International Journal of Multiphase Flow 29 (2003) 1237–1247

www.elsevier.com/locate/ijmulflow

Pool boiling of nano-fluids on horizontal narrow tubes

a,*
Sarit K. Das , Nandy Putra b, Wilfried Roetzel b

a
Heat Transfer and Thermal Power Laboratory, Mechanical Engineering Department,
Indian Institute of Technology, Madras 600036, India
b
Institut f€ur Thermodynamik, Universit€ at der Bundeswehr, Hamburg D-22039, Germany
Received 1 October 2002; received in revised form 14 May 2003

Abstract
The search for new cooling medium does not limit itself to liquids alone. Liquid–solid suspensions have
got a good promise in convective cooling applications. Suspension of common fluids with particles of the
order of nanometers (typically 10–100 nm) in size are called Ônano-fluidsÕ which have been found to enhance
the heat transfer capability of the base fluid to a considerable extent. With very small volume fraction, such
particles are capable of increasing the thermal conductivity and convective heat transfer significantly
without the known problems encountered in common slurries such as clogging, erosion, sedimentation and
increase in pressure drop. A recent study on pool boiling on a tube of large diameter (20 mm) shows that
the nano-particles degrade the boiling performance with increasing particle concentration pushing up the
wall superheat for a given heat flux. The present investigation focuses on an experimental study of pool
boiling in water–Al2 O3 nano-fluids on horizontal tubes of small diameter. Tubes of small diameter are of
interest in efficient cooling applications such as those in electronic modules or LASER devices where
miniaturisation is taking place at a rapid pace. However pool boiling of narrow horizontal tubes (4 and 6.5
mm diameter) is qualitatively different from the large diameter tubes due to difference in bubble sliding
mechanism. It is found that at the range of narrow tubes the deterioration in performance in boiling is less
compared to large industrial tubes which makes it less susceptible to local overheating in convective ap-
plications. Thus, the present study on boiling of nano-fluids can act as a guidance for the use of these
engineered fluids in the above applications.

2003 Elsevier Ltd. All rights reserved.

1. Introduction

In recent times heat transfer technology is confronted with increasing demand of cooling ap-
plications of miniaturised high heat flux components. Applications such as LASER diagnostics,

*
Corresponding author. Tel.: +91-44-4458550.
E-mail address: [email protected] (S.K. Das).

0301-9322/03/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0301-9322(03)00105-8
1238 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247

superconducting magnets and most importantly superfast computing are posing tremendous
challenge to thermal management. Usually air based cooling systems are more common and re-
liable but they are inadequate for high heat flux applications where liquid cooling is preferred. The
cooling liquids usually used are chilled water, refrigerants or cryogens depending on the re-
quirement of heat removal. Looking at the requirement, it makes sense to consider alternatives
such as fluid suspensions of ultrafine solid particles. The fluids with suspended solid particles have
not been considered as an alternative for heat transfer applications so far due to associated
technological problems such as sedimentation, clogging, erosion, fouling and increase of pressure
drop. However in recent times a fresh look has been cast on these fluid–solid suspensions with
particles of nano-meter size which have been named as Ônano-fluidsÕ by Choi (1995). The erosion,
clogging and pressure drop problems are also greatly reduced due to small particles and the small
volume fraction (usually 1–5%) required and the stability of such fluids against sedimentation is
remarkably improved. Lee et al. (1999) reported a substantial enhancement of thermal conduc-
tivity of water and ethylene glycol based nano-fluids with Al2 O3 or CuO nano-particles at room
temperature. In a recent study (Das et al., 2001) the present authors have shown that the en-
hancement of thermal conductivity of nano-fluids increases even more at elevated temperature
which makes it more attractive for cooling at high heat flux applications. This enhancement of
thermal conductivity received an impressive breakthrough when Eastman et al. (2001) reported an
increase of thermal conductivity by an outstanding 40% with only 0.04% of nano-particles of pure
copper having average size less than 10 nm. The above works indicate that usual theories of
thermal conductivity of suspensions such as the Hamilton and Crosser (1962) model fail in case
of nano-fluids. A satisfactory theory is yet to evolve.
However, for heat transfer applications, the enhancement of thermal conductivity is not the
only concern, the real worth of such fluids as coolants can only be examined under convective
conditions. Ahuja (1975) and Liu et al. (1988) have shown that performance of suspensions even
with micrometer size particles are encouraging under convective conditions. The proposition of
dispersion model by Xuan and Roetzel (2000) can be a useful tool in theoretical modelling of
nano-fluids under convective conditions. While using nano-fluids for convective cooling, one
cannot overlook the need of proper knowledge of its boiling characteristics. This is because during
convective heat transfer with high heat flux local boiling condition may be reached. It is important
to know the behaviour of nano-fluids under such conditions to avoid unwanted effects if any. The
present authors have carried out an experimental study (Das et al., 2003) of pool boiling char-
acteristics of water–Al2 O3 nano-fluid under atmospheric conditions on a tube of large diameter
(20 mm). The study shows that the nano-particles degrade the boiling performance systematically
with increasing particle concentration resulting in an increase of wall superheat for a given heat
flux. The deterioration in boiling performance is observed to be more drastic at a higher surface
roughness.
The present paper is aimed at understanding the pool boiling of nano-fluids in the regime of
small diameter tubes (4 and 6.5 mm) which are on the one hand more important for application
in miniaturised heat sources, on the other hand they are different from the usual tubes with
respect to pool boiling due to the closeness of bubble and tube diameter and a consequent
deviation in bubble sliding mechanism. Thus, the present study can act as a guidance for the
use of these fluids in the applications where high heat flux is associated with smaller dimensions
of the components.
 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247 1239

2. Characterisation of nano-fluid

Although a number of combinations of base fluid and particle can be used for nano-fluids, in
the present investigation, water–Al2 O3 particles nano-fluids have been used. This is due to the fact
that the boiling characteristics of the base fluid water is widely known and the enhancement of
thermal conductivity of water–Al2 O3 nano-fluids has already been studied (Das et al., 2001). The
particles were supplied by nano-phase Technologies Corporation, IL (USA) produced by physical
vapour deposition technique. In the powder state they form loose agglomerates as shown by
transmission electron microscopy (TEM) photo in Fig. 1. The dispersion of the particles in water
was carried out by using ultrasonic vibration for 4 h. It was observed that they break down to
smaller sizes when dispersed in water. Fig. 2 shows the TEM picture of dispersed particles by
drying a dilute dispersed suspension on silicon wafers. The volume weighted average particle size
was found to be 58.4 nm. The distribution of particle size is shown in Fig. 3. It was found that the
resulting suspension is quite stable. At higher concentration (4%) very nominal amount of sedi-
mentation occurs after a time (6 h) much greater than the time required for boiling experiment.

Fig. 1. TEM of nano-particles agglomerates.

Fig. 2. TEM of dispersed nano-particles.

1240 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247

14

12

10
Volume (%)
8

0
11.6 15.4 20.6 27.4 36.6 48.8 65 86.7 116 154 206 274 366
Diameter (nanometers)

Fig. 3. Particles size distribution of the nano-particles.

To characterize the produced nano-fluid, first a rheological study was made. The measurement
with a disc type rotating rheometer confirms a Newtonian behaviour of the nano-fluids used here
but the viscosity was found to be higher compared to water as shown by Das et al. (2003). This
rotating disc type rheometer contains a spring mounted base disc on which the fluid for which
viscosity is to be measured is spread as a thin film and is touched by a rotating disc at the top. The
top disc is attached to a spindle rotating with the help of a motor. For a given disc size and spindle
speed the deflection of the spring will be proportional to the fluid viscosity. The instrument was
first calibrated with water at different temperatures and finally measurement was made for nano–
fluids. After each measurement it was ensured that particles were not sticking to the discs. Since
surface tension is an important parameter in pool boiling, it was also measured by ring method
and found to be unaffected by the presence of the particles.

3. Pool boiling experiment

3.1. Experimental setup

A simple experimental setup was designed keeping watch that the experiments for different
nano-fluids and water were performed under identical conditions. The test section is shown in Fig.
4. It consists of 120 mm · 100 mm · 200 mm rectangular stainless steel vessel (1) with thick in-
sulation (2) outside. The vessel has two cooling arrangements cascaded together. The first one (3)
is a counter current copper condenser which on one hand connects the vessel directly to atmo-
sphere maintaining an atmospheric pressure in it, on the other hand it also serves the purpose of
after-cooling of any vapour which may try to escape as well as act as a vent to non-condensable
gases. The cooling water from this vertical condenser is then circulated through an oval shaped
copper coil (4) which performs the task of condensing the bulk of the vapour produced. This coil
hangs from the roof of the vessel and is designed to cool the entire vapour at the maximum rate of
 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247 1241

Fig. 4. Sketch of experimental apparatus.

evaporation. A pressure gauge (5) mounted at the top of the vessel checks the pressure at which
boiling takes place. As boiling surface, cylindrical cartridge heaters (6) of different diameters are
used. To observe the boiling characteristics during water experiments, round windows (7), with
double walled glass (inner 8 mm and outer 6 mm thickness) were built on both the side walls. A
sheathed 0.5 mm thick Chromel–Alumel (K-type) thermocouple (8) was inserted to observe the
bulk liquid temperature during boiling. To measure temperature on the heating cartridge 10
K-type thermocouples of 0.1 mm thickness were welded at different radial and axial locations as
indicated in Fig. 5. The radial locations are a, b, c and d and axial locations are 1, 2, and 3. The
thermocouples were planted at locations 1a, 1b, 2a, 2b, 2c, 3c and 3d. The leads of the thermo-
couples were taken out from tip of the heater in a bunch (9). The vessel was mounted on a se-
condary disc type heater (10) to supply additional heat during the experiment with the smaller
heaters with less power rating. The power supply to the primary heater was varied by a trans-
former and the power was recorded with a wattmeter. The thermocouples, pressure gauge and
wattmeter were connected to a data logger which was in turn connected to a PC for recording and
storage of data.

Fig. 5. Location of thermocouples on the heater.

1242 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247

In the present experiment stainless steel heaters of different diameters are used the surface of
which are machine drawn. The surface characteristics of the heater were measured using a
profilometer having a diameter tip of 2 lm and a sensitivity of 0.02 lm. The major parameters for
characterisation of surface roughness are Ra and Rq (DIN 4762) which are defined as
Z
1 L
Ra ¼ jZðxÞj dx ð1Þ
L 0
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Z
1 L 2
Rq ¼ ZðxÞ dx ð2Þ
L 0

A typical surface profile for the heaters is shown in Fig. 6. The typical measured values of
roughness for the different heater surfaces are of the same order (0.37–0.45 lm). The surface
roughness is measured before the boiling experiments. It is known from literature that the surface
structure changes after boiling experiments with water because a lot of deposits are found. One
may also have to consider the above phenomenon when one wants to measure the decrease of heat
transfer due to nano-particle deposits on the surface. Stationary conditions were observed during
the present experiments. Hence, the changes in the surface structure will be perhaps only in the
microstructure and have no influence on the heat transfer results.

3.2. Experimental procedure and uncertainty

At the beginning a set of runs with doubly distilled water was made to ensure that the ex-
perimental results conform to those in literature for boiling on horizontal tubes as well as to check
the repeatability of the experiment. The other objective of the water experiment was to observe the
boiling process visually and get an idea of the change in the physical process with reducing di-
ameter. During the experiments, temperatures were recorded at a number of thermocouple lo-

3
Ra = 0.387 µ m
Rq = 0.458 µ m
2

1
z [ µ m]

-1

-2

-3
0.00 0.25 0.50 0.75 1.00 1.25
x- position [mm]

Fig. 6. Surface characteristics of the heater.

 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247 1243

cations and an average was taken to designate the wall temperature. Before inserting, an infrared
camera was used to check the uniformity of heating by switching on the heating in air. Prior to
each experiment, the liquid was boiled for half an hour to drive out any dissolved gas. Experiment
was first performed from the lowest to the highest power input and then it was carried out in the
reverse direction to eliminate the possibility of any hysteresis using a sampling rate of 20 readings
per second. During experiments with the nano-fluids, runs with the same nano-fluid were repeated
with a run of boiling with pure water in between after cleaning the heater surface with a water jet
to ensure that the particles were not sticking to the heating surface to change the surface char-
acteristics.
The thermocouples on the heater surface with 0.1 mm diameter have an accuracy of 0.1 K and
that for fluid temperature measurement with 0.5 mm diameter (for better stability within boiling
fluid) is of 0.2 K accuracy. The wattmeter used for recording the power has got an accuracy of 10
W which is 1.25% at the lowest and 0.4% at the highest heat flux. The uncertainty in diameter
measurement was 0.05 mm. A systematic error analysis was made to ascertain the measurement
uncertainty which turned out to be 4.01% for surface heat flux. The uncertainty of temperature
difference remains to be restricted to a maximum of 4%.

4. Results and discussion

Previous study (Das et al., 2003) shows that for nano–particle concentration of present range a
very small decrease of 0.4 K in the boiling point occurs which is of the order of error in tem-
perature measurement and can be neglected. The visual observation for pure water shows that for
tube of large diameter (20 mm) bubbles sliding bubble regimes similar to those observed by a
series of studies by Cornwell and Sch€ uller (1982), Cornwell (1990) and the bubbles coalesce
forming larger bubbles similar to those observed by Chun and Kang (1998). This shows that
physically the boiling behaviour in the present apparatus is similar to that in the literature.
However, the mechanism for the tubes of narrow diameter was found to be considerably different.
For boiling on tubes of 4 and 6.5 mm diameter there seems to be less importance of sliding
mechanism for larger bubbles which are comparable to the size of bubbles of boiling on 20 mm
tube. This is because of the relatively small size of the tube which produces a large curvature of the
surface which does not allow the sliding of larger bubbles but induce direct departure. However a
large number of smaller bubbles are produced in a sustainable way here and they slide but to a
relatively smaller distance. For these experimental runs with water, Nu–Reb data are plotted in
Fig. 8 (this figure is placed after Fig. 7 for convenience of comparison with Fig. 9). The figure
shows that for tube of larger diameter the data fits quite well to convective type correlation as
suggested by Cornwell and Houston (1994) given by
Nu ¼ C Re0:67
b Pr
0:4
ð3Þ
as well as the correlation suggested by Gorenflo (1997) in VDI W€armeatlas. Here for both Nu and
Reb the tube diameter was used as the characteristic length and the measured viscosity of the
nano-fluid at saturation temperature was used in the calculation of Re. The plot shows that for
tube diameter between 4 and 6.5 mm the correlation does not perform very well which is expected
because the above correlations were not developed for narrow tubes. The above results clearly
1244 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247

water
Nano 1%
150 Nano 2%
Nano 4%

100
q [kW/m 2 ]

50

6.5 mm diameter tube 20 mm diameter tube

4 mm diameter tube

0
7 13 7 9 11 7 13
Tw - Ts [K]

Fig. 7. q–DT curves for nano-fluids on tubes of different diameters.

300

250

200
Nu

150
Diameter 6.5 mm
Diameter 4 mm
100 Diameter 20 mm
Diameter 20 mm
Diameter 20 mm
50 Cornw ell
Gorenflo VDI

0
0 0.5 1 1.5 2 2.5 3 3.5
Re

Fig. 8. Boiling characteristics of water on narrow tubes.

 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247 1245

200

180

160

140

120

100
Nu

Nano 1% (diam. 6.5 mm)

80 Nano 1 % (diam 4 mm)
Nano 4 % (diam 6.5 mm)

60 Nano 4 % (diam 4 mm)

Nano 1 % (diam 20 mm)

40 Nano 4 % (diam 20 mm)

20

0
0 0.5 1 1.5 2 2.5 3 3.5
Re

Fig. 9. Nu–Reb characteristics of nano-fluids.

indicate that the present apparatus is in good agreement with established boiling characteristics on
horizontal tube whereas it reaffirms the need to develop correlations for boiling on narrow tubes.
Subsequently, experiments were carried out to evaluate pool boiling with nano-fluids of 1%, 2%
and 4% Al2 O3 nano-particle concentration in water. Repeatable boiling characteristics of these
nano-fluids in the q–DT form are presented in Fig. 7 for different tube diameters. These plots
clearly indicate that, in general, the boiling performance of the base fluid deteriorates with the
addition of nano-particles pushing the boiling curves to the right which means that the nano-fluid
can cause harm to cooled surface if boiling occurs because it will give a higher surface temperature
compared to water at the same heat flux similar to that observed earlier (Das et al., 2003). It has
been observed that the shift of the curve to the right is not proportional to the particle concen-
tration and it is strongly dependent on the tube diameter even for the similar values of surface
roughness. For narrower heaters (4 and 6.5 mm) the shift of the curve is considerable and is
almost of the same order over the entire range of heat fluxes. For 20 mm tube from 1% to 4%
concentration a regular shift of the curve was observed at lower heat fluxes but at the upper part
of the curves the difference between wall superheats for various particle concentrations was found
to increase with increasing heat flux.
This depicts a regular tendency of deterioration of boiling character for nano-fluids which
generally increase with particle concentration but the nature of deterioration is different in the
narrow tube regime compared to the domain of large diameter tubes. The present study indicates
that in the region of narrow tubes, the tube diameter plays a crucial role to decide the nature of
this deterioration presumably for the change in bubble diameter and sliding bubble mechanism.
In the range of narrow diameter the deterioration seems to be independent of flux which for the
1246 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247

larger diameter is strongly flux dependent. Also, for pure fluids the heat transfer increases with
increase in viscosity but in the present case two competing phenomena are taking place, viz.,––
increase of heat transfer due to viscosity, decrease of heat transfer due to decrease of nucleation
site density by plugging off (micro) surface cavities by nano-particles. The results indicate that the
latter effect is dominant over the former because the increase in viscosity is very marginal.
To further understand the effect of nano-particles on heat transfer dimensionless Nu–Reb plot is
presented in Fig. 9. This figure indicates that for each particle concentration the Nu–Reb char-
acteristics are different and shifted downwards. This is a general observation for all the tubes
which indicates that the change in boiling characteristics of nano-fluids can neither be explained in
terms of property change nor of changes in Nu and Reb due to change in characteristic length
(diameter). Particularly interesting to note are the results at low Reb . The deterioration in boiling
seems to be reducing between 20 and 6.5 mm but it reappears rather strongly at the lowest di-
ameter of 4 mm. This is different from that at higher Reb where the deterioration is rather strong at
larger diameter. This effect can be explained by noting that for given tube diameter decrease in Reb
indicates decrease in heat flux and a consequent decrease in nucleation site density. With the
decrease in nucleation sites the probability of plugging of these sites by nano-sized particles also
decreases resulting in a decrease in the boiling deterioration. This happens at lower Reb for lower
diameter because at lower diameter the same Reb corresponds to a higher heat flux giving a higher
nucleation site density. At the intermediate diameter of 6.5 mm this effect is somewhat smaller
because of the dominance of curvature effect which is known to give highest heat transfer coef-
ficient near 8 mm diameter and decreases both above and below that diameter (Cornwell, 1990).
In general, the change in Nu–Reb correlations is more drastic at higher Reb for large diameter tubes
than for narrow tubes. This indicates the danger of local overheating is smaller on narrow tubes
compared to larger tubes for high heat flux applications when boiling point is reached for a nano-
fluid. These observations are important from the application points of view. Nano-fluids are
generating interest in heat transfer community for cooling of high heat flux applications. How-
ever, these fluids perform poorly when boiling occurs. The present study shows that the degra-
dation in boiling is a strong function of tube diameter for narrow tubes.

5. Conclusion

The use of nano-fluids for cooling of high heat flux devices in modern electronics, computing
and optical technology has been claimed to be a new possibility due to their enhanced thermal
conductivity and capability of further enhancing convective process through particle dispersion.
They have been found to be much improved compared to common slurries with respect to sedi-
mentation, clogging and pressure drop. However under phase change conditions this possibility
has been conclusively negated due to the fact that the presence of nano-particles deteriorates
boiling performance with increase in particles concentration. In the present study it has been
shown that the nature of this deterioration is different for the narrow tube regime compared to the
large tube regime. From the application point of view, the present study shows that for convective
cooling with nano-fluids, the extent of degradation in heat transfer is less for narrow tubes
compared to large diameter tubes particularly at high heat flux if local boiling limit is reached.
This makes nano-fluids more compatible to cooling narrow components with high heat flux. The
 S.K. Das et al. / International Journal of Multiphase Flow 29 (2003) 1237–1247 1247

results can also be used for specific applications such as heat treatment of small components where
a higher temperature of heated surface can be maintained and can be controlled by changing
particle concentration or particle size.

Acknowledgement

This research work has been carried out during the stay of the first author in Hamburg under
the Humboldt Research Fellowship which is gratefully acknowledged.

References

Ahuja, A.S., 1975. Augmentation of heat transport in laminar flow of polystyrene suspension experiments and results.
J. Applied Physics 46, 3408–3416.
Choi, U.S., 1995. In: Siginer, D.A., Wang, H.P., (Eds.), Enhancing thermal conductivity of fluids with nano-particles,
developments and applications of non-Newtonian glows, FED-vol. 231/MD-vol. 66. ASME, New York, pp. 99–105.
Chun, M., Kang, M., 1998. Effects of heat exchanger tube parameters on nucleate pool boiling heat transfer. J. Heat
Transfer, Trans. ASME 120, 468–476.
Cornwell, K., 1990. The role of sliding bubbles in boiling on tube bundles. In: Proceedings of Int. Heat Transfer
Conference, Jerusalem, pp. 455–460.
Cornwell, K., Houston, S.D., 1994. Nucleate pool boiling on horizontal tubes: a convection-based correlation. Int. J.
Heat and Mass Transfer 37 (suppl. 1), 303–309.
Cornwell, K., Sch€uller, R.B., 1982. A study of boiling outside a tube bundle using high speed photography. Int. J. Heat
and Mass Transfer 25, 683–690.
Das, S.K., Putra, N., Thiesen, P., Roetzel, W., 2001. Temperature dependence of thermal conductivity enhancement for
nano-fluids, J. Heat Transfer, Transactions of the ASME, in press.
Das, S.K., Putra, N., Roetzel, W., 2003. Pool boiling characteristics of nano–fluids. Int. J. Heat and Mass Transfer 46,
851–862.
Eastman, J.A., Choi, U.S., Li, S., Yu, W., Thompson, L.J., 2001. Anomalously increased effective thermal
conductivities of ethylene glycol-based nano-fluids containing copper nano-particles. Applied Physics Letter 78,
718–720.
Gorenflo, D., 1997. Beh€altersieden, section Hab, VDI-W€ armeatlas, eighth ed. VDI Verlag GmbH, D€ usseldorf.
Hamilton, R.L., Crosser, O.K., 1962. Thermal conductivity of heterogeneous two component systems. I & EC
Fundamentals 1, 187–191.
Lee, S., Choi, U.S., Li, S., Eastman, J.A., 1999. Measuring thermal conductivity of fluids containing oxide nano-
particles. ASME J. Heat Transfer 121, 280–289.
Liu, K.V., Choi, U.S., Kasza, K.E., 1988. Measurement of pressure drop and heat transfer in turbulent pipe flows of
particulate slurries, Argonne National Laboratory Report, ANL-88-15.
Xuan, Y., Roetzel, W., 2000. Conceptions for heat correlation of nano-fluids. Int. J. Heat and Mass Transfer 43, 3701–
3707.