JOURNAL OF THERMOPHYSICS AND HEAT TRANSFER

Vol. 8, No. 2, April-June 1994

Thermodynamic Aspects of Heat Pipe Operation

Robert Richter*
California Institute of Technology, Pasadena, California 91104
and
Joseph M. Gottschlicht
Wright Laboratory, Wright-Patterson Air Force Base, Ohio 45433

In this article the general operation and performance of heat pipes is being approached from fundamental
thermodynamic considerations. This is in contrast to the classic heat pipe design analysis which equates an
available "capillary pressure" with the two pressure drops associated with the circulation of the working fluid
in the vapor and liquid phase and body forces where applicable, and erroneously attributing the circulation of
the working fluid in a heat pipe to "capillary pumping." This article shows that the working fluid circulates
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

in a heat pipe as the result of a thermodynamic cycle in which thermal energy is converted to kinetic energy.
The basic analysis which is presented, and which can be extended in future research, identifies the total internal
temperature difference over which the heat pipe operates and the heat pipe operating temperature as the two
key operating parameters in full agreement with all observed heat transfer phenomena. For given physical
characteristics of a heat pipe, i.e., length, internal and external diameters, wick, artery or groove structure,
the thermal transport performance below its operating limits is shown to be entirely a function of these two
parameters as it is for any other heat conductor. The thermodynamic cycle clarifies the often substantial
discrepancy between the predicted and the actually realized thermal power transfer capacity of a heat pipe.

Nomenclature that were provided by outside contractors performed short of

C(l = configuration constant, m2 their design specifications. Some devices actually underper-
Cr = constant formed by an order of magnitude. For many years an intensive
Ev = vaporization energy, J/mol and systematic research program was conducted on a double
k(T) = heat transfer coefficient, W/m2-K wall artery heat pipe, which was an in-house design. The
m = flow rate, kg/s operation of this heat pipe generated some of the most com-
P = thermal transfer power, W plete and valuable data which finally shed new light on the
ph = body forces on the fluid, N/m3 actual performance parameters of heat pipes. This copper/
pc = available capillary pressure, N/m2 water heat pipe demonstrated a thermal transport capacity of
pv2 = vapor pressure at maximum internal operating 2.8 kW-m at an operating temperature of about 210°C, al-
pressure of the heat pipe, T2, as found from though its theoretical capacity as based on the standard heat
Eq. (9), N/m2 pipe design theory should have reached a performance of 6.0
Q = total heat tansfer, W kW-m at 100°C (Ref. 3). However, the experimental data
R = gas constant, 8.3144 J/mole-K showed that in every respect the heat transfer of this piece of
rc = effective pore radius, m material followed the laws of heat transfer of all other thermal
T = temperature, K conductors, i.e.
T2 = internal evaporator temperature, K
Q = Cdk(T)±T (1)
A///g = latent heat of vaporization, J/kg
Ap7/ = liquid pressure drop, N/m2 A casual review of the heat pipe literature reveals that large
A/?r/ = vapor pressure drop, N/m2 discrepancies between predicted performances based on the
AT1 = temperature difference, K capillary pumping design analysis and experimental perfor-
a = surface tension, N/m mances are no rare occurrences. Even the "monogroove" heat
</> = attachment angle, deg pipe as shown in Fig. 1 apparently carries only about 50% of
its theoretically predicted heat load.4 Other designs, such as
Introduction the "cat's eye" artery concept (Fig. 1), failed to match the-
T HE actual performance of heat pipes of many varied
designs has failed to match the expected performance
predictions which were based on the currently accepted heat
oretical predictions by over an order of magnitude. It was
therefore necessary to arbitrarily multiply theoretically pre-
dicted thermal transport capacities by a factor that is consid-
pipe design relations as compiled in Refs. 1 and 2 and many erably less than unity when sizing heat pipes for spacecraft
other publications. For a considerable number of years heat applications.5 A study of the initial failure to remove thermal
pipes have been tested at Wright Laboratory. Many heat pipes waste heat by the 0-g Space Station heat pipe advanced ra-
diator element (SHARE)6 during a 1990 space flight exper-
iment, and the subsequent modifications in the heat pipe de-
Presented as Paper 90-1772 at the AIAA/ASME 5th Joint Ther- sign, as well as in the 0-g operation,7 can illustrate clearly that
mophysics and Heat Transfer Conference, Seattle, WA, June 18-20, currently applied heat pipe design techniques do not correctly
1990; received Oct. 18,1991; revision received Aug. 3,1992; accepted
for publication July 28, 1993. This paper is declared a work of the
predict either the performance or the operation of these high
U.S. Government and is not subject to copyright protection in the technology arterial heat pipes.
United States. The theory of heat pipe design as presented in Refs. 1 and
*Member of the Technical Staff, Jet Propulsion Laboratory, 4800 2, and many other publications, bases the design of a heat
Oak Drive, M/S 157-102. pipe principally on the stipulation that the combined pressure
tProject Engineer, WL/POOS-3. drops of the vapor and the liquid while circulating in the heat
334
 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION 335

x Entrainment Limit
D
Wicking or
MONOGROOVE CAT'S EYE DOUBLE WALL .Capillary Limit

12,900 W-m dem 400 W-m dem 2.8kW-m dem 0

Q) Boiling Limit
30,000 W-m pred 15,000 W-m pred 6.2 kW-m pred
AMMONIA AMMONIA WATER I Sonic Velocity
Limit
GRUMMAN VOUGHT WRIGHT LABORATORY

Fig. 1 Heat pipe designs. Viscous Limit

LIQUID SATURATED WALL WICK

Temperature
Fig. 3 Operating limits of heat pipes.

V) 2500
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

-t-*
O

- 2000
Theory

O
Q_

1000

LOCATION
Fig. 2 Generic heat pipe operation. 500

pipe and the pressure associated with body forces are equal
to the available capillary pressure which is produced in a 0 50 100 150 200 250 300 350
porous structure. The operating temperature enters into the o
analysis only for the determination of the temperature-de- Temperature, C
pendent property values of the working fluid. No consider- Fig. 4 Experimentally determined heat transport capacity of a heat
ation is given to temperature differences at all. The circulation pipe.
of the working fluid is attributed to capillary pumping. Special
variations of heat pipes are actually called "capillary pumped The thermal transport limit of a heat pipe is manifested by
loops." the "dry-out" of the evaporator. Dry-out of the evaporator
Thus, the fundamental heat pipe design relation is ex- is the final, but not necessarily the immediate, manifestation
pressed by of working fluid returning to the evaporator in the liquid state
at a lower rate than leaving in the vapor state. The delay
pc = ph A/?v/ (2) between the onset of the imbalance between working fluid
flowing into and out of the evaporator and the failure of the
The available capillary pressure is assumed to be equal to heat pipe is determined by the amount of fluid that can be
the pressure as determined by the Laplace relation contained in the evaporator. Since failure of a heat pipe op-
eration will not coincide instantaneously with a change in
pc = (2cr/rc,)cos (3) operating conditions, a correlation between failure and the
detrimental operating condition is illusive and almost impos-
Furthermore, this heat pipe design analysis must postulate sible to establish.
that at the same axial location along a heat pipe the pressure The accepted heat pipe operating model that postulates the
of the liquid that returns in the wick, arteries, or grooves is capillary pressure of a porous structure to be the driving force
less than the pressure of the vapor as shown in Fig. 2. for circulating the working fluid in a heat pipe has failed to
While the capillary pressure is accepted as the driving force predict the performance of heat pipes correctly regardless of
for the circulation of the operating fluid, it is equally consid- their design as shown in Fig. 1. Artery heat pipes that were
ered an operating limit with respect to the total pressure drop. sized to carry relatively large amounts of thermal power as
No relations are presented for predicting heat pipe operation needed in future thermal management space applications have
below this capillary pressure limit. Four additional limits for especially fallen short of their expected performance. The
the operation of a heat pipe have been recognized that can observed shortcomings have stimulated this re-evaluation of
cause a heat pipe to operate in certain temperature regimes the accepted heat pipe design theory.
below its thermal transport capability as determined by the
capillary pressure. The four limits that are shown in Fig. 3— Analysis
as taken from Ref. 2—are given as viscous limit, sonic ve- A heat pipe is designed to permit circulation of a working
locity limit, entrainment limit, and boiling limit. Although fluid in its vapor and liquid phase for the transport of thermal
these limits are based on valid physical phenomena, experi- energy in the form of latent heat of vaporization over a desired
mental data have rarely fallen on any of these limits. Typi- distance. Three main processes must occur internally of the
cally, measured thermal transport rates are found located as heat pipe: 1) evaporation, 2) mass flow in the vapor and liquid
indicated in Fig. 4, which was taken from Ref. 8. It is the state, and 3) condensation of the working fluid. In every
objective of this article to show that the operation of the heat aspect of engineering it is accepted that fluid circulates only
pipe within those limits can be determined by the thermo- if and when potential energy is converted to kinetic energy.
dynamic aspect of heat pipe operation. Since heat is the only available potential energy for pumping
 336 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION

is the result of a temperature difference between the evap-

orator and the condenser

AT = T2 - T3 (5)
LU
C£
The pressure drop of the liquid returning to the evaporator
in its flow passage is

LJ 4P// = ~ Pi (6)
o_
^ Several processes are associated with the flow of the liquid:
LU EVAPORATOR (A-B-C)
"ADIABATIC" VAPOR FLOW (C-D)
1) friction, 2) conduction of heat into the liquid from the
CONDENSER (D-E) adjacent vapor, 3) heat conduction from the evaporator struc-
"ADIABATIC" LIQUID FLOW (E-A) ture to the incoming liquid, and 4) the heat conduction from
CYCLEHP.DC2 WATER the liquid to the heat pipe wall and from the wall to the environ-
ment. Processes 1, 2, and 3 cause a rise in the temperature
of the liquid unless rejection of heat in process 4 is possible.
ENTROPY Effects which increase the entropy in the liquid will force the
Fig. 5 Basic thermodynamic cycle in a heat pipe. liquid towards the two phase region unless the liquid is cooled.
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

The capillary pressure pc, Eq. (3), enters into the heat pipe
operation solely, but importantly, in that it has to be equal
the working fluid in a heat pipe, a thermodynamic cycle must to or larger than the sum of the pressure drops associated
be an integral aspect of the operation of a heat pipe, a con- with the liquid and vapor flow so that the required pressure
sideration that is entirely absent from the current heat pipe difference between the liquid and vapor phase
design theory. The current heat pipe design theory solely
stipulates that the vapor and the liquid flow are the result of = P2 ~ Pi (7)
"capillary suction" of a liquid film stretched over a porous
structure that can be a wick, groove, or artery. The suction as shown in Fig. 2 can be maintained. The enclosed area of
pressure is assumed to be equal to the capillary pressure as the T-S diagram represents naturally the work performed on
determined by the effective pore size and the surface tension the working fluid necessary for its circulation in the heat pipe.
of the fluid. It should be clear that the required work for pumping the
In reality, the circulation of the working fluid in a heat pipe working fluid around the entire cycle is not solely a function
is the consequence of a pumping process that results from the of the total heat transfer, but also of the physical design of
conversion of potential thermal energy into kinetic energy in the heat pipe, i.e., its length, bends, internal and external
a well-defined thermodynamic cycle. The fluid membrane diameters, wick, artery or groove structure, and whatever
stretched across the porous structure solely facilitates this other aspects cause pressure drops in fluid circulation. These
pumping process by separating the liquid from the vapor as physical parameters determine the value of Cd of Eq. (1) for
needed for the flow in the liquid and the vapor flow passages. a heat pipe.
For the thermodynamic cycle to produce the pumping power The discussion of the thermodynamic cycle clearly indicates
commensurate with the product of the flow rate and the total that with the circulation of the working fluid, pressure drops
pressure drop of the vapor and liquid within the confines of will occur as has been the primary consideration of the cur-
the heat pipe, a specific minimum temperature difference has rently accepted heat pipe design theory. But pressure differ-
to exist at a given operating temperature. ences can only be established by compatible temperature dif-
The most basic thermodynamic cycle that can support heat ferences that have to be allowed to exist in the heat pipe. The
pipe operation is shown in Fig. 5. The working fluid enters work performed on the working fluid in the form of pumping
the evaporator at a temperature T{ and is raised to the upper energy is directly a function of the temperature difference
operating temperature T2 along the path A-B. It should be T2 - T3 that exists inside the heat pipe as can be seen from
clear that the maximum internal temperature T2 is lower than Fig. 5. Like in any other thermodynamic cycle, conversion of
the externally applied and measured temperature of a heat thermal energy to kinetic energy is associated with heat re-
pipe evaporator. In the detailed analysis of a specific heat jection at a temperature below the upper temperature of the
pipe which starts with the basic thermodynamic cycle analysis cycle, and occurs with a conversion efficiency of less than
presented in this article the temperature drops at the evap- 100%.
orator as well as at the condenser across the heat pipe struc- From the above thermodynamic considerations it becomes
tures have to be accounted for. Evaporation and expansion clear that a heat pipe cannot operate at a uniform tempera-
from the liquid volume to the vapor volume occurs along path ture. The higher the thermal transport, the larger the internal
B-C. The vapor flows under the influence of the pressure temperature difference has to be between the evaporator and
difference between the evaporator and the condenser through the condenser at a given operating temperature in order to
the "adiabatic" section of the heat pipe along path C-D. The achieve the fluid circulation that is commensurate with the
vapor condenses in the condenser section along path D-E. desired thermal transfer power as given by
The lowest temperature measured on the outer surface of a
heat pipe must naturally be lower than the lowest internal P = m x A//A (8)
temperature that establishes the thermodynamic cycle be-
cause of the temperature drop across the heat pipe structure With the understanding of the interrelation between total
at the condenser. Finally, the liquid working fluid returns to pressure drop and temperature difference in the heat pipe,
the condenser along path E-A. The cycle shown in Fig. 5 is the minimum temperature difference can be determined that
the most basic thermodynamic cycle and is subject to varia- has to be maintained for a given pressure difference. Figure
tions as imposed by the hardware design and the application 6 shows the correlation between saturation temperature and
like any other thermodynamic cycle. The location of path vapor pressure for water. The vapor pressure is related to the
C-D depends on the radial heat fluxes that might occur along saturation temperature by an exponential function (Glausius-
the assumed adiabatic section. The pressure difference Clapeyron Equation)

V = P2 ~ (4) exp(-EJRT) (9)

 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION 337

p (atm) = 5.7955*10 exp(-4974/T)

0.01
300 350 400 450 500
400 450 500 550
Operating Temperature T, K

Temperature T, K Fig. 7 Temperature difference requirement for water heat pipes to

operate at their capillary limit.
Fig. 6 Vapor pressure of water.
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

180

By differentiation of Eq. (9) it is easily seen that a vapor 160

o
pressure difference A/? proves to be a function of A T as well 140 Temperature Difference AT, C -
as the absolute temperature T: ' 0.5
' 0.4

Apt, = Cv x (Ev/RT2)exp(-Ev/RT) x AT (10) 100 -

0.3

The AT required for providing a desired vapor pressure

difference A/?,, can therefore be closely approximated for small
differences by the relation 40
2
AT = Apv x (RT IEVCV) x exp(Ev/RT) (11) 20

0
For larger pressure differences the needed pressure differ- 50 100 150 200 250
o
ence Ap for the desired flow rate is Operating Temperature T, C

Fig. 8 Generated pressure difference as a function of operating tem-

Ap = pv2 ~ pvl (12) perature and total temperature difference.

and the actual temperature difference is found from this limit which is entirely a function of the thermodynamic
parameters, operating temperature, and temperature differ-
AT - (EV/R) x [!//„ (Cv/pv2) - \l/*< (Cv/pvl)] (13)

or Discussion
Mercury has been considered quite often as the working
AT = T2 - (EJR)l/,f (Cv/pvl) (14) fluid for applications in the 440 K temperature regime, be-
cause its temperature-vapor pressure relationship appeared
However, the pressure difference cannot be larger than that to be more favorable than water, i.e., water having an ex-
which can be tolerated across the fluid membrane that sep- cessively high vapor pressure at that operating temperature.
arates the liquid from the vapor in the evaporator. The pres- However, it has also been observed that a correctly filled
sure difference is limited by the capillary pressure, which is mercury heat pipe, i.e., having no liquid working fluid in the
a function of the characteristics of the porous structure and vapor flow volume that permits the heat pipe to operate in
the temperature dependent surface tension as determined by the reflux mode in a Ig environment, would have to reach a
Eq. (3). minimum operating temperature before heat transport could
Equation (11) indicates that the higher the operating tem- be detected, and a considerably higher operating temperature
perature, the lower the A Twill be to achieve a required Ap. than its design temperature, to come close to its design ca-
This relation between pressure difference and temperature pacity. It can now easily be shown that the limitation on the
difference is shown graphically in Figs. 7 and 8 for water. lower operating temperature is imposed by the thermody-
From the very beginning of heat pipe development it was namic cycle. By the currently accepted heat pipe design theory
observed that for the same working fluid the thermal transport that is based on capillary pumping, a 1.41-m-long mercury
capacity of a given heat pipe improved with increasing tem- heat pipe with a 1.3-cm o.d. and a wall wick made of three
perature. The test data which were presented in Ref. 3 and layers of 90 Mesh wire cloth should transfer 52 W of thermal
are reproduced in Fig. 4, substantiate this observation. power when operating at 440 K, at which the vapor pressure
Instead of being a capillary pressure driven device, clearly would be 7.0 x 10~3 atm. Based on the surface tension of
a heat pipe operates under the influence of a thermodynamic mercury at that operating temperature, a "suction pressure"
process, the two principle parameters of the cycle being the of pc = 6435 N/m 2 would be achieved. The design of this heat
total temperature difference and the operating temperature. pipe satisfies fully the standard pressure balance as required
The fluid membrane drawn across the pores of a porous struc- by the accepted design theory, i.e., the combined pressure
ture or a small groove solely functions as a diode, preventing drops of the vapor and the fluid is equal to the suction pres-
vapor from expanding into the liquid volume while permitting sure.
liquid to evaporate from its surface. The capillary pressure The correlation of vapor pressure and temperature for mer-
sets only an upper limit to the thermal transport capacity. It cury is presented in Fig. 9. It shows that at an operating
does not determine or influence the heat transfer rate below temperature of 440 K, a pressure difference of 6435 N/m 2
 338 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION

could be achieved only by a temperature difference consid- As an example, the senior author of this article operated
erably higher than the absolute operating temperature. This a 4.572-m (15-ft) sodium heat pipe that was insulated with
means that at an operating temperature of 440 K no ther- MinK® at a nominal operating temperature of 1113 K in the
modynamic cycle can be established to generate a sufficient laboratory environment, 20°C.10 Of a total of 6 kW of input
pumping pressure differential for circulating the working fluid, thermal power at the evaporator, about 1 kW was lost along
mercury, at a rate commensurate with the predicted heat the adiabatic section. A 3 K temperature must have existed
transport of 52 W. Figure 9 illustrates that to achieve the between the vapor and the liquid to achieve this heat loss.
design heat transfer rate, the minimum operating temperature This temperature difference was sufficient to maintain the
would have to be at least T2 — 506 K. The thermodynamic return fluid in the wall wick in the liquid phase, and thus
analysis makes it now quite obvious that at the design oper- permitted the heat pipe to operate to its full design capacity
ating temperature the temperature difference for generating of 6 kW.
the pumping power needed to circulate the amount of working The same temperature distribution that was shown to be
fluid at a rate commensurate to a heat transfer of 52 W could required for wall wick heat pipes must be imposed on heat
not be achieved; although the porous structure provides suf- pipes which are designed with arteries. Whenever heat pipe
ficient capillary suction pressure. Naturally, at an operating action failed in artery heat pipes, "depriming" of the arteries
temperature of 440 K, the heat pipe will operate below its has usually been blamed as the cause of failure. As previously
operating limit according to relation 1. This relation between noted, there are effects that tend to increase the entropy of
heat transfer, temperature difference, and operating temper- the liquid in return passages. These can cause the generation
ature was well demonstrated in Ref. 9. of vapor in the artery with a partial or total vapor block of
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

The thermodynamic considerations point to additional de- the liquid flow. Depriming, therefore, is not the displacement
sign and operating conditions that must be observed for a of the working fluid by a gas, but is the blockage of the liquid
common cylindrical heat pipe whose liquid flow passages are flow by the vapor of the working fluid. The vapor does not
located along the inner wall. To maintain the flow, a pressure flow into the liquid flow passages, but is generated in it. The
gradient has to exist in the vapor flow as well as in the liquid vapor generation that causes failure is the result of a local
return flow as was shown in Fig. 2. Thus, a pressure difference mismatch between temperature and pressure in the liquid that
between vapor and liquid has to exist at the very same axial is incompatible with the thermodynamic cycle that supports
location including the assumed adiabatic section of the heat the heat pipe operation.
pipe. If the adiabatic section were truly devoid of heat trans- The thermodynamic principle of heat pipe operation makes
fer, no radial heat flux would be present. The vapor and liquid it difficult to consider designs in which arteries are located in
would be of the same temperature unless the condensate had the center of a heat pipe, like in the once proposed cat's eye
been highly supercooled at the condenser. A uniform tem- artery design (Fig. 1). In this configuration the liquid return
perature of vapor and liquid at the same axial location, how- passage is entirely surrounded by vapor in the adiabatic sec-
ever, would violate one of the thermodynamic prerequisites tion, resulting in heat transfer from the vapor to the liquid
for the circulation of the working fluid. Vapor would form in with no cooling of the liquid possible. This design greatly
the porous structure, arteries, or grooves that could reduce enhances the probability of vapor generation in the artery,
or even totally block the return flow of liquid to the evapo- causing burnout of the pipe at a relatively low power transfer
rator. Before the argument can be advanced that superheating level unless a high degree of subcooling of the liquid is achieved
of the fluid is a factual possibility in the return flow, it has to at the condenser, as indicated by path E-A in Fig. 5. This
be convincingly shown that indeed the fluid can remain su- evaluation has been fully borne out by the low performance
perheated while flowing through the tortuous path of a wick. of this type of heat pipes,4 and other similar designs.
Furthermore, the allowable level of the superheating would The effect of subcooling of the liquid return flow was in-
have to be sufficient to prevent vapor development. Naturally, vestigated during the performance testing of a sideflow heat
nobody has ever been able to measure superheating of a flow- pipe.11 The heat pipe was tested with and without subcooling,
ing liquid inside a wick structure located inside a hermetically with the artery in various attitudes, and at various adverse
sealed heat pipe. tilts. The test results are shown in Fig. 10. The performance
In the real thermodynamic world, the liquid can be pre- at 90-deg rotation, and with a 0.635-cm (i-in.) tilt, are par-
vented from vaporizing in the wick by permitting a radial ticularly interesting. Without subcooling, the heat pipe reached
temperature gradient commensurate with the pressure gra- 1.2 kW of thermal power transfer just prior to burnout. With
dient. This condition requires radial heat transfer in the adi- 10°C subcooling, which was achieved by extracting 100 W of
abatic section of the heat pipe. Fortunately, the insulation thermal power from the liquid return flow, the heat pipe
material used for wrapping the adiabatic section of heat pipes transferred 2.1 kW of thermal power without impending burn-
was most of the time limited in its effectiveness, thereby al- out. The 2.1-kW thermal power transfer limit was imposed
lowing the radial heat flux needed for maintaining the essen- by the cooling capacity of the laboratory heat sink. Thus, 100
tial radial temperature gradients. W of subcooling provided an increase of at least 900 W in
heat transfer capacity to the heat pipe. After subcooling was
removed, the heat pipe failed within minutes due to a vapor
200 lock in the liquid return tube.
From the above discussion it becomes clear that cooling
* 100 and/or subcooling of the liquid return has to be sufficient in
70
any heat pipe configuration anywhere in the liquid return
^ path—whether the return path is located internally or exter-
nally of the vapor path—so that the liquid is prevented from
I 30 moving into the vapor region not only in the return path
0) itself—whether this path be coupled or decoupled thermally
3
"o Mercury
from the vapor flow path—but also at the entrance to the
0)
Q. p c = 6,435 N/rr/ evaporator. For the latter condition the required amount of
E 10 subcooling is very much a function of the heat conduction
from the evaporator structure through the container wall to
the liquid return path.
500 520 540 560 580 600
The understanding of the thermodynamic aspects of heat
Operating Temperature T, K
pipe operation can now explain the difficulty in correlating
Fig. 9 Temperature difference requirement for a mercury heat pipe. test data and evaluating heat pipe operation. The thermo-
 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION 339

SUMMARY OF SIDEFLOW HEAT PIPE TEST RESULTS

TRANSPORT CAPACITY (kW)
ferences. Unfortunately, the temperature differences were
Tilt"
Rotation0 not recorded. At the time of testing the total temperature
0° 90° 180°
(in.) w/oSC w/oSC with SC with SC difference had not been identified yet as one of the two major
1/8 2.1 a 2.1 a — 1.8 operating parameters.
1/4 2.3 a 1.2 2.1 a ——
1/2 2.3 a — —— —
2 2.1 — —— —— Conclusions
2-1/2 1.8 — —— —
2-3/4 1.5 — —— —
The previously accepted heat pipe design theory cannot
predict the performances of heat pipes. The fluid circulation
NOTE:
in a heat pipe is the result of a thermodynamic cycle that
a) Heat sink limited
operates over a temperature difference. The capillary pres-
sure supported by the porous structure in a heat pipe and the
b) Tilt
temperature dependent surface tension of the working fluid
only facilitates the conversion of thermal energy to kinetic
energy. It provides the fluid diode action which prevents vapor
c) Rotation from penetrating into the liquid return passages while per-
mitting liquid to be vaporized and expand into the vapor
space. The rate at which the working fluid can be circulated,
and thus thermal power can be transferred, is determined by
the internal temperature difference that can be achieved and
Downloaded by INDIAN INST OF TECHN. BOMBAY on June 28, 2020 | [Link] | DOI: 10.2514/3.543

o the absolute operating temperature.

A heat pipe has been shown to obey all the observed laws
Fig. 10 Sideflow heat pipe subcooling data. of heat transfer like any other heat conductor. Its high heat
conductivity (super thermal conductivity) is a very strong
function of the operating temperature in a limited tempera-
10 2500 ture range as determined by the working fluid. When the
O
operating range is exceeded either at the lower end or at the
upper end, the working fluid loses its ability to contribute to
- 200° Limits the total heat conduction of the device by circulating in its
two phases, liquid and vapor. At the low temperature end
-£ 1500
the vapor pressure of the working fluid is insufficient, while
O
above the upper, temperature limit the capillary pressure is
Q. inadequate to sustain the pressure difference. This variation
1000 in thermal conductivity of a heat pipe structure is almost
exactly the same as that found in solid material structures,
where in different temperature ranges different physical phe-
S 500 nomena contribute to the total heat transfer to different de-
grees.12-13
This understanding of heat pipe operation should not only
0 50 100 150 200 250 300 350 lead to better predicting heat pipe performances, but might
o point to new thermal transport concepts that can carry thermal
Temperature, C
energy over longer distances with components that weigh less
Fig. 11 Presentation of experimental heat pipe data based on ther- and perform with a lower loss in thermal potential than the
modynamic considerations. present heat pipes. It is hoped that the theoretical consider-
ations that were presented will stimulate their full experi-
mental verification and further refinement.
dynamic analysis has identified the two principle parameters
that determine the thermal transport capacity, the total tem- Acknowledgments
perature difference, and the operating temperature. The cap-
illary pumping pressure, which previously was considered the The research described in this article was carried out by
parameter which establishes the operation of the heat pipe, the Jet Propulsion Laboratory, California Institute of Tech-
obviously defines only one of the operating limits. Neither nology and by Wright Laboratory, Wright Patterson Air Force
one of the two thermodynamic parameters alone is sufficient Base. It was sponsored by the U.S. Air Force and the National
Aeronautics and Space Administration.
to characterize the heat pipe transport capacity. With the
presented thermodynamic relations, the test data of Fig. 4
can now correctly be interpreted. The authors of Ref. 8 sug- References
gested that these data were located on the viscous and/or sonic
velocity limits as shown in Fig. 3, with some undetermined ^hi, S. W., "Heat Pipe Theory and Practice, A Source Book,"
Hemisphere, New York, 1976, pp. 33-77.
offset due to the idealization of the theoretical model. How- 2
Dunn, P. D., andReay, D. A., Heat Pipes, Pergamon, New York,
ever, a comparison between Figs. 4 and 8 indicates that the 1978, pp. 19-98.
test data follow the thermodynamically imposed relation be- 3
Ponnappan, R., Ramalingam, M. L., Beam, J. E., and Mahefkey,
tween the actually generated pressure difference, operating E. T., "Remedial Measures for Condenser Quenching in a Long
temperature, and temperature difference. Water Heat Pipe," AIAA Paper 87-1646, June 1987.
4
As in all flow devices, the flow rate of the working fluid, Alario, J. P., Haslett, R., and Kosson, R., "The Monogroove
which is equivalent to heat transport in a heat pipe, is directly High Performance Heat Pipe," AIAA Paper 81-1156, June 1981.
5
a function of the available pressure difference. The exact Benner, K., and Feild, Z., private communication, Lockheed Mis-
relation between flow rate and pressure drop is naturally de- siles and Space Co., discussion at Wright Research and Development
termined by the physical design parameters of a heat pipe as Center, Nov. 1989.
6
Kosson, R., Brown, R., and Unger, E., "Space Station Heat Pipe
mentioned previously. Thus, the test data of Ref. 8 would Advanced Radiator Element (SHARE) Flight Test Results and Anal-
have been properly plotted as indicated in Fig. 11, accounting ysis," AIAA Paper 90-0059, Jan. 1990.
for the additional parameter AT. Since, during testing, many 7
Brown, R., Unger, E., and Cornwell, J., "Flight Test Results of
different condenser chill block configurations were used, the the SHARE II Monogroove Heat Pipe," AIAA Paper 92-2886, July
heat pipe was exposed to a large range of temperature dif- 1992.
INST OF TECHN. BOMBAY on June 28, 2020 | [Link]

340 RICHTER AND GOTTSCHLICH: HEAT PIPE OPERATION

<s
Ponnappan, R., and Mahefkey, E. T., "Development of a Dou- Wright-Patterson
H
Air Force Base, OH, Nov. 1974.
ble-Wall Artery High Capacity Heat Pipe," Spacecraft Thermal Con- Fleischman, G. L., Bascules, A., and Gottschlich, J. M., "Ad-
trol, Design and Operation, edited by P. E. Bauer and H. E. Collicott, vanced Heat Pipe Components for High Power Spacecraft Thermal
Vol. 86, Progress in Astronautics and Aeronautics, AIAA, New York, Management,"
12
AIAA Paper 90-0058, Jan. 1990.
1983,
9
pp. 202-221. Touloukian, Y. S., et al., "Thermophysical Properties of Matter
Deverall, J. E., "Mercury as a Heat Pipe Fluid," Los Alamos Volume 1: Thermal Conductivity Metallic Elements and Alloys," IFI/
Scientific Lab., LA-4300-MS, Los Alamos, NM, October 1969. Plenum,
13
New York-Washington, 1970
'"Richter, R., "Solar Collector Thermal Power System, Volume Touloukian, Y. S., et al., "Thermophysical Properties of Matter
III Basic Study and Experimental Evaluation of Thermal Train Com- Volume 2: Thermal Conductivity Nonmetallic Solids," IFI/Plenum,
ponents," Air Force Applied Physics Lab., AFAPL-TR-74-89-3, New York-Washington, 1970.