi/Pa)

Fig. 3A. Effect of inlet pressure ratio on the temperature splitting effect for various
cold fractions.

inlet temperature would change with inlet pressure (due to compressor work), and it would be appropriate to use the inlet temperature to scale the temperature splitting effect. Temperature
splitting effect is dened as
Temperature splitting effect T h  T c

More properly, we should use input energy (which is compressor

work) to non-dimensionalize the temperature gradients as shown
in Fig. 3B. Input energy is the compression work required to compress air from atmospheric pressure (Patm) to the inlet pressure
(Pi) isentropically which is given by
(
)
k1
k
k
Pi
CW
 1 J=kg;
2
 R  T atm 
Patm
k1

1.00

Energy separation/ CW

0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
1.00

4. Results and discussion

0.5

0
1.00

cp T h  T c
CW

1.50

2.00

2.50

90% cf

77% cf

60% cf

50% cf

37% cf

15% cf

3.00

In Fig. 4, energy separation efciency is plotted against the cold

fraction for various cold end orice diameters [dr = (dc/D) = 0.18
0.66] at inlet mass ow rate of 0.45 kg/min. The energy separation
efciency can be dened as the ratio of energy separation effect
(either cold end or hot end) to the isentropic compression work
per unit mass.
Energy separation efficiency cold end

cp T i  T c
CW

Inlet mass ow rate is already non-dimensionalized in the form of

cold fraction (mc/mi). So by dening the energy separation efciency and the cold fraction, dimensional quantities like temperatures, inlet pressure, and inlet and cold mass ow rates are nondimensionalized.
From Fig. 4, it can be observed that hot side energy separation
efciencies are almost linear except at very high cold fractions
(above 60%). Certainly, energy separation efciencies for different
orice diameters collapse on each other more at low cold fraction
showing, what we believe to be, similar uid mechanics. On the
other hand, cold side energy separation efciencies should keep
increasing with drop in cold fraction as observed by Stephan et al.
[8]. But on the contrary, below 60% cold fraction energy separation
efciencies are either constant or decreasing asymptotically.
Above observations indicate that the effect of orice diameter
on energy separation efciency is negligible for cold fraction less
than 60%.
Another important observation is that at 50% cold fraction both
cold and hot end energy separation efciencies are almost equal as
expected. But at the same time, at 50% cold fraction almost every
curve representing a different orice diameter is passing through
the same point. That means, for 50% cold fraction vortex tube
shows similar energy separation performance irrespective of orice diameter.
To take into consideration the effect of the mass ow rates on
the performance of the vortex tube, let us dene the energy ux
separation efciency. It is dened as the ratio of cold end energy
ux (or hot end energy ux) to the isentropic compressor work.
Energy flux separation effciency cold end cf 

3.50

Inlet Presure Ratio (Pi/Pa)

Fig. 3B. Appropriate non-dimensionalization of the energy splitting effect with
isentropic compression work.

cp T i  T c
CW

Fig. 5 shows the dependence of energy ux separation efciency

of the vortex tube on the cold fraction with the orice diameter at
the inlet mass ow rate of 0.45 kg/min. As discussed earlier, the

512

S.U. Nimbalkar, M.R. Muller / Applied Thermal Engineering 29 (2009) 509514

1.00

Energy Separation Efficiency

0.80
0.60
0.40

Ec for dc/D= 0.18

Eh for dc/D= 0.18
Ec for dc/D= 0.37
Eh for dc/D= 0.37
Ec for dc/D= 0.50
Eh for dc/D= 0.50
Ec for dc/D= 0.66
Eh for dc/D= 0.66

0.20
0.00
0
-0.20

10

20

30

40

50

60

70

80

90

100

-0.40
-0.60
-0.80

Cold fraction (cf) %

Fig. 4. Effect of orice diameter on the energy separation.

effect of inlet mass ow rate is washed out because of its presence

in the cold fraction.
The results are showing that below 60% cold fraction the energy
uxes are varying almost linearly, and the effect of orice diameter
is very insignicant. Beyond 60% cold fraction, the effect of orice
diameter is dominant. Firstly, energy ux separation efciency is
dropping with an increase in cold fraction and secondly for constant cold fraction it is increasing with orice diameter.
Hence, it can be concluded that while operating the vortex tube
for cold fraction above 60%, selecting larger orice diameter will
give better energy ux separation efciency.
4.1. Explaining the drop in efciency beyond 60% cold fraction
Ahlborns [7] theory of the existence of a secondary circulation
and resulting mixing of cold air with hot air or vice versa inside
vortex tube can be used to explain performance degradation of
the vortex tube beyond 60% cold fraction. In the next section, the
phenomenon of secondary circulation and the mechanism of mixing are briey discussed.

implies the existence of a secondary circulation ow in the vortex

tube. He also found the presence of a backow core even at cf = 0,
when the cold end is closed, and as the cold mass ow rate increases the radius of backow core also increases.
Secondary circulation ow appears to be the characteristics of
cold end orice diameter as conrmed by our studies and could
be a performance degrading mechanism. The degradation could
be due to the transfer of colder uid elements near the cold end
exit zone through the swirling secondary loop to the warmer ow
region causing a decrease in the hot-end temperature and the
transfer of warmer ow elements back to the cold end exit zone
causing an increase in cold end exit temperature of the ow.
Let us assume, dc1,2,3 = different cold end orice diameters,
dcold = diameter of the backow core, and D = Vortex tube inner
diameter (see Fig. 6 and Table 2).
Parametric relation between energy ux separation efciency,
dc/D and dcold/D can be dened as follows:
Energy flux separation effciency cold end
(

2 )
dc  dcold
;
Fudge Factor  1 
D

where dcold/D is the function of inlet pressure, nozzle dimensions

and uid properties. Above equation (Eq. (6)) shows that energy
ux separation efciency is maximum when dc = dcold. The effect
of orice diameter on energy ux separation efciency and cold
end temperature splitting for different mass ow rates is shown
in Fig. 7.
4.3. Optimum cold fraction
As discussed before, Fig. 5 shows that the maximum performance factor was always reachable at a 60% cold fraction irrespective of the orice diameter and the inlet pressure. Through
application point of view it is necessary to investigate why maximum energy separation (equal for hot and cold sides) occurs at
60% ow fraction irrespective of the orice diameter of the
generator.

4.2. Secondary circulation and mixing theory

Ahlborn and Groves [7] proposed the so-called secondary circulation model based on his experimental results. He found that the
cumulative mass ow over the cross section of the vortex tube in
the cold-end direction is larger than the cold exhaust ow, which

Fig. 5. The effect of orice diameter on the energy ux separation efciency

(mi = 0.45 kg/min). Error bars in the graph indicate 5% of experimental error.

Fig. 6. The ow pattern inside the vortex tube.

S.U. Nimbalkar, M.R. Muller / Applied Thermal Engineering 29 (2009) 509514

Table 2
Secondary circulation and mixing hypothesis
dc1 < dcold

dc2 = dcold
dc3 > dcold

The transfer of colder uid elements near the cold end exit zone
through the swirling secondary loop to the warmer ow region causing
a decrease in the hot end temperature
Cold stream through cold end and hot stream through hot end. No
recirculation
The transfer of warmer ow elements back to the cold end exit zone
causing an increase in cold end exit temperature of the ow

Fig. 7. The effect of orice diameter on the energy ux separation efciency (mi =
0.45, 0.68 and 0.82 kg/min). Error bars in the graph indicate 5% of experimental
error.

4.4. Explaining the 60% cold fraction

In the year 1995, Zaidkov and co workers [12] found that the
maximum refrigeration capacity of the vortex tube falls within
the range from 60% to 70% cold fraction. Later in year 2003, Poshernev and Khodorkov [13] found that within their range of the input
parameters the refrigerating capacity and the adiabatic efciency
have a distinct maximum at cold fraction = 50%60%. As stated before, our results also show that the maximum value of performance
factor is always reachable at a 60% cold fraction irrespective of the
orice diameter and the inlet pressure. This phenomenon can be
explained on the basis of pressure balance inside the vortex tube
discussed by Love [9] and Piralishvili and Fuzeeva [10].
As shown in Fig. 8, the total pressure drop across vortex tube is
the addition of pressure drop at the inlet (a), pressure drop due to
the generator (b), the ow (c), pressure drop through the cold-end
orice (d) and due to the hot fraction control valve (e). Pressure
drop across the generator takes into account the geometry of the
generator and the number of nozzles machined on its surface. Pressure drop due to the ow will take into consideration not only the
decay of the vorticity in the swirling ow but also the reduction or

513

reversal of the axial velocity in the core of the vortex. Pressure drop
across the hot fraction control valve decides the cold fraction and
the location of axial stagnation point.
As stated by Brendon et al. [14] and Beran and Culick [15]
depending upon the initial swirl intensity, the swirl decay rate
and various pressure drops inside the tube for similar geometric
and thermo-physical conditions, different types of ow structures
can be observed inside the vortex tubes. But for a constant inlet
pressure and xed orice diameter, ow structure inside the vortex tube would completely depend upon the pressure drop across
the hot fraction control valve. And as the pressure drop across the
hot fraction control valve denes the amount of cold fraction (cf),
cold fraction becomes a better parameter to relate the ow structure with the performance factor of the vortex tube. We hypothesize that when there is an increase in hot fraction (or decrease in
cold fraction), axial stagnation point moves towards the hot end,
and due to the stretching of the central recirculating core, radial
stagnation point moves towards the axis of the tube. On the contrary, when the hot fraction decreases, axial stagnation point
moves towards the cold end, and radial stagnation point travels
to the wall of the tube. But for the ideal separation of cold and
hot air streams, there are xed critical locations for the axial and
the radial stagnation points. We believe that at 60% cold fraction,
the vortex tube achieves aforementioned ideal operating conditions, and hence the performance factor reaches to its maximum
value irrespective of orice diameter and inlet pressure.
5. Conclusions
Secondary circulation ow appears to be characteristics of the
cold end orice diameter as conrmed by our studies and could
be a performance degrading mechanism. The degradation could
be due to the transfer of colder uid elements near the cold end
exit zone through the swirling secondary loop to the warmer ow
region causing a decrease in the hot end temperature and the
transfer of warmer ow elements back to the cold end exit zone
causing an increase in cold end exit temperature of the ow.
Isentropic compression work seems an appropriate scale to
non-dimensionalize the energy separation effect. Energy separation and energy ux separation efciencies are suitable for measuring the parameters to recover the characteristic properties of
the vortex tube.
The experimental results indicate that there is an optimum
diameter of cold-end orice for achieving maximum energy separation. It was observed that for cold fraction 660%, the effect of
cold end orice diameter is negligible and above 60% cold fraction
it becomes prominent. The results also show that the maximum
value of performance factor was always reachable at a 60% cold
fraction irrespective of the orice diameter and the inlet pressure.
Our claim is that the cold fraction is the crucial parameter to relate
the ow structure inside the vortex tube to its performance factor.

Fig. 8. Characterizing pressure drops in vortex tubes. (a) Pressure drop at the inlet, (b) pressure drop through the generator, (c) pressure drop due to ow, (d) pressure drop
through the cold end orice, and (e) pressure drop across the hot fraction control valve.

514

S.U. Nimbalkar, M.R. Muller / Applied Thermal Engineering 29 (2009) 509514

References
[1] G.J. Ranque, Expriences sur la dtente giratoire avec production simultane
dun chappement dair chaud et dun chappement froid, J. Phys. Radium 4
(1933) 112.
[2] R. Hilsch, The use of the expansion of gases in a centrifugal eld as cooling
process, Rev. Sci. Instrum. 18 (2) (1947) 108.
[3] H. Takahama, Studies on vortex tubes: 1st and 2nd report, Bull. JSME 8 (31)
(1965) 433440.
[4] H. Takahama, H. Yokosawa, Energy separation in vortex tubes with a divergent
chamber, J. Heat Transfer (1981) 103.
[5] J. Marshall, Effect of operating conditions, physical size and uid
characteristics on the gas separation performance of a LinderstromLang
vortex tube, Int. J. Heat Mass Transfer 20 (1977) 227231.
[6] M.H. Saidi, M.S. Valipour, Experimental modeling of vortex tube refrigerator,
Appl. Therm. Eng. 23 (2003) 19711980.
[7] Boye Ahlborn, Stuart Groves, Secondary ow in a vortex tube, Fluid Dyn. Res.
21 (1997) 7386.
[8] K. Stephan, S. Lin, et al., An investigation of energy separation in a vortex tube,
Int. J. Heat Mass Transfer 26 (3) (1983) 341348.
[9] Wiilliam J. Love, Prediction of pressure drops in straight vortex tube, AIAA J. 12
(7) (1974).
[10] A. Piralishvili, A.A. Fuzeeva, Hydraulic characteristics of RanqueHilsch energy
separators, High Temp. 43 (6) (2005) 900907.

[11] R.T. Balmer, Pressure driven RanqueHilsch temperature separation in liquids,

Trans. ASME J. Fluids Eng. 110 (1988) 161.
[12] V.V. Nikolaev, V.P. Ovchinnikov, M.A. Zhidkov, Experience from the
operation of a variable vortex tube in a gas separating station, Gaz. Prom.
10 (1995) 13.
[13] N.V. Poshernev, I.L. Khodorkov, Experience from the operation of a conical
vortex tube with natural gas, Chem. Pet. Eng. 39 (2003) 910.
[14] Brendon Adams, Michael Jones, et al., Hysteresis in the open pipe or with
vortex breakdown, in: Second International Conference on CFD in the
Minerals and Process Industries, CSIRO, Melbourne, Australia, December
1999.
[15] P.S. Beran, F.E. Culick, The role of non-uniqueness in the development of vortex
breakdown in tubes, J. Fluid Mech. 242 (1992) 491527.
[16] S. Eiamsa-ard, P. Promvonge, Review of RanqueHilsch effects in vortex tubes,
Renew. Sust. Energy Rev. (2007).
[17] Y.T. Wu, Y. Ding, Y.B. Ji, C.F. Ma, M.C. Ge, Modication and experimental
research on vortex tube, Int. J. Refrig. 30 (2007) 10421049.
[18] C.M. Gao, et al., Experimental study on a simple Ranque Hilsch vortex tube,
Cryogenics 45 (3) (2005) 73183.
[19] H. Takahama, N. Soga, Studies on vortex tubes: 2nd report, Bull. JSME 9 (33)
(1966) 121137.
[20] J.P. Hartnett, E.R.G. Eckert, Experimental study of the velocity and temperature
distribution in a high velocity vortex tube ow, Trans. Am. Soc. Mech. Eng.,
Series C, J. Heat Transfer 79 (1976) 751758.