You are on page 1of 13

International Journal of Heat and Mass Transfer 133 (2019) 548–560

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer


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

Pool boiling of HFE-7200 on nanoparticle-coating surfaces: Experiments


and heat transfer analysis
Zhen Cao a, Zan Wu a, Anh-Duc Pham a,b, Yanjie Yang a,c, Sahar Abbood a, Peter Falkman b,
Tautgirdas Ruzgas b, Cathrine Albèr b, Bengt Sundén a,⇑
a
Department of Energy Sciences, Lund University, Box 118, SE-22100 Lund, Sweden
b
Department of Biomedical Science, Malmö University, SE-205 06 Malmö, Sweden
c
College of Aerospace Science and Engineering, National University of Defense Technology, Changsha 410073, PR China

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

Article history: In the present study, an electrophoretic deposition method was employed to modify copper surfaces with
Received 4 June 2018 Cu-Zn (100 nm) nanoparticles. Pool boiling heat transfer of HFE-7200 on the modified surfaces was
Received in revised form 18 December 2018 experimentally studied. The results showed that the heat transfer coefficient on the modified surfaces
Accepted 21 December 2018
was significantly enhanced compared with that on a smooth surface, e.g., a maximum 100% enhance-
Available online 29 December 2018
ment, while the maximum superheat on the modified surfaces was around 20 K lower than that on
the smooth surface. However, the critical heat flux (CHF) was not improved considerably, and supple-
Keywords:
mentary tests indicated that the wickability of HFE-7200 was almost the same on the modified surfaces
Pool boiling
Heat transfer
and the smooth surface. The departure diameters of bubbles were recorded by a high speed camera,
Nanoparticle which were compared with several models in literature. Active nucleation site sizes were evaluated by
Bubble dynamics the Hsu nucleation theory and active nucleation site densities were estimated by appropriate correla-
tions. In addition, a heat transfer model, considering natural convection, re-formation of thermal bound-
ary layer and microlayer evaporation, was formulated to predict the heat transfer on the modified
surfaces and the smooth surface. A relatively good prediction was achieved.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction machining [12], laser machining [13–15], electrodeposition


[9,10,16,17] and composite methods [18]. In these studies, differ-
High-heat-flux surfaces widely exist in many applications, e.g., ent liquids were tested, including water [3,5–9,12,13,16–18], etha-
power plants, electronics and refrigeration systems [1]. Accord- nol [3], n-pentane [7,15], acetone [11] and dielectric liquids, e.g.,
ingly, cooling is of great importance for these applications to oper- FC-72 [3,14] and PF-5060 [10]. All these studies indicated that both
ate normally. Up to date, many cooling approaches have been the heat transfer coefficient and critical heat flux were augmented.
developed, like single-phase cooling, jet impingement cooling, The mechanisms could be summarized as: i) the increase of heat
spray cooling, heat pipes, indirect cooling with phase change mate- transfer area, ii) the increase of active nucleation sites, iii) surface
rials, and flow and pool boiling [2]. Pool boiling is one of the wettability and iv) bubble dynamics modulation, e.g., large bubble
promising methods, which can dissipate high heat fluxes with departure frequency.
low superheats. The performance of pool boiling is associated with With the progress in nano technologies, surface modifications
the interactions between surfaces and liquids, e.g., roughness and with nano structures have caught the attention all over the world.
wettability. Therefore, many efforts have been paid to surface According to the Hsu nucleation theory [19], the conditions that
modifications for pool boiling enhancement. Numerous micro cavities can be activated depend on the wettability and the sizes
structures have been processed on boiling surfaces, e.g., micro of the cavities. Nano structures, on one hand, change the wettabil-
pin fins [3,4], micro pillars [5], micro channels [6], copper foams ity, especially for water, e.g., smaller static contact angle of
[7,8] and microporous coatings [9,10], while various techniques water on nanotextured surfaces [20,21]; On the other hand,
have been applied, e.g., sintering [11], wire electrical discharge nanostructures create more cavities that can be activated [2,22].
At present, various nano technologies have been used, e.g., electro-
spraying [2,20,23], electroplating [24,25], electrostatic deposition
⇑ Corresponding author. [26], chemical vapor deposition [27,28] and electron beam physical
E-mail address: bengt.sunden@energy.lth.se (B. Sundén). vapor deposition [29,30], layer-by-layer deposition [31]. Sinha-Ray

https://doi.org/10.1016/j.ijheatmasstransfer.2018.12.140
0017-9310/Ó 2018 Elsevier Ltd. All rights reserved.
Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 549

Nomenclature

Ac cross section area of the microcapillary tube, (m2) Vme volume of microlayer, (m3)
CHF critical heat flux, (W/cm2) V’ absorbed flow rates, (m3/s)
Cpl specific heat of liquid, (J/kgK) w boiling surface
D diameter of the test surfaces, (m)
Dd bubble departure diameter, (m) Greek symbols
Di diameter of bubble influenced area, (m) Di = 2Dd a thermal diffusivity of liquid, (m2/s)
EPD electrophoretic deposition b thermal expansion coefficient, (K1)
f bubble departure frequency, (s1) d thickness of thermal layer, (m)
g acceleration of gravity, (m/s2) h static contact angle
hnc heat transfer coefficient in natural convection, (W/m2K) k thermal conductivity, (W/mK)
Dh liquid height drop, (m) m kinematic viscosity, (m2/s)
HTC heat transfer coefficient, (W/cm2K) q density, kg/m3
ifg latent heat, (J/kg) r surface tension, N/m
Ja⁄ modified Jacob number, ql Cpl(Tw - Tf)/(qvifg),
Na active nucleation site density, (sites/m2)
Subscript
P liquid pressure, (N/m2) c copper
Pr Prandtl number cond conduction
q heat flux, (W/m2)
f fluid
R radii of nucleation cavities, (m) l liquid
Ra roughness, (m) me microlayer evaporation
SS smooth surface nc natural convection
t time, (s)
sat saturated
td bubble departure time, (s) v vapor
T temperature, (K) w wall or waiting
Tw temperature of the boiling surface, (K)
Tsat saturated temperature, (K)

et al. [2] compared pool boiling of water and HFE-7300 on copper In the present study, copper surfaces were modified by an elec-
surfaces with nano fibers. The results indicated that the experi- trophoretic deposition method, depositing Cu-Zn alloy nanoparti-
ments with water did not show as strong enhancement of the heat cles (100 nm, Sigma-Aldrich) on the copper surfaces. Four
transfer coefficient as the experiments with HFE-7300, because the modified surfaces were prepared with the nanoparticles of
size of HFE-7300 bubbles is smaller and thus affected by the nano 0.3 mg (EPD-1), 0.6 mg (EPD-2), 0.9 mg (EPD-3) and 1.2 mg (EPD-
structures. The heat transfer enhancement was due to the increase 4), respectively. Pool boiling of HFE-7200 was experimentally stud-
of active nucleation sites. Sahu et al. [25] analyzed pool boiling of ied, because this liquid has not been studied widely and it has a
HFE-7300 and self-rewetting fluid on surfaces with nanofibers, and lower global warming potential than many other dielectric liquids,
the pool boiling performance i.e., heat transfer coefficient and crit- such as FC-72, HFE-7000, HFE-7100, HFE-7300 and PF5060. This
ical heat flux, were improved considerably. Jo et al. [23,24], Cao study tries to give a quantitative analysis of the mechanisms of
et al. [26], Jaikumar et al. [27] and Dharmendra et al. [28] studied enhanced heat transfer on the nanoparticle-coated surfaces, i.e.,
pool boiling of water on nano-structured surfaces, in which both active nucleation site density, bubble departure diameter and bub-
the heat transfer coefficient and critical heat flux were enhanced, ble departure frequency, and formulate a heat transfer model, con-
while Das et al. [29,30] reported an enhancement of heat transfer sidering natural convection, re-formation of the thermal boundary
in water without reaching the critical heat flux. In these studies, layer and microlayer evaporation, in order to predict the heat
generally, the mechanism about heat transfer enhancement was transfer on the nanoparticle-coated surfaces based on the previous
contributed to more active nucleation sites and faster bubble quantitative analysis.
departure due to the nanostructures. However, a quantitative anal-
ysis about the active nucleation site density of nano-structured
surfaces is still lacking, even though a few studies are already avail- 2. Experimental method and setup
able [20,32]. Regarding the critical heat flux enhancement, the
mechanisms were classified as (i) the small contact angle and cap- 2.1. Preparations of surfaces
illarity [23,24,26], (ii) large contact angle hysteresis [27] and (iii)
the increase in the effective surface area for heat transfer [28]. Smooth surfaces were prepared by successive polishing with
Especially, Jaikumar et al. [27] compared the mechanisms of sandpapers of grit P220, P600, P1000, P1500 and P2000. Then,
enhanced pool boiling performance on microscale and nanoscale the surfaces were cleaned by an ultrasonic bath with acetone (20
coated surfaces, which indicated that larger critical heat fluxes min) and ethanol (5 min), then flushed with DI water and dried
were due to contact angle hysteresis, roughness and enhanced at the end. Cu-Zn nanoparticles were dispersed in water by the
microlayer evaporation on microscale coated surfaces, and contact ultrasonic bath for two hours, with a concentration of 20 mg/ml.
angle hysteresis on nanoscale coating surfaces. Although, numer- Then the desired amount of solution was dropped in ethanol where
ous studies about pool boiling on nano-structured surfaces have an electric field was applied. The nanoparticles could then be
been carried out, quantitative analyses to illustrate the mecha- deposited on smooth surfaces due to the electric field. By zetapo-
nisms of the enhanced pool boiling performance are still few. In tential analysis, the Cu-Zn nanoparticle is positively charged and
addition, water is the most common working liquid in the above the nanoparticles move to the cathode which is the smooth copper
reviewed studies, but in many applications, e.g., electronics, water surface. In the present study, the electrophoretic deposition was
is not a suitable option. implemented for 30 min on each surface with a voltage of 9.5 V.
550 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

Table 1
Test surfaces.

Surface Weight of nanoparticles (mg) Estimated thickness of nano-coating (nm) Roughness Ra (nm) Contact angle (h)
SS * * 108.6

EPD-1 0.3 443 540.5

EPD-2 0.6 886 549.8

EPD-3 0.9 1329 1276.4

EPD-4 1.2 1772 1434.9

Four modified surfaces were prepared with 0.3 mg, 0.6 mg, 0.9 mg large number of cavities from several hundreds of nanometers to
and 1.2 mg nanoparticles, namely EPD-1, EPD-2, EPD-3 and EPD-4, several micrometers can been seen. This size range is preferable
respectively. Table 1 summarizes the information of the test sur- for nucleation. Considering EDX, oxygen was detected on the
faces, e.g., the thickness of the nanoparticle-coating, roughness smooth surface besides copper (see Fig. 1(a)). This is because cop-
and contact angle. The thickness was estimated assuming a hexag- per easily oxidizes. The EDX before and after the experiments is
onal packing of nanoparticles. The roughness was measured by a compared in Fig. 1(b) and Fig. 1(c), from which no apparent differ-
3D optical profiler (Dektak 6 M). Because the coating is not exactly ences can be detected.
uniform, the roughness was averaged with 10 measured values for
every surface. The contact angle was measured by an in-house 2.2. Experimental apparatus and test section
developed setup with an uncertainty of around ± 10%.
Fig. 1 provides SEM images of the smooth surface (SS) and the Fig. 2(a) provides the components of the setup, including a
nanoparticle-coated surface (EPD-2). Because the morphology of heating system, a temperature control system and a data acquisi-
the four modified surfaces is quite similar, only EPD-2 is shown tion system. The heating system has a transformer (KIEA 8, Tufvas-
here. Additionally, EDX analysis was conducted to compare the sons Transformator), two digital meters, five cartridge heaters
chemical elements before and after the experiments. As shown in (CIR-10121, OMEGA) and a copper rod. The cooling system
Fig. 1(a), some cavities and scratches exist on the smooth surface, involves a temperature controller (MAXVU16), an auxiliary heater
which are preferable locations for nucleation. However, many of and a cooler to provide a saturated state. The data acquisition sys-
them are around several tens of nanometers, which probably are tem consists of a data acquisition board (Agilent 34970A), a high
too small for nucleation according to the Laplace equation and speed camera (Phantom v611) and a computer. Additionally, a ves-
the Clausius-Clapeyron equation. However, on the modified sur- sel made of optic glass works as the boiling chamber. The size of
face as shown in Fig. 1(b), numerous porous structures appear. A the vessel is 100 mm  100 mm  300 mm in length, width and

Fig. 1. Surface characterization including SEM and EDX: (a) smooth surface before experiments, (b) EPD-2 before experiments, (c) EPD-2 after experiments.
Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 551

Fig. 2. Schematics of (a) pool boiling facility, (b) heating system and (c) wickability measurement.

height, respectively. The copper rod is covered with PTFE to mini- 12 mm. A tin paste (BERA-FIX soldering paste) is used to solder the
mize the heat loss. The heat loss can be neglected in the present test substrate on the copper rod. Five cartridge heaters (CIR-10121
study, which is discussed in details in the supplementary docu- from OMEGA) are inserted into the bottom of the rod to heat it up.
ment. A relief valve is mounted to provide atmospheric surround- The distribution of the heaters is also shown in Fig. 2(b), with one
ings during the experiments. Fig. 2(b) shows details of the copper in the center surrounded by the other four symmetrically. The dis-
rod, i.e., the distribution of the thermocouples and cartridge hea- tance between the central one and the others is 17 mm.
ters. The copper rod is a cylinder with a bottom diameter of
60 mm and a top diameter of 12 mm. At the height of 60 mm, 2.3. Experimental procedure
the diameter of the rod transits smoothly from 60 mm to 12 mm
with an angle of 135°. To evaluate the heat flux (q) and the boiling Saturated pool boiling of HFE-7200 was studied. About 1 L HFE-
surface temperature (Tw), four K-type thermocouples are mounted 7200 was poured in the boiling vessel and vigorously boiled for
on the rod vertically with 15 mm intervals and a K-type thermo- half an hour to degas non-condensable gases. The properties of
couple is embedded in the test substrate at 8 mm from the boiling HFE-7200 at the saturated state are shown in Table 2. During the
surface. The test substrate has a height of 10 mm and a diameter of experiments, the voltage was increased by 3 V in each case until

Table 2
Properties of HFE-7200 at the saturated state, T = 349.15 K.

ql (kg/m3) qv (kg/m3) r (mN/m) ifg (kJ/kg) kl (W/mK) Cpl (J/kgK) m (mm2/s) b (K1)
1303 10.3 9.2 119 0.0555 1220 0.26688 0.0016
552 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

"  2  2 #1=2
the critical heat flux (CHF) was reached, i.e., when an abrupt y2 q
increase of the temperatures appeared. All data were collected at dT w ¼ ðdT 5 Þ2 þ dq þ dy2 ð7Þ
kc kc
a steady state when the recorded temperatures deviated within
±0.2 °C during 5 min. For each run, it took around 40 min to reach
"   2  2 #1=2
steady state and the temperatures recorded during 2 min were dHTC dq
2
dT w dT f
averaged as the final results. The experiments were repeated three ¼ þ þ ð8Þ
HTC q Tw  Tf Tw  Tf
times with an interval of 24 h between two experiments. Consis-
tent with [33], the nanoparticle-coated surfaces performed the
Therefore, the maximum uncertainties of q and HTC are assumed to
best in the experiments of the first time. However, after the exper-
occur when (T4 –T1) and (Tw –Tf) are minimum. Because the present
iments of the first time for typically 8–10 h, the nanoparticle coat-
study focuses on boiling heat transfer and the minima of (T4 –T1)
ing becomes conditioned and the boiling curves are reproducible.
and (Tw –Tf) represent the measured minimum heat flux. Accord-
Therefore, only the results of the second time are shown in the fol-
ingly, the maximum uncertainties of q and HTC are calculated at
lowing analysis. More details about the repeatability and durability
the minimum heat flux measured in the nucleate boiling region.
of the test surfaces are available in the supplementary document.
However, it should be noted that the minimum heat flux here does
not mean the incipient heat flux and it is only the first test case
2.4. Data reduction and uncertainty analysis
when the nucleate boiling occurs in the experiments. In fact, the
minimum heat flux varies on the smooth surface (SS) and the mod-
The one-dimensional Fourier law is employed to calculate the
ified surfaces (EPDs). The minimum heat fluxes on the SS, EPD-1,
heat flux (q). According to the recorded temperatures, T1, T2, T3
EPD-2, EPD-3 and EPD-4 are 5.33 W/cm2, 3.29 W/cm2,
and T4, the heat flux can be expressed as
3.02 W/cm2, 5.60 W/cm2 and 3.56 W/cm2, respectively. Correspond-
T4  T1 ingly, the maximum uncertainties of q on the SS, EPD-1, EPD-2,
q ¼ kc ð1Þ
3y1 EPD-3 and EPD-4 are 4.9%, 7.8%, 8.4%, 4.7% and 7.2%, respectively,
while the maximum uncertainties of HTC on the SS, EPD-1, EPD-2,
where y1 = 15 mm is the distance between two neighboring thermo-
EPD-3 and EPD-4 are 5.1%, 8.1%, 8.8%, 5.2% and 7.5%, respectively.
couples and kc = 400 W/mK is the thermal conductivity of copper.
Accordingly, the boiling surface temperature Tw is estimated as
follow 3. Results and discussion
q  y2
Tw ¼ T5  ð2Þ 3.1. Pool boiling on the nanoparticle-coated surfaces and the smooth
kc
surface
where y2 = 8 mm is the distance between the thermocouple T5 and
the boiling surface. To analyze the effect of nanoparticle coating on nucleate pool
The liquid temperature Tf is measured by the thermocouple T6, boiling, the boiling curves and heat transfer coefficients of satu-
then the heat transfer coefficient (HTC) is calculated as rated HFE-7200 are illustrated in Fig. 3(a) and (b), respectively.
q Obviously, the boiling curves of the nanoparticle-coated surfaces
HTC ¼ ð3Þ shift to the left very much compared with the boiling curve of
Tw  Tf
the smooth surface. The maximum superheats of EPD-1, EPD-2,
Here, the uncertainties of heat flux (q) and heat transfer coefficient EPD-3 and EPD-4 are 18.1 K, 15.4 K, 16.8 K and 18.9 K, respectively,
(HTC) mainly come from the uncertainty of temperature measure- while the maximum superheat of the smooth surface (SS) is 39.8 K.
ment and tolerances of manufacturing. The uncertainty of the K- However, the CHFs of EPD-1, EPD-2, EPD-3 and EPD-4 are
type thermocouples is ±0.2 °C, while the tolerance of the space 19.1 W/cm2, 18.6 W/cm2, 19.6 W/cm2 and 18.5 W/cm2, respec-
between thermocouples is estimated as ±0.2 mm. Then the uncer- tively, while the CHF of the SS is 19.2 W/cm2. This indicates that
tainties of q and HTC were estimated using the uncertainty analysis the CHF of HFE-7200 is not enhanced by the nanoparticle coatings,
method in [34]. The uncertainty in the result F is a function of the probably because HFE-7200 is already superhydrophilic on the
independent variables Xi and written as smooth surface with a contact angle of around 24.4° ± 2°. The wet-
F ¼ FðX i Þ ð4Þ tability is not enhanced very much by the nanoparticle coatings.
More discussion about CHF will be provided in the later section.
The uncertainty of F is expressed as
To provide a more informative and clear picture, the boiling
"  2 #1=2 curves have been redrawn as the heat transfer coefficient (HTC)
X
n
@F
dF ¼ dX i ð5Þ versus the heat flux (q), as shown in Fig. 3(b). Evidently, the heat
@X i
i¼1 transfer coefficient of EPD-1, EPD-2, EPD-3 and EPD-4 is much
where dF=@X i and dX i are the sensitivity coefficient and uncertainty higher than that of SS at a given heat flux. For example, at
level associated with the variable Xi. dX i was obtained by a root- q = 14.2 W/cm2 the heat transfer coefficients of SS, EPD-1, EPD-2,
mean square combination of the precision uncertainty of the instru- EPD-3 and EPD-4 are 0.46 W/cm2K, 0.92 W/cm2K, 1.0 W/cm2K,
ments and the unsteadiness uncertainty. dF=@X i represents the 0.93 W/cm2K and 0.90 W/cm2K, respectively. Thus, the maximum
measurement resolution, instrumentation variance, geometric enhancement of heat transfer is over 100%. This is probably
uncertainty, substrate conduction loss and calibration error. The because of more active nucleation sites and larger effective heat
variable Xi to be included in the calculation of the total uncertainty transfer area on the nanoparticle-coated surfaces. Interestingly,
level of F depends on the purpose of the analysis. The uncertainties the performances of EPD-1, EPD-2, EPD-3 and EPD-4 are quite
in the present study are calculated as follows: close, which means the thickness does not affect the heat transfer
The uncertainties of the heat flux (q) and heat transfer coeffi- remarkably at least on the present test surfaces. However, a larger
cient (HTC) are formulated as: thickness has better performance in the studies [29,30], while a
" larger thickness induces a larger thermal resistance and degrada-
2  2  2 #1=2 tion of heat transfer in the study [35]. More discussion about the
dq dT 4 dT 1 dy1
¼ þ þ ð6Þ mechanisms of heat transfer enhancement will be provided in
q T4  T1 T4  T1 y1
the following sections.
Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 553

20 3.2.2. Bubble departure diameter


SS To quantitatively compare bubble departure diameters, the
EPD-1
EPD-2
bubble diameters were measured at the isolated-bubble regime.
15 EPD-3 As bubble grows, the buoyancy force eventually overcomes the
EPD-4 other forces, i.e., interfacial tension and viscous drag. With the high
speed visualization, at least five bubbles were detected among 200
q (W/cm )
2

10 images. The bubble diameters were measured by averaging these


bubbles. In the image processing, the diameter of the surface
worked as the calibrated length L = 12 mm which includes various
5 pixels (x) on the various surfaces, depending on the size of view
during the visualization, e.g., the pixels  are 207 pixels and 180
pixels on the SS and EPD-1, respectively. Then, the bubble depar-
0 ture diameter (Dd) can be measured by counting the pixels of the
0 10 20 30 40 bubble (y), which is Dd = (L/x)y. With the method of uncertainty
Tw-Tf(K) analysis in Section 2.4, the uncertainty of Dd can be expressed as
(a)
"   2  2 #1=2
2
dDd dL dx dy
1.5 ¼ þ þ ð9Þ
Dd L x y
SS
EPD-1
1.2 EPD-2 The uncertainties dx and dL were evaluated as ±4 pixels and
EPD-3 ±0.2 mm, respectively. The counted pixels of the bubbles range 3–
HTC (W/cm K)

EPD-4 10 pixels with a maximum uncertainty dy of ±1 pixels. Then the


2

0.9 uncertainty of the bubble departure diameter dDd / Dd can be eval-


uated with Eq. (9). Based on the uncertainty analysis, the maximum
0.6 and minimum uncertainties of the bubble departure diameters are
36.9% and 6.8%, respectively. Fig. 5 compares the bubble departure
diameters on the nanoparticle-coated surfaces and the smooth sur-
0.3
face. Apparently, the nanoparticle-coated surfaces have smaller
bubble departure diameters than the smooth surface at a given heat
0.0 flux, e.g., the bubble departure diameter on EPD-1 is around 291 mm
0 5 10 15 20 at q = 5.0 W/cm2 while the bubble departure diameter on SS is
q (W/cm )
2
(b) around 727 mm at q = 5.3 W/cm2.
Up to date, many models have been proposed to predict the
Fig. 3. Saturated pool boiling heat transfer of HFE-7200 on the teste surfaces: (a) bubble departure diameter in pool boiling. These were reviewed
boiling curves, and (b) heat transfer coefficients.
in [36]. Fritz proposed a bubble departure diameter correlation
for pure liquids and liquid mixtures [37], considering the force bal-
ance between buoyancy force and surface tension force. It is fre-
quently and widely used by researchers with or without surface
3.2. Bubble dynamics
modifications. The formula reads
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3.2.1. Bubble visualization r
Dd ¼ 0:0208h ð10Þ
Bubble dynamics was visualized by a high speed camera (Phan- g ðql  qv Þ
tom v611) with 1280  800 maximum resolution and 1000 Hz fre-
quency. Fig. 4(a–e) shows the bubble visualization on SS, EPD-1, where h, ql, qv, r and g are the contact angle (degree), liquid den-
EPD-2, EPD-3 and EPD-4, respectively, at low, moderate and high sity, vapor density, surface tension and acceleration of gravity,
heat fluxes. It is evident that the onset of nucleation on the respectively.
nanoparticle-coated surfaces is earlier than on the smooth surface. Cole [38] modified the Fritz model by involving the superheat
As shown in Fig. 4(a) bubble nucleation does not occur on the and replacing the contact angle with a new constant. The model
smooth surface at q = 3.2 W/cm2, but it occurs on the other four was validated by numerous liquids, e.g., water, acetone, methanol
nanoparticle-coated surfaces at nearly the same heat flux as seen and n-pentane. It is expressed as follows
in Fig. 4(b–e). However, the exact superheat condition when the
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r
incipient boiling appears is not especially detected in the experi- Dd ¼ 0:04Ja ð11Þ
g ðql  qv Þ
ments. At a moderate heat flux, the test surfaces are all covered
with bubbles, like q = 8.7 W/cm2 on SS (Fig. 4(a)), q = 7.6 W/cm2 where Ja⁄ = ql Cpl(Tw  Tf)/(qvifg), and Cpl and ifg are the specific
on EPD-1 (Fig. 4(b)), q = 7.8 W/cm2 on EPD-2 (Fig. 4(c)), heat and latent heat of the liquid, respectively.
q = 7.4 W/cm2 on EPD-3 (Fig. 4(d)) and q = 7.1 W/cm2 on EPD-4 Phan et al. [39] modified the Fritz model by considering the
(Fig. 4(e)). For these conditions, most of the bubbles are still iso- energy factor as the contribution of the wetting effect. The energy
lated and the bubble coalescence is not intensive. In general, the factor is simply the volume ratio between a truncated sphere of
bubbles on SS have larger departure diameters and lower depar- which the contact angle is h and a full sphere which has the same
ture frequencies than those on the nanoparticle-coated surfaces. diameter. The model is validated by water and it is given as
As the heat fluxes increase, bubbles begin to collide and coalesce,   rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 þ 3 cos h  cos3 h r
e.g., q = 14.2 W/cm2 on SS (Fig. 4(a)), EPD-1 (Fig. 4(b)) and Dd ¼ 0:626977  ð12Þ
EPD-2 (Fig. 4(c)), q = 13.5 W/cm2 on EPD-3 (Fig. 4(d)) and
4 g ðql  qv Þ
q = 14.0 W/cm2 on EPD-4 (Fig. 4(e)). The coalesced bubbles depart Kim et al. [40] performed pool boiling experiments of R 113 for sub-
from the surfaces almost periodically. cooled, saturated and superheated pool conditions. A dimensionless
554 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

Fig. 4. Bubble visualizations at low, moderate and high heat fluxes on the test surfaces.

"
rffiffiffiffiffiffi rffiffiffiffiffi#2
analysis of bubble growth and departure was carried out, based on 27 ql

which, a new model was proposed. For the saturated pool boiling, Dd ¼ 25  c  Ja  a ð13Þ
2 r
the model is formulated as
Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 555

2.0
SS
EPD-1
EPD-2
EPD-3
1.5 EPD-4
Fritz (1935)
Dd (mm) Cole (1967) 2mm
1.0
Phan et al. (2009) q = 5.3 W/cm2
Kim et al. γ =0.7 (2006)
Kim et al. γ =0.9 (2006)

0.5
2mm
q = 5.0 W/cm2
0.0
5 10 15 20 25 30
Tw-Tf (K)

Fig. 5. Average bubble departure diameter on the test surfaces, with comparisons to several models.

where c is a constant, ranging from 0 to 1, and a is the thermal dif- sient heat conduction to the liquid during the waiting time, but
fusivity of the liquid. with different boundary conditions and initial conditions. Both
Fig. 5 compares the measured departure diameters and several derived a size range of active nucleation sites. In the present study,
models. It is seen that the model proposed by Kim et al. [40] can the size range of active nucleation sites is estimated by the Hsu
predict the present results quite well with c = 0.7 for the smooth model [19], as shown below
surface and c = 0.9 for the nanoparticle-coated surfaces. Accord- " sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi#
ingly, the bubble departure frequency can be predicted with the d 4C 2 2rT sat
Rmin=max ¼ 1 1  ð15Þ
model proposed by Kim et al. [40] as well, even though the bubble 2C 1 dðT w  T f Þ ifg qv
departure frequency cannot be measured well in the present
experiments due to the large number of bubbles on the surfaces where d is the thermal layer thickness which can be estimated as
and the resolution of the visualization. The model proposed by kf/hnc. C1 and C2 are constants as 2 and 1.6, respectively. In saturated
Kim et al. [40] to predict the bubble departure frequency is boiling, Tf = Tsat. Fig. 6 compares the size range of active nucleation
expressed as sites on the nanoparticle-coated surfaces and the smooth surface.
Based on the superheats in the experiments, the solid lines repre-
1 2r sent the nucleate boiling regime, while the dashed lines represent
f ¼ ¼ ð14Þ
td 135c  a  Ja  ql  Dd the natural convection region (no nucleation) or the region reaching
CHF. The results show that the maximum size is around 10 mm to
where td is the departure time. From Eq. (14), it can be deduced that
15 mm which is much larger than the cavities observed in the SEM
smaller departure diameters represent higher departure frequen-
images. However, the minimum size is between 100 nm and
cies. As discussed previously, the nanoparticle-coated surfaces have
500 nm for all surfaces, depending on the superheat, which means
smaller bubble departure diameters than the smooth surface at a
that the nucleation site with a size smaller than 100 nm is ineffec-
given heat flux. Therefore, the nanoparticle-coated surfaces have
tive. The previous analysis on the SEM images of the smooth surface
higher bubble departure frequencies. This is one reason why the
and the nanoparticle-coated surfaces shows that the nanoparticle-
heat transfer is enhanced on the nanoparticle-coated surfaces.
coated surfaces have much more sites larger than 100 nm than
the smooth surface. This is to say that the number of active nucle-
3.3. Heat transfer analysis

3.3.1. Size of active nucleation sites 16


Here, the heterogeneous boiling is studied, which means that No nucleation
some nucleation sites, e.g., cavities should exist on the surfaces. 14 on SS
However, the conditions for which the nucleation sites can be acti-
vated or not depend on many factors, e.g., the size and geometry of 12
the nucleation sites and superheat. Accordingly, it is necessary to
study when and how the bubble nucleation occurs. Nowadays, Rmax
R (μm)

bubble nucleation has been studied extensively, but only some 10


0.9
pioneering works are focused on in this part. Courty and Foust
[41] were the first ones to formulate a simple model of bubble 0.5 Rmin CHF arrives on EPDs SS
nucleation by considering an idealized conical cavity under EPD-1
EPD-2
isothermal conditions. Similarly, Griffith and Wallis [42] analyzed
EPD-3
bubble nucleation in a conical cavity, involving the contact angle. EPD-4
Combining the Laplace equation and the Clausius-Clapeyron equa-
0.1
tion, the conditions for bubble nucleation were derived. It is well 5 10 15 20 25 30 35 40 45
known that a bubble period includes waiting time and growth Tw-Tsat(K)
time. The waiting time can be interpreted as the time to be ready
for nucleation. Hsu [19] and Han & Griffith [43] both studied the Fig. 6. Evaluation of the range of sizes of active nucleation sites based on Hsu
bubble nucleation, assuming a problem of one dimensional tran- nucleation model [19].
556 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

Z Di =2
ation sites on the nanoparticle-coated surfaces is much larger than p 2 
on the smooth surface. Therefore, heat transfer is enhanced.
qcond ¼ Na  f  2pdrdr þ Di  D2d dd
R 4
Z 1
3.3.2. Heat transfer model  ðT  T sat Þql  C pl dx
0
Currently, numerous models have been developed to predict " #
nucleate boiling heat transfer. In general, the models can be D2i D2d
¼ Na  f  2ql  C pl ðT w  T sat Þ dd  ðdd  dc Þ ð19Þ
divided into three types, i.e., analogy models by Foster and Zuber 4 12
[44], and Rohsenow [45], hydrodynamic models by Tien [46] and
where R is the radius of the cavity. dd and dc are the thickness of the
Zuber [47] and mechanistic models [48–52]. In the present study,
thermal layer at the departure time and the waiting time, which are
a mechanistic model is formulated to predict the heat transfer on
defined as (ptda)1/2 and (ptwa)1/2, respectively.
nanoparticle-coated surfaces and the smooth surface. It is worth
Mikic and Rohsenow [49] modified the model of Han and Grif-
pointing that the mechanistic model does not consider the interac-
fith, assuming only pure conduction to the liquid occurred in the
tions between bubbles, e.g., bubble coalescence. It has been gener-
influenced area. The qcond over the area of influence in a unit area
ally accepted that the heat flux in nucleate boiling is attributed to
is modeled as
three heat transfer mechanisms, as shown in Fig. 7 including:
Z 1
kl ðT w  T sat Þ
f

(1) The heat transferred by microlayer evaporation qme. qcond ¼ Na  pD2d  f pffiffiffiffiffiffiffiffiffi dt
0 pat
(2) The heat transferred by re-formation of thermal boundary pffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffi

layer qcond. ¼ 2 p kl  ql  C pl f D2d ðT w  T sat Þ  Na ð20Þ
(3) The heat transferred by natural convection from the heating
surface not influenced by the bubbles qnc. Later, Benjamin and Balakrishnan [51] assumed that only pure con-
duction to the liquid occurred in the influenced area during the
Heat flux due to natural convection (qnc) can be defined as [48] waiting time, then Eq. (20) was modified as
Z
1 kl ðT w  T sat Þ
tw
Laminar range105 < Ra < 2  107 : qcond ¼ Na  pD2d  pffiffiffiffiffiffiffiffiffi dt
" #1=4 tw 0 pat
bg ðT w  T sat Þ5 a3 sffiffiffiffiffi
qnc ¼ 0:54  ql  C pl ð16Þ pffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 2
mD ¼ 2 p kl  ql  C pl D ðT w  T sat Þ  Na ð21Þ
tw d

Turbulent range2  107 < Ra < 3  1010 : However, Moghaddam and Kiger [53] studied the physical mecha-
" #1=3 nisms of heat transfer during single bubble nucleate boiling of FC-
bg ðT w  T sat Þ4 a2 72 under saturation conditions, indicating that the transient heat
qnc ¼ 0:14  ql  C pl ð17Þ
m conduction to the liquid happens not only during the waiting time,
but also throughout the whole period of bubbles. In addition, Thia-
where Ra, b, m and D is the Rayleigh number, thermal expansion garajan et al. [54] studied bubble dynamics and nucleation pool
coefficient, kinematic viscosity and diameter of the test surfaces, boiling of HFE-7100, finding that the growth time in HFE-7100 is
respectively. In the present study, nucleate boiling is in the turbu- much longer than the waiting time. This finding is different from
lent range, Thus Eq. (17) is used to estimate qnc. The qnc over the the observation during pool boiling of water where the waiting time
uninfluenced area in a unit area is calculated as is much longer than the growth time [55]. Because the properties of
" #1=3 "  2 # HFE-7200 is quite similar to those of HFE-7100, Eq. (20) is used to
bg ðT w  T sat Þ4 a2 Di calculate qcond while the bubble frequency is provided by Eq. (14)
qnc ¼ 0:14  ql  C pl 1  Na  p ð18Þ
m 2 in the present study.

where Di is the diameter of the influenced area, defined as 2Dd. Na is  Heat flux by microlayer evaporation (qme)
the active nucleation site density. Voutsinos and Judd [56] studied the microlayer evaporation
phenomenon using laser interferometry and high speed photogra-
 Heat flux due to the re-formation of thermal boundary layer phy. They derived an equation to estimate the volume of liquid
(qcond) evaporated from the microlayer. This equation reads
Han and Griffith [48] considered qcond equal to the rate of the Z Dd =2
increased internal energy in the influenced area of bubbles while V me ¼ 2p ½dðr; 0Þ  dðr; t Þ rdr
assuming that the thickness of thermal layer was linear in r. Then 0
" rffiffiffiffi #
the qcond over the area of influence in a unit area reads 2p 3 qv 3 ql  C pl ðT sat  T f Þ
¼ ðDd =2Þ 1þ ð22Þ
3 ql p 60qv ifg

If it is at saturated condition, then Tsat = Tf. Therefore, Eq. (22) can be


Di simplified as
2p q
V me ¼ ðDd =2Þ3 v ð23Þ
Dd 3 ql
x Eq. (23) tells that the liquid mass of the microlayer is only 50% that
r of the bubble, which means that the latent heat from the microlayer
qnc qcond+qme qnc evaporation is around 50% of the required heat for bubble growth.
This is consistent with the results in [40]. Then, the qme in a unit
Fig. 7. Schematic of heat transfer mechanisms. area is expressed as
Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 557

2p is reasonable to consider this model to predict the active nucle-


qme ¼ Na  ql  V me  ifg  f ¼ Na  ðDd =2Þ3 qv  ifg  f ð24Þ
3 ation site density on nanoparticle-coated surfaces in the present
Benjamin and Balakrishnan [51] also proposed an equation to calcu- study. The model is defined as
late the volume of liquid evaporated from the microlayer, but this is ( !)
h2 k0
not discussed here. Na ¼ N 1  exp  02
exp f ðqþ Þ 1
8h L
Combining Eq. (18), Eq. (20) and Eq. (24), the composed heat
flux can be modeled as
" #13 2r½1 þ ðqv =ql Þ P
L¼  ð27Þ
bg ðT w  T sat Þ4 a2 h i
exp ifg ðT w  T sat Þ=ðRT w T sat Þ  1
q ¼ qnc þ qcond þ qme ¼ 0:14  ql  C pl 1  Na  p D2d
m
pffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffipffiffiffi f ðqþ Þ ¼ 0:01064 þ 0:48246qþ  0:22712ðqþ Þ þ 0:05468ðqþ Þ
2 3
þ 2 p kl  ql  C pl f D2d ðT w  T sat Þ  Na
2p  
þ Na  ðDd =2Þ3 qv  ifg  f ð25Þ ql  qv
3 qþ ¼ log
qv
As indicated in Eq. (25), Na, Dd and f are essential to predict the heat
transfer. In Section 3.2.2, Eq. (13) has been validated by the exper- where, hˈ is a static parameter, 0.722 rad, kˈ is also a static param-
imental results to evaluate the bubble departure diameter Dd. eter, 2.50  106 m and R is the gas constant based on a molecular
Accordingly, Eq. (14) works to estimate the bubble departure fre- weight, 31.48 J/kgK. N is the average site density, suggested as
quency f. Now, the problem is how to calculate the active nucleation 4.72  105 sites/m2 on the smooth surface in [59], but as
site density Na. Due to the large number of bubbles on the test sur- 3  4.72  105 here, because the size of the bubbles on the
faces, it is difficult to get the active nucleation site density from the nanoparticle-coated surfaces is smaller than that on the smooth
high speed photography. However, many correlations for the active surface, meaning a larger average site density. N is also adjusted
nucleation site density are available in literature [36]. It is worth in [54] to predict the active nucleation site density on the microp-
noting that some correlations are based on water only, e.g., Wang orous surfaces. At the moment, the active nucleation sites on the
and Dhir [57], while some correlations are considering different liq- nanoparticle-coated surfaces are predicted by Eq. (27) and the heat
uids, e.g., water, carbon tetrachloride, n-hexane, and acetone in flux as modeled by Eq. (25) can be calculated. The results in Table 3
[58]. Because the HFE-7200 in the present study is totally different indicate that the present modeled heat fluxes agree with the exper-
from water, Benjamin and Balakrishnan’s model [58] is firstly used imental heat fluxes on the nanoparticle-coated surfaces. Therefore,
to estimate Na as Eq. (26) and Eq. (27) are suitable to predict the active nucleation site
  density on the smooth surface and the nanoparticle-coating surface,
kl  ql  C pl
Na ¼ 218Pr1:63 W0:4 ðT w  T sat Þ3 ; respectively, in the present study.
kc  qc  C pc
   2 Based on the previous analysis, Fig. 8 compares the predicted
RP RP active nucleation site density on the smooth surface and the
W ¼ 14:4  4:5 a þ 0:4 a ð26Þ
r r
where, qc, kc and Cpc are the density, thermal conductivity and 7
specific heat of copper; P is the liquid pressure.
10 Nucleate
center
With the active nucleation sites predicted by Eq. (26), the heat Namax Di
1 1
flux modeled by Eq. (25) can be predicted. Table 3 compares the Namax =
2 3( Di / 2) 2
=
2 3( Dd ) 2
modeled heat flux and the experimental heat flux on SS, EPD-1
6
and EPD-2, showing that the modeled heat fluxes agree with the 10 Namax
Na (sites/m )
2

experimental heat fluxes on SS, but disagree with the experimental


heat fluxes on the nanoparticle-coated surfaces. This is probably
because Benjamin and Balakrishnan’s model is validated for speci-
fic conditions, e.g., 2.2 < w < 14. The smooth surface can almost sat- 5
isfy these conditions, but the nanoparticle-coated surfacse are far 10 SS
away from the condition 2.2 < w < 14 due to the larger roughness. EPD-1
EPD-2
Therefore, other models should be considered to estimate Na on
EPD-3
the nanoparticle-coated surfaces. Thiagarajan et al. [54] investi- EPD-4
gated the active nucleation density of HFE-7100 on copper sur- 4
10
faces. It is found that the model proposed by Hibiki and Ishii [59] 5 15 25 35
can predict Na of HFE-7100 well by adjusting the coefficient in Tw-Tsat(K)
the equation. Because the properties of HFE-7200 is similar to
those of HFE-7100, e.g., surface tension, viscosity and density, it Fig. 8. Comparison of active nucleation site density on the test surfaces.

Table 3
Validation of the selected models about active nucleation site density.

DT (K) Dd (mm) (exp.) f (s1) (Eq.14) Na (m2) (Eq.26) qmodel (W/cm2) Na (m2) (Eq.27) qmodel (W/cm2) qexperiment (W/cm2)
SS 22.2 727 204.2 214,634 4.7 5.3
25.6 810 159.5 326,028 8.4 8.7
28.4 919 126.4 448,498 12.1 11.4
EPD-1 12.7 291 626.2 78,475 0.5 1,287,939 4.7 5.0
13.4 366 472.1 91,854 0.7 1,587,979 8.3 7.6
EPD-2 11.4 271 748.5 56,252 0.4 925,822 3.0 3.0
12 331 581.1 65,942 0.5 1,116,482 4.9 5.4
558 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

25 Fig. 9. It is seen that at low and moderate heat fluxes, i.e.,


EPD-1 (experiment) EPD-2 (experiment)
EPD-1 (model) EPD-2 (model) q < 12.0 W/cm2, the modeled curves agree with the experimental
curves well. However, at high heat fluxes (q > 12.0 W/cm2), an
20 obvious deviation appears between the experimental results and
the modeled results, e.g., over prediction on SS. This is because
15 the formulated model does not consider the interactions between
q (W/cm )

the bubbles, e.g., bubble coalescence. In reality, the bubble coales-


2

cence probably delay bubble departure, which decreases the heat


10 transfer. In future studies, on one hand, the heat transfer at low
SS (experiment)
and moderate heat fluxes (partial nucleate boiling) will be further
SS (model) studied. On the other hand, the heat transfer at high heat fluxes
5
(fully nucleate boiling) will consider bubble interactions, e.g., bub-
ble coalescence.
0
5 10 15 20 25 30 35 40 3.4. CHF analysis
Tw-Tsat(K)
As is shown previously, CHF of HFE-7200 is not enhanced effec-
Fig. 9. Comparison of boiling curves between experiments and models. tively on the nanoparticle-coated surfaces. This part intends to
explore why CHF is not enhanced. Fig. 10 shows the bubble behav-
ior at CHF. The behavior actually is quite similar on the smooth
surface and the nanoparticle-coated surfaces. The bubble revolu-
nanoparticle-coated surfaces. It is evident that the active nucle- tion process on SS and EPD-1 is provided in Fig. 10(a) and (b),
ation site density becomes large with increasing superheat. How- respectively. In general, at CHF, bubbles coalesce quite intensively
ever, there should exist a maximum value (Namax) beyond which and a vapor blanket (red arrow) can be recognized on the surfaces.
the active nucleation site density is assume to be constant even Then a vapor column stems from the vapor blanket and grows up,
the superheat is increased. Han and Griffith [48] suggested the e.g., from t = 0 ms to t = 30 ms in Fig. 10(a) and from t = 0 ms to
pffiffiffi
maximum Na as 1=ð2 3D2d Þ, assuming that the influenced area is t = 40 ms in Fig. 10(b). At the end, the column gets rid of the vapor
hexagonally close-packed. Fig. 8 provides a quantitative evidence blanket and departs from the surface, e.g., t = 40 ms on SS, and
that the nanoparticle coatings increase the number of the active t = 50 ms on EPD-1. This process is almost periodic.
nucleation sites, which enhances heat transfer. In addition, with Kwark et al. [21] and Im et al. [60] investigated pool boiling heat
the predicted Na, Dd and f, the modeled boiling curves (Eq. (25)) transfer on nanocoated surfaces, and claimed that CHF enhance-
are compared with the experimental boiling curves, as shown in ment was primarily due to the capillary wicking. In the present

Fig. 10. Bubble visualizations on the test surfaces at CHF.


Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560 559

"   2 #1=2
2
dDh dx dy
¼ þ ð30Þ
Dh x y

With Eq. (29) and Eq. (30), the uncertainty of the liquid absorbed
volume V can be estimated. The maximum and minimum uncer-
tainties of V are 30.3% and 6.8%, respectively, depending on the
number of pixels.
Fig. 11(a) records the process of liquid absorption on SS and
Fig. 11(b) shows the liquid absorbed flow rate. Evidently, the wick-
ability of HFE-7200 is not enhanced effectively by the nanoparticle
coatings. This is probably because the surface tension of HFE-7200
is quite small and it already has quite good wetting on the smooth
surface. In future studies, the configuration of nanoparticle deposi-
tion must be considered and then CHF may show an enhancement
from the perspective of reducing the hydrodynamic instability.

4. Conclusions

In the present study, pool boiling heat transfer of HFE-7200 was


experimentally studied on a smooth surface (SS) and several
nanoparticle-coated surfaces. The nanoparticle-coated surfaces
were prepared by an electrophoretic deposition method with dif-
ferent weight of nanoparticles, i.e., 0.3 mg (EPD-1), 0.6 mg (EPD-
2), 0.9 mg (EPD-3) and 1.2 mg (EPD-4). A quantitative analysis
about the active nucleation site density, bubble departure diameter
and bubble departure frequency was conducted. Accordingly, a
heat transfer model was formulated to predict the heat transfer.
In addition, supplementary tests were carried out to explore why
CHF was not enhanced effectively in the present study. The conclu-
sions are summarized as follows:

 The maximum superheat is significantly decreased on the


nanoparticle-coated surfaces, around 20 K lower than that on
the smooth surface.
Fig. 11. Experimental characterization of wickability, (a) high-speed imaging of
capillary wicking from a micro-capillary tube onto SS, (b) absorbed volume as a  Heat transfer on the nanoparticle-coated surfaces is enhanced
function of time for the test surfaces. by around 100% compared with the smooth surface. For exam-
ple, at q = 14.2 W/cm2 the heat transfer coefficient of SS, EPD-1,
EPD-2, EPD-3 and EPD-4 is 0.46 W/cm2K, 0.92 W/cm2K, 1.0 W/
study, the wickability of HFE-7200 on the tested surfaces was mea- cm2K, 0.93 W/cm2K and 0.90 W/cm2K, respectively. However,
sured with the method introduced by Rahman et al. [61]. However, CHF is not enhanced effectively.
a glass microcapillary tube with an inner diameter of 340 mm was  The number of active nucleation sites is considerably increased
used in this study. The wickability is characterized by the absorbed by the nanoparticle coatings. The bubble departure diameter
flow rate defined as and frequency on the nanoparticle-coating surfaces is smaller
      and faster than that on the smooth surface, respectively, at a
dV DV A c Dh
V0 ¼ ¼ lim ¼ lim ð28Þ given heat flux.
dt t¼0 Dt!0 Dt t¼0 Dt!0 Dt t¼0
 A heat transfer model was formulated to predict the heat trans-
where Ac is the cross section area of the microcapillary tube and Dh fer, and it predicted the experimental results well, especially for
is the liquid height drop in time Dt. V is the liquid absorbed volume low and moderate heat fluxes.
by the surface.  The wickability of HFE-7200 is not enhanced effectively by the
In terms of the uncertainty of the liquid absorbed volume V, it nanoparticle coating, which is the reason why CHF is not
can be estimated as enhanced.
" 2 #1=2
dV d Dh d Dh
¼ ¼ ð29Þ Conflicts of interest
V Dh Dh
None.
Here, only the uncertainty due to the liquid height measurement
was considered. In the measurement of the liquid height in the
microcapillary tube, a ruler was used as the calibrated scale, which Acknowledgments
had 110 pixels (x) in 5 mm, with an uncertainty of ±4 pixels. Then
the liquid height could be measured by counting the number of pix- The work is supported by a STINT-NSFC joint project and the
els, and the liquid height drop Dh could be calculated, considering Swedish Scientific Council (VR). Also the China Scholarship Council
the difference of the number of pixels (y) at two time points which (CSC) is acknowledged. Special thanks to Mr. Martin Carlsson and
varied from 4 pixels to 18 pixels with an uncertainty of ±1 pixels. Mr. Mats Bengtsson for the help on the experimental rig and mate-
The uncertainty of Dh could be estimated as rial preparation.
560 Z. Cao et al. / International Journal of Heat and Mass Transfer 133 (2019) 548–560

Appendix A. Supplementary material [27] A. Jaikumar, A. Gupta, S.G. Kandlikar, C.Y. Yang, C.Y. Su, Scale effects of
graphene and graphene oxide coatings on pool boiling enhancement
mechanisms, Int. J. Heat Mass Transf. 109 (2017) 357–366.
Supplementary data associated with this article can be found, in [28] M. Dharmendra, S. Suresh, C.S. Sujith Kumar, Q. Yang, Pool boiling heat
the online version, at https://doi.org/10.1016/j.ijheatmasstransfer. transfer enhancement using vertically aligned carbon nanotube coatings on a
copper substrate, Appl. Therm. Eng. 99 (2016) 61–71.
2018.12.140.
[29] S. Das, B. Saha, S. Bhaumik, Experimental study of nucleate pool boiling heat
transfer of water by surface functionalization with crystalline TiO2
nanostructure, Appl. Therm. Eng. 113 (2017) 1345–1357.
References [30] S. Das, B. Saha, S. Bhaumik, Experimental study of nucleate pool boiling heat
transfer of water by surface functionalization with SiO2 nanostructure, Exp.
[1] D.E. Kim, D.I. Yu, D.W. Jerng, M.H. Kim, H.S. Ahn, Review of boiling heat transfer Therm. Fluid Sci. 81 (2017) 454–465.
enhancement on micro/nanostructured surfaces, Exp. Therm. Fluid Sci. 66 [31] M. Tetreault-Friend, R. Azizian, M. Bucci, T. McKrell, J. Buongiorno, M. Rubner,
(2015) 173–196. R. Cohen, Critical heat flux maxima resulting from the controlled morphology
[2] S. Sinha-Ray, W. Zhang, R.P. Sahu, S. Sinha-Ray, A.L. Yarin, Pool boiling of Novec of nanoporous hydrophilic surface layers, Appl. Phys. Lett. 108 (2016) 243102.
7300 and DI water on nano-textured heater covered with supersonically- [32] Y.Y. Li, Z.H. Liu, B.C. Zheng, Experimental study on the saturated pool boiling
blown or electrospun polymer nanofibers, Int. J. Heat Mass Transf. 106 (2017) heat transfer on nano-scale modification surface, Int. J. Heat Mass Transf. 84
482–490. (2015) 550–561.
[3] R. Pastuszko, Pool boiling heat transfer on micro-fins with wire mesh - [33] M.S. El-Genk, A.F. Ali, Enhancement of saturation BOILING of PF-5060 on
Experiments and heat flux prediction, Int. J. Therm. Sci. 125 (2018) 197–209. microporous copper dendrite surfaces, J. Heat Transf. 132 (2010) 071501.
[4] Z. Cao, B. Liu, C. Preger, Z. Wu, Y.H. Zhang, X.L. Wang, M.E. Messing, K. Deppert, [34] Z. Cao, An Experimental Study on Liquid-Liquid Flow in Microchannels and
J.J. Wei, B. Sundén, Pool boiling heat transfer of FC-72 on pin-fin silicon Pool Boiling Heat Transfer on Nanostructured Surfaces, Licentiate thesis, Lund
surfaces with nanoparticle deposition, Int. J. Heat Mass Transf. 126 (2018) University, 2018.
1019–1033. [35] I.S. Kiyomura, L.L. Manetti, A.P. da Cunha, G. Ribatski, E.M. Cardoso, An analysis of
[5] K.H. Chu, R. Enright, E.N. Wang, Structured surfaces for enhanced pool boiling the effects of nanoparticles deposition on characteristics of the heating surface
heat transfer, Appl. Phys. Lett. 100 (2012) 241603. and ON pool boiling of water, Int. J. Heat Mass Transf. 106 (2017) 666–674.
[6] A. Walunj, A. Sathyabhama, Comparative study of pool boiling heat transfer [36] R.L. Mohanty, M.K. Das, A critical review on bubble dynamics parameters
from various microchannel geometries, Appl. Therm. Eng. 128 (2018) 672–683. influencing boiling heat transfer, Renew. Sustain. Energy Rev. 78 (2017) 466–494.
[7] L. Zhou, W. Li, T. Ma, X. Du, Experimental study on boiling heat transfer of a [37] J.G. Collier, J.R. Thome, Convective Boiling and Condensation, Oxford
self-rewetting fluid on copper foams with pore-density gradient structures, University Press, 1994.
Int. J. Heat Mass Transfer 124 (2018) 210–219. [38] R. Cole, Bubble frequencies and departure volumes at subatmospheric
[8] Y.P. Yang, X.B. Ji, J.L. Xu, Pool boiling heat transfer on copper foam covers with pressures, AIChE J. 13 (1967) 779–783.
water as working fluid, Int. J. Therm. Sci. 49 (2010) 1227–1237. [39] H.T. Phan, N. Caney, P. Marty, S. Colasson, J. Gavillet, How does surface
[9] A.M. Gheitaghy, H. Saffari, D. Ghasimi, A. Ghasemi, Effect of electrolyte wettability influence nucleate boiling?, Comptes Rendus Mécanique 337
temperature on porous electrodeposited copper for pool boiling enhancement, (2009) 251–259
Appl. Therm. Eng. 113 (2017) 1097–1106. [40] J. Kim, B.D. Oh, M.H. Kim, Experimental study of pool temperature effects on
[10] M.S. El-Genk, A.F. Ali, Enhanced nucleate boiling on copper micro-porous nucleate pool boiling, Int. J. Multiphase Flow 32 (2006) 208–231.
surfaces, Int. J. Multiphase Flow 36 (2010) 780–792. [41] G. Courty, A. Foust, Surface variables in nucleate boiling, Che. Engr. Progr.
[11] X.B. Ji, J.L. Xu, Z.W. Zhao, W.L. Yang, Pool boiling heat transfer on uniform and Symposium 51 (1955) 1.
non-uniform porous coating surfaces, Exp. Therm. Fluid Sci. 48 (2013) 198– [42] P. Griffith, J.D. Wallis, The role of surface conditions in nucleate boiling, Tech.
212. Rept. MIT 14 (1958).
[12] Y. Tang, J. Zeng, S. Zhang, C. Chen, J. Chen, Effect of structural parameters on [43] C.Y. Han, P. Griffith, The mechanism of heat trasnfer in nucleate pool boiling-
pool boiling heat transfer for porous interconnected microchannel nets, Int. J. Part I, Int. J. Heat Mass Transf. 8 (1965) 887–904.
Heat Mass Transfer 93 (2016) 906–917. [44] H.K. Forster, N. Zuber, Dynamics of vapor bubbles and boiling heat transfer,
[13] C.M. Kruse, T. Anderson, C. Wilson, C. Zuhlke, D. Alexander, G. Gogos, S. Ndao, AIChE J. 1 (1955) 531–535.
Enhanced pool-boiling heat transfer and critical heat flux on femtosecond [45] W.M. Rohsenow, A method of correlating heat transfer data for surface boiling
laser processed stainless steel surfaces, Int. J. Heat Mass Transfer 82 (2015) of liquids, Tech. Rept. MIT 5 (1951).
109–116. [46] C.L. Tien, A hydrodynamic model for nucleate pool boiling, Int. J. Heat Mass
[14] J.Y. Ho, K.K. Wong, K.C. Leong, Saturated pool boiling of FC-72 from enhanced Transf. 5 (1962) 533–540.
surfaces produced by Selective Laser Melting, Int. J. Heat Mass Transf. 99 [47] N. Zuber, Nucleate boiling. The region of isolated bubbles and the similarity
(2016) 107–121. with natural convection, Int. J. Heat Mass Transf. 6 (1963) 53–78.
[15] B. Liu, Z. Cao, Y.H. Zhang, Z. Wu, A.D. Pham, W.J. Wang, Z.X. Yan, J.J. Wei, B. [48] C.Y. Han, P. Griffith, The mechanism of heat trasnfer in nucleate pool boiling-
Sundén, Pool boiling heat transfer of N-pentane on micro/nanostructured Part II, Int. J. Heat Mass Transf. 8 (1965) 905–914.
surfaces, Int. J. Therm. Sci. 130 (2018) 386–394. [49] B.B. Mikic, W.M. Rohsenow, A new correlation of pool boiling data including
[16] Y.Q. Wang, J.L. Luo, Y. Heng, D.C. Mo, S.S. Lyu, Wettability modification to the effect of heating surface characteristics, J. Heat Transf. 91 (1969) 245–250.
further enhance the pool boiling performance of the micro nano bi-porous [50] R.L. Judd, K.S. Hwang, A comprehensive model for nucleate pool boiling heat
copper surface structure, Int. J. Heat Mass Transf. 119 (2018) 333–342. transfer including microlayer evaporation, J. Heat Transf. 98 (1976) 623–629.
[17] P. Xu, Q. Li, Y.M. Xuan, Enhanced boiling heat transfer on composite porous [51] R.J. Benjamin, A.R. Balakrishnan, Nucleate pool boiling heat transfer of pure
surface, Int. J. Heat Mass Transf. 80 (2015) 107–114. liquids at low to moderate heat fluxes, Int. J. Heat Mass Transf. 39 (1996)
[18] D. Deng, J. Feng, Q. Huang, Y. Tang, Y. Lian, Pool boiling heat transfer of porous 2495–2504.
structures with reentrant cavities, Int. J. Heat Mass Transf. 99 (2016) 556–568. [52] Y.Y. Li, Z.H. Liu, G.S. Wang, A predictive model of nucleate pool boiling on
[19] Y.Y. Hsu, On the size range of active nucleation cavities on a heating surface, J. heated hydrophilic surfaces, Int. J. Heat Mass Transf. 65 (2013) 789–797.
Heat Transf. 84 (1962) 207–213. [53] S. Moghaddam, K. Kiger, Physical mechanisms of heat transfer during single
[20] H.S. Jo, T.G. Kim, J.G. Lee, M.W. Kim, H.G. Park, S.C. James, J. Choi, S.S. Yoon, bubble nucleate boiling of FC-72 under saturation conditions. II: theoretical
Supersonically sprayed nanotextured surfaces with silver nanowires for analysis, Int. J. Heat Mass Transf. 52 (2009) 1295–1303.
enhanced pool boiling, Int. J. Heat Mass Transf. 123 (2018) 397–406. [54] S.J. Thiagarajan, R. Yang, C. King, S. Narumanchi, Bubble dynamics and nucleate
[21] S.M. Kwark, G. Moreno, R. Kumar, H. Moon, S.M. You, Nanocoating pool boiling heat transfer on microporous copper surfaces, Int. J. Heat Mass
characterization in pool boiling heat transfer of pure water, Int. J. Heat Mass Transf. 89 (2015) 1297–1315.
Transf. 53 (2010) 4579–4587. [55] C. Gerardi, J. Buongiorno, L.W. Hu, T. McKrell, Study of bubble growth in water
[22] S. Das, D.S. Kumar, S. Bhaumik, Experimental study of nucleate pool boiling pool boiling through synchronized, infrared thermometry and high-speed
heat transfer of water on silicon oxide nanoparticle coated copper heating video, Int. J. Heat Mass Transf. 53 (2010) 4185–4192.
surface, Appl. Therm. Eng. 96 (2016) 555–567. [56] C.M. Voutsinos, R.L. Judd, Laser interferometric investigation of the microlayer
[23] H.S. Jo, M.W. Kim, K. Kim, S. An, Y.I. Kim, S.C. James, J. Choi, S.S. Yoon, Effects of evaporation phenomenon, 97 (1975) 88 – 92.
capillarity on pool boiling using nano-textured surfaces through [57] C.H. Wang, V.K. Dhir, Effect of surface wettability on active nucleation site
electrosprayed BiVO4 nano-pillars, Chem. Eng. Sci. 171 (2017) 360–367. density during pool boiling of water on a vertical surface, J. Heat Transf. 115
[24] H.S. Jo, S. An, H.G. Park, M.W. Kim, S.S. Al-Deyab, S.C. James, J. Choi, S.S. Yoon, (1993) 659–669.
Enhancement of critical heat flux and superheat through controlled [58] R.J. Benjamin, A.R. Balakrishnan, Nucleation site density in pool boiling of
wettability of cuprous-oxide fractal-like nanotextured surfaces in pool saturated pure liquids: effect of surface microroughness and surface and liquid
boiling, Int. J. Heat Mass Transf. 107 (2017) 105–111. physical properties, Exp. Therm. Fluid Sci. 15 (1997) 32–42.
[25] R.P. Sahu, S. Sinha-Ray, S. Sinha-Ray, A.L. Yarin, Pool boiling of Novec 7300 and [59] T. Hibiki, M. Ishii, Active nucleation site density in boiling systems, Int. J. Heat
self-rewetting fluids on electrically-assisted supersonically solution-blown, Mass Transf. 46 (2003) 2587–2601.
copper-plated nanofibers, Int. J. Heat Mass Transf. 95 (2016) 83–93. [60] Y. Im, C. Dietz, S.S. Lee, Y. Joshi, Flower-like CuO nanostructures for enhanced
[26] Z. Cao, C. Preger, Z. Wu, S. Abbood, M.E. Messing, K. Deppert, B. Sundén, Pool boiling, Nanosc. Microsc.e Thermo. 16 (2012) 145–153.
boiling heat transfer of water on copper surfaces with nanoparticles coating, [61] M.M. Rahman, E. Olceroglu, M. McCarthy, Role of wickability on the critical heat
ASME IMECE (2017), V008T10A065-V008T10A065. flux of structured superhydrophilic surfaces, Langmuir 30 (2014) 11225–11234.

You might also like