Experimental Thermal and Fluid Science 113 (2020) 110025

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

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

Pool boiling heat transfer of HFE-7100 on metal foams T

a b a
Leonardo Lachi Manetti , Gherhardt Ribatski , Reinaldo Rodrigues de Souza ,
⁎
Elaine Maria Cardosoa,
a
UNESP – São Paulo State University, School of Engineering, Post-Graduation Program in Mechanical Engineering, Av. Brasil, 56, 15385-000, Ilha Solteira, SP, Brazil
b
Heat Transfer Research Group, Department of Mechanical Engineering, Escola de Engenharia de São Carlos (EESC), University of São Paulo (USP), Brazil

A B S T R A C T

The search for new techniques to increase boiling heat transfer has been driven by more efficient and compact heat exchangers, especially in microelectronics and
equipment with high thermal loads. Two-phase cooling systems are a promising thermal management technology for high-heat dissipation. In this context, the
present study investigated the performance of modified heating surfaces consisting of metal foams of nickel (Ni) and copper (Cu). Pool boiling tests were performed
using HFE-7100 as working fluid, at saturation conditions. The metal foams surfaces provided a higher heat transfer coefficient compared to plain surfaces and
prevented thermal overshoot at the onset nucleate boiling. The Cu foam provided the best performance for the entire boiling curve. In general, for low and moderated
heat fluxes, there is a combined effect of surface area and thermal conductivity of foams; the high surface area of Ni foam provides a barrier for the departure of the
vapor bubble, inhibiting the cooling effect of the heating surface. For the Cu foam, no significant vapor trapped effect was observed, and the highest heat transfer
coefficient was 12.4 kW/m2∙K for a heat flux around 270 kW/m2; in addition, the thermal behavior is a function of the permeability and wickability behaviors of the
surfaces.

1. Introduction under pool boiling conditions were evaluated and are summarized in
the broad literature review presented by Shojaeian and Koşar [6].
Efficient thermal management solutions are crucial to maintaining Among the techniques presented in this study, engineering the surface
new electronic devices within the operating temperature limits [1]. morphology was one of the methods commonly used to enhance the
Thermal management systems based on phase change such as pool and heat transfer coefficient (HTC) and the critical heat flux (CHF). Such
flow boiling, jet impingement, and sprays cooling are promising solu- goals can be achieved by coating the surface with metal foams (high
tions to attend such high power dissipation needs [2]. porosity surfaces) in order to create a porous structure that allows fluid
Pool boiling heat transfer with the cooling device contacting di- inflow to keep nucleation sites active and protrusions that enlarge the
rectly the coolant eliminates the contact resistance associated to a boiling surface area [7,8].
significant temperature jump between the heat sink and the cooled El-Genk and Parker [9] investigated pool boiling of HFE-7100 on a
device. However, such solution demands fluids with low electric con- graphite foam with porosity ε = 60%, under saturation conditions. At
ductivity, e.g., water cannot be used. In this way, the fluorochemical high heat fluxes, they observed for the porous surface a reduction
fluids as fluorocarbons (FC-72, FC-82) and hydrofluoroether (HFE- around 9 K of the wall superheating compared to the plain surface. In
7100, HFE-7200, HFE-7300) become suitable solutions due to their addition, the temperature wall overshoot at the incipient boiling was
dielectric and inert properties; moreover, they are stable, non- not observed. Xu et al. [10] used acetone as working fluid on copper
flammable, and non-reactive, i.e., chemically compatible [3]. Besides, foam (ε ≥ 88%) with three different porous densities (porous per inch,
FC-72 and HFE-7100 present a low boiling point (Tsat ≈ 60 °C at 1 atm) PPI): 30, 60, and 90. For all foams configurations, the wall temperature
to maintain the electronic components at low operating tempera- excursion at boiling incipience was negligible. The authors associated
tures, < 85 °C [4]. Additionally, the hydrofluoroether fluid, HFE-7100 the HTC enhancement to the higher liquid contact area, higher density
for instance, possesses superior environmental properties (ozone de- of active sites, and capillary-assist liquid flow towards foam cells. Yang
pleting potential, ODP = 0; global warming potential, GWP100 = 320) et al. [11] used the same surfaces and experimental methodology of Xu
[5]. Finally, fluorochemical fluids are highly wetting liquids, which et al. [10] for water as working fluid. Based on the results of both
suppress nucleate boiling and highlights the demand for engineered studies, Yang et al. [11] pointed out that the HTC enhancement on the
surfaces in order to keep high heat transfer rates. metal foams depends on the balance between the liquid suction cap-
In this context, several combinations of liquid-surface operating ability towards the porous structure and the vapor release resistance to

⁎
Corresponding author.
E-mail address: elaine.cardoso@unesp.br (E.M. Cardoso).

https://doi.org/10.1016/j.expthermflusci.2019.110025
Received 9 July 2019; Received in revised form 8 December 2019; Accepted 15 December 2019
Available online 26 December 2019
0894-1777/ © 2019 Elsevier Inc. All rights reserved.
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Nomenclature Ra average roughness, [µm]

Rp maximum peak height of surface roughness, [µm]
Alphabetic S square cross-section area, [m2]
V̇ wicked volume, [m3]
asf Area density, [m2/m3] T temperature, [K]
b Exponent relative to roughness effects, [−] t time, [s]
cp Specific heat capacity, [J/kg∙K] ui uncertainty
Csf Surface-fluid coefficient, [−] u fluid velocity, [m/s]
Dd Bubble departure diameter, [m]
df Fiber diameter, [m] Greek symbols
dh liquid column variation, [m]
dp pore diameter, [m] δ foam thickness, [m]
F inertia coefficient, [-] ΔT Temperature difference, [K]
fw heat surface material parameter , [-] ε porosity, [−]
h heat transfer coefficient, [W/m2∙K] ηfin pin-fin efficiency, [−]
k thermal conductivity, [W/m∙K] ηfoam foam efficiency, [−]
kfoam foam effective thermal conductivity, [W/m∙K] θ static contact angle, [degree]
kCu copper thermal conductivity, [W/m∙K] µ dynamic viscosity, [kg/m∙s]
kmat foam material thermal conductivity, [W/m∙K] ρ density, [kg/m3]
K permeability, [m2] σ surface tension, [N/m]
L copper block distances, [m]
Lc corrected fin/foam length, [m] Subscripts
M molar mass, [kg/kmol]
m exponent of heat flux in Eq. (15); fin/foam efficiency 1, 2, or 3 thermocouples position
coefficient in Eq. (18) and Eq. (19), [−] air air properties
ṁ mass flow rate, [kg/s] atm atmospheric condition
n Stephan exponent, [−] l or liq liquid
P pressure, [Pa] s surface square cross-section
Pr Prandtl number, [−] sat saturated state
pr reduced pressure, [−] vap vapor
q”measured heat flux measured at the copper block, [W/m2] w surface wall

the bulk liquid. Xu and Zhao [12] investigated saturated pool boiling that porous structures increase the bubble nucleation site density due to
heat transfer of deionized (DI) water on a copper foam surface with V- the larger surface area and capillary-assisted liquid suction which im-
shape grooves. They reported that large foam thickness offers high proves liquid replenishment; however, a larger surface area corre-
surface areas, however, it creates considerable flow resistance to the sponds to an increase in the form drag and, consequently, in the bubble
bubble scape. In this sense, grooves in the foam structure could help the evacuation resistance. The bubble evacuation resistance increases for
vapor bubbles escape. In the same way, Li et al. [13] investigated the higher heat fluxes, structure height and for smaller unit cell sizes.
influence of copper foam (fixed at ε = 0.98, 110 PPI and δ = 2 mm) Doretti et al. [14] carried out tests for aluminum foams (ε = 92%),
with and without channels on boiling heat transfer of R141b at atmo- 10 mm thick and porous densities of 5, 10, and 40 PPI. The boiling
spheric pressure. The results showed that the copper foam without curves for the surfaces covered with foam presented remarkably lower
channels enhanced the HTC at heat fluxes lower than 180 kW/m2, but it wall superheating compared to the plain surface. Moreover, 10 PPI and
deteriorated the heat transfer at relatively high heat fluxes compared to 40 PPI foams exhibited the same behavior, which was slightly different
the plain surface. On the other hand, HTCs for copper foams with from that for 5 PPI. In the case of 5 PPI and for the range of experi-
channels were all improved. The results were different according to the mental conditions evaluated by them, the HTC increased continuously
channel diameter and number; more channels and larger channels while for 10 PPI and 40 PPI the HTC increased according to an almost
diameter reduced the bubble escaping resistance at high heat fluxes, constant gradient until a heat flux of 500 kW/m2. Then, for heat fluxes
which resulted in higher HTCs. However, when the channel number higher than this value, the gradient of the heat transfer coefficient is
and channel diameter were too large, the heat transfer enhancement significantly reduced. At low heat fluxes, the vapor bubbles are small
was reduced due to the decreasing heat transfer area and capillary and can easily escape from the foam. As the heat flux increases, the
force. Recently, Zhou et al. [8] investigated copper foams (with 10, 50, bubbles size grow, experiencing larger flow resistance through the
and 110 PPI) with a pore density gradient in their structure and using porous layer: the highest the PPI, the largest flow resistance.
water with and without n-butanol solution at 6 wt% as working fluid. In this paper, the results of an experimental investigation con-
According to their results, the metal foams increase the HTC perfor- cerning pool boiling of HFE-7100 on metal foams of copper (Cu) and
mance due to the increase of porous density for both fluids tested. In nickel (Ni) are presented. The present work aims to: (i) provide lit-
addition, for high heat fluxes, the surface tension of the n-butanol so- erature with new heat transfer data; (ii) understand the vapor bubble
lution within the porous matrix decreases due to the increase of its dynamics associated to the heat transfer performance based on si-
concentration. Therefore, the bubble departure diameter is also re- multaneous high-speed filming and heat transfer coefficient measure-
duced, allowing the vapor bubbles release. On the other hand, for water ments; (iii) and provide new insights to the development of new heat
without n-butanol solution, the gradient structure seems to present less transfer enhancement surfaces operating with fluorochemical fluids.
effect on the HTC enhancement due to the foam caging effect. Wong
and Leong [7] printed a homogeneous metal foam structure by using
2. Foams parameters
Selective Laser Melting (SLM) in order to study the effect of porous
diameter and foam thickness on pool boiling of FC-72. They concluded
The porous surfaces used in the present study were purchased from

2
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Nanoshel® in 500 × 500 mm2 panels. They are open-cell metal foam, obtained from Forchheimer–Darcy′s equation, re-written in this text as
with 3 mm thick, fabricated by using metal deposition in a cellular follows:
preform as detailed by Ashby et al. [15] and Banhart [16]. They were ρ F μ
dP 1
cut in 16 × 16 mm2 panels by using a wire electrical discharge ma- − ∙ = air u + air = a∙u + b
dz u K K (1)
chining (wire-EDM).
Their porosity (ε) was evaluated by weighting seven samples with where u is the fluid velocity at the inlet of the foam (based on the cross-
the same size in an analytical balance and comparing the foam weight section of the empty structure, S),
with that of a solid sample with an equivalent volume maiṙ
(16 × 16 × 3 mm3). Moreover, the following methods were used to u=
ρair ∙S (2)
characterize the metal foams: (i) optical images by using a stereo mi-
croscopy Zeiss® SteREO Discovery.V8 and scanning electron micro- and F is the dimensionless inertia coefficient. Moreover, µair is the air
scopy (SEM) by using an EVO LS15 Zeiss®, Tab. 1; (ii) microcomputer dynamic viscosity and, ρair, the air density (both have been evaluated at
tomography (µCT) by using a Skycan 1272 at a resolution of 15 µm the inlet). The air density was calculated by using the equation of state
(100 kV X-ray source voltage); (iii) permeability based on air flow for an ideal gas - by considering the values of the pressure and tem-
through the foam; and (iv) wickability based the wicked volume in a perature in the upstream flow - while the air dynamic viscosity was
capillary tube (1 mm diameter), Tab. 3. The implementation of these calculated by using the equation used by Gasche et al. [24], as follows,
two last methods is described in items 2.1 and 2.2, respectively.
μair = (0.872 + 7.029 × 10−2∙T − 3.81 × 10−5∙T 2) × 10−6 (3)
The number of pores per inch (PPI) also is a parameter that influ-
ences the heat transfer. This parameter was measured as proposed by In the present study, as presented by Mancin et al. [23], the test
Athreya et al. [17] and Zhu et al. [18]. In order to perform the PPI facility schematically illustrated in Fig. 3 was built in order to evaluate
measurements by using the Stereo images, seven lines in each direction the variation of the pressure drop gradient with the air flow velocity,
– horizontal and vertical, were traced and the number of porous in- and, then, from the curve given by Eq. (1), to extract the foams per-
tercepted by the lines was counted; so, an average yields the PPI of the meability.
open-cell metal foam. For Cu and Ni open-cell metal foam, the average The experimental tests were carried out using air from a reservoir
PPI values were 31.75 and 62.72, respectively. Moreover, the µCT containing compressed air at 9 bar. The mass flow rate was controlled
images were used for measuring the foam porous diameter and fibers by a needle valve and measured through a Coriolis mass flow meter
diameters (dp and df, respectively), and area density (asf). First, the µCT (Rheonik RHM 04) with an operating range of 5.0 kg/min and max-
virtual slices were input in the Imoph software [19] in order to measure imum uncertainty equal to ± 0.2% of the reading. A differential
the surface granulometry by using the aperture map of both phases transducer with a measurement range from 0 to 0.1 bar (error of 0.1%
(solid and porous) as explained by Vicente et al. [20]. Fig. 1 shows the of the full scale, FS) was used to measure the pressure drop across the
histogram and the normal distribution that allows defining a mean pore foam surface. A barometer with 0.05 kPa resolution measured the at-
diameter for each foam sample. Next, the µCT virtual slices were input mospheric pressure. The upstream pressure was evaluated by gauge
in the vmtk® software to measure the area density. vmtk® uses pressure transducer with 1 bar FS and uncertainty of ± 0.1% FS. A
‘marching cubes’ algorithm to extract interfaces between the porous PT100 sensor (as part of the Coriolis mass flow meter) measured the
and solid phases, creating a polygonal mesh surface (see Fig. 2). In temperature of the flow [24]. Fig. 4a shows the variation of the pressure
order to distinguish the two phases, a grayscale threshold value was drop gradient and Fig. 4b shows the pressure drop gradient - multiplied
selected such that the porosity of the reconstructed 3D volume matched by the factor 1/u, as shown on the left side of the Eq. (1) - plotted
with the measured foam porosity (Table 1) as explained by Sarangi against the air velocity.
et al. [21]. After setting the correct grayscale, the area density was From the linear curves (Fig. 4b) it was possible to calculate the
obtained by the ratio between the surface solid area (foam area) and its permeability for both metal foams. It was obtained K equal to
total volume (solid volume + porous volume). The results of ε, dp, df, 0.66 × 10−7 ± 0.43 × 10−7 m2 and
with the respective standard deviation according to a confidence in- 0.15 × 10−7 ± 0.005 × 10−7 m2; and, F equal to 0.275 ± 0.091 and
terval of 95%, and asf are given in Table 2. 0.136 ± 0.003 for Cu foam and Ni foam, respectively. As expected, the
Cu foam presents a higher permeability due to the larger pore diameter
2.1. Foams permeability than the Ni foam.

The permeability (K) is a foam characteristic associated to the ca- 2.2. Metal foams capillary wicking
pacity of flow conductance through a porous media and is expressed in
m2. According to Mancin et al. [23], the foam permeability can be In the present study, the foam capillary wicking behavior was

Fig. 1. Pore diameter distribution: (a) Cu foam; (b) Ni foam.

3
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Fig. 2. Rendered tridimensional surfaces: (a) Cu foam; (b) Ni foam.

evaluated for HFE-7100 since authors [7,10,11,14] associated the heat 3. Experimental facility and data reduction
transfer enhancement provided by the foam to capillary wicking effect.
Ahn et al. [25], Rahman et al. [26], Cao et al. [27,28] carried out a 3.1. Experimental facility
capillary-wicking test where the porous surface is slowly raised to
contact a pendant fluid droplet attached to a small diameter capillary The pool boiling tests were performed in the apparatus shown in
tube. As the surface contacts the liquid droplet, the fluid is wicked into Fig. 6, which consists of a rectangular vessel (120 × 100 × 200 mm3)
the porous structure and the volumetric flow rate is measured by made of glass with wall thickness of 5 mm. The upper and lower bases
monitoring the liquid meniscus in the tube. consisted of two stainless steel plates, 200 × 200 × 10 mm3. The re-
The experimental layout is shown in Fig. 5 where the capillary tube quired tightness was obtained with nitrile rubber and silicone; ad-
had 1 mm in diameter. The z-axis raised the surface up to it touches the ditionally, screws were passed through the upper and bottom stainless
tube while the high-speed camera recorded the meniscus displacement. steel plate in order to press the glass box against the nitrile rubber.
After that, tracking image software was used to measure the liquid Water from a thermal bath circulated through a cooling coil, located at
column variation inside the capillary tube and, then, the volume wicked the top of the boiling chamber. This heat exchanger was used to con-
was calculated. Table 3 shows the volume wicked by the surface at the densate the vapor generated by the heaters and to control the saturation
first 50 ms. pressure inside the vessel. An auxiliary heater – a cartridge resistance
One may observe that the Cu foam showed the best capillary with a maximum power of 250 W at 220 V – submerged in the working
wicking behavior with a total volume of 1.4 µL absorbed during the first fluid was used to maintain the liquid temperature near the saturation
50 ms while the Ni foam absorbed 0.84 µL. The plain surface showed state. This heater was powered through a variable transformer. Two K-
the worst wicking result of 0.51 µL. According to Xu et al. [10], foams type thermocouples, Tliq and Tvap, located in the liquid and vapor re-
with small pore size present a large capability of pumping liquid due to gions of the vessel, respectively, were used to monitor the test fluid
their higher capillary pressure (2σ∙cosθ/dp). Thus, according to this temperature. An absolute pressure transducer Omega PXM309-2A
hypothesis, it was expected a higher wickability for the Ni foam. measured the pressure inside the boiling chamber. The experiments
However, the fact that the Ni foam presents a higher contact area with were performed under conditions close to the local atmospheric pres-
the liquid, increasing the flow friction, seems to explain the higher sure, patm = 98 kPa. The pressure uncertainty, according to the cali-
wickability of the Cu foam. Wu et al. [29] reported similar results, bration curve, is ± 0.05 kPa.
where surfaces with small pores could facilitate the liquid wicking, The test section, illustrated in Fig. 7, consisted of a copper piece
while the surface irregularities could inhibit the liquid spreading as with a square plate on the upper surface (16 × 16 × 3 mm3) of the
compared to surfaces with a larger number of micro-pores and micro- copper cylinder. The test section was machined from a unique copper
structures (which assists the liquid spreading and wicking). piece in order to avoid thermal contact resistance among its compo-
nents. Three K-type thermocouples (T1, T2, and T3) with hot junction
diameters of 0.5 mm were embedded within the cylinder. The ther-
mocouple junctions were fixed at the center axis at the end of radial

Table 1
Metal foams Stereo and SEM images.
Material Stereo (top view) SEM (top view) SEM (fiber cross section)

Cu

Ni

4
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Table 2
Metals foams characteristics.
Material Weight Foam density Relative density1(%) ε dp df asf
(kg × 10−3) (kg/m3) (%) (mm) (mm) (m2/m3)

Cu 0.697 ± 0.022 908.1 ± 28.63 10.0 ± 0.32 90.0 ± 0.32 0.52 ± 0.22 0.09 ± 0.04 2166
Ni 0.106 ± 0.010 138.0 ± 14.12 1.6 ± 0.15 98.4 ± 0.15 0.22 ± 0.10 0.04 ± 0.02 5133

1
Pure material density: ρCu = 8960 kg/m3; ρNi = 8900 kg/m3 [22].

Fig. 3. Layout of the experimental apparatus for permeability measurements.

holes (1 mm diameter), machined in the cylinder. After positioning the reached. Then, the chamber was feed with the working fluid. Before
thermocouple in the cylinder, the holes were filled with copper powder each series of tests, to eliminate non-condensable gases dissolved in the
in order to ensure low contact resistance and avoid empty spaces, which refrigerant, the auxiliary heater was powered and the working fluid
could affect the temperature field in the copper piece. This cylinder is boiled during a period of one hour.
fixed on a second copper block containing a heater cartridge with a The test conditions were adjusted by monitoring the pressure and
maximum power of 300 W at 220 V responsible for heating the test the temperature inside the boiling chamber. For each metal foam test,
section. This resistance is powered by a variable DC power source. The the experiment was carried out at least twice under similar conditions
test section was thermally insulated from the environment by a radial to ensure that the results were repeatable.
polytetrafluoroethylene (PTFE) layer with 30 mm thick. A data acquisition system (Agilent 34970A) was used to acquire all
The tests were performed using HFE-7100 (3MTM NovecTM) as the data signals (power, pressure, and temperature) and, then, they
working fluid at a saturation pressure of 98 ± 1 kPa. Results for a plain were registered in a personal computer using the Agilent Benchlink
surface (Ra = 0.14 µm, Fig. 8a) polished according to the procedure Data Logger.
presented by Manetti et al. [30] were used as reference. To prevent the The heating effect was imposed by increasing the electrical power
natural parasite sites at the interface between the PTFE piece and the according to heat flux steps of 10 kW/m2 until a condition close to the
test section surface, two-component epoxy resin was used to fill the gap dryout heat flux corresponding to a maximum footprint heat flux of
as shown in Fig. 8a. The metal foams with a thickness (δ) of 3 mm 375 kW/m2. Data were recorded for each heat flux step after the es-
(Fig. 8b and c) were welded on the copper block using tin-lead (0.1 mm tablishment of steady-state conditions, characterized by variations in
thick) as solder, in order to ensure a low thermal resistance between the the measured temperatures within the uncertainty of their measure-
foam and the test section. The junction was carried out by a soldering ments ( ± 0.3 °C). At least 100 data points were recorded, corre-
process. First, a tin-lead (50–50%) bar was rolling to obtain a layer with sponding to 500 s of steady-state.
thickness close to 0.1 mm. Next, an organic flux was applied in the
copper plate in order to prevent oxides and to ensure good tin-lead 3.2. Data regression procedure and uncertainties estimation
contact with the surface. Thus, the copper block was heated by a car-
tridge resistance up to the tin alloy melting point (≈200 °C); then, a tin- Fig. 9 shows the temperature profiles along the copper cylinder
lead layer was added on the copper plate and the metal foam was central axis based on the measurements of the thermocouples T1, T2,
placed on it. Finally, a dead weight was put on the foam to pressure it and T3 for different heat fluxes. According to this figure, linear curves
on the plain surface. The assembly was cooled and the organic flux was fitting with R-square higher than 0.99 are obtained. Therefore, the as-
removed with alcohol and acetone. sumption of negligible heat losses in the radial direction seems rea-
Before charging the vessel with the HFE-7100, a vacuum pump was sonable, as previously reported by Wu et al. [29] and Kiyomura et al.
used in the boiling chamber until an absolute pressure of 10 kPa was [31].

Fig. 4. Permeability tests: (a) experimental pressure gradient and (b) Regression analysis.

5
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Fig. 5. Schematic of the experimental device used for the wickability measurements.

Table 3
Metal foams capillary wicking.
Plain

0 ms 10 ms 50 ms
Cu foam

0 ms 10 ms 50 ms Fig. 6. Pool boiling apparatus: (1) cooper block; (2) auxiliary heater; (3)
pressure transducer; (4) vacuum/feed valve; (5) cooling coil; (6) glass chamber;
Ni foam
(7) stainless steel plate.

'' ''
qmeasured qmeasured
h= =
Tw − Tsat (pint ) ΔTsat (5)
where Tsat(pint) corresponds to the saturation temperature of the HFE-
7100 at the pressure inside de boiling chamber estimated as follows [5]:
3641.9
0 ms 10 ms 50 ms lnPsat (atPascal) = − + 22.415
Tsat (atKelvin) (6)
and Tw is the surface wall temperature also estimated according to
Thus, the heat flux was estimated based on the Fourier law con-
Fourier′s law and given as follows:
sidering the thermocouples T1 to T3 as follows:
''
qmeasured 4 q ''
π ΔT Tw = T1 − ∙ L1s − measured ∙Lsw
''
qmeasured = ∙kCu ∙ 13 kCu π kCu (7)
4 L13 (4)
where the second term in the right-hand side is associated to the linear
where L13 and ΔT13 are the distance from the thermocouples 1 to 3 and temperature profile at the end of circular section (L1s = 5 mm) and the
the temperature difference between them equal to 12 mm, as shown in third term is associated to the linear temperature profile along the
Fig. 7. square section (Lsw = 3 mm) as shown Fig. 7.
The HTC was calculated based on Newton′s law of cooling given by: The experimental uncertainties (u) were calculated by using the
method described by Moffat [32] where the uncertainty in the result R
is a function of the independent variables Xi as follow:

6
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Fig. 9. Verification of the one-dimensional Fourier law at the circular cross-

section (solid lines) and linear temperature profiles used to estimate the wall
Fig. 7. Distance between the thermocouples and the surface in contact with the temperatures at the square cross-section (dotted lines).
working fluid.

4. Results and discussion

1
n 2 2
⎡⎛ ∂R ⎞⎤
uR = ⎢ ⎜∑ u Xi ⎟ ⎥ 4.1. Validation of the experimental apparatus
⎢ ∂Xi
⎣⎝ i ⎠⎥⎦ (8)
In order to check the accuracy of the pool boiling apparatus and the
Therefore, the relative uncertainty for the heat flux between the data regression procedure, tests were carried out for HFE-7100 on the
thermocouples 3 and 1 was given by: plain surface, comparing them with prediction values based on well-
uq '' 2 2 1/2 known correlations from the literature (Fig. 10). The re-written version
measured
=⎡ ⎛ uΔT13 ⎞ ⎛ u L13 ⎞ ⎤ of the Rohsenow′s correlation [33] is given:
⎢ ΔT13 + L13 ⎥
⎜ ⎟ ⎜ ⎟
''
qmeasured ⎣⎝ ⎠ ⎝ ⎠⎦ (9)
1/3
1 q''cpl ⎡ μl hlv g (ρl − ρv ) ⎤
where the uncertainty of the K-type thermocouples was ± 0.3 °C (cor- h=
Csf Prl hlv ⎢
1.7
⎣ q
'' σ ⎥
⎦ (12)
responds to the thermocouples uncertainties after the calibration); the
uncertainty of the thermocouple junction position was ± 0.03 mm, and where µl, hlv, cpl, and Prl are the liquid dynamic viscosity (kg/m∙s), the
the wall superheating uncertainty was given by: latent heat of vaporization (J/kg), the liquid specific heat (J/kg∙K), and
2
the Prandtl number of the liquid, respectively. In addition, σ is the
⎡ −(4/ π ) ∙L1s − Lsw surface tension, ρl and ρv are the liquid and vapor densities, respec-
uΔTsat = ⎢uT21− Tsat + ⎛⎜ uq '' ⎞
⎟

⎢ ⎝ kCu measured
⎠ tively, and g is the gravitational acceleration. Csf is a coefficient that
⎣
'' 2 '' 2 1/2 depends on the fluid and the heating surface material. The thermo-
⎛ −(4/ π ) ∙qmeasured ⎞ ⎛q ⎞⎤ physical properties of the HFE-7100 were obtained from 3 M Company
+⎜ u L1s ⎟ + ⎜ measured u Lsw ⎟ ⎥
kCu kCu [5] and Raush et al. [34] at psat = 98 kPa.
⎝ ⎠ ⎝ ⎠⎥⎦ (10)
Cooper [35] proposed a correlation to predict the boiling heat
Finally, the HTC uncertainty was given by: transfer coefficient taking into account the surface roughness and re-
duced pressure of the liquid,
2 1/2
uh ⎡ uΔT 2 ⎛ uq '' ⎞ ⎤
= ⎢ ⎛ sat ⎞ + ⎜ ''measured ⎟
⎜ ⎟
⎥ h = 55prb (−logpr )−0.55 M−0.5q''0.67 (13)
h ΔT q
⎣⎝
⎢ ⎠ ⎥
sat
⎝ measured ⎠ ⎦ (11)
where b = (0.12 − 0.2logRp) , and pr, M and Rp represent the reduced
Therefore, the experimental uncertainty of the heat transfer coeffi- pressure, the molar mass of the working fluid, and the maximum peak
cient is higher for low heat fluxes, decreasing as heat fluxes increase. height of surface roughness, respectively. In the present study, the
For all surfaces tested, the experimental uncertainty for the heat flux correlation was evaluated based on the average surface roughness, Ra,
and the heat transfer coefficient varied from 18.3% to 3.3% and from whereas Cooper′s correlation includes the parameter, Rp. Thus, for the
18.4% to 3.8%, respectively. present analysis, the following expression suggested by Gorenflo [36]

Fig. 8. Heating surfaces: (a) plain; (b) Cu foam; and (c) Ni foam.

7
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

cavities in the surface. Additionally, Wong and Leong [7] reported that
lower boiling incipience superheat is due to the larger surface area and
more active nucleation sites (due to the porous structure). According to
them [7], the incipience nucleation is highly sensitive to the nucleation
sites available on the surface. For the porous structure, the intersections
of struts form cavities, which can be suitable for nucleation, thus al-
lowing lower boiling incipience superheat as compared to the plain
surface. Xu et al. [10] also reported that miniature bubbles nucleate at
the foam ligaments and fiber junctions in foam cells. The hypothesis of
El-Genk and Parker [9] associated to the effect of entrapped air is also
supported by [7,10] because even they did not mention it to justify the
decrease of the thermal overshooting, the phenomenon of the air en-
trapped is cited by them, i.e., the nucleation sites in the intersections of
struts and at the foam ligaments and fiber junctions in foam cells can
occur due to the vapor embryo caused by the air entrapped. In addition,
in the present study bubble nucleation was also observed at the foam
Fig. 10. Pool boiling apparatus validation with HFE-7100. ligaments, although a vacuum pump has been used to eliminate the air
within the chamber and surface (air can remain in the porous struc-
was used: ture).
For moderated heat fluxes (≈120 kW/m2), the HTC enhancement
Ra = 0.4Rp (14)
becomes evident by a decrease in the wall superheating of 11.4 K for Cu
Likewise, Ribatski and Jabardo [37] proposed an empirical corre- foam and 12.1 K for Ni foam, as compared to the plain surface for a
lation to estimate the boiling heat transfer in terms of reduced pressure. similar heat flux. This result implies a HTC augmentation of 100% and
Their correlation was based on the experimental data for saturated pool 82% promoted by the Cu and Ni foams, respectively. For heat flux
boiling of halocarbon refrigerants on cylindrical surfaces of different values higher than 120 kW/m2, the vapor within the Ni foam structure
materials: began to create an unstable boiling pattern, probably promoting in-
termittent dry patches within the foam structure, which increases the
h = fw pr0.45 (−logpr )−0.8 M−0.5Ra0.2 q''m (15) wall temperature and, consequently, degrades the heat transfer coeffi-
where m = 0.9 − 0.3pr0.2
and fw is the heat surface material parameter, cient. For high heat fluxes (≈200 kW/m2), the wall superheating for Cu
equal 100 to copper. and Ni foams is about 16.6 K and 22.5 K, respectively. For the Ni foam,
Finally, the curve fitting proposed by Stephan [38] based on the the HTC decreases for heat fluxes higher than 125 kW/m2, indicating
relation between the heat transfer coefficient and the heat flux was the occurrence of the dryout phenomenon associated with the vapor
considered: captured by the foam structure, which inhibits the liquid replenishment
effect, degrading the heat transfer coefficient (Fig. 12). In addition,
h = Cq"n (16)
Wong and Leong [7] noted that the smaller the pore cell, the higher is
where C is a coefficient dependent on the surface-fluid interaction and the bubble evacuation resistance. This causes a large liquid-vapor
reduced pressure, and n is an exponent of the heat flux. counter-flow since the liquid replenishment is inhibited, leading to a
The predicted values agree reasonably well with the experimental lower heat transfer coefficient and an earlier occurrence of dryout.
data providing a mean absolute percentage error (MAPE) of 11% for the As shown in Table 2 (Section 2), smaller pore diameters and higher
prediction method of Rohsenow [33] using the value of Csf = 0.005 as PPI implies increasing the surface area density. Hence, it can be con-
proposed by Cao et al. [39]; Cooper′s correlation [35] presents a MAPE cluded that the Ni foam presents a higher contact area between its ef-
of 26% and, Ribatski and Jabardo [37] correlation presents the lower fective surface and the fluid than the Cu foam, which should increase
mean absolute percentage error, equal to 10.6%. The exponent n ob- the HTC. However, the thermal conductivity of Cu is approximately 5
tained by fitting the curve h vs. q′′ as proposed by Sthepan′s was around times higher than the Ni thermal conductivity; in addition, the metal
0.7, agreeing with the values from literature which generally lies be- foam thermal conductivity is different from the solid material one, due
tween 0.6 and 0.8. It is worth mentioning that the last two points in to the porous medium. Thus, the metal foams thermal conductivity was
Fig. 10 are close to the onset of dryout heat flux; thus, it is reasonable calculated by using the model of Calmidi and Mahajan [41] that con-
that the pool boiling prediction method does not capture the trend of siders the porous medium formed by two-dimensional array of
the heat transfer coefficient under such condition. According to Cardoso
et al. [40], the onset of dryout was defined as the condition at which the
heat transfer coefficient presents a maximum value.

4.2. Metal foams pool boiling curves

The metal foams boiling curves are shown in Fig. 11 and the cor-
responding curves displaying the HTC behaviors in Fig. 12. Both figures
include the results for the plain surface. One may observe that both
metal foams eliminated the thermal overshoot necessary to the onset of
nucleate boiling (ONB) that is observed in the magnification of Fig. 11.
In fact, nucleation sites are already activated for the lowest evaluated
heat flux (≈10 kW/m2) and superheating lower than 5 °C. For the plain
surface, the sites remain non-actives until a heat flux of 15 kW/m2 and
superheating higher than 15 °C was attained as can be seen again in the
magnification of Fig. 11. El-Genk and Parker [9] also detected it for
graphite porous surface, attributing this phenomenon to the entrapped
air in the high volume interconnected porosity and the re-entrant Fig. 11. HFE-7100 pool boiling curves at saturation temperature.

8
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Fig. 12. HTC performance for HFE-7100 on plain and metal foam surfaces. Fig. 13. Bubble departure diameter distribution.

hexagonal cells. The model yields a kCu;foam = 13.27 W/m∙K and 4.3. Vapor bubble dynamic visualization
kNi;foam = 0.83 W/m∙K, corresponding to a Cu foam thermal con-
ductivity 16 times higher. In order to understand the metal foam Bubbles dynamics were visualized by using a high-speed camera
thermal conductivity values in the foam temperature distribution, it (Photron FASTCAM SA3) with 1024 × 1024 maximum resolution at
was considered the metal foam fiber as a pin–fin with diameter equal to 1000 fps. Through videos and image tracking software, the bubble de-
df and length equal to δ. Hence, it was used the fin efficiency equation, parture diameter (Dd) was calculated by averaging three diameter
measurements of a spherical single-bubble immediately after its de-
tanhmLc
ηfin = tachment from the surface, according to the procedure presented by
mLc (17)
Thiagarajan et al. [44]. For each heat flux, the bubble diameters were
where m = (4h/ kfoam df )1/2
and Lc = δ + (df /4) . By considering a HTC measured for at least three different bubble departure sites for a re-
value of 10 kW/m2∙K, the efficiency was equal to 5.7% and 0.1% for Cu cording period of 1 s. Then, the arithmetic average diameters of all the
foam and Ni foam, respectively. Despite these considerations, the Cu evaluated bubbles for a certain experimental condition were calculated.
foam presented one order of magnitude higher than the Ni foam. Fig. 13 shows the results of the bubble departure diameter for five
Mancin et al. [42] proposed a model for foam surfaces efficiency different walls superheating, corresponding to heat fluxes from 10 to
based on their data regression: 25 kW/m2. In this figure, the error bars correspond to the standard
deviations of the measured values. The coalescence of the vapor bub-
tanhmLc
1+ mLc
∙asf ∙δ bles makes measurements above 25 kW/m2 difficult and not reliable.
ηfoam = According to Fig. 13, the procedure used to evaluate the bubble de-
1 + asf ∙δ (18)
parture diameter is satisfactory because a reasonable agreement is
where, found among the experimental results of the present study and those of
Thiagarajan et al. [44], Borishanskiy et al. [45] and Phan et al. [46], for
0.5 0.52
4∙h ⎞ ⎛ kmat ⎞ plain surfaces. It is worth mentioning that none of the prediction
m = ⎜⎛ ⎟ ∙⎜ ⎟
⎝ kmat ⎠ ⎝ kliq ⎠
d f (19) methods for bubble departure diameter considers surface morphology;
however, Phan et al. [46] incorporates the influence of fluid properties
and and an energy factor as the contribution of the wetting effects, while
Fritz [47] and Cole and Rohsenow [48] correlations only take into
Lc = 1055∙δ 1.18∙PPI∙ (0.0254 − df ∙PPI)0.66 (20)
account a balance between buoyancy and surface tension forces, not
From Mancin et al. [42] model, the foam efficiency was equal to being able to predict the HFE-7100 bubble departure diameter. Re-
16.04% and 6.30% for Cu and Ni foams, respectively (as found out cently, Wang et al. [49] also found that the Fritz [47], and Cole and
previously, the Cu foam presented one order of magnitude higher than Rohsenow correlation [48] not well predict their data, while Phan et al.
the Ni foam). Therefore, the foam thermal conductivity plays an im- correlation [46] predicted their experimental data within an error band
portant role in the boiling heat transfer process. It was also observed by of ± 30%, which is similar to our findings. Moreover, Wang et al. [49]
Xu and Qin [43] who reported that more nucleation sites were activated proposed a correlation based on a modified force balance among the
on copper foam in comparison with nickel foam because of the copper surface tension - based on the advancing contact angle as suggested by
higher thermal conductivity, assuring a more uniform wall super- Mukherjee and Kandlikar [50] – the viscous drag force, the liquid in-
heating along with the foam matrix. ertial force, the pressure difference force, the buoyancy force, and the
For q” > 120 kW/m2, it is reasonable to infer that the vapor mass Marangoni force; in addition, they considered the asymmetry of the
within the Ni foam structure increases the bubble evacuation resistance bubble shape due to the coalescence of smaller bubbles and the sur-
due to the larger heat transfer surface area, and consequently, the rounding liquid motion. The correlation proposed by Wang et al. [49],
smaller area available for the vapor release as also indicated by Li et al. by using an advancing contact angle of 42° that was measured by the
[13]. Moreover, the bubble evacuation resistance is more pronounced authors, is also plotted with our experimental data. The predicted data
as the heat flux increase; in addition, Wong and Leong [7] pointed out is close to our experimental data, which validated the modified force
that it also occurs as the foam height increases. balance analysis proposed by Wang et al. [49].
For Cu foam, the highest HTC occurs at heat fluxes close to 270 kW/ The vapor bubbles growing and coalescence within the foam
m2. The maximum HTC for the Cu foam at the highest heat flux com- structure lead to larger bubble departure diameters for the foam sur-
pared to the Ni foam is explained by the capillary-assisted behavior of faces than for the plain surface, as the results displayed in Fig. 13 and
the Cu foam, which improves liquid replenishment in the porous matrix the bubbles images shown in Fig. 14. These larger bubbles are similar to
delaying the dryout occurrence. “cage bubbles” reported by Xu et al. [10]. From the analyses of Figs. 13

9
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

Fig. 14. Vapor bubbles pattern at different heat flux values.

and 14, for low heat fluxes, the Cu foam provides larger bubbles de- forces acting within the porous medium; the denser and smaller the
parture diameter due to its larger porous diameters that allow bubbles pore sizes, the smaller the vapor bubbles and vice-versa.
to rise easily to the foam top. It is similar to the behavior reported by According to Fig. 14, as the heat flux increases from 18 kW/m2 to
Sadaghiani et al. [51], who attribute the vapor bubbles size to the 45 kW/m2, the bubble departure diameter from the foam surface is still

10
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

bigger; however, a much higher density of bubble departure sites is also decreases with the foam pore diameter decrease (the liquid cannot
noted on the plain surface. Further heat flux increases, up to the con- flows into the foam structure, deteriorating the HTC) [53].
dition for which the HTC presents its maximum value, show that the
bubbles flow not only through the foam upper surface but also through
5. Conclusions
its lateral sides. This behavior is more pronounced for Ni foam due to its
smallest pore diameter. The highest bubble evacuation resistance is due
In this work, we presented a pool boiling experimental work by
to the larger foam contact area with the fluid, increasing the flow
using HFE-7100, at saturation conditions, and metal foams of copper
friction and blocking the liquid-vapor counter-flow, decreasing the HTC
(Cu) and nickel (Ni) as heating surfaces. The following conclusions can
as also reported by Li et al. [13]. In fact, for the highest HTC values,
be drawn from the present study:
vapor bubbles coalesce very quickly, forming vapor columns and
• The metal foams are able to absorb and spread the fluid more
leading to an unstable boiling pattern, which causes an earlier dryout
efficiently, being able to increase the HTC as compared to the plain
than when the boiling pattern is stable at the same heat flux, as also
surface. Moreover, these surfaces prevent thermal overshoot at the
reported by Emery and Kandlikar [52] for a smooth surface and, Wong
onset nucleate boiling.
and Leong [7] for a porous surface.
• The metal foam thermal conductivity plays a key role in the boiling
curve; Cu foam shows better performance as compared to Ni foams,
4.4. Influence of foams parameters on boiling of dielectric fluid even with the smallest surface area.
• The high area density of Ni foam provides a barrier for the de-
Based on the authors′ results and literature review, a summary can parture of the vapor bubble, inhibiting the cooling effect of the heating
be drawn (Fig. 15). The Cu foam shows better boiling performance for surface (incipience dryout occurs earlier).
all heat fluxes than the Ni foam even the latter providing the highest • The capillary wicking of the copper foam contributed to the stable
effective surface area in contact with the fluid. By foams parameters HTC at high heat fluxes.
analyses, Cu presents higher thermal conductivity than Ni; therefore, It is worth mentioning that our recent studies are focused on the
the thermal conductivity plays a key role in the boiling performance thickness of the foam in order to find the optimum foam thickness re-
with foam surface due to the thermal gradient that occurs in the foam, lated to the foam pore diameter and thermal conductivity.
such as a fin surface [42]. Moreover, the porous surfaces present dif- Funding
ferent pore diameters, which are linked to the area density (asf) and Funding was received for this work.
permeability (K) as also reported by Calmidi and Mahajan [41]. The Intellectual property
smaller the pore diameter of an open-cell metal foam, the higher the We confirm that we have given due consideration to the protection
area density, which improves the nucleate boiling and natural con- of intellectual property associated with this work and that there are no
vection heat transfer [10,17]; moreover, as the bubble nucleation impediments to publication, including the timing of publication, with
strongly depends on the liquid superheat and presence of trapped gas, respect to intellectual property. In so doing we confirm that we have
the intersections of foam fibers can form cavities with entrapped gas, followed the regulations of our institutions concerning intellectual
which is suitable for nucleation, thus eliminating the ONB overshoot as property.
compared to the plain surface [9]. However, as the heat flux increases Authorship
the number of active nucleation sites also increases, leading the vapor We confirm that the manuscript has been read and approved by all
bubbles to stay trapped in the foam cell [7,10]; the foam permeability named authors.

Fig. 15. Effect of open-cell metal foam parameters on pool boiling performance.

11
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

We confirm that the order of authors listed in the manuscript has metal foams, Prog. Mater Sci. 46 (2001) 559–632, https://doi.org/10.1016/S0079-
been approved by all named authors. 6425(00)00002-5.
[17] B. Athreya, R. Mahajan, S. Sett, Pool Boiling of FC-72 over Metal Foams: Effect of
Contact with the editorial office Foam Orientation and Geometry, In: 8th AIAA/ASME Jt. Thermophys. Heat Transf.
The Corresponding Author declared on the title page of the manu- Conf. (2002) 1–10. doi:10.2514/6.2002-3214.
script is: [18] Y. Zhu, H. Hu, S. Sun, G. Ding, Heat transfer measurements and correlation of re-
frigerant flow boiling in tube filled with copper foam, Int. J. Refrig. 38 (2014)
This author submitted this manuscript using his/her account in 215–226, https://doi.org/10.1016/j.ijrefrig.2013.04.012.
EVISE. [19] E. Brun, J. Vicente, F. Topin, R. Occelli, IMorph: A 3D morphological tool to fully
All authors understand that this Corresponding Author is the sole analyse all kind of cellular materials, Cellular Metals Struct. Funct. Appl. (2008).
[20] J. Vicente, F. Topin, J.-V. Daurelle, Open celled material structural properties
contact for the Editorial process (including EVISE and direct commu- measurement: from morphology to transport properties, Mater. Trans. 47 (2006)
nications with the office). He/she is responsible for communicating 2195–2202, https://doi.org/10.2320/matertrans.47.2195.
with the other authors about progress, submissions of revisions and [21] S. Sarangi, J.A. Weibel, S.V. Garimella, Quantitative evaluation of the dependence
of pool boiling heat transfer enhancement on sintered particle coating character-
final approval of proofs.
istics, J. Heat Transfer 139 (2016) 021502, , https://doi.org/10.1115/1.4034901.
[22] MatWeb, LLC. MatWeb, Material Property Data. http://www.matweb.com/.
Declaration of Competing Interest [23] S. Mancin, C. Zilio, A. Cavallini, L. Rossetto, International journal of heat and mass
transfer pressure drop during air flow in aluminum foams, Int. J. Heat Mass Transf.
53 (2010) 3121–3130, https://doi.org/10.1016/j.ijheatmasstransfer.2010.03.015.
The authors declared that there is no conflict of interest. [24] J.L. Gasche, D.M. Arantes, T. Andreotti, Pressure distribution on the frontal disk for
turbulent flows in a radial diffuser, Exp. Therm Fluid Sci. 60 (2015) 317–327,
Acknowledgments https://doi.org/10.1016/j.expthermflusci.2014.10.005.
[25] H.S. Ahn, G. Park, J.M. Kim, J. Kim, M.H. Kim, The effect of water absorption on
critical heat flux enhancement during pool boiling, Exp. Therm Fluid Sci. 42 (2012)
The authors are grateful for the financial support from the PPGEM – 187–195, https://doi.org/10.1016/j.expthermflusci.2012.05.005.
UNESP/FEIS, from CAPES, from the National Council of Technological [26] M.M. Rahman, E. Ölçeroglu, M. McCarthy, Role of wickability on the critical heat
flux of structured superhydrophilic surfaces, Langmuir 30 (2014) 11225–11234,
and Scientific Development of Brazil (CNPq grant number 458702/ https://doi.org/10.1021/la5030923.
2014-5) and from São Paulo Research Foundation – FAPESP (grant [27] Z. Cao, B. Liu, C. Preger, Z. Wu, Y. Zhang, X. Wang, M.E. Messing, K. Deppert,
number 2013/15431-7; 2017/13813-0; 2019/02566-8). We also extend J. Wei, B. Sundén, Pool boiling heat transfer of FC-72 on pin-fin silicon surfaces
with nanoparticle deposition, Int. J. Heat Mass Transf. 126 (2018) 1019–1033,
our gratitude to Prof. Dr. Alessandro Roger Rodrigues, Prof. Dr. Tito https://doi.org/10.1016/j.ijheatmasstransfer.2018.05.033.
José Bonagamba (Escola de Engenharia de São Carlos/EESC-USP) and [28] Z. Cao, Z. Wu, A.-D. Pham, Y. Yang, S. Abbood, P. Falkman, T. Ruzgas, C. Albèr,
Iago Lessa de Oliveira (UNESP/FEIS) for their important contribution to B. Sundén, Pool boiling of HFE-7200 on nanoparticle-coating surfaces: Experiments
and heat transfer analysis, Int. J. Heat Mass Transf. 133 (2019) 548–560, https://
this work.
doi.org/10.1016/j.ijheatmasstransfer.2018.12.140.
[29] Z. Wu, Z. Cao, B. Sundén, Saturated pool boiling heat transfer of acetone and HFE-
References 7200 on modified surfaces by electrophoretic and electrochemical deposition, Appl.
Energy 249 (2019) 286–299, https://doi.org/10.1016/j.apenergy.2019.04.160.
[30] L.L. Manetti, T.S. Mogaji, P.A. Beck, E.M. Cardoso, Evaluation of the heat transfer
[1] K.C. Leong, J.Y. Ho, K.K. Wong, A critical review of pool and flow boiling heat enhancement during pool boiling using low concentrations of Al2O3 -water based
transfer of dielectric fluids on enhanced surfaces, Appl. Therm. Eng. 112 (2017) nanofluid, Exp. Therm Fluid Sci. 87 (2017) 191–200, https://doi.org/10.1016/j.
999–1019, https://doi.org/10.1016/j.applthermaleng.2016.10.138. expthermflusci.2017.04.018.
[2] M.S. El-Genk, Immersion cooling nucleate boiling of high power computer chips, [31] I.S. Kiyomura, L.L. Manetti, A.P. da Cunha, G. Ribatski, E.M. Cardoso, An analysis of
Energy Convers. Manag. 53 (2012) 205–218, https://doi.org/10.1016/j.enconman. the effects of nanoparticles deposition on characteristics of the heating surface and
2011.08.008. on pool boiling of water, Int. J. Heat Mass Transf. (2016), https://doi.org/10.1016/
[3] G. Liang, I. Mudawar, Review of pool boiling enhancement by surface modification, j.ijheatmasstransfer.2016.09.051.
Int. J. Heat Mass Transf. 128 (2019) 892–933, https://doi.org/10.1016/j. [32] R.J. Moffat, Describing the Uncertainties in Experimental Results (1988) 3–17.
ijheatmasstransfer.2018.09.026. [33] W.M. Rohsenow, A method of correlating heat transfer data for surface boiling of
[4] E. Teodori, A.S. Moita, A.L.N. MOREIRA, Empirical and modeling-based correla- liquids, Trans. ASME – J. Heat Transf. 74 (1952) 969–976.
tions for pool boiling on microstructured surfaces, Interfacial Phenom. Heat Transf. [34] M.H. Rausch, L. Kretschmer, S. Will, A. Leipertz, A.P. Fröba, Density, surface ten-
2 (2014) 273–292. doi:10.1615/InterfacPhenomHeatTransfer.2015011663. sion, and kinematic viscosity of hydrofluoroethers HFE-7000, HFE-7100, HFE-7200,
[5] 3M product brochure, 3M Novec™ engineered fluid HFE-7100 for heat transfer, HFE-7300, and HFE-7500, J. Chem. Eng. Data 60 (2015) 3759–3765, https://doi.
(2002). org/10.1021/acs.jced.5b00691.
[6] M. Shojaeian, A. Koşar, Pool boiling and flow boiling on micro- and nanostructured [35] M.G. Cooper, Heat Flow Rates in Saturated Nucleate Pool Boiling-A Wide-Ranging
surfaces, Exp. Therm Fluid Sci. 63 (2015) 45–73, https://doi.org/10.1016/j. Examination Using Reduced Properties, in: 1984: pp. 157–239. doi:10.1016/S0065-
expthermflusci.2014.12.016. 2717(08)70205-3.
[7] K.K. Wong, K.C. Leong, Saturated pool boiling enhancement using porous lattice [36] D. Gorenflo, Pool Boiling VDI Heat Atlas, VDI-Verlag, Dusseldorf, Germany, 1993.
structures produced by selective laser melting, Int. J. Heat Mass Transf. 121 (2018) [37] G. Ribatski, J.M.S. Jabardo, Experimental study of nucleate boiling of halocarbon
46–63, https://doi.org/10.1016/j.ijheatmasstransfer.2017.12.148. refrigerants on cylindrical surfaces, Int. J. Heat Mass Transf. 46 (2003) 4439–4451,
[8] L. Zhou, W. Li, T. Ma, X. Du, International journal of heat and mass transfer ex- https://doi.org/10.1016/S0017-9310(03)00252-7.
perimental study on boiling heat transfer of a self-rewetting fluid on copper foams [38] K. Stephan, Heat transfer in condensation and boiling, Springer-Verlag, 1992.
with pore-density gradient structures, Int. J. Heat Mass Transf. 124 (2018) [39] Z. Cao, Z. Wu, B. Sundén, Heat transfer prediction and critical heat flux mechanism
210–219, https://doi.org/10.1016/j.ijheatmasstransfer.2018.03.070. for pool boiling of NOVEC-649 on microporous copper surfaces, Int. J. Heat Mass
[9] M.S. El-Genk, J.L. Parker, Enhanced boiling of HFE-7100 dielectric liquid on porous Transf. 141 (2019) 818–834, https://doi.org/10.1016/j.ijheatmasstransfer.2019.
graphite, Energy Convers. Manag. 46 (2005) 2455–2481, https://doi.org/10.1016/ 07.036.
j.enconman.2004.11.012. [40] E.M. Cardoso, O. Kannengieser, B. Stutz, J.C. Passos, FC72 and FC87 nucleate
[10] J. Xu, X. Ji, W. Zhang, G. Liu, Pool boiling heat transfer of ultra-light copper foam boiling inside a narrow horizontal space, Exp. Therm Fluid Sci. 35 (2011)
with open cells, Int. J. Multiph. Flow 34 (2008) 1008–1022, https://doi.org/10. 1038–1045, https://doi.org/10.1016/j.expthermflusci.2011.02.006.
1016/j.ijmultiphaseflow.2008.05.003. [41] V.V. Calmidi, R.L. Mahajan, The effective thermal conductivity of high porosity
[11] Y. Yang, X. Ji, J. Xu, Pool boiling heat transfer on copper foam covers with water as fibrous metal foams, J. Heat Transf. (1999), https://doi.org/10.1115/1.2826001.
working fluid, Int. J. Therm. Sci. 49 (2010) 1227–1237, https://doi.org/10.1016/j. [42] S. Mancin, C. Zilio, A. Diani, L. Rossetto, Air forced convection through metal
ijthermalsci.2010.01.013. foams: experimental results and modeling, Int. J. Heat Mass Transf. 62 (2013)
[12] Z.G. Xu, C.Y. Zhao, Thickness effect on pool boiling heat transfer of trapezoid- 112–123, https://doi.org/10.1016/j.ijheatmasstransfer.2013.02.050.
shaped copper foam fins, Appl. Therm. Eng. 60 (2013) 359–370, https://doi.org/ [43] Z.G. Xu, J. Qin, Pool boiling investigation on gradient metal foams with double
10.1016/j.applthermaleng.2013.07.013. layers, Appl. Therm. Eng. 131 (2018) 595–606, https://doi.org/10.1016/j.
[13] T. Li, X. Wu, Q. Ma, Pool boiling heat transfer of R141b on surfaces covered copper applthermaleng.2017.12.040.
foam with circular-shaped channels, Exp. Therm Fluid Sci. 105 (2019) 136–143, [44] S.J. Thiagarajan, R. Yang, C. King, S. Narumanchi, Bubble dynamics and nucleate
https://doi.org/10.1016/j.expthermflusci.2019.03.015. pool boiling heat transfer on microporous copper surfaces, Int. J. Heat Mass Transf.
[14] L. Doretti, H. Sadafi, G. Righetti, K. Hooman, G.A. Longo, S. Mancin, Aluminum 89 (2015) 1297–1315, https://doi.org/10.1016/j.ijheatmasstransfer.2015.06.013.
Foams for Water Pool Boiling Enhancement (2019) 8075–8082, https://doi.org/10. [45] A.V. Borishanskiy, G.P. Danilova, A.V. Kupriyanova, Correlation of data on heat
1615/ihtc16.pma.022819. transfer in, and elementary characteristics of the nucleate boiling mechanism, Heat
[15] M.F. Ashby, T. Evans, N.A. Fleck, J.W. Hutchinson, H.N.G. Wadley, L.J. Gibson, Transfer Sov. Res 13 (1981) 100–116.
Metal foams: a design guide, Elsevier, 2000. [46] H.T. Phan, N. Caney, P. Marty, S. Colasson, J. Gavillet, How does surface wettability
[16] J. Banhart, Manufacture, characterization and application of cellular metals and influence nucleate boiling? Comptes Rendus Mécanique 337 (5) (2009) 251–259,

12
L.L. Manetti, et al. Experimental Thermal and Fluid Science 113 (2020) 110025

https://doi.org/10.1016/j.crme.2009.06.032. 127–138, https://doi.org/10.1016/j.ijheatmasstransfer.2006.06.037.

[47] W. Fritz, Berechnung des maximalvolumes von dampfblasen, Physik Zeitschr 36 [51] A.K. Sadaghiani, A.R. Motezakker, S. Kasap, I.I. Kaya, A. Koşar, Foamlike 3D gra-
(1935) 379–384. phene coatings for cooling systems involving phase change, ACS Omega (2018)
[48] R. Cole, W.M. Rohsenow, Correlation of bubble departure diameters for boiling of 2804–2811, https://doi.org/10.1021/acsomega.7b02040.
saturated liquids, Chem. Eng. Prog. Symp. Ser. 65 (1968) 211–213. [52] T.S. Emery, S.G. Kandlikar, Pool boiling with four non-ozone depleting refrigerants
[49] X. Wang, Z. Wu, J. Wei, B. Sundén, Correlations for prediction of the bubble de- and comparison with previously established correlations, Exp. Therm Fluid Sci. 85
parture radius on smooth flat surface during nucleate pool boiling, Int. J. Heat Mass (2017) 132–139, https://doi.org/10.1016/j.expthermflusci.2017.02.027.
Transf. 132 (2019) 699–714, https://doi.org/10.1016/j.ijheatmasstransfer.2018. [53] S. Moghaddam, M. Ohadi, J. Qi, Pool boiling of water and FC-72 on copper and
12.029. graphite foams, Int. Electron. Packag. Tech. Conf. Exhib. ASMEASME 2 (2003)
[50] A. Mukherjee, S.G. Kandlikar, Numerical study of single bubbles with dynamic 675–680, https://doi.org/10.1115/IPACK2003-35316.
contact angle during nucleate pool boiling, Int. J. Heat Mass Transf. 50 (2007)

13