The document discusses research on enhancing critical heat flux (CHF) through the use of nanofluids and surfaces with nanostructures. Key findings include:
1) Prior studies found nanofluids and nanoparticle-coated surfaces can significantly increase CHF, with effects attributed to enhanced wettability, liquid spreadability, and multi-scale surface geometry from nanoparticles.
2) The authors fabricated test surfaces with micro-posts, zinc oxide nanorods, or a combination of both, to further study how wettability, liquid spreadability, and multi-scale geometry influence CHF.
3) Preliminary results showed surfaces with both micro-posts and zinc oxide nanorods had the lowest contact angle and thus highest wet
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
ppt
1. Effects of nano-fluid and surfaces with nano structure on the increase of CHF
Seontae Kim, Hyung Dae Kim, Hyungmo Kim, Ho Seon Ahn, Hangjin Jo, Joonwon Kim, Moo Hwan Kim *
Dept. of Mechanical Engineering, Pohang University of Science and Technology, San 31, Hyoja-dong, Namgu, Pohang, Kyoungbuk 790-784, Republic of Korea
a r t i c l e i n f o
Article history:
Received 5 November 2008
Received in revised form 26 May 2009
Accepted 26 May 2009
Keywords:
Critical heat flux (CHF)
Nano-fluid
Micro-posts
Nano structure
Pool boiling
Wettability
Liquid spreadability
a b s t r a c t
Critical heat flux (CHF) has necessitated inconvenient compromises between economy and safety in most
industries related to thermal systems. Recent development of nanotechnology has enabled synthesis of
nano-sized particles and development of new heat transfer fluids with suspended nano-sized particles,
i.e., nanofluids. When nanofluids were used in boiling heat transfer cooling, anomalous increase of CHF
was reported. Subsequently, nanoparticle deposition on the boiling surface was revealed to contribute
to CHF enhancement. Research on surface characteristics determined that three major characteristics
affect CHF: wettability, liquid spreadability and multi-scale geometry. We fabricated artificially modified
surfaces with arrays of octagonal micro-posts, or ZnO nanorods, or both, and measured their performance
in enhancing CHF. The presence of three major characteristics enhanced CHF most.
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
The problem of cooling has become an increasingly critical prob-
lem in nuclear industry and electrical chip cooling. One cooling
method is boiling heat transfer, which exploits the latent heat of
vaporization during the liquid-to-gas phase change, and is the most
effective way to cool thermal systems running at high temperatures.
However, boiling heat transfer has an inherent limitation: CHF
(critical heat flux). CHF is the maximum heat flux where boiling
heat transfer sustains its high cooling efficiency. When a surface
reaches CHF, it becomes coated with a vapor film, which interferes
with contact between the surface and the ambient liquid, and de-
creases heat transfer efficiency. System temperature rises, and if it
exceeds the limits of its constituent materials, system failure oc-
curs. For this reason, every system incorporates a safety margin
by running at heat flux much lower than CHF, but this approach re-
duces the system’s efficiency. This compromise between safety and
efficiency is an important problem in industry. For this reason, to
guarantee a large safety margin while extending operation region,
there have been diverse approaches to understand the mechanism
of CHF and also to enhance CHF point.
1.1. Approaches to CHF enhancement
Many methods for CHF enhancement have been suggested.
We will focus only on methods of changing conditions at the
heat exchange surface. These include surface modification by
oxidation, etching, sanding, polishing, machining and attaching
foreign materials to the surface. Surface modification controls
surface characteristics: surface chemistry and geometry. Usually
contact angle was measured for estimating surface energy and
openly stated as level of wettability, which is proportional to
CHF. For controlling surface chemistry, various techniques
including oxidation [1–4], etching, UV irradiation [5] were con-
ducted and showed noticeable CHF enhancement with higher
wettability. Another approach is to increase surface roughness.
Some studies have shown enhanced CHF [4], but some have
not [6]. So, the usefulness of increasing surface roughness to en-
hance CHF is doubtful. But it is certain that changing roughness
affects surface characteristics, like contact angle to some extent
[7]. The third approach to enhance CHF is surface machining. Re-
search on this approach has yielded inconsistent results. Some
reports showed increase of CHF [8–10] and some showed no ef-
fect [8]. Though many studies showed that increased density of
nucleation sites may enhance heat transfer enhancement in the
nucleate boiling region, they can not guarantee CHF enhance-
ment. The fourth approach is to attach foreign materials to the
surface. In some cases, the foreign materials are mainly com-
posed of wicking material [11,12] which induces capillary inflow
of the cooling liquid to the boiling area. In other cases, the for-
eign material is a porous layer [13,14] composed of particles of
various sizes and materials, which is intended to affect hydrody-
namics or to induce capillary inflow of liquid. And recently,
some research groups developed artificial structures which has
multiple of these approaches.
0894-1777/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.expthermflusci.2009.05.006
* Corresponding author. Tel.: +82 54 279 2006; fax: +82 54 279 3199.
E-mail address: mhkim@postech.ac.kr (M.H. Kim).
Experimental Thermal and Fluid Science 34 (2010) 487–495
Contents lists available at ScienceDirect
Experimental Thermal and Fluid Science
journal homepage: www.elsevier.com/locate/etfs
2. 1.2. CHF enhancement with nanofluids
Nanotechnology has made it possible to produce many kinds of
nano-scaled particles (nanoparticles). Materials downscaled to
nano-size showed superior properties (including mechanical, opti-
cal, electrical, and thermal) than when they were of conventional
size. In 1995, Choi introduced the concept of nanofluids: ‘‘. . . a
new class of nanotechnology-based heat transfer fluids engineered
by dispersing and stably suspending nanoparticles with typical
length scales on the order of 10 nm in traditional heat transfer flu-
ids” [15]. Many researchers have since utilized the outstanding
properties of nanofluids. One finding that is valuable in the heat
transfer field is anomalous CHF enhancement with nanofluids.
You et al. [16] reported 2–3-fold enhancement in CHF when a very
small amount of alumina nanoparticles was suspended in water.
Many subsequent studies reported similar results with various
nanoparticles [17,18]. By producing the same or higher CHF
enhancement with a nanoparticle-coated wire heater in pure
water compared with nanofluids (Fig. 1), our group revealed that
CHF enhancement is solely dependent on nanoparticle deposition
on the boiling surface [19]. We proposed that roughness, wettabil-
ity and capillary wicking contribute to delaying CHF phenomena
[19]; this idea was later confirmed [20]. Our group also suggested
another important parameter, capillary wicking [21,22]. We no-
ticed fast and wide liquid spreading on nanoparticle-coated wire
surfaces. So we measured the capillary wicking length on vertically
erected wires in a water reservoir. After accounting for the capil-
lary wicking effect, the remaining data which could not be ex-
plained by wettability became clear. Consequently, it became
clear that wettability and liquid spreadability (capillary wicking)
are both important parameters that contribute to CHF increase
nanoparticle-coated surfaces.
The effects of roughness on CHF enhancement were not easily
determined. The surface has a very complex and fractal geometry
with micro and nano scales in hierarchical surface topography
(Fig. 2). Because it is known that surface roughness can control
the degree of wettability [23,24], the wettability (moreover, liquid
spreadability) of nanoparticle-coated surfaces must have been
amplified by its complex geometry. For these reasons, it can be said
that the nanoparticle-coated surface is a kind of well modified sur-
face with amplified critical surface characteristics which may have
major effects on CHF enhancement.
1.3. Next steps for better surface
From this reasoning, we deduced two important facts. The first
is that wettability and liquid spreadability are critical surface
parameters which may enhance CHF. The second is that multi-
scaled surface structure amplifies the characteristics of these sur-
face parameters.
Although nanoparticle-coated surfaces show all of these useful
characteristics and although nano-fluid pool boiling also can be
used as a new surface coating technique, nano-fluid pool boiling
processes and nanoparticle-coated surfaces have inherent limits
in practical uses and real applications, because they are difficult
to control and are dependent on natural phenomena. So, we
decided to make artificial surfaces which have the characteristics
of wettability, liquid spreadability, and multi-scaled geometry, to
confirm their contributions to CHF and to determine the optimum
conditions of these characteristics. As a first test we fabricated four
simple and basic artificial surfaces using the microelectromechan-
ical systems (MEMS) technique and obtained some useful results
by comparing their CHF performances.
2. Experiments and results
2.1. Test heater fabrication
To facilitate both CHF test and surface modification, we embed-
ded a thin film heater on one side of a silicon wafer and created
artificial surfaces on the other side of the wafer by using the MEMS
technique. Although the fabrication processes of the film heater
side and surface modification were conducted simultaneously,
we will describe them separately for convenience of explanation.
The test heater is a rectangular silicon wafer plate (Fig. 3). The
substrate silicon plate is 25 mm  20 mm, and has a SiO2 layer
for electrical insulation on both the top and bottom. Because Joule
Nomenclature
q00
heat flux, kW/m2
q00
CHF Zuber critical heat flux by Zuber’s correlation, kW/m2
h heat transfer coefficient, kW/m2
K
V loaded electric voltage, V
Icircuit current passing through a test sample, A
Aheater area of a test sample, m2
R electric resistance, ohm
T temperature, °C
a(b) constant
hlg latent heat of vaporization, kJ/kg
r surface tension, N/m
q density, kg/m3
Subscripts
reference properties of a reference resistance
heater properties of heater
sat properties at saturation
lg latent quantity
l liquid
g gas
Fig. 1. Comparison of CHF enhancements of pure water on a nanoparticle-coated
heater, nanofluids on a bare heater, and nano-fluids on a nanoparticle-coated heater
[19].
488 S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495
3. heating was selected for this experiment, insulation layers were
needed to exclude the possibility of electrical interference on both
modified surface structures and of hydrodynamics on the top. For
heating part, a titanium thin film of about 1000 Å thickness was
layered on the bottom of the substrate using an E-beam evapora-
tor. The completed titanium film heater has ‘H’ shape. The center
bar of ‘H’ is 15 mm  10 mm with a vertical bar at each end. The
center bar is the main heat generating area, and we confirmed by
simple calculation that it generates more than 99% of the total
heat. The vertical bars at ends are used for wire connection. Be-
cause it is impossible to use lead soldering to connect copper wires
to a titanium film, an additional gold (Au) film (1500 Å thickness)
was deposited on each vertical bar using an E-beam evaporator.
With lead soldering and Au film on the vertical bars, their resis-
tance becomes smaller than that of the center bar and guarantees
that heat generation is focused in the center bar.
2.2. Test surface fabrication
Four different surfaces (F, M, N, NM) were fabricated by MEMS
techniques on top of the test heaters. F is a reference surface which
is flat, without special geometry or surface modification. This sur-
face was used as the standard to which properties and perfor-
mance of modified surfaces were compared.
M is a modified surface with micro-scale geometry (Fig. 4). This
image was obtained using a Scanning electron microscope (JSM-
7401F /JEOL). The M surface was prepared to create multi-scaled
geometry, and also to determine its CHF enhancement ability by
comparing it with other surfaces. For the M surface, silicon bulk
wet etching was applied to a (1 0 0) silicon substrate using tetra-
methyl ammonium hydroxide etchant. M surface is covered by
an array of micro-scaled posts. Each post has an octagonal cross
section which narrows with height. The longest diagonals of the
posts are 26 lm at the base and 21 lm at the top. The slopes of
post are discontinuous due to limitations of the fabrication process.
The posts are 6 lm heights and are separated by 50 lm in both
directions on the surface.
N is a modified surface with nano-scale geometry (Fig. 5). The N
surface was prepared to create nano-scale geometry and to deter-
mine how the wettability and liquid spreadability of this surface
affect CHF. For the N surface, an array of ZnO nanorods was se-
lected. The rods were grown as described in [25]. To fabricate the
nanorod array, a ZnO layer was prepared on the top of the test hea-
ter. This layer was used as a seed layer for fabrication of ZnO nano-
rods. Next, zinc nitrate hexahydrate (Zn(NO3)2Á6H2O, 98%) and
ammonium hydroxide (28 wt% NH3 in water, 99.99%) were used
to grow ZnO nanorods. ZnO nanorods on the N surface grow very
close to each other. The heights of the nanorods range from several
hundred nanometers to over 1 lm, with a mode around 800–
900 nm. Because they are grown in solution and are approximately
perpendicular to the seed layer, ZnO nanorods become narrower
with height.
Finally, NM is a surface with multi-scaled geometry. It has an
array of micro-scale posts on the SiO base, as on the M surface,
and nano-scale ZnO nanorods which totally cover these micro-
posts (Fig. 6). NM is microscopically the same as M surface, and
nanoscopically the same as N surface. Especially, ZnO nanorods oc-
cur on the slopes of micro post as well as on its top and bottom.
The side slope coverage is distinguished character of ZnO nanorods
array by growth method from ordinary tries like nanostructure
deposition by adhesives. Dimensions of multi-scaled structures
on NM surface are the same as those of the M and N surfaces.
So, four different test surfaces are prepared for estimating and
analyzing the surface effects on CHF. However, the end materials
of each test surface, which make contact with operating liquid
and provide boiling sites, are different. F and M surfaces have Si
on it, while N and NM surfaces have ZnO for their nanorods array.
To make same surface chemical condition, F and M surfaces were
covered with an additional ZnO layer.
2.3. Characteristics of test surfaces
We compared wettability, liquid spreadability and surface
geometry (multi-scale geometry) of each test surfaces to analyze
how each of these parameters affects CHF. Roughness and area
enhancement were also measured.
To estimate the surface wettability of each surface, static con-
tact angle was measured by using a drop-shape analysis system
Fig. 2. TiO2 nanoparticle-coated NiCr wire after pool boiling CHF experiment with nanofluids [21].
Fig. 3. Schematic of the test heater (bottom and side views).
S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495 489
4. (KRUSS DSA 100). In all measurements, 2 ll of water was used to
form a droplet on the test surface. Static contact angles differed
on the various surfaces (Fig. 7). All the values are averaged over five
measurements, and uncertainty is about ±3°. The F and M surfaces
have almost the same wettability. This means that the array of mi-
cro-scale posts does not affect the wettability of the surface. The
contact angle values of water droplets on the N and NM surfaces
could not be measured. As soon as the water droplet contacts these
surfaces, it spread out. In the case of the N surface, because water
spreading was not complete, the water droplet formed a gentle
curve (Fig. 7c). On the NM surface, the water droplet spread uni-
formly over the whole surface. The N and NM surfaces both include
ZnO nanorods, whereas the F and N surfaces do not. Therefore, the
high wettability is probably caused by these nanorods.
The second parameter is liquid spreadability. No conventional
technique to measure this quantity exists. The capillary wicking
height measuring method [22] did not work well with nanoparti-
cle-coated wires. Because the test surface is a 2-D plate, the irreg-
ular shape of the liquid head mark could not be defined accurately.
Therefore, we measured spreadability qualitatively by the amount
by which a 1 ll of water drop spread in 1 sec, after contact with the
surfaces (Table 1). The data were obtained by capturing images
with a digital camcorder. The contact direction was vertical and
the positioning of surface was horizontal. For the F and M surfaces,
there was no noticeable water spreading in 1 min. On the N sur-
face, the drop spread to cover about half of the test surface. On
the NM surface, the drop spread out until it covered the whole test
surface. Moreover, liquid spreading velocity was also much larger
Fig. 4. SEM images of M surface (a) 500 lm, (b) 2500 lm.
Fig. 5. SEM images of N surface (a) 35,000 lm, (b) 100,000 lm.
Fig. 6. SEM images of NM surface (a) 500, (b) 2500, (c) 5000, (d) 25,000 lm.
490 S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495
5. on NM surface than on the N surface. We postulate that liquid
spreading is the result of surface characteristic amplification in
multi-scale geometry on the NM surface.
Surface roughness was measured using an atomic force micro-
scope (AFM) (Dimension 3100 VEECO DI). The root mean square
(Rq) roughness value depended on whether or no the surface in-
cluded ZnO nanorods (Table 1). Because the F and M surfaces were
surface-coated with ZnO by sputtering, they inherently have clean
surfaces of 0.00 lm roughness. Meanwhile, the N and NM surfaces
have ZnO nanorods on their surfaces and these result in 0.07 lm
roughness for both surfaces.
Finally, surface area enhancement was calculated manually (Ta-
ble 1). Approximate surface area was calculated from the scale bar
and tilting angle information of each surface’s SEM images. The
geometry of the M surface is simple, so the calculated surface area
increase due to the micro-scale posts array is rather exact, but the
area calculation of the N and NM has uncertainty due to the more
complex geometry of the array of ZnO nanorods. This is not a prob-
lem, because the enhancement ratio is so much larger than the M
surface. The contribution of the array of ZnO nanorods on surface
area increase (50Â the surface area of the F surfaces) was much lar-
ger than that of the array of micro-scale posts (1.2Â). To estimate
increase of the NM surface, these two increase ratios were
multiplied.
2.4. Test facility and measurement system
The test facility (Fig. 8) is designed for a horizontal upward ori-
entation pool boiling experiment under atmospheric pressure with
water, and consists of a test sample jig, a main test pool, and a lid
with an immersion heater and a condenser. The test sample jig is
designed to hold the test sample in a test pool and to make chang-
ing the test sample easy. The jig (Fig. 9) is consists of an anodized
aluminum cylinder base and a polyetheretherketone (PEEK) test
sample frame. PEEK is a thermoplastic which has high thermal
resistance and is compatible with an aqueous environment. The
base and the frame are connected by bolts and O-rings. The test
sample frame has a framework for sample mounting. Test sample
was fixed and waterproofed with adhesive sealant (PermatexÒ
clear RTV silicon) and 2-tone epoxy (DuralcoTM 4461).
The main test pool is an octagonal aluminum cylinder bath with
four glass windows, one on each side wall. It is 165 mm high and
has 3 l capacity. Through a large hole on the bottom of pool, the
test sample jig can be interconnected and separated before and
after each test. The main test pool is connected to the aluminum
lid plate by bolts and O-rings. The lid is an aluminum plate which
has an immersion heater below it and a reflux condenser above it.
The immersion heater is operated by a proportional integral-deriv-
Fig. 7. Mean (n = 5) static contact angle images of various test surfaces. (a) F
surface, (b) M surface, (c) N surface, (d) NM surface. The contact angle value of each
test surface is marked on the upper right of each image. Uncertainty is about (+/À)
3°.
Table 1
Characteristics of surfaces tested.
Characteristic Sample Surface
F M N NM
Spread diameter (mm) 2.1 2.1 4.3 10.8
Surface roughness (lm) 0.0 0.0 0.07 0.07
Surface area ratio (vs. F) 1 1.15 $50 $55
Surface temp. at onset of nucleate boiling
(°C)
1 113 112 114 112
2 116 113 110 114
Critical heat flow (kW/m2
) 1 1052 1563 2016 2346
2 1190 1740 1990 2305
Fig. 8. Schematic of pool boiling test facility and measurement system (a) test
sample jig component, (b) aluminum test pool, (c) immersion heater, (d) reflux
condenser, (e) cooling water line to condenser, (f) K-type thermocouple for PID
control, (g) immersion heater PID controller unit, (h) T-type thermocouple for pool
temperature measurement, (i) data acquisition system, (j) reference resistance in
constant temperature beaker, (k) DC power supply unit.
(a)
(b)
(g)
(h)
(e)
(d)
(c)
(f)
(i)
(j)(j)
(a)
(b)
(g)
(h)
(e)
(d)
(c)
(f)
(i)
(j)(j)
Fig. 9. Detailed schematic of test sample jig component (a) test sample, (b) test
sample frame, (c) O-ring, (d) 2-tone epoxy, (e) RTV sealant, (f) bolts, (g) aluminum
base, (h) power line, (i) voltage-measure line, (j) wire passage holes.
S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495 491
6. ative (PID) controller to maintain pool temperature at saturation
temperature. A reflux condenser prevents water loss from the test
pool by evaporation, and maintains atmospheric pressure inside
the pool.
The measurement system consists of a power supply, a refer-
ence resistance unit and a data acquisition system. Because we
use Joule heating to generate heat flux through the test sample,
an ‘HP Agilent 6575A DC power supply (120 V/18A)’ was used to
supply power. The heat flux is calculated using the following
equation
q00
¼
VheaterIcircuit
Aheater
ð1Þ
Vheater is the voltage loaded on each side of the test sample. It is
measured directly using soldered voltage-measure wires. Icircuit is
the current running through the series circuit, which has the test
sample and the reference resistance unit as components. It is calcu-
lated using the equation
Icircuit ¼
Vreference
Rreference
ð2Þ
Vreference is the voltage loaded over the reference resistance unit and
Rreference is the total resistance value of the reference resistance unit
(1.0223 0.0005 ohm). Generally the resistor’s resistance increases
with surrounding temperature, whether it comes from the environ-
ment or its own heat generation. To guarantee constant reference
resistance, the reference resistance unit consists of nine precision
wire wound resistors, a cooling coil and a constant temperature
beaker. The resistors have a very low temperature coefficient of
resistance of 5 ppm/°C and they are immersed in a constant temper-
ature beaker containing R-113 refrigerant with a cooling coil which
carries 10 °C water from a constant temperature bath. This refer-
ence resistance unit guarantees precise measurement of current
by detecting the voltage over it during real operation. A data acqui-
sition system (HP Agilent 34970A Data Acquisition/Switch Unit)
was used to gather the data from the test facility. This system reads
inputs of Vheater, Vreference and Tbulk. Tbulk is the pool temperature and is
measured using a T type thermocouple to check the saturation state
of the pool.
2.5. Operating procedure
Before the pool boiling test, we derived a temperature–resis-
tance conversion equation for all the test samples. This equation
was used to estimate the surface temperature of the test sample
during pool boiling by measuring the sample’s resistance. A con-
stant temperature convection oven was used to establish various
constant temperatures. The calibration was conducted for 5–6
temperatures between 100 and 150 °C, on samples which were
in thermal equilibrium. The calibration Eq. (3) is linear over the
range of interest. From this process, we obtained the values of con-
stants and in Eq. (3).
Rheater ¼ aTheater þ b ð3Þ
Rheater can be calculated directly from Ohm’s law using measured
and calculated values of Vheater and Icircuit. Then, the wall superheat
(DTheater) is calculated using the following equation
DTheater ¼ Theater À Tsat ¼
1
a
ðRheater À RsatÞ ð4Þ
Rsat is the resistance at the saturated temperature. By using Eq. (4),
the wall superheat of test sample during the whole test can be eas-
ily estimated. The heat transfer coefficient is calculated by Newton’s
law of cooling
h ¼ Q00
=DTheater ð5Þ
After temperature-resistance calibration and preparation of the
test facility, the pool boiling test was conducted. First the working
fluid (de-ionized water) was degassed by heating it to saturation
temperature (100 °C) and maintaining it at that temperature for
2–3 h prior to the test. Thereafter, the test was conducted by
increasing the electric power supplied to the test sample in small
steps up to the CHF point. Every data point obtained for plotting
the boiling curve is the average of more than 100 measurements
of heat flux and temperature at the thermal equilibrium state.
The power was increased, and at a certain point temperature of
test surface increased rapidly. We judged that this was the CHF
point and shut down the power before burning out the test sample.
Tests were conducted twice for each surface type; new samples
were used each time. The uncertainty in the measurement was
estimated as less than 10 kW/m2
in heat flux and 0.5 °C in surface
temperature.
2.6. Experimental results
For all surfaces, the heat transfer characters in the free convec-
tion regime are similar. We conjecture that the modified micro or
nano structures could not act as effective fins, because they are too
small to extend beyond the thermal boundary layer.
There is no significant difference among surfaces in onset of
nucleate boiling (ONB) (Table 1). Temperature data are scattered
around 110–115 °C and there is no noticeable tendency among
them.
For nucleate boiling, the M surface shows better heat transfer
characteristic than the F surface, but the boiling curves of the N
and NM surfaces lie between those of the M and F surfaces (Figs.
10 and 11). Therefore, it can be considered that a surface transfers
heat better in the nucleate boiling regime when micro post arrays
are added. The nano-scaled structure (surfaces N and NM) certainly
has some good effect on boiling heat transfer, but it is less than the
effect of the M surface.
Every modified surface showed noticeable CHF enhancement
compared with the reference F surface (Fig. 10). The average CHF
was 1121 kW/m2
for F, 1652 kW/m2
for M, 2003 kW/m2
for N
and 2326 kW/m2
for NM (Table 1). Differences in CHF estimates
on samples of the same surface differed by <15%. The NM surface
showed the most outstanding enhancement (107% increase in
CHF compared to the F surface), followed by the N (79% increase)
and M surfaces (47% increase). The array of micro-scaled posts
caused CHF enhancement of 531 kW/m2
(from F surface to M sur-
face) and 323 kW/m2
(from N surface to NM surface). The nanorods
NM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
q"(kW/m2)
Theater ( C)
CHFNM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
q"(kW/m2)
Theater ( C)
CHF
Zuber’s correlation
NM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
q"(kW/m2)
Theater ( C)
CHFNM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
q"(kW/m2)
Theater ( C)
0 10 20 30 40 50 60
0
500
1000
1500
2000
2500
CHF
Zuber’s correlation
Fig. 10. Boiling curves of the four test surfaces. (The origin points of arrows are the
CHF points of each test sample.)
492 S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495
7. caused a CHF enhancement of 882 kW/m2
(from F surface to N sur-
face) and 674 kW/m2
(from M surface to NM surface). The test re-
sults were compared with the well known CHF models of Zuber
[26]. Zuber’s correlation for an upward-facing horizontal flat plate
is
q00
CHF Zuber ¼ 0:131hlgq0:5
lg ½rgðql À qgÞŠ0:25
ð6Þ
It predicts 1108 kW/m2
and agrees quite well with the CHF va-
lue (1121 kW/m2
) of the flat plate sample (F surface). Because Zu-
ber’s correlation does not consider the effects of surface
modification on test samples, it cannot explain the CHF enhance-
ment of the other test samples.
3. Discussion
Wettability, as measured by contact angle in Fig. 7 was better in
surfaces with nanorods (N, NM surfaces) than in those without (F,
M surfaces). Both the N and NM surfaces show very small contact
angles, indicating complete wetting. The surfaces with nanorods
also show a marked CHF enhancement in Table 1. Interestingly,
the increments are similar between pairs with the same base sur-
faces (F surface vs. N surface: 882 kW/m2
; M surface vs. NM sur-
face: 674 kW/m2
).
CHF increased with the nanorod structure which made contact
angle decrease on each base surface (F and M surface). Many pre-
vious experiments [4,11] have shown a negative dependency of
CHF on the contact angle and have reported that good wettability
enhances CHF. Most of these previous studies used the contact an-
gle as a parameter of wettability. Surfaces with a large contact an-
gle generally show bubble coalescence at low heat flux, and earlier
burnout. Costello and Frea [11], Fong et al. [27] said small contact
angle facilitates the approach of the liquid to the bubble base and
bubble departure, and results in smaller bubble departure diame-
ters; all of these phenomena are effective in delaying growth of
dry spots and enhancing CHF.
In 2001, Kandlikar [28] made a CHF correlation which consid-
ered wettability with contact angle and orientation. By Kandlikar’s
CHF correlation, the expected CHF enhancement by contact angle
decrease from 80° to 0° is about 800 kW/m2
, from 775 kW/m2
at
80° to 1572 kW/m2
at 0°. Fig. 12 plots experimental data of this
study together with Kandlikar’s correlation, showing a general
agreement with some degree of shift. Fig. 12 also shows plots of
modified curves. They are shifted with 400 and 800 kW/m2
to
match the reference points of CHF at around contact angle 80°.
Then, interestingly, each modified correlation is well matched with
the experimental data. Experimental data followed the general
trends between CHF and contact angle well showing good agree-
ment with Kandlikar’s correlation. However, there were questions
about the gap between the N and NM surfaces, and between the F
and M surfaces. Some answers for that question could be obtained
through following discussions.
In the case of wettability, there has been general consensus
about its effect on CHF. But, wettability is not a parameter which
can be measured directly. Although contact angle has been used
to describe wettability as a measurable parameter, it just touches
on one facet of wettability itself. Kim and Kim [22] reported that
there was something which could not be explained only with con-
tact angle, and suggested that capillary wicking influences CHF. It
is said that capillary wicking induces lateral liquid spreading to re-
wet the dry region beneath the vapor bubbles. Then, it can be said
that the important surface characteristic is liquid spreadability.
Capillary wicking would rather be one of the main mechanisms
which bring liquid spreading. Thus, the second surface characteris-
tic addressed in this study is liquid spreadability.
Surfaces with nanorods showed not only small contact angles
but also good liquid spreadability. As previously described, during
contact angle measurements, we observed N and NM surfaces
showed rapid spreading, with the water covering part of the sur-
face within several seconds. This property may be related to the
good liquid adhesion arising from the highly wettable nature of
nanorods structures and the excellent water spreading condition
due to a morphology that facilitates capillary action.
Liquid inflow to the heating surface is definitely important.
There are a few CHF mechanisms in the mainstream. Even though
they have their own approaches to CHF phenomena, there is no
dispute that the instability of liquid supply to the heating surface
brings CHF. At high heat flux, the ability to rewet the dry area
formed by vapor patches is very important. If the system is power
controlled, as heat flux emitted from heater surface increases, the
bubble generation will become more violent and severe. Around
CHF, vapors must already have been large mushrooms with coa-
lesced vapor columns. There must be dry areas under large mush-
rooms, and it becomes hard for ambient coolant to contact the
heating surface. Therefore, if the surface does not have enough
ability to rewet the dry area during the very short time period after
the departure of large mushrooms, it will be harder for that surface
to resist the burn out failure by hot spot generation. On the other
hand, if a surface has great ability to spread coolant liquid in a very
0
500
1000
1500
2000
2500
0 20 40 60 80 100 120 140 160 180 200
Contact angle ( )
CHF[kW/m2]
Kandlikar’s (2001)
Kandlikar’s + 400 kW/m2
Kandlikar’s + 800 kW/m2
F
M
NM
N
Fig. 12. Plot of CHF correlation of Kandlikar [28] and experimental data of various
surfaces.
h(W/m2K)
q" (kW/m2)
0
10000
20000
30000
40000
50000
60000
NM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
h(W/m2K)
q" (kW/m2)
0 500 1000 1500 2000 2500
0
10000
20000
30000
40000
50000
60000
NM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
NM surface - 1
NM surface - 2
N surface - 1
N surface - 2
M surface - 1
M surface - 2
F surface - 1
F surface - 2
Fig. 11. Heat transfer coefficient (h) vs. heat flux (q00
) for the four test surfaces.
S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495 493
8. short time period, the surface will have a greater margin for the
CHF occurrence.
According to Kandlikar’s CHF correlation, there must be al-
most no CHF enhancement with contact angle change below
10°, but Kim and Kim [22] showed that there is something more
beyond that. They explained this by the liquid spreading effect of
capillary wicking. Then, the CHF difference between N and NM
surfaces, which have almost the same contact angle, can be ex-
plained with their different liquid spreadability. At now, there
is no proper method which can measure liquid spreadability
quantitatively. However, it is certain that it has its own effect
on CHF which is distinguishable from contact angle and its exis-
tence has been proved experimentally. Therefore, developing a
good method to measure this parameter will be the next prob-
lem to be solved.
The last question is the CHF difference between F and M sur-
faces. The CHF increment of the M surface could not be explained
by wettability or spreadability at this time. The only difference
was surface geometry. Some previous research has hypothesized
about the mechanism of CHF enhancement with micro-size
structures, but none of them could explain this exactly. Anderson
and Mudawar [6] said that CHF enhancement with micro
structures comes from changed hydrodynamics by geometry, like
enhanced turbulence. Liter and Kaviany [29] used a micro-size
porous-layer coating to modulate hydrodynamics above the hea-
ter surface. On the basis of their researches, it was possible to
conjecture that the CHF difference between F and M surfaces
comes from the geometry effect on hydrodynamics modulation.
It is highly probable, but the effect of the micro-scale structures
on the CHF enhancement should be explained not only with the
F and M surfaces, but also with the N and NM surfaces. The CHF
difference between N and NM surfaces was explained with liquid
spreading which came from geometrical difference. Until now,
there is no explanation which can relate the effects of the mi-
cro-scale structures with both the hydrodynamics
modulation and liquid spreading. To support these contentions,
more delicate visualization of boiling phenomena in high heat
flux and CHF tests with various micro-size structures will be
necessary.
4. Conclusions
1. Boiling heat transfer is a promising heat dissipating method for
systems operating under thermally critical conditions. Critical
heat flux is one of the most significant design criteria for boiling
heat transfer and is also a very important factor in economy and
safety.
2. Numerous studies have dealt with experimental and theoretical
approaches to enhance CHF. These include changing surface
chemistry and geometry.
3. After development of nanotechnology, attempts were made to
apply it to thermal engineering. In 2003, You et al. [16] reported
great enhancement in CHF of nanoparticle-dispersed liquid
(nano-fluid) and several subsequent studies confirmed their
results.
4. The CHF enhancing performance of nano-fluid was revealed to
be a result of nanoparticle deposition on boiling surfaces [19].
Thereafter, roles of wettability and liquid spreadability were
suggested as important contributors to this phenomenon.
5. We used MEMS to fabricate artificial surfaces and evaluate how
their wettability, spreadability and surface geometry affect CHF.
The multi-scaled surface with higher wettability and spread-
ability produced more (107%) in CHF enhancement than the
other surfaces.
6. Wettability and liquid spreading effect was shown as CHF
enhancement difference between F surface and N, NM surface.
Liquid spreading effect was shown as CHF enhancement differ-
ence between N surface and NM surface. And, multi-scaled
geometry effect was shown as difference between liquid
spreading characteristics of N and NM surface and indirectly
affected CHF enhancement.
7. From this research, we measured the effects of wettability,
liquid spreadability and geometry on CHF. Additional experi-
ments and analysis of surface characteristics and various sur-
face-modification methods are required for optimum condition.
Acknowledgments
This research was supported by KOSEF (Korea Science and Engi-
neering Foundation) via Nuclear R&D Programs.
References
[1] E. Hahne, T. Diesselhorst, Hydrodynamic and surface effects on the peak heat
flux in pool boiling, in: Proceedings of the Sixth International Heat Transfer
Conference, vol. 1, 1978, pp. 209–214.
[2] S.K.R. Chowdhury, R.H.S. Winterton, Surface effects in pool boiling, Int. J. Heat
Mass Transfer 28 (1985) 1881–1889.
[3] S. Liaw, V.K. Dhir, Effect of surface wettability on transition boiling heat
transfer from a vertical surface, in: Proceedings of the Eighth International
Heat Transfer Conference, 1986, pp. 2031–2036.
[4] K. Ferjancic, I. Golobic, Surface effects on pool boiling CHF, Exp. Therm. Fluid
Sci. 25 (2002) 565–571.
[5] Y. Takata, S. Hidaka, M. Masuda, T. Ito, Pool boiling on a superhydrophilic
surface, Int. J. Energy Res. 27 (2003) 111–119.
[6] T.M. Anderson, I. Mudawar, Microelectronic cooling by enhanced pool boiling
of a dielectric fluorocarbon liquid, J. Heat Transfer 111 (1989) 752–759.
[7] R.N. Wenzel, Surface roughness and contact angle, J. Phys. Colloid Chem. 53
(1949) 1466.
[8] P.J. Marto, Lt.V.J. Lepere, Pool boiling heat transfer from enhanced surfaces to
dielectric fluids, J. Heat Transfer 104 (1982) 292–299.
[9] A.D. Messina, E.L. Park Jr., Effects of precise arrays of pits on nucleate boiling,
Int. J. Heat Mass Transfer 24 (1981) 141–145.
[10] C.K. Yu, D.C. Lu, T.C. Cheng, Pool boiling heat transfer on artificial micro-cavity
surfaces in dielectric fluid FC-72, J. Micromech. Microeng. 16 (2006) 2092–
2099.
[11] C.P. Costello, W.J. Frea, The roles of capillary wicking and surface deposits in
the attainment of high pool boiling burnout heat fluxes, AIChE J. 10 (1965)
393–398.
[12] J.C. Corman, M.H. McLaughlin, Boiling augmentation with structured surfaces,
ASHRAE Trans. (1976) 906–918.
[13] N.H. Afgan, L.A. Jovic, S.A. Kovalev, V.A. Lenykov, Boiling heat transfer from
surfaces with porous layers, Int. J. Heat Mass Transfer 28 (1985) 415–422.
[14] G.S. Hwang, M. Kaviany, Critical heat flux in thin, uniform particle coatings, Int.
J. Heat Mass Transfer 49 (2006) 844–849.
[15] S.U.S. Choi, Nanofluids: a new filed of scientific research and innovative
applications, Heat Transfer Eng. 29 (2008) 429–431.
[16] S.M. You, J.H. Kim, K.H. Kim, Effect of nanoparticles on critical heat flux of
water in pool boiling heat transfer, Appl. Phys. Lett. 83 (2003) 3374–3376.
[17] P. Vassallo, R. Kumar, S. D’Amico, Pool boiling heat transfer experiments in
silica–water nano-fluids, Int. J. Heat Mass Transfer 47 (2004) 407–411.
[18] D. Milanova, R. Kumar, Role of ions in pool boiling heat transfer of pure and
silica nanofluids, Appl. Phys. Lett. 87 (2005) 233107.
[19] H. Kim, J. Kim, M.H. Kim, Experimental study on CHF characteristics of water–
TiO2 nano-fluids, Nuclear Eng. Tech. 38 (2006) 61–68.
[20] S.J. Kim, I.C. Bang, J. Buongiorno, L.W. Hu, Effects of nanoparticle deposition on
surface wettability influencing boiling heat transfer in nanofluids, Appl. Phys.
Lett. 89 (2006) 153107.
[21] H. Kim, J. Kim, M.H. Kim, Effect of nanoparticles on CHF enhancement in pool
boiling of nanofluids, Int. J. Heat Mass Transfer 49 (2006) 5070–5074.
[22] H. Kim, M.H. Kim, Effect of nanoparticle deposition on capillary wicking that
influences the critical heat flux in nanofluids, Appl. Phys. Lett. 91 (2007)
014104.
[23] P. Gennes, F. Brochard-Wyart, D. Quere, Capillarity and Wetting Phenomena,
Springer, 2003 (Chapters 1, 4, 9).
[24] J. Bico, U. Thiele, D. Quere, Wetting of textured surfaces, Colloid Surface A 206
(2002) 41–46.
[25] Y. Tak, K. Yong, Controlled growth of well-aligned ZnO nanorods array using a
novel solution method, J. Phys. Chem. B 109 (2005) 19263–19269.
[26] N. Zuber, Hydrodynamic aspects of boiling heat transfer, Ph.D. thesis,
University of California, Los Angeles, CA, 1959
[27] R.W.L. Fong, T. Nithenandan, C.D. Bullock, L.F. Slater, G.A. McRae, Effect of
oxidation and fractal surface roughness on the wettability and critical heat
494 S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495
9. flux of glass-peened zirconium alloy tubes, in: Proceedings of the Fifth
International Conference on Boiling Heat Transfer, 2003.
[28] S.G. Kandlikar, A theoretical model to predict pool boiling CHF incorporating
effects of contact angle and orientation, J. Heat Transfer 123 (2001) 1071–
1079.
[29] S.G. Liter, M. Kaviany, Pool-boiling CHF enhancement by modulated porous-
layer coating: theory and experiment, Int. J. Heat Mass Transfer 44 (2001)
4287–4311.
S. Kim et al. / Experimental Thermal and Fluid Science 34 (2010) 487–495 495