Energy Conversion and Management 49 (2008) 3398–3404

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Heat pipe for cooling of electronic equipment

Yu-Wei Chang, Chiao-Hung Cheng, Jung-Chang Wang, Sih-Li Chen *
Department of Mechanical Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan

a r t i c l e i n f o a b s t r a c t

Article history: This article experimentally investigates the thermal performance of the heat pipe cooling system with the
Available online 24 June 2008 thermal resistance model. Evaporator and condenser, which are the two main devices, connect to each
other to form a closed system. The liquid water absorbs heat from heat source and evaporates in the
Keywords: evaporator. The evaporating fluid moves toward the condenser, and then condenses in the condenser.
Thermal resistance The experimental parameters are different evaporation surfaces, fill ratios of working fluid and input
Thermal performance heating powers. The result shows that the evaporation resistance and the condensation resistance both
Heat pipe
grow with increasing heating power and decreasing fill ratio. Flooding is found at the fill ratio of 20% with
Flooding phenomenon
the evaporation surface noted Etched Surface 2 when heating power is above 120 W. Flooding phenom-
enon is caused by the opposite flow direction of vapor and liquid in a closed two-phase system. According
to the result, the lowest total thermal resistance is 0.65 °C/W by the evaporation surface noted Etched
Surface 2 at 30% fill ratio.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction experiments were performed with heating power in the range of

60–140 W with increment of 20 W.
Air cooling solution which comprises a fan and heat sink is em-
ployed to remove heat generated by electronic device for stability 2. Experimental investigation
and lifespan [1,2]. To solve the growing heat generated by elec-
tronic devices, two-phase change devices (eg. heat pipe, loop heat 2.1. Apparatus
pipe, thermosyphon cooling system) become main cooling tech-
nologies in the electronic industry [3]. This article experimentally As shown in Fig. 1, evaporator and condenser which are the two
investigates the thermal performance of the heat pipe cooling sys- main devices in the cooling system are connected to form a closed
tem for electronic equipments. cooling system. The liquid working fluid absorbs heat from heat
Two-phase cooling device has been proved as a promising heat source and evaporates in the evaporator, and then moves toward
transfer device with higher effective thermal conductivity than the condenser. In the condenser the vapor releases heat to the
over 200 times of copper [4]. With the advantage of high thermal ambiance and condenses into liquid. Finally the condensed work-
conductivity, two-phase cooling devices are used in many applica- ing fluid returns to the evaporator by gravity. After installing the
tions such as electronic cooling, etc. Some works has been done to evaporation surface, the experiments are done with different fill
investigate the thermal performance of electronic cooling systems ratio under different input heating power.
[5,6]. The heating surfaces of evaporator are not flat in the past This research adopts water as working fluid. The flat vapor
studies, however, a flat plate heating surface is more practical in chamber is set above the heater. The check valve I is used to control
electronic cooling application. In the presented study, a vapor the amount of working fluid into the evaporator after vacuuming
chamber was used as an evaporator and the evaporation structure the cooling system via check valve II. A pressure transducer is used
was etched to enhance the boiling performance. Three major fac- to measure the vapor pressure inside the evaporator. Thermal cou-
tors that influence the thermal performance of the cooling system ples are used to measure the temperature of working fluid inside
were studied. They are different evaporation surfaces, fill ratios of the condenser. The vapor chamber is made of copper, it consists
working fluid and heating powers. For an organized analysis, this two flat plates and a frame. The etched evaporation surface is man-
study develops a thermal resistance network to quantify the ther- ufactured on the flat plate contacting with heater. The size of evap-
mal performances under different experimental conditions. The orator is 100  80  15 mm3 and the interior volume of evaporator
is 43.2 cm3. The condenser employs two copper pipes to reach high
fin efficiency. The inner and outer diameters are 5.15 and 6.35 mm,
* Corresponding author. Tel.: +886 2 23631808; fax: +886 2 23631755. respectively. There are 60 pieces of fins soldered on the pipes. The
E-mail address: slchen01@ntu.edu.tw (S.-L. Chen). total length of fin array is 88.8 mm since the distance of 25 mm

0196-8904/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2008.05.002
 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404 3399

Nomenclature

Ag contact surface area of thermal grease, m2 base base of evaporator

Cl specific heat of liquid, kJ/kg °C c contact surface
Ck defined as Eq. (9) cnv condensation
D Inner diameter of tube, m con convection
g gravity, 9.81 m/s2 cw condenser wall
hfg latent heat, kJ/kg die dummy heater
Q heating power, Watt evap evaporator
R thermal resistance, °C/W f actual fill volume in the evaporator
T temperature, °C if interface between heater and base
V volume, c.c. l liquid
tg thickness of thermal grease, m v vapor
kg conductivity of thermal grease, °C/mW
Non-dimensional numbers
Greek symbols c fill ratio, as Eq. (1)
m kinetic viscosity, m2/s Bo Bond number, as Eq. (10)
q mass density, kg/m3 Re Reynolds number, [UD/v]
r surface tension, N/m

Subscripts
amb ambience
ave average

Fig. 1. Heat pipe cooling system for electronic equipment.

Fig. 2. Geometry of the etched evaporation surface.

from the evaporator. The fins are made of aluminum. The area of
fin is 50 mm2 and the thickness is 0.3 mm. The area of contact sur- Table 1
face is 15  15 mm between heater and evaporator. The configura- Three etched evaporation surface
tion of etched evaporation surface is shown in Fig. 2. It is
Symbol P (mm) H (mm)
manufactured by chemical etched process. There are three etched
1 PLATE 0 0
evaporation surfaces with different pitch as shown in Table 1.
2 Etched Surface 1 0.6 0.2
The etched area for every etched evaporation surface is 3 Etched Surface 2 1 0.2
80  60 mm. There are four fill ratios (eg. 20%, 30%, 40%, 50%) in
this experiment. The fill ratio is defined as
Vf
c¼  100% ð1Þ the contact resistance Rif, the evaporation resistance Revap, the con-
V evap
densation resistance Rcon and the convection resistance Rcnv, which
where Vf is the actual fill volume of working fluid, and Vevap is the is mathematically expressed as
interior volume of evaporator.
T die  T amb
Rt ¼ ¼ Rif þ Revap þ Rcon þ Rcnv ð2Þ
Q
2.2. Thermal resistance network
In Eq. (2), contact resistance Rif is defined as the temperature
The thermal resistance network of the cooling system is pre- difference across the interface (the temperature of heat source Tdie
sented in Fig. 3. The total thermal resistance is the summation of minus the temperature at the center of the lower surface of the flat
3400 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404

Tamb of condensation liquid recovers, the vapor temperature drops. This

is so called flooding phenomenon. When flooding happens, the
Rcnv vapor temperature and pressure oscillate with time. The occurrence
of flooding phenomenon limits the performance of two-phase
Tcw,ave
closed system. Eq. (7) is the correlation for prediction flooding lim-
Rcon itation. Assuming that the density of liquid is much larger than vapor
[7], it is shown as
Tevap  2  1=2
~ 3=2 Ck qv
Revap Rel P D ; ð7Þ
Bo1=4 ql
Tbase where
 1=3
Rif
e ¼D g
D ð8Þ
Tdie
2
l m
pffiffiffiffiffiffiffi
C k ¼ 3:2 tanhð0:5Bo1=4 Þ ð9Þ
Q q g 1=2
Bo ¼ D l ð10Þ
r
Q
As shown in Eqs. (7)–(10), the flooding limitation is a function of
Fig. 3. Thermal resistance network. tube inner diameter, liquid viscosity, liquid density and vapor den-
sity. Because liquid viscosity, liquid density and vapor density of
water are physical properties, they are the functions of tempera-
plate Tbase) divided by total Q. The contact resistance can be mini- ture. Therefore, the flooding limitation is the function of tube inner
mized by applying thermal grease, such as silicon oil, on the con- diameter, the condensation temperature and the evaporation tem-
tact surface before assembling evaporator and heater. The perature. According to the numerical evaluation of Eq. (7), the lim-
thickness of the thermal grease is tg, the thermal conductivity is itation is boosted with a large tube inner diameter. The limitation is
kg and the contact surface area is Ag, which are related to the ther- also boosted by the high ratio of vapor density to liquid density,
mal contact resistance by which is affected by the evaporation temperature and condensation
T die  T base tg temperature. The evaporation temperature and the condensation
Rc ¼ ¼ ð3Þ temperature result from the evaporation ability and condensation
Q k g Ag
ability. Hence, the ratio varies at different conditions.
Evaporation resistance is defined as the temperature difference be-
tween the center of the lower surface of base plate (Tbase) and the 3. Results and discussion
vapor temperature of the chamber of base plate (Tevap) divided by
Q. This thermal resistance is resulted from the thickness of base The contact resistance Rif and the convection resistance Rcnv do
plate, the thermal properties of its material, and the boiling heat not vary seriously for all operating conditions. The average contact
transfer capability. It can be expressed as resistance is 0.11 °C/W, and the average convection resistance is
T base  T evap 1 0.13 °C/W. Since the contact resistance Rif and convection resis-
Revap ¼ ¼ ð4Þ tance Rcnv are almost constant, the evaporation resistance Revap
Q hevap Aevap
Condensation resistance is defined as the temperature difference
between the vapor temperature (Tevap) and the average temperature 1
of condensation (Tcw,ave) divided by Q. The thickness of condensa-
tion tube, the thermal properties of its material and the condensing 20%
heat transfer of condensation result in this thermal resistance. It can 30%
0.9 40%
be expressed as
50%
T evap  T cw;ave 1
Rcon ¼ ¼ ð5Þ
Q hcon Acon
Convection resistance is defined as the temperature difference be- 0.8
Rt (ºC/W)

tween the average temperature of condensation (Tcw,ave) and the

ambient temperature (Tamb) divided by the heat transfer rate Q.
The thermal resistance is expressed as
0.7
T cw;ave  T amb 1
Rcnv ¼ ¼ ð6Þ
Q hcnv Acnv

2.3. Flooding phenomenon 0.6

When the pressure inside the evaporator is high, the high pres-
sure at the entrance of condenser pipe blocks the condensation
working fluid to reflow to the evaporator. The vapor temperature 0.5
40 60 80 100 120 140 16 0
inside the evaporator will increase while the condensation liquid
Q (Watt)
is trapped in the condenser pipes. When the gravitation force of
the liquid is large enough to overcome the shear stress, the liquid Fig. 4. Total thermal resistance with plan plate surface under different input
would be able to reflow to the evaporator. As soon as the reflowing heating powers and fill ratios.
 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404 3401

and condensation resistance Rcon are the major causes to affect the ing powers and fill ratios when the plan plate evaporation surface
variation of the total thermal resistance Rt. is used. The total percentages of evaporation resistance and con-
Fig. 4 shows the total resistance of the system with plan plate densation resistance are over 65%. The percentage of evaporation
evaporation surface at different input heating powers and fill ra- resistance Revap is dropping from 44% to 39% with increasing input
tios. The total thermal resistance Rt decreases from 60 W to heating power from 60 W to 140 W at fill ratio of 20%, from 40% to
140 W at different fill ratios, and the decreasing tendency is more 37.5% at fill ratio of 30%, from 38% to 34.5% with at fill ratio of 40%
obvious at high fill ratio. The Rt drops from 0.84 °C/W to 0.73 °C/W and from 34% to 29% at fill ratio of 50%. The percentage of conden-
at fill ratio of 50%, from 0.72 °C/W to 0.66 °C/W at fill ratio of 40%, sation resistance Rcon is growing from 22% to 27.5% with increasing
from 0.69 °C/W to 0.66 °C/W at fill ratio of 30%, and from 0.66 °C/W input heating power from 60 W to 140 W at fill ratio of 20%, from
to 0.65 °C/W at fill ratio of 20%. 28.5% to 33% at fill ratio of 30%, from 30% to 34.5% at fill ratio of 40%
Figs. 5 and 6, respectively, illustrate the percentage of evapora- and from 37% to 40% at fill ratio of 50%. The condensation
tion resistance and condensation resistance at different input heat-

0.5
60
Fill Ratio Fill Ratio: 20%
20%
Surface
30%
50 Plate
40%
Etched surface 1
50% 0.4
Etched surface 2

40
Percentage (%)

Revap (ºC/W)
0.3
30

20
0.2

10

0.1
0 40 60 80 100 120 140 160
60 80 100 120 140
Q (Watt) Q (Watt)

Fig. 5. The percentage of the evaporation resistance with plan plate surface under Fig. 7. Evaporation resistance with different input heating power and etched
different input heating powers and fill ratios. evaporation surfaces at 20% fill ratio.

60 0.5
Fill Ratio
Fill Ratio: 30%
20%
30% Surface
50 40% Plate
50% 0.4 Etched surface 1
Etched surface 2
40
Percentage (%)

Revap (ºC/W)

30 0.3

20

0.2

10

0 0.1
60 80 100 120 140 40 60 80 100 120 140 160
Q (Watt) Q (Watt)

Fig. 6. The percentage of the condensation resistance with plan plate surface under Fig. 8. Evaporation resistance with different input heating power and etched
different input heating powers and fill ratios. evaporation surfaces at 30% fill ratio.
3402 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404

resistance grows with increasing input power because the thick- 140 W. Because the boiling mechanism is intensive at high heating
ness of condensing liquid film in the condenser also increases. power, the evaporation resistances decrease with increasing heat-
The evaporation resistances Revap of evaporation surfaces of ing power. As indicated by Figs. 7–10, the Revap of Etched Surface 2
plan plate and Etched Surface 1 decrease with increasing input achieves the lowest value among all etched evaporation surfaces.
heating power at different fill ratios, the tendency is shown in Figs. Thus, Etched Surface 2 is the most effective structure to enhance
7–10. The Fig. 7 shows that the Revap by Etched Surface 2 drops boiling performance.
from 0.25 °C/W to 0.20 °C/W at the heating power from 60 W to Figs. 11–13, respectively, show the evaporation resistances of
100 W and then grows from 0.26 °C/W to 0.47 °C/W at the heating the three evaporation surfaces under different input heating pow-
power from 120 W to 140 W. This reverse is caused by dry-out. Be- ers and fill ratios. As shown in Fig. 11, the evaporation resistance of
cause the fill ratio is too low, the condensing liquid cannot plan plate decreases with increasing fill ratio. Generally speaking,
promptly reflow at such high input heating power of 120 W and the same trend is also shown by Fig. 12 with the Etched Surface
1. However, the evaporation resistance of Etched Surface 1 at fill

0.5 0.5
Fill Ratio: 40%
Plate
Surface
Plate Fill Ratio
Etched surface 1 20%
0.4 0.4 30%
Etched surface 2
40%
50%
Revap (ºC/W)

Revap (ºC/W)
0.3 0.3

0.2 0.2

0.1 0.1
40 60 80 100 120 140 160 40 60 80 100 120 140 160
Q (Watt) Q (Watt)

Fig. 9. Evaporation resistance with different input heating power and etched Fig. 11. Evaporation resistance of plan plate with different input heating power and
evaporation surfaces at 40% fill ratio. fill ratios.

0.5 0.5

Fill Ratio: 50% Etched surface 1

Surface Fill Ratio
Plate 20%
0.4 Etched surface 1 0.4 30%
Etched surface 2 40%
50%
Revap (ºC/W)
Revap (ºC/W)

0.3 0.3

0.2 0.2

0.1 0.1
40 60 80 100 120 140 160 40 60 80 100 120 140 160
Q (Watt) Q (Watt)

Fig. 10. Evaporation resistance with different input heating power and etched Fig. 12. Evaporation resistance of Etched Surface 1 with different input heating
evaporation surfaces at 50% fill ratio. power and fill ratios.
 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404 3403

0.5 above 100 W. The fill ratio between 30% and 40% is the most appro-
priate value for Etched Surface 2 for the stable performance during
Etched surface 2 the heating powers.
Fill Ratio Figs. 14–17 show condensation resistance as different evapora-
20% tion surfaces employed under different input heating powers and
0.4 30% fill ratios. The condensation resistance Rcon increases with increas-
40% ing heating power. And the condensation resistance Rcon of Etched
50% Surface 2 has the lowest value among all the evaporation surfaces.
As predicted by Eq. (7), flooding limitation might occur with
Revap (ºC/W)

Etched Surface 2 at the condition of fill ratio 20% with input heat-
0.3 ing power above 120 W. The prediction matches the experimental
result. To prove the occurrence of flooding phenomenon, the
dry out

0.3
0.2
Fill Ratio: 30%
Surface
Plate
Etched surface 1
0.25
0.1 Etched surface 2
40 60 80 100 120 140 160
Q (Watt)

Rcon (ºC/W)
Fig. 13. Evaporation resistance of Etched Surface 2 with different input heating
power and fill ratios. 0.2

ratio of 20% drops below that of 30% at 140 W. The drop also hap-
pens on the evaporation resistance of Etched Surface 2 at the fill ra-
tio of 20% and 100 W as shown in Fig. 13. Note that the drop occurs 0.15
just before dry-out with Etched Surface 2. Similarly, dry-out might
happen with Etched Surface 1 at fill ratio of 20% above 140 W. The
dry-out point of Etched Surface 1 is higher than Etched Surface 2
because the pitch of Etched Surface 1 is smaller. It results in a high-
er capillary force to draw water toward heating area. As shown in 0.1
Fig. 13, the evaporation resistances at the fill ratio of 30% and 40% 40 60 80 100 120 140 16 0
Q (Watt)
are almost the same under different heating powers. At the fill
ratio of 50%, the evaporation resistance of Etched Surface 2 is larger Fig. 15. Condensation resistance with different input heating powers and etched
than that at 30% and 40% when heating power is below 100 W, but evaporation surfaces at 30% fill ratio.
becomes less than that at 30% and 40% when heating power is

0.3 0.3

Fill Ratio: 40%

Fill Ratio: 20%
Surface
Surface Plate
Plate Etched surface 1
0.25 Etched surface 1 0.25
Etched surface 2
Etched surface 2
Rcon (ºC/W)
Rcon (ºC/W)

0.2 0.2

0.15 0.15

0.1 0.1
40 60 80 100 120 140 160 40 60 80 100 120 140 160
Q (Watt) Q (Watt)

Fig. 14. Condensation resistance with different input heating powers and etched Fig. 16. Condensation resistance with different input heating powers and etched
evaporation surfaces at 20% fill ratio. evaporation surfaces at 40% fill ratio.
3404 Y.-W. Chang et al. / Energy Conversion and Management 49 (2008) 3398–3404

0.3 40

Fill Ratio: 50%

35
Surface Q=60W Q=80W Q=100W Q=120W Q=140W
Plate
0.25 Etched surface 1 30
Etched surface 2

Pressure (Kpa)
25
Rcon (ºC/W)

0.2 20

15

0.15 10

0.1 0
40 60 80 100 120 140 160 0 1800 3600 5400 7200 9000
Q (Watt) Time (sec)

Fig. 17. Condensation resistance with different input heating powers and etched Fig. 19. Pressure of Etched Surface 2 with times at different input heating powers.
evaporation surfaces at 50% fill ratio.

tance increases with increasing input heating power. The lowest

total resistance is 0.65 °C/W at 30% fill ratio. Flooding phenomenon
180
Tdie is caused by the opposite flow direction of vapor and liquid in the
Tbase condenser. As soon as flooding phenomenon occurs, the evapora-
160 tion resistance and condensation resistance simultaneously grow
Q=60W Q=80W Q=100W Q=120W Q=140W with increasing heating power and decreasing fill ratio.
140
4. Conclusion
120
Temperature (ºC)

The thermal resistance model for the heat pipe cooling system
100 is developed in the presented study. Through the thermal resis-
tance analysis, the evaporation resistance Revap and the condensa-
80 tion resistance Rcon are the most influential factors to the total
resistance Rt of the cooling system. The result shows that the struc-
ture of etched evaporation surface has significant influence on the
60
thermal performance of the cooling system. According to the above
results, Etched Surface 2 shows the best performance among the
40 evaporation surfaces. Flooding phenomenon occurs as the system
operated at low fill ratio and high input heating power. The flood-
20 ing operation point for this system has been predicted by Eq. (7),
and the prediction is closed to the experimental result in the pre-
0 sented study.
0 1800 3600 5400 7200 9000
Time (sec)
References
Fig. 18. Temperatures of Etched Surface 2 with times at different input heating
powers. [1] Haper PR, Brown WB. Mathematical equations for heat conduction in the fins of
air cooled engines. NACA Report 158; 1923. p. 32.
[2] Bejan A, Sciubba E. The optimal spacing of parallel plates cooled by forced
convection. Int J Heat Mass Transfer 1992;35(12):3259–64.
experimental record data are shown in Figs. 18 and 19. Fig. 18 [3] Leonard L. Vasiliev. Heat pipes in modern heat exchanger. Appl Therm Eng
illustrates the temperatures of evaporator with Etched Surface 2, 2005;25:1–19.
the oscillations is obvious at 120 W and 140 W. Meanwhile, the [4] Noie SH. Heat transfer characteristics of a two-phase closed thermosyphon.
Appl Therm Eng 2005;25:495–506.
vapor pressure also oscillates with the temperatures at 120 W
[5] Nguyen-Chi H, Groll M, Dang-Van T. Experimental investigation of closed
and 140 W, which is shown in Fig. 19. Therefore, the experimental two-phase thermosyphons. In: AIAA 14th thermophysics conference; 1979.
results of this study are close to the prediction by Eq. (7). p. 239–46.
The experimental results are summarized as following. The [6] Nitipong S, Terdtoon P, Tantakom P, Polchai A. A performance limit model of an
inclined two-phase closed thermosyphon. In: Proceeding 6th international heat
total resistance and evaporation resistance generally decreases pipe symposium; 2000. p. 258–68.
with increasing input heating power while the condensation resis- [7] Tien CL, Chung KS. Entrainment limits in heat pipes. AIAA J 1979;17(6):643–6.