You are on page 1of 11

Acta Astronautica 61 (2007) 16 – 26

www.elsevier.com/locate/actaastro

Heat pipes with binary mixtures and inverse Marangoni effects for
microgravity applications
Raffaele Savinoa,∗ , Nicola di Francescantonioa , Raimondo Fortezzab , Yoshiyuki Abec
a Dipartimento di Ingegneria Aerospaziale, Sezione Spazio Luigi G. Napolitano, Universita’ di Napoli Federico II, Napoli, Italy
b MARS Center, Napoli, Italy
c AIST—National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan

Available online 6 March 2007

Abstract
Numerical and experimental investigations have been carried out to understand the role of Marangoni effects in heat pipes
with aqueous solutions of long-chain alcohols, proposed to replace common pure liquids used in conventional heat pipes.
Laboratory measurements of surface tensions have been carried out for a number of binary solutions at different temperatures and
concentrations, to identify the most promising two-component mixtures with a nonlinear dependence of the surface tension with
temperature. Fundamental studies on Marangoni flows in binary mixtures, for different experimental configurations of interest,
show the potential advantages of these new working fluids.
A number of tests have been performed with grooved heat pipes, filled with aqueous solutions with the appropriate alcohol
concentrations. The results show improved performances of these heat pipes, when compared to conventional commercial water
heat pipes.
© 2007 Elsevier Ltd. All rights reserved.

1. Introduction different sides of the spacecraft due to solar radiation or


internal dissipation of electronic devices and surfaces
Capillary-based devices are important for several ap- oriented toward the deep space. It implies that capil-
plications. Capillary flows in micro-pipes can substitute lary motion could be intrinsically driven by existing
electromagnetic pumps in cooling loops and radiators. conditions without the use of specific devices power
Pumps are subject to failures; they are heavy and re- consuming.
quire electrical power to work. Capillary flow is natural The second class of application is represented by
and does not need any power to be activated. Fluid- evaporation-based heat transfer devices (e.g. heat pipes)
dependent pipe geometry (for capillary flow) or tem- that can be successfully applied to control the temper-
perature difference between the two ends of the pipe ature of spacecrafts components, satellites, electronic
(thermocapillary flow) is the only requirement to en- devices, propulsion, energy recovery systems and other
sure the flow of the fluid inside sealed pipes. Very often instruments. After the pioneering publication by Grover
in space applications, temperature difference exists in and Cotter [1] research on heat pipes spread worldwide
as they offer many advantages because of their low-
weight penalty, zero maintenance, and reliability [2–5].
∗ Corresponding author. Fax: +39 81 5932044. By using latent heat of vaporization of the fluid to
E-mail address: rasavino@unina.it (R. Savino). transfer heat efficiently at a nearly constant temperature,
0094-5765/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.actaastro.2007.01.002
R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26 17

one can transfer several hundred times the amount of The influence of surface tension-driven flows unbal-
heat energy compared to solid copper conductive device ance in heat transfer devices has received attention in
for a given temperature difference. the literature [6–12]. In general, for all the pure work-
It is recognized that one of the most critical per- ing fluids used in conventional heat pipes, the surface
formance limitations of these passive thermal devices tension is a decreasing function of the temperature.
is the pumping capability (i.e. providing enough liq- Therefore, surface motions due to a surface tempera-
uid return to the evaporator to ensure continuous flow) ture gradient are directed toward the cold regions of
under microgravity conditions. Although heat pipe per- the surface, which may be unfavorable for the return
formances and operations are strongly dependent on the of the liquid to the evaporator. Eninger and Marcus
shape, working fluid, and wick structure, the surface [6] proposed the Marangoni effect as the primary
tension and wetting angle of the fluid play an extremely mechanism responsible for the observed degradation in
important role in determining the amount of heat that axially grooved heat pipes with non-condensible gases.
can be transferred by a given heat pipe. The main speculation was that the temperature-induced
In normal gravity conditions, for a grooved heat pipe, surface tension gradient in the condenser region could
the liquid remains in the condenser due to a stratification drive a recirculatory flow with consequent pressure drop
effect. In this case, the upper grooves are not sufficiently that negatively affects the capillary pressure available
fed and heating on the upper side of the evaporator is to pump the condensate return at the evaporator. The
not efficient. authors proposed a very simple idealized model, based
In gravity-assisted devices heat added to the bottom on the assumption of one-dimensional flow and pure
region vaporizes the working fluid. Because the vapor liquid that proved to correctly predict the heat transfer
in the evaporator region is at a higher temperature and reduction observed during experimental measurements.
hence at a higher pressure the vapor rises and flows Although the basic aspects of the Marangoni effect
to the cold condenser where it releases the latent heat are clear, its influence in axially grooved heat pipes
of vaporization (buoyancy forces assist this process). still needs for a more realistic mathematical model
Gravitational forces then cause the condensate film to and corresponding experimental research. Preliminary
flow back down along the heat pipe wall to the evapora- microgravity experiments were performed in a drop
tor where it can again be vaporized. Although the inner tower in Japan [7], using a water–ethanol solution as
surface of a thermosyphon may occasionally be lined working fluid in place of distilled water. The results
with grooves or with a porous structure to promote re- pointed out that the condensed liquid in the condenser
turn of the condensate to the evaporator thermosyphons section is turned back more to the heating section
principally rely upon the local gravitational acceleration when the ethanol concentration is relatively low. These
for the return of the liquid from the evaporator to the results could be explained by the Marangoni effect
condenser. induced by concentration gradients. Experiments with
On the contrary, heat pipes utilize the capillary binary mixtures as working fluids have been also con-
wicking structure to promote the liquid flow from the sidered by several authors to improve thermo-syphon
condenser to the evaporator and as a result they can operations. For a summary of the mixture boiling
be used in a horizontal orientation, in microgravity work done prior to 1990, refer to the chapter “En-
environments, or even for applications where the capil- hanced Boiling of Mixtures” in the book by Thome [8]
lary structure is able to “pump” the liquid against grav- where the results for boiling of various water–alcohol
ity from the evaporator to the condenser. To improve mixtures are reported. Some experimental results for
the heat transfer, one would like to identify additional contact angles for methanol–water mixtures are also
mechanisms other than capillary and gravitational able provided there. Mc Gillis et al. [9] published results
to pump the liquid phase from the cold to the hot side of experimental research on boiling and condensing
of the pipe. water–alcohols binary mixtures in thermo-syphons at
sub-atmospheric pressures.
2. Role of surface tension in heat pipes
Pratt and Hallinan [10] investigated the role of sur-
For binary mixtures the surface tension is a func- face tension gradients in heat transfer devices. They
tion of the temperature and of the concentration. In performed a very preliminary study in a simple config-
this case, two additional forces can be identified at the uration consisting in a meniscus laterally heated in a
film–vapor interface: thermal Marangoni; concentration cylindrical capillary, showing that surface tension driven
Marangoni. Both these effects can oppose or favor the flows can enhance the heat transfer effectiveness of de-
film motion towards the hot end. vices that rely on evaporative phase change. Pratt and
18 R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26

Kihm [11] investigated a binary mixture of pentane and concentrations. Identification of the most promising
decane and demonstrated that the addition of about 1.5% two-component mixture.
decane significantly enhanced the heat transport com- (b) Fundamental studies on Marangoni flows in binary
pared to pure pentane. This procedure also increased mixtures for different experimental configurations
the stability of the interfacial film producing a more of interest (non-isothermal liquid films in closed
steady heat transfer performance. The increased effi- cuvettes). The results show potential advantages of
ciency could be explained by the concentration gra- a number of working fluids.
dients associated with the distillation in binary fluid (c) Characterization of operation and performances of
mixtures that counteract the thermocapillary stresses heat pipes under normal gravity conditions in prepa-
arising from the liquid-vapor interfacial temperature ration of microgravity experiments
gradient.
Zhang and Chao [12] performed a theoretical and 3. Two component mixtures and inverse
experimental investigation to show that boiling heat Marangoni flow
transfer can be enhanced in binary mixtures, taking ad-
vantage from both surface concentration and tempera- As indicated in Fig. 1, the liquid mass flow rate be-
ture gradients. They presented a model predicting the tween the condenser and the evaporator can be strongly
bubble detachment diameter, the nucleation boiling heat enhanced in microgravity if a surface tension gradient
transfer coefficient and the critical heat flux taking into can be induced on the liquid film surface and if the max-
account the thermal Marangoni effect. However, solutal imum value of  takes place in correspondence of the
Marangoni effect was not included in their analysis. evaporator. Marangoni effect will then drive the liquid
The present authors have recently proposed to use towards the evaporator increasing the pumping effect
heat pipes with dilute aqueous solutions of higher and therefore the liquid mass flow rate inside the heat
carbon (number of carbons4) that exhibit a surface pipe. This is an additional effect that helps the basic heat
tension increases with the temperature in suitable tem- pipe liquid flow due to the capillary action toward the
perature [13–15]. In dilute compositions alcohol is the evaporator. The substances that have been examined are
more volatile component and the Marangoni flow due water mixtures with small additions of long chain alco-
to the concentration gradient, coupled with the thermo- hols. The surface tension derivative (j/jx)z=0 along
capillary flow is considered to induce rather strong x (the heat pipe axis) can be expressed in terms of the
liquid inflow from the cold to the hot region. One of temperature and concentration derivatives (T , c ) of
the most promising applications is, therefore, the use the surface tension ():
of these working fluids in wickless heat pipes under j jT jc
reduced gravity conditions. In this work, experimental = T + c , (1)
jx jx jx
and theoretical studies have been carried out to select
suitable two-component working fluids as best candi- where c is the concentration of the solute dissolved in
dates to improve heat pipe transfer capabilities, due water.
to Marangoni effects (caused by concentration and/or To appreciate the driving action of the surface tension
temperature differences). unbalance one may write at the liquid surface the bal-
Research activities have been carried out including: ance between shear stress and surface tension gradient:

(a) Measurements of surface tensions for a number jV j


 = , (2)
of binary solutions at different temperatures and jz jx

Evaporator Adiabatic Vacuum


q
∂σ
q Evaporation Condensati o Vacuum
∂x
Hot

Cold

+
q Liquid
MARANGONI EFFECT

wick
structure

Fig. 1. Schematics of the basic idea for the enhancement of heat pipes performances by Marangoni.
R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26 19

where  is the liquid viscosity, from which an order of concentration, i.e. c (T , c) < 0. Therefore, it is expected
magnitude of the velocity may be evaluated: to observe surface flows directed from the cold region
  to the hot side, for temperatures higher than that of the
 T c minimum of the surface tension.
V
T + c , (3)
 L L The measurements have been performed using the
Dataphysics OCA 15 tensiometer (Fig. 2), based on the
 being the (average) film thickness, L the channel (heat pendant drop method (see e.g. [17]). The experimen-
pipe) length; T = Te − Tc > 0 (and c = ce − cc ) is tal apparatus includes an experimental cell, an optical
the difference between temperature (and concentration) system for the illumination and the visualization of the
at the evaporator and at the condenser. drop shape and a data acquisition system. Using an elec-
The proposed new working liquid should be charac- tronic injection system, a pendant drop is formed, whose
terized by T > 0. contour is analyzed to infer the surface tension using a
The value of c c should also be greater than zero suitable software. The whole process of analysis of the
(to be concurrent with the thermal Marangoni effect). drop shape consists of four main steps: (1) the image
In this case we expect to observe motions going from of the pendant drop is captured and digitalized using a
a cold region to a hot region for temperatures higher computer with a frame grabber; (2) the drop contour is
than that of the minimum of the equilibrium surface extracted to evaluate the radius of curvature at the apex
tension. necessary for the calculation of the surface tension; (3)
In addition, any imposed temperature difference smoothing of the extracted contour is carried out using
creates a surface concentration gradient if evaporation polynomial regression; (4) the experimental drop shape
occurs at the hot region and condensation in the cold is compared with theoretical results obtained solving the
one. Bashforth and Adams equation [18] by a fourth order
Runge–Kutta method.
4. Surface tension measurements
A thermocouple was used to control the temperature
Laboratory measurements have been performed with with an accuracy of 0.1–0.2 ◦ C.
a pendant drop apparatus to evaluate the values of The surface tension measurements are summarized in
T and c for a number of long-chain alcohol solu- Fig. 3. For all the investigated alcohol solutions, when
tions, i.e. water/butanol, water/pentanol, water/exanol, the alcohol concentration is very low, the surface ten-
water/heptanol, water/octanol. sion decreases when the temperature increases; increas-
According to previous works [16], for these alco- ing the alcohol content, it is possible to find suitable
hol solutions, in some ranges of concentrations, the concentrations such that a positive surface tension gra-
surface tension as a function of the temperature goes dient with the temperature is found when the tempera-
through a minimum and there is a range of temperature ture exceeds a certain value. Typically, the temperature
above which T (T , c) > 0. Typically, at a fixed temper- corresponding to the minimum of the surface tension is
ature, the surface tension is a decreasing function of the an increasing function of the number of Carbon atoms.

Fig. 2. Tensiometer (left) and corresponding drop sample image (right).


20 R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26

systems were investigated in both the horizontal and


vertical configurations without additional cooling units
(the only mechanism for heat transfer at the cold sides
of the pipes was radiation and natural convection).
The diagnostic system includes an infrared ther-
mocamera FLIR SC-3000 and thermocouples with
diameter of 0.3 mm, to measure the temperature at the
evaporator sections.
Fig. 4a shows that, in the vertical configuration (grav-
ity assisted), when the power supplied to the heater is
5 W, the heat pipes filled with water and with the binary
mixture exhibit the same heat transfer performances.
This result can be explained because in this config-
Fig. 3. Surface tensions of alcohol solutions as a function of the uration the gravity force prevails over surface tension
temperature. forces and represents the main “driving” action. On
the contrary, in the horizontal configuration, the perfor-
5. Fundamental studies on wickless heat pipes mances of the pipe filled with the binary mixture are
better, as qualitatively shown in Fig. 4b.
Preliminary experiments have been carried out us- The heat pipe filled with a binary mixture shows a
ing horizontal glass cuvettes partially filled with a layer temperature difference between the two opposite sides
of water or with alcohol solutions, to study the liquid of about T
50 ◦ C compared with a T
90 ◦ C of the
behavior under a prescribed temperature difference at heat pipe filled with pure water. The equivalent thermal
the cell sides. The surface temperature distribution has conductivity of the first heat pipe is almost twice than
been detected with an infrared thermocamera. A CCD the heat pipe containing pure water. In order to better
camera has been utilized to observe the motion of par- investigate the Marangoni effect, the temperature distri-
ticle tracers at the layer interface. Experimentally, the bution along the outer surfaces of glass heat pipes was
average surface velocities in the binary mixtures with detected looking at the horizontal systems from below
thickness of few millimetres are of the order of 1 cm/s and from above, as shown in Fig. 5.
in a horizontal cuvette of length 4 cm, in presence of In the heat pipe filled with pure water, the isotherms
T
20 ◦ C, while for pure water the velocities are in are distorted towards the condenser in the top view;
the order of 1 mm/s at the same conditions. According to and towards the evaporator, in the bottom view. On the
Eq. (3), these findings witness the presence of a strong contrary, in the binary solution, the distortions are in
Marangoni effects in the favourable directions. the same directions condenser in both cases.
It is, however, difficult to appreciate the sign of c A numerical model has been prepared that takes into
(i.e. if at the film surface the concentration of alcohol account the complex interplay between liquid, solid,
in correspondence of the evaporator is less than that vapor in presence of heating, phase changes, viscous
corresponding to the condenser); some of the experi- flow, capillary effects and Marangoni flows. A mass flow
mental tests and preliminary theoretical models suggest rate (2.2 × 10−6 kg/s) was imposed at the evaporator
that the alcohol is the more volatile component and that and condenser sections, based on the measured power
c < 0 so that also c c > 0 and the thermal and solu- supplied to the heater (5 W) and on the latent heat of
tal Marangoni flows are concurrent. vaporization of water (540 cal/g). According to the ex-
The substances experimentally investigated in the cu- periments the temperatures imposed at the end walls are
vette have been also utilized to fill enter vacuum glass 90 ◦ C in the heater zone (3.5 cm length) and 30 ◦ C at the
pipes with inner diameter of 10 mm, wall thickness of condenser. Heat conduction across the tube thickness
1 mm and length of 165 mm, to show their behavior in has been taken into account in the numerical model.
comparison with pure water. Finally, a surface tension driven flow at liquid–vapor
In particular, one pipe was filled with bi-distilled interface has been imposed according to the measured
water and the other one with a binary solution with values of surface tension.
proper concentration. A copper wire was wrapped The computed temperature distributions on the
around the outer surface of each tube for a length of outer surface of the pipes are in qualitative agree-
3.5 cm, and connected to a power supply in order to ment with the experimental profiles obtained with the
establish a temperature gradient along the pipe. The thermocamera.
R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26 21

Fig. 4. Thermographic images of pipes filled with water and binary mixture: (a) vertical configuration; (b) horizontal configuration.

Fig. 5. Thermographic images of pipes filled with water and binary mixture in the horizontal configuration. (a) Top view; (b) bottom view.

Fig. 6. Numerical simulation of the piple filled with water in the horizontal configuration.

In the case of the pipe filled with water (Fig. 6), By continuity, the liquid returns toward the
the velocity vectors in the mid-section perpendicular to evaporator in the bottom regime below the in-
liquid–vapor interface show that in the pipe filled with terface, and this explains why the isotherms are
pure water, the vapor flow entrains a fluid flow at the deformed toward the evaporator zone in the top
interface directed from the evaporator to the condenser. view.
22 R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26

Fig. 7. Numerical simulation of the pipe filled with a binary mixture in the horizontal configuration.

For the case of the pipe filled with the binary mixture conductivity material. The only heat transfer mecha-
(Fig. 7), the computations show that there is a strong nisms, from the condenser to the ambient air, are radi-
liquid flow directed from the cold side towards the hot ation and natural convection.
region (i.e. an inverse Marangoni flow). On the other The temperature distributions along the heat pipes
hand, the liquid returns to the cold region along the side and on the surface of the aluminium panels are detected
walls, causing an evident distortion of the isotherms as by thin thermocouples (0.25 mm diameter) and by the
shown by the thermographic images and by the numer- infrared thermocamera. A Keithley data logger is used
ical results. for the acquisition of the temperature signals, as well
In both cases, the vapor flow from the hot side to the as of the current and voltage.
cold side explains why the isotherms on the top wall of During each test a prescribed electrical power is sup-
the pipes are directed from the evaporator toward the plied to the cartridge heaters; when temperature steady-
condenser zone. state conditions are established, the electrical power is
increased to the next step and this procedure is repeated
6. Heat pipes performances until dry-out conditions are achieved, with a sudden in-
crease of the temperature difference between evaporator
A number of heat transfer performances tests have and condenser.
been carried out to optimize the composition of the The experiments pointed out that heat pipes filled
working fluid using a set of commercial heat pipes, filled with alcohol solutions exhibit better heat transfer
with different aqueous solutions. performances than the ones filled with pure water.
The experiments have been performed with axi- Thermographic images of the experiments, for the ver-
ally grooved copper heat pipes, with 25 cm length tical (gravity assisted) and horizontal configurations,
and diameters of 4 and 8 mm. The heat pipes were and electrical power of 4 W, are shown in Fig. 9. In the
filled with 0.41 cc (or 1.9 cc) of pure water and with horizontal configuration (Fig. 9a), for the same electri-
the same amount of water solutions of long chain cal power, the temperature of the heat pipe filled with
alcohols. the water/alcohol solution is lower at the evaporator
Fig. 8 shows an artistic layout and the pictures section and higher at the condenser section.
of the different sub-components of the experimental The behavior of the two heat pipes, in the grav-
setup. ity assisted configuration, are almost the same
The evaporator is a copper block of 30 mm. Cartridge (Fig. 9b).
heaters are introduced inside holes and are heated by a On the contrary, as shown in Fig. 9a, the thermal re-
power supply. The heat pipes condensers are embedded, sistance of the heat pipe filled with water, defined by the
for a length of 6 cm, into thin black painted aluminium ratio between the temperature difference at the evapo-
panels (8 cm × 6 cm). The heat pipes are mounted in rator and condenser sections and the supplied power, is
the horizontal configuration and the adiabatic region, relatively large compared to the case of the heat pipe
of about 16 cm, is properly insulated by a low thermal filled with the binary mixture. The behavior found in
R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26 23

Fig. 8. Experimental setup. (a) Artistic layout; (b) evaporator block; (c) evaporator block with cartridge heatres; (d) heat pipe-evaporator
assembly; and (e) top view of two assembled heat pipes with black painted radiator.

Fig. 9. Thermographic images of the heat pipes assemblies for Q = 4 W. (a) Horizontal configuration; (b) vertical configuration. Maximum
temperature: 52 ◦ C, minimum temperature 25 ◦ C.
24 R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26

Fig. 10. Time history of the temperatures during heat pipes tests with the horizontal (a) and vertical (b) configurations.

Figs. 9a, b is qualitatively the same already discussed All these results confirm that heat pipes filled with
in the case of glass tubes filled with the same liquids alcohols solutions with unusual Marangoni effect ex-
(Figs. 4a, b). hibit better performances when compared to conven-
The time histories of the temperatures at the evapo- tional commercial water heat pipes.
rator and condenser sections are shown in Fig. 10. For
relatively low electrical power (1 or 2 W), the heat pipe 7. Microgravity experimentation
filled with pure water exhibit a relatively low thermal re-
sistance. However, increasing the power, dry-out occurs The experiments proposed for the ESA parabolic
in the water heat pipe at smaller electrical powers (4 W) flight campaign onboard the Airbus A-300 include tests
compared to the heat pipe filled with the water/alcohol on heat pipes filled with different fluids and tests on in-
solution (8–10 W). novative flexible/inflatable/deployable radiators based
R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26 25

Fig. 13. Sketch of the heat pipes setup.


Fig. 11. Setup of radiator panel.

suitable material transparent in the infrared band of the


instrument (BaF2) will be inserted in the wall of the
chamber (see Fig. 12). Heating will be provided by re-
sistant heaters with power of hundreds of Watts, to en-
sure a quick heating process compatible with the short
time available during the parabola.
As a cooling device a water jacket in combination
with Peltier element will be used for the same reason.
The diagnostic system will include a CCD camera, an
IR camera, thermocouples, power measurements units
and a data acquisition system with a PC.
Ground experiments will be carried out to investigate
the liquid behavior and the temperature distributions for
horizontal and vertical orientation (thermosiphon).
Observation of the working fluid behavior in the
Fig. 12. Vacuum chamber.
polyamide channel in the panel will be carried out in
combination with measurements of the temperature
on the same fluids. Concerning the first part of the distribution of the panel and fluid temperature using
experimentation a number of commercial or wickless both the IR camera and sheet-thermocouples.
heat pipes will be filled with aqueous solutions with The working fluid behavior will be related to the
the appropriate alcohol concentrations and tested with “negative” temperature dependence of the surface ten-
a simple apparatus in order to compare their perfor- sion in a low-temperature region.
mances with the corresponding performances under Concerning the tests on heat pipes filled with dif-
normal gravity conditions. ferent fluids, each day the behavior of heat pipes,
In addition, heat pipe radiator panels, fabricated by properly filled with different liquid mixtures, will be in-
the vacuum hot-press procedure, will be subjected to vestigated. An artistic layout of the experimental setup
a series of thermal performance tests in the reduced with four heat pipes is shown in Fig. 13, where it is pos-
gravity condition. sible to see the adiabatic region of the heat pipes, the
The panels material is a Polyamide multi-layer lam- black condensers (cooled by circulating water) and the
inate sheet. It is manufactured with many tubes filled metal blocks with the wires connected to the cartridge
with suitable liquids (i.e. different alcohol aqueous), see heaters. The experimental procedure will be optimized
e.g. Fig. 11. The experiments will be carried out to study in such a way that different electrical powers will be
the distribution of the working fluid in the polyamide provided to the heat pipes heaters before the parabolas,
channels in the panel and the temperature distribution to measure the evolution of the axial temperature dis-
of the panel under different heating conditions. tributions along the pipes by fine thermocouples during
A transparent vacuum chamber (Fig. 12) will be used the microgravity period. The main objective will be to
to reduce heat transfer to the surrounding ambient. compare 1-g and micro-g performances of the selected
Since the surface temperature distribution will be de- heat pipes and the improvements obtained with binary
tected by an infrared thermocamera, windows with a mixtures.
26 R. Savino et al. / Acta Astronautica 61 (2007) 16 – 26

8. Conclusions [5] G.P. Peterson, An Introduction to Heat Pipes. Modeling Testing


and Applications, Wiley, New York, 1994.
The experimental and numerical results presented in [6] J.E. Eninger, D.B. Marcus, The Marangoni effect and capacity
degradation in axially grooved heat pipes, Third International
this paper show that suitable two-component working Heat Pipe Conference, Palo Alto, CA, May 22–24, 1978, AIAA
fluids can be identified to improve heat pipe transfer Technical PaperA78-35576, pp. 14–34.
performances. The inverse Marangoni effect, caused by [7] M. Kuramae, M. Suzuki, Two components heat pipes utilizing
a surface tension increasing with temperature, provides the Marangoni effect, Chemical Engineering of Japan 26 (1993)
an additional mechanism for liquid return from the con- 230–231.
[8] J.R. Thome, Enhanced Boiling Heat Transfer, Hemisphere
denser to the evaporator, other than forces like capil- Publishing Co., New York, 1990.
lary and gravitational. The numerical results show that [9] Mc Gillis, et al., Western research laboratory Technical Note
this particular surface tension behavior, induced by the TN-23, 1992.
evaporation at liquid–vapor interface, is associated to [10] D.M. Pratt, K.P. Hallinan, Thermocapillary effects on the
an increase of the heat transfer. The experimental re- wetting characteristics of a heated curved meniscus, Journal of
Thermophysics and Heat Transfer 11 (1997) 519–525.
sults confirm that heat pipes filled with long-chain alco- [11] D.M. Pratt, K.D. Kihm, Binary fluid mixture and thermo-
hol solutions at suitable concentrations perform better capillary effects on the wetting characteristics of a heated curved
than heat pipes filled with pure water. This was found meniscus, Journal of Heat Transfer 125 (1997) 867–874.
both in wickless glass pipes and in conventional copper [12] N. Zhang, D. Chao, Models for enhanced boiling heat transfer
grooved heat pipes. by unusual Marangoni effects under microgravity conditions,
International Communications in Heat Mass Transfer 20 (8)
Future experiments will be carried during parabolic (1999) 1081–1090.
flights and on board an Italian satellite as a microgravity [13] R. Savino, R. Monti, Heat pipes for space applications, Space
platform to assess the performances of the newly pro- Technology 25 (1) (2005).
posed heat pipes and of innovative wickless heat pipes [14] Y. Abe, About self-rewetting fluids, Possibility as a new working
in microgravity conditions. fluid, Thermal Science and Engineering 8 (4) (2003).
[15] Y. Abe, K. Tanaka, Heat transfer devices with self rewetting
fluids, in: Proceedings of IMECE 2004, Aaheim, November
References 2004.
[16] R. Vochten, G. Petre’, Journal of Colloid and Interface Science
[1] G.M. Grover, T.P. Cotter, G.F. Erickson, Structures of very high 42 (1973) 320.
thermal conductance, Journal of Applied Physics 35 (6) (1964) [17] E.Y. Arashiro, N.R. Demarquette, Use of the Pendant drop
1990–1991. method to measure interfacial tension between molten polymers,
[2] T.P. Cotter, Principles and prospects for micro heat pipes, in: Materials Research 2 (1) (1999) 23–32.
Proceedings of the Fifth International Heat Pipes Conference, [18] S. Bashforth, J.C. Addams, An attempt to test the theory of
Tsukuba, Japan, 1984, pp. 328–335. capillary action, Cambridge University Press and Deighten, Bell
[3] A. Faghri, Heat pipe Science and Technology, Taylor and and Co, London, 1882.
Francis, Washington, 1995.
[4] P.D. Dunn, D.A. Reay, Heat Pipes, Pergamon Press, Oxford,
1994.

You might also like