SUBCOOLED BOILING OF HFE-7100 ... - Eurotherm 2008

5th European Thermal-Sciences Conference, The Netherlands, 2008
SUBCOOLED BOILING OF HFE-7100 DIELECTRIC LIQUID
ON COPPER SURFACES WITH CORNER PINS
J. L. Parker, M. S. El-Genk
Chemical and Nuclear Engineering Department and Institute for Space and Nuclear Power Studies,
University of New Mexico, Albuquerque, NM, USA
Abstract
This paper reports the results of experimental investigations of pool boiling of HFE-7100 dielectric liquid on
copper surfaces measuring 10 x 10 mm and having 3 x 3 mm corner pins, 2, 3 and 5 mm tall. The pool
boiling curves and the Critical Heat Flux (CHF) for saturation and 10 K, 20 K, and 30 K subcooling and
surface inclinations from 0 o (upward-facing) to 180 are presented. Results show significant increases in the
total thermal power removed from the surface, compared to plane copper of the same footprint. Thermal
power removed at CHF from the surface with 5 mm tall corner pins in the upward-facing orientation (θ = 0 o )
at ΔTsub = 30K is 93 W, decreasing only to 86 W at 180 o . CHF increases linearly with increased liquid
subcooling, and for this surface its rate of increase rises from 2.4%/K at 0 o to 3.03%/K at 180 o . The
developed natural convection correlation is within +12% of the present data. The developed CHF correlation
for plane and micro-structured surfaces accounts for both surface orientation and liquid subcooling and is
within +12% of present data and that reported for other micro-structured surfaces
Nomenclature
AR ratio of geometrical & footprint areas
CHF critical heat flux (W/cm 2 )
g acceleration of gravity (9.81 m/s 2 )
hfg
latent heat of vaporization (J/kg)
P thermal power removed (W)
q heat flux based on footprint (W/cm 2 )
T temperature (K)
ΔTp surface superheat, (Tw – Tp)
ΔTsat surface superheat, (Tw – Tsat)
ΔTsub surface superheat, (Tsat – Tp)
ρ density (kg/m 3 )
σ surface tension (N/m)
1 Introduction
θ inclination angle (º)
Subscripts
NC natural convection
p liquid pool
plane plane surface
l liquid
Sat saturation
Sub subcooling
w boiling surface or footprint area
v vapor
The challenge to future development of faster and more powerful computer chips is removing the
ever increasing dissipating thermal power while maintaining the junction temperature below 85 o C,
and 125 o C in special high temperature applications. The increases in the heat dissipation flux from
the surface of the chips are typically associated with increased non-uniformly and the development
of hot spots. To deal with these issues, the interest in investigating cooling high power chips using
nucleate boiling of dielectric liquids such as FC-72 and HFE-7100 has been increasing. In addition
to their chemical compatibility with structure materials of interest and being environmentally
friendly, these liquids have low boiling points for keeping the junction temperature below the
requirement values. However, the high wetting of dielectric liquids, when used with commercial
grade surfaces like copper and silicon, causes large temperature excursions up to 40 K prior to

5th European Thermal-Sciences Conference, The Netherlands, 2008
boiling incipience, which is undesirable for computer chip cooling. To increase the heat removal
rate by nucleate boiling and reduce or eliminate the temperature excursion prior to boiling
incipience of dielectric liquids, many researches have investigated increasing the surface roughness
or covering the surface with a micro-porous coating or porous or micro-porous materials (e.g., Jiang
et al. 2001; Webb 2004; Parker and El-Genk 2006; Arik, Bar-Cohen, and You 2007; El-Genk and
Parker 2005).
Further increases in the thermal power removed by nucleate boiling can be achieved by increasing
the geometrical surface area in contact of the boiling liquid, for the same footprint, using microfins,
micro- and macro-structures, and pins of various lengths, cross section and density(e.g.,
Nakayama et al. 1980; Anderson and Mudawar 1989; Misale et al. 1999; Rainey and You 2000;
Wei and Honda 2003; Al-Hajri et al. 2005; Launay et al. 2006; Yu and Lu 2007; Rajulu et al. 2004).
Micro-structures are generally those with a length scale < ~ 0.5 mm and spacing between pins or
fins < 0.5 mm, while macro-structured surfaces have larger length scale and spacing (Parker and El-
Genk, 2007). The reported investigations showed measurable increases in the thermal power
removal by nucleate boiling and at CHF. In addition, with these surfaces, the temperature excursion
prior to boiling incipience was either reduced or eliminated. Most of the reported work with micro-
and macro-structured surfaces has been for saturation boiling of dielectric liquids in the upwardfacing
orientation (θ = 0 o ). Only a few investigators has reported results in the vertical orientation
(90 o ) and on the effect of the subcooling on nucleate boiling and CHF (El-Genk and Bostanci 2003;
Parker and El-Genk, 2006; Watwe et al., 1997); and no results could be found for subcooled boiling
on micro- and macro-structured surfaces at different orientations.
This paper investigates the effects of subcooling and orientation on nucleate boiling and CHF of
HFE-7100 on 10 x 10 mm plane copper and Cu surfaces with 3.0 x 3.0 mm square corner pins, 2, 3,
and 5 mm tall. The liquid subcooling, ΔTsub, in the experiments varies from 0 K (saturation) to 30
K in 10 K increments, and the inclination angle varies from θ = 0 o to 180 o , in 30 o increments. A
general CHF correlation for HFE-7100 on plane and macro-structured surfaces is developed based
on the present and reported data by others (Rainey and You, 2000; Yu and Lu, 2007), which
accounts for the effects of both subcooling and surface orientation. The present natural convection
data are also correlated and compared to those for plane surfaces (Parker and El-Genk, 2005).
2 Experiments
The experiment setup is detailed elsewhere (El-Genk and Bostanci 2003; El-Genk and Parker 2005;
Parker and El-Genk 2005 and 2006) and briefly described in section 3.1 (Figures 1a and 1b). The
test section consists of a Teflon block with a 1.0 mm deep square cavity at the centre of the top
surface for the heating element. The 1.6 mm thick, 10 x 10 mm copper block, with either a plane
surface or corner pins, is soldered to the top of the heating element. A two-part epoxy adhesive fills
the shallow top cavity of the test section (30 x 30 mm) that is encased in a Lexan frame with a
closed bottom.
The two K-type thermocouples for measuring the average temperature of the plane copper surface,
and the four for measuring the average temperature of the Cu with corner pins are inserted into 0.6
mm diameter horizontal holes, ~ 0.8 mm below the footprint surface. The average reading of these
thermocouples, after accounting for the temperature drop due to conduction to the footprint surface
(< 1.2 K), is taken as the surface temperature, Tw, in the present boiling curves. The heat losses
through the sides, top and bottom of the assembled test section are negligible; thus, the removed
power by HFE-7100 boiling in the experiments is very close to that generated by the heating
element. The latter equals the applied voltage across the heating element multiplied by the electric
current provided by the DC power supply.

5th European Thermal-Sciences Conference, The Netherlands, 2008
3.1 Setup and Procedures
The experiments used degassed HFE-7100 and the test vessel (Figure 1b) is sealed tight and placed
in a water bath with a submerged electrical heater to maintain the HFE-7100 liquid in the vessel at
saturation or at the desired subcooling, with the help of two cooling coils immersed in the liquid
pool. Because of the elevation of Albuquerque, NM (~ 1680 m above sea level), the saturation
temperature of HFE-7100 in the experiments is ~ 54 o C. The pool temperature is monitored using
four submerged thermocouples. In the upward facing orientation, θ = 0°, the height of the liquid
above the test section surface is kept at ~ 8 cm. The water-cooled copper coil, fastened to the inside
of the top cover of the test vessel, condenses the vapour generated in the experiments, thus
maintaining a constant pressure in the vessel (Figure 1b).
Figure 1: The experimental setup and boiling vessel.
In the experiments, the electrical power to the test section is increased in small increments to avoid
burning the heating element when reaching CHF. An increase of > 30 K in two subsequent steady
state measurements of the average temperature of the footprint surface is taken as an indication of
reaching CHF, and the experiments are terminated. The measurements uncertainties are ~ + 0.2 K
for temperatures, + 2% for input power, ~ 3% for the thermal power removed at CHF, and up to +
3 K in the surface temperature at CHF.
4 Results
Figures 2a and 2b present the measured boiling curves for 30 K subcooling on a plane Cu and a Cu
surface with 5 mm tall corner pins. With the former, there is a 25 K excursion in the surface
temperature prior to boiling incipient, which is absent with the surface with corner pins. The boiling
curves have 4 distinct regions: natural conversion (NC), extending up to incipient boiling, discrete
bubbles nucleate boiling (I), fully developed nucleate boiling (II), and lateral coalescence boiling
(III). In region I, there is little coalescence of departing bubbles and not all potential nucleation sites
on the surface are active, except near the transition to region II. The surface temperature in the
discrete bubble region increases with increasing the input power, at a higher rate than in region II, in
which nucleate boiling heat transfer rate is the highest and the slope of the boiling curve is the
steepest. Near the end of this region, lateral coalescence of departing bubbles begins to affect the
ebullition cycle and decrease the heat removal rate from the surface. In the lateral coalescence

5th European Thermal-Sciences Conference, The Netherlands, 2008
region (III), as the bubbles continue to coalesce and grow larger near the heated surface, the boiling
heat transfer rate progressively decreases and the surface temperature increases, as indicated by the
gradually decreasing slope of the boiling curves in region III. Eventually, the formation of batches
of vapour on and near the heated surface causes a surge in surface temperature by as much as 45 K,
marking CHF; CHF is indicated by the last points on the boiling curves.
Figure 2: Boiling curves on plane Cu and Cu with 5 mm tall corner pins for subcooled HFE-7100.
4.1 Natural convection
Figure 2 shows the natural convection data on plane surfaces and Cu surfaces with corner pins.
Since the 10 x 10 mm footprint area is uniformly heated, the removed thermal power by natural
convection, PNC, is proportional to the surface superheat, ΔTp, raised to the 1.2 power. The results
in Figure 2 show that the heat removal rate by natural convection from the surfaces with 3 x 3 mm
corner pins is on average 67.5% higher than from plane surfaces having the same foot print area (10
x 10 mm) at the same surface superheat, ΔTp. These results are important to the cooling of
computer chips in the standby mode, in which the thermal power dissipation is typically < 3 W/cm 2 .
At such heat fluxes, the surfaces with corner pins will be ~ 20 K cooler than plane surfaces.
Figure 3: Natural convection data and correlations for plane Cu and Cu with 3 x 3 mm corner pins.
4.2 Effect of liquid subcooling and inclination angle
The obtained boiling curves for θ = 0 o are shown in Figures 4a – 4d for saturation and 10 K, 20 K,
and 30 K subcooling. The CHF values increase with increased subcooling and so does the
corresponding surfaces temperature. For the surface with 5 mm pins, CHF is as much as 93 W/cm 2

5th European Thermal-Sciences Conference, The Netherlands, 2008
at ΔTsub = 30 K, compared to only 53 W/cm 2 for saturation boiling (Figure 4d). As indicated
earlier, for the Cu surfaces with corner pins there is no excursion in the footprint temperature prior
to boiling incipience (Figures 4b – 4d). Figures 5a and 5b compare the pool boiling curves for
saturation boiling of HFE-7100 liquid on Cu with 2 mm and 5 mm tall corner pins.
Figure 4: Effect of liquid subcooling on pool boiling on plane Cu and Cu with corner pins.
Figure 5: Effect of inclination angle on pool boiling on Cu surfaces with corner pins.
Generally, the removed power by nucleate boiling at low surface temperatures (65 to 68 o C)
increases slighly with increased surface inclination; the highest values are for the downward-facing
orientation (θ = 180 o ). For the surface with 5 mm pins, CHF at 30 K subcooling decreases from 93
W/cm 2 at 0 o to only 86 W/cm 2 at 180 o (Figure 5d). Conversely, at higher surface temperatures, the

5th European Thermal-Sciences Conference, The Netherlands, 2008
heat removed by nucleate boiling increases with decreased inclination; the highest is in the upwardfacing
orirntation (θ = 0 o ). CHF increases with decreased surface inclination and are highest at θ =
0 o and corresponding surface temperature is sensative to the surface orientation, with a relatively
large uncertainties of + 3 o C. Considering such uncertaities, the results in Figure 5a and 5b suggest
that the surface temperature at CHF changes slightly with surface inclination, but increases as the
height of the corner pins increases; the same is also true for CHF. Figures 5a and 5c reveal the
effect of subcooling on the boiling curves for the surface with 2 mm tall corner pins and Figures 5b
and 5d for the surface with 5 mm tall pins. Increasing the liquid subcooling increases not only the
power removed by nucleate boiling, at the same average temperature of the footprint area, but also
CHF. Increasing liquid subccooing also increases the average temperature of the footprint area at
CHF by ~ 3 – 4 o C for the surface with 2 mm tall corner pins and by as much as 15 o C for the
surface with 5 mm tall corner pins.
4.3 Critical heat flux
The present values of saturation boiling CHF based on the footprint area of 10 x 10 mm, on plane
copper at θ = 0 o are correlated, using a form similar to that suggested by Kutateladze (1961):
sat
o
o 0.
5
( 0 ) C ( 0 ) ρ h σ g ( ρ − )
0.
25
[ ] .
CHF = l ρ
(1)
sat , plane
v fg
v
The coefficient, Csat, plane (0 o ) determined from the least squares fit of CHF data is 0.196 for HFE-
7100 and 0.166 for FC-72 (Figure 6a). To account for effects of θ and ΑΡ, Equation (1) becomes:
o
( , AR)
CHF ( 0 ) F ( θ ) F ( AR)
.
CHFsat sat sat sat
θ = (2)
Fsat (θ) (Figure 6b) accounts for the effect of inclination and is correlated from present data as:
−5
−7
2
−7
3
( θ ) = 1− 6.
31x10
θ − 4.
97x10
θ −1.
25x10
θ
F sat , for plane Cu (AR = 1). (3a)
−11
−6
2
−8
3
( θ ) = 1− 8.
27x10
θ − 8.
26x10
θ − 3.
42x10
θ
F sat , for Cu with 2 mm pins. (3b)
−5
−6
2
−8
3
( θ ) = 1− 4.
03x10
θ − 5.
12x10
θ − 2.
51x10
θ
F sat , for Cu with 3 mm pins. (3c)
−7
−6
2
−11
3
( θ ) = 1− 1.
22x10
θ − 4.
76x10
θ −1.
59x10
θ
F sat , for Cu with 5 mm pins. (3d)
Fsat (AR) in Equation (2) accounts for the effect of the corner pins on saturation CHF (Figure 6c); it
is unity for plane copper and for macro-stuctured Cu surfaces (Rainey and You, 2000; Yu and Lu,
2007) and present surfaces with 3 x 3 mm corner pins, it is correlated as:
F sat
F sat
( AR)
0. 868 + ( 0.
139 AR)
= , for FC-72 liquid. (4a)
( AR)
0. 655 + ( 0.
348 AR)
= , for HFE-7100 Liquid. (4b)
For subcooled boiling CHF, Equation (2) is rewritten as:
CHF θ , AR,
Δ T = CHF θ , AR 1 + C ( θ ) ΔT
. (5)
sub
( ) ( ) ( )
sub
sat
CHF increases linearly with increased subcooling, ΔTsub. and the subcooling coefficent, Csub (θ),
(Figure 6d) is correlated based on the present data for plane Cu and Cu surfaces with corner pins as:
−6
2
−9
3
−11
4
Csub ( θ ) = 0.
024 + 1.
04x10
θ − 7.
36x10
θ + 1.
52x10
θ . (6)
The value of this coefficent in the upward-facing orientation, Csub( 0 o ), obtained from the least
square fit of the present data is 0.024 K -1 and increases with increased inclination to 0.0308 K -1 at
180 o . Equation 5 is within + 10% of the present and reported data for macro-structured surfaces.
sub
sub

5th European Thermal-Sciences Conference, The Netherlands, 2008
Figure 6: Effects of inclination angle, areas ratio, and subcooling on CHF for HFE-7l00 Liquid.
5 Summary
The conducted experiments investigated saturation and 10 – 30 K subcooled boiling of HFE-7100
dielectric liquid on Cu surfaces with 3 x 3 mm corner pins, 2, 3 and 5 mm tall. The effect of
surface orientation from 0 o to 180 o (downward-facing) on nucleate boiling heat transfer and CHF
are also investigated. Results show significant increases in the total thermal power removed from
the surfaces with corner pins, compared plane copper of the same footprint. The thermal power
removed at CHF from the surface with 5 mm tall corner pins in the upward-facing orientation (θ =
0 o ) with 30 K subcooled liquid is 93 W, decreasing only to 86 W at 180 o . The values of CHF on
plane Cu and the surfaces with 3 x 3 mm corner pins increases linearly with increased liquid
subcooling. The rate of increase in CHF in the surface with 5 mm tall corner pins rises from
2.4%/K at 0 o to 3.03%/K at 180 o . These values of the subcooling coefficient are independent of the
height of the corner pins.
A correlation for natural convection heat transfer prior to boiling incipience is developed based on
the present data and compared to that for plane surfaces of the same footprint. This correlation,
which is within + 12% of the data, shows a 67.5% increase in the heat removal rate, compared to
plane surfaces. The CHF correlation developed for plane and micro-structured surfaces accounts
for both surface orientation and liquid subcooling and is within + 12% of the present data and that
reported for other investigators for micro-structured surfaces.

5th European Thermal-Sciences Conference, The Netherlands, 2008
References
Al-Hajri et al., 2005, Thermal Performance of Micro-Structured Evaporation Surfaces-Application
to Cooling of High Flux Microelectronics, Proceedings ASME International Mechanical
Engineering Congress and Exposition, 799-805.
Anderson, T. M. and Mudawar, I., 1989, Microelectronic Cooling by Enhanced Pool Boiling of a
Dielectric Fluorocarbon Liquid, J. Heat Transfer, 111, 752 - 759.
Arik, M., Bar-Cohen, A. and You, S. M., 2007, Enhancement of Pool Boiling Critical Heat Flux in
Dielectric Liquids by Microporous Coatings, Int. J. Heat & Mass Transfer, 50, 997-1009.
Chang, J. Y. and You, S. M., 1996, Heater Orientation Effects on Pool Boiling on Micro-Porous-
Enhanced Surfaces in Saturated FC-72, J. Heat Transfer, 118, 937-943.
El-Genk, M. S. Bostanci, H., 2003, Combined Effects of Subcooling and Surface Orientation on
Pool Boiling of HFE-7100 from a Simulated Electronic Chip, J. Experimental Heat Transfer, 16,
281-301.
El-Genk, M. S. and Parker, J. L., 2005, Enhanced Boiling of HFE-7100 Dielectric Liquid on Porous
Graphite, J. Energy Conversion and Management, 46, 2455 – 2481.
Jung, J.-Y., and Kwak, H.-Y., 2006, Effect of Surface Condition on Boiling Heat Transfer from
Silicon Chip with Submicron-Scale Roughness, Int. J. Heat & Mass Transfer, 49, 4543-4551.
Kutateladze, S. S. "Boiling Heat Transfer," Int. J. Heat & Mass Transfer, 4, 31- 45, 1961.
Launay, S., et al., 2006, Hybrid Micro-Nano Structured Thermal Interfaces for Pool Boiling Heat
Transfer Enhancement, Microelectronics Journal, 37, 1158-1164.
Misale, M., Guglielmini, G., Frogheri, M. and Bergles, A. E., 1999, FC-72 Pool Boiling from
Finned Surfaces Placed in a Narrow Channel: Preliminary Results, Heat and Mass Transfer, 34,
449-452.
Nakayama, W., et al., 1980, Dynamic Model of Enhanced Boiling Heat Transfer on Porous
Surfaces Part I: Experimental Investigation, J. Heat Transfer, 102, 445-450.
Parker, J. L. and El-Genk, M.S., 2005, Enhanced Saturation and Subcooled Boiling of FC-72
Dielectric Liquid, Int. J. Heat and Mass Transfer, 48, 3736 – 3752.
Parker, J. L. and El-Genk, M.S., 2006, Subcooled Boiling of Dielectric Liquids on Porous Graphite
at Different Orientations, Proceedings 13 th Int. Heat Transfer Conference, Paper No. BOI-36,
Sydney, Australia.
Parker, J. L., and M. S. El-Genk, M. S., 2007, Enhanced Saturation Boiling of HFE-7100 Dielectric
Liquid on Extended Copper Surfaces, Proceedings ASME-JSME Thermal Engineering Summer
Heat Transfer Conference, Vancouver, British Colombia, Canada.
Rainey, K. N. and You, S. M., 2000, Pool Boiling Heat Transfer from Plain and Microporous,
Square Pin-Finned Surfaces in Saturated FC-72, J. Heat Transfer, 122, No. 3, pp. 509-516.
Rajulu, K. G., et al., 2004, Enhancement of Nucleate Pool Boiling Heat Transfer Coefficient by
Reentrant Cavity Surfaces, Heat and Mass Transfer, 41, 127 – 132.
Watwe, A. A., Bar-Coen, A., and McNeil, A., 1997, Combined pressure and Subcooling Effects on
Pool Boiling From a PPGA Chip Package, J. Electronics Packaging, 119, 95 – 105.
Webb, R. L., 2004, Odyssey of Enhanced Boiling Surface, J. Heat Transfer, 126, 1051-1059.
Wei, J. J. and Honda, H., 2003, Effects of Fin Geometry on Boiling Heat Transfer from Silicon
Chips with Micro-Pin-Fins Immersed in FC-72, Int. J. Heat & Mass Transfer, 46, 4059-4070.
Yu, C. K. and Lu, D. C., 2007, Pool Boiling Heat Transfer on Horizontal Rectangular Fin Array in
Saturated FC-72, Int. J. Heat and Mass Transfer, 50, 3624-3637.