International Journal of Heat and Mass Transfer 79 (2014) 816–828
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer
journal homepage: www.elsevier.com/locate/ijhmt
Pool boiling enhancement through microporous coatings selectively
electrodeposited on fin tops of open microchannels
Chinmay M. Patil, Satish G. Kandlikar ⇑
Department of Mechanical Engineering, Rochester Institute of Technology, USA
a r t i c l e
i n f o
Article history:
Received 17 August 2014
Accepted 25 August 2014
Available online 18 September 2014
Keywords:
Pool boiling
Heat transfer enhancement
Microchannel
Microporous
Critical heat flux
Electrodeposition
a b s t r a c t
Open microchannels and microporous coatings have been individually employed by previous investigators for enhancing pool boiling heat transfer. In this paper, their combined effect is investigated by electrodepositing microporous coatings on the fin tops of microchannels. The microporous coatings were
applied using the optimal electrodeposition parameters developed in an earlier study. The effect of
microchannel geometry on heat transfer performance for water boiling at atmospheric pressure on
10 mm 10 mm copper chips is reported here. A maximum critical heat flux (CHF) of 3250 kW/m2
was obtained for Chip 9 with fin width = 200 lm, channel width = 500 lm and channel depth = 400 lm
at a wall superheat of 7.3 °C. A maximum value of heat transfer coefficient (HTC) of 995 kW/m2 °C was
achieved for Chip 12 with a different channel width of 762 lm for a heat flux of 2480 kW/m2 at a wall
superheat of 2.5 °C. Bubble growth and heat transfer processes are altered when nucleation takes place
preferentially on the fin tops. Visual studies indicate a microconvective mechanism in which bubbles
leaving from the fin tops induce strong localized liquid circulation currents in the microchannels. A liquid
microcirculation based theoretical model is developed to predict heat transfer under this mechanistic
description. The preliminary results are in good agreement with the experimental data.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
There is an increased demand for improved functionality and
reliability of microelectronic devices in many diverse applications.
Since performance of these devices is adversely affected by higher
temperatures, thermal management is becoming an important
consideration. Conventional air cooling systems do not meet the
cooling needs of these devices due to low heat transfer performance associated with air-cooled systems. Compared to other
cooling techniques, pool boiling is attractive due to its ability to
remove large amounts of heat at low wall superheats, and absence
of any moving parts. Improving pool boiling performance in various heat exchangers is also beneficial in other applications such
as power generation, petrochemical, chemical, pharmaceutical
and process industries. Improvement in heat transfer will result
in lower equipment sizes and higher power generation or process
efficiency. The enhancement techniques can be classified into
active devices like ultrasonic vibrations, electrostatic fields, etc.,
⇑ Corresponding author. Address: Department of Mechanical Engineering,
Rochester Institute of Technology, 76 Lomb Memorial Dr., Rochester, NY 14623,
USA. Tel.: +1 585 475 6728; fax: +1 585 475 7710.
E-mail addresses: [email protected] (C.M. Patil), [email protected] (S.G. Kandlikar).
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.08.063
0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
and passive devices like fins, porous/microporous surfaces, structured surfaces like open microchannels (microgrooves), finned or
knurled surfaces, etc. Passive enhancement techniques reduce
overall bulk of the equipment, and improved heat transfer will
result in reduced size of the equipment. Since the enhancement
in both critical heat flux (CHF) and heat transfer coefficient (HTC)
is desired in many applications, it is the primary focus of the work
presented here.
2. Literature review and hypothesis
Pool boiling over microporous surfaces has been widely studied
in literature. Recently, Patil and Kandlikar [1] presented a comprehensive review of plain and structured microporous surfaces used
in enhancing pool boiling heat transfer. Some of the enhancement
techniques that provided significant improvement are briefly discussed here, while a more in-depth review on various pool boiling
enhancement techniques is presented in [2]. Wang et al. [3] created sintered surfaces on carbon steel and conducted pool boiling
experiments with water. They observed a maximum heat flux
enhancement of 800–1400% in CHF over a plain surface at a wall
superheat of about 80 °C. Mori and Okuyama [4] obtained a CHF
of 2.50 MW/m2 at a wall superheat of 50 °C using a honeycomb
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
817
Nomenclature
Abbreviations
CHF
critical heat flux
CNC
computer numerically controlled
HTC
heat transfer coefficient
SEM
scanning electron microscope
Symbols
Dh
h
Nux,h
Pv
P1
Pe
Pr
Rc,min
Rc,max
hydraulic diameter
heat transfer coefficient
local Nusselt number
vapor pressure inside the bubble
pressure
Peclet number
Prandtl number
minimum radius of the cavity to initiate nucleation
maximum radius of the cavity to initiate nucleation
structure. Although CHF is enhanced, the high wall superheat
(hence a low HTC) is of concern. Li and Peterson [5] studied the
effect of thickness of wire mesh and obtained a CHF of 3500
kW/m2 at a wall superheat of approximately 40 °C. Cooke and
Kandlikar [6,7] obtained a CHF of 2440 kW/m2 at a wall superheat
of 9 °C, yielding a high HTC of 269 kW/m2°C. Thus, it can be seen
from the literature that researchers have recently shifted their
focus to enhancing both CHF and HTC during pool boiling.
2.1. Enhanced boiling over porous surfaces
Several authors studied heat transfer enhancement mechanisms on porous surfaces. O’Neil et al. [8] proposed that vapor is
generated in a porous network and is squeezed out of an open
pore. The escaped vapor creates an open space, and liquid is supplied through other pores that act as liquid supply channels. Bergles and Chyu [9] concluded that there is steady vapor formation
inside the porous matrix. Wang et al. [3] suggested that the numerous cavities in the porous network act as nucleation sites. As the
bubbles depart, the smaller nucleation sites underneath the
departing bubbles become active. The higher the nucleation frequency, the higher is the heat dissipated from the porous surface.
The rapid evolution of bubbles creates a large vapor column and
turbulent convective flow, further enhancing the heat transfer rate.
The porous matrix structure also promotes upward squirt effect.
Fig. 1 shows a representation of this mechanism based on the work
of earlier researchers [10–12]. Mankovskij et al. [11] identified the
convection process inside liquid filled capillaries as a dominant
Fig. 1. Schematic representation of boiling mechanism in porous surfaces proposed
by earlier researchers [10–12].
Rc,critical
Re
DTsat
DTsub
Vm
x
x⁄
critical radius of the cavity
Reynolds number
wall superheat
liquid subcooling
mean velocity of the liquid
distance from the entrance region of the flow
non-dimensional x [x⁄ = x/(DhPe)]
Greek symbols
r
surface Tension
qv
density of vapor
Hr
receding contact angle of water in degrees
factor in enhancing pool boiling heat transfer. The convective heat
transfer in pores is inversely proportional to the hydraulic diameter of the flow passage. According to the model proposed by Smirnov [12], there are vapor chimneys in the porous matrix, and part
of the liquid returns back to the nucleation sites like a reflux, as in
the chimney walls. He proposed that the formation and motion of
vapor in the matrix is determined by the processes taking place
near the wall. Cieśliński [13] observed that CHF is set by the rate
at which the vapor escapes without obstructing the liquid flow
paths.
2.2. Boiling over microchannel surfaces
Microstructures such as pin fins have been studied extensively
for pool boiling enhancement [1]. Open microchannels were investigated by Cooke and Kandlikar [6] for pool boiling on
10 mm 10 mm square copper chips with degassed water at
atmospheric pressure. They observed boiling at relatively low heat
fluxes and noted that nucleation occurred preferentially on the top
surfaces (fin top regions) of the microchannel fins. The channels
remained flooded with water, and provided a pathway for water
to the nucleation sites on the fin tops. Fig. 2 shows a schematic
Fig. 2. Mechanism of bubble dynamics on open microchannel surfaces as proposed
by Cooke and Kandlikar [6].
818
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Fig. 3. A 3-dimensional representation of the test chip showing its (a) top and (b) bottom side.
of bubble dynamics observed by the authors. However, at higher
heat fluxes it was difficult to clearly identify the nucleation site
locations.
With the combination of the two techniques of open microchannels and microporous coatings on the fin tops, bubble nucleation is expected to occur preferentially at the fin tops at even
lower wall superheats compared to Cooke and Kandlikar [7]
results. The microporous coating is expected to provide preferred
nucleation at the top of the fins, as opposed to possible nucleation
at the bottom of the channels, even at high heat fluxes.
2.3. Hypothesis
Earlier researchers have proposed that in porous media there
are vapor escape channels, and liquid is replenished from the surrounding channels. They concluded that there is a steady vapor formation inside the porous matrix. The use of microchannel surfaces
with their fin tops coated with a microporous layer is seen as an
efficient way to enhance heat transfer without introducing resistance to the liquid pathways experienced in the porous matrix.
The present study thus focuses on pool boiling heat transfer
enhancement using a multiscale passive enhancement technique,
namely microporous coatings over microstructures such as open
microchannel surfaces to improve both CHF and HTC.
3. Test section
Pool boiling tests were conducted over enhanced copper chips
with microchannels and microporous coatings using saturated,
distilled and degassed water at atmospheric pressure. The test
chips were made of a 3 mm thick copper plate. Dimensions of
the test section are 20 mm 20 mm. The lower side, referred to
as the heater side of the chip, had a 2 mm wide and 2 mm deep slot
on the underside of the 10 mm 10 mm square pool boiling surface as shown in Fig. 3. This slot reduces heat losses to the surrounding and promotes 1-dimensional heat conduction in the
chip. On one of the sides of the test chip, a 0.3 mm diameter,
10 mm deep hole was drilled at the center, 1.5 mm from the top
surface, and a thermocouple was inserted to measure temperature
close to the test surface. Open microchannels were machined on
the boiling surface using a CNC machine. This side is referred to
as the boiling side.
After machining the microchannels, the chips were cleaned
with isopropyl alcohol (IPA) and distilled water, and further dried
using pressurized air. The region with microchannels was left
exposed, while the remaining area was electrically insulated for
Fig. 4. Process steps to deposit microporous coating on microchannel fins.
electrodeposition on the central 10 mm square heat transfer region
only. The microchannels were then filled with a paraffin wax as an
electrical insulating material ensuring that only the top surfaces of
the fins were exposed to the electrolytic bath. The microchannel
chip was connected to cathode and another copper chip of the
same dimension was connected to the anode. The electrolytic bath
consisted of 0.8 M CuSO4 and 1.5 M H2SO4. The current density,
based on an earlier study [2], was set on the software operating
the potentiostat. After the electrodeposition process, a microporous coating was obtained only on the top of the microchannel fins.
The paraffin wax, used as the sacrificial material, was then
removed from the microchannels. Fig. 4 gives a schematic of the
process for depositing a microporous layer on the top surfaces of
the microchannel fins. Table 1 lists the microchannel dimensions
and the electrodeposition parameters employed in preparing the
test chips. The chips were then tested on pool boiling setup as
described in [2] using a similar experimental procedure.
Uncertainty analysis was conducted prior to testing the chips
for their pool boiling heat transfer performance. Detailed description of uncertainty analysis is given in [2]. From this analysis it
was observed that at lower heat fluxes there is a higher uncertainty
of 17% associated with heat flux estimation. This error reduces to
5% at heat fluxes higher than 300 kW/m2. A heat loss study was
conducted to understand the losses of heat from the copper heater.
819
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Table 1
Test matrix showing dimensions of the microchannels, number of channels, parameters for electrodeposition process, and coating thickness.
Sr. No.
Fin width (lm)
Channel
width (lm)
Channel
depth (lm)
Number of
channels
Current density
(mA/cm2)
Time (s)
Current density
(mA/cm2)
Time (s)
Thickness of
deposit (lm)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
200
200
200
200
200
200
200
200
200
200
200
200
500
1000
300
300
300
400
400
400
500
500
500
762
762
762
762
762
200
300
400
200
300
400
200
300
400
200
300
400
400
400
20
20
20
17
17
17
15
15
15
10
10
10
8
5
400
400
400
400
400
400
400
400
400
400
400
400
400
400
15
15
15
15
15
15
15
15
15
15
15
15
15
15
40
40
40
40
40
40
40
40
40
40
40
40
40
40
2500
2500
2500
2500
2500
2500
2500
2500
2500
2500
2500
2500
2500
2500
76.3
68.6
81.1
67.5
70.3
56.9
79.7
72.6
73.3
80.9
81.4
63.9
61.9
59.6
Fig. 5. SEM images of the test chips showing identical cauliflower-like structures, (a) Chip 1 (b) Chip 2 (c) Chip 6 (d) Chip 9 (e) Chip 10 and (f) Chip 12.
The heat losses to the surroundings were found to be less than 1%
as compared to the heat transferred to the boiling region of the
chip.
4. Results and discussion
Prior to conducting the heat transfer experiments the contact
angles for all the test surfaces were measured. The surfaces were
found to be superhydrophilic with contact angles in the range of
2–9°. Scanning electron microscope (SEM) images of the surfaces
were taken to compare their morphology. Fig. 5 shows the SEM
images of Chip 1, Chip 2, Chip 6, Chip 9, Chip 10 and Chip 12. All
SEM images were almost identical as they were prepared without
any agitation using the same electrolytic bath composition, current
density and time of deposition. All the other chips also had similar
SEM images, thus confirming the morphology of the microporous
820
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Fig. 6. Pool boiling curves for test chips, using projected heater area with saturated
distilled water at atmospheric pressure.
Fig. 8. Heat transfer coefficient variation of test chips with channel depth.
surfaces on all the microchannel fin tops to be identical and similar
to the cauliflower-like structures obtained using the optimal
process parameters developed in [2]. Pool boiling experiments
were then conducted to study the combined effects of the microporous coatings and microchannels for these chips.
Fig. 6 shows a comparison of pool boiling curves for the Chips
1–14 with different microchannel geometries and one plain chip
tested in this investigation. Wall superheats for all tested chips
were less than 15 °C. The CHF and HTC values are reported on
the basis of the projected surface area of the heat transfer surface.
A maximum CHF of 3250 kW/m2 was attained by Chip 9 at a wall
superheat of 7 °C. This corresponds to an enhancement of 175% in
CHF over a plain chip. Chip 8 attained a CHF of 3000 kW/m2 at a
wall superheat of 9.0 °C, showing an enhancement of 150% as compared to the plain chip. The lowest CHF obtained for a coated
microchannel chip was 1760 kW/m2 for Chip 14, which translates
into an enhancement in CHF of 75% over a plain chip.
Another observation can be made regarding the trend in the
pool boiling curve for some of the test chips. It is seen that the wall
superheat actually decreases at higher heat fluxes. This is believed
to be due to a combination of additional nucleation occurring
within the porous structure and the increased liquid motion within
the microchannel passages.
Fig. 7 shows the variation of HTC plotted as a function of heat
flux. The general trend indicated that the HTC increased with
increasing heat fluxes. The best performing chip is Chip 12,
yielding a record HTC of 995 kW/m2 °C at CHF, representing an
enhancement of 2300% over a plain chip at its CHF. Chip 3
performed similar to Chip 12 yielding an HTC of 867 kW/m2 °C
and a CHF of 2400 kW/m2 at a wall superheat of 2.8 °C. Chip 10
yielded the lowest value of HTC at 152 kW/m2 °C and a CHF of
2310 kW/m2 at wall superheat of 15.2 °C.
4.1. Effect of channel depth
Fin depth has significant effect on heat transfer performance. At
higher channel depths, the wall superheat decreased and CHF
improved. Fig. 8 shows the variation of HTC with increasing channel depths. It is seen from plots for Chips 1, 2 and 3 that with an
increase in channel depth, the HTC increased. There is a 10%
increase in HTC by increasing the depth from 200 lm to 300 lm
as seen from the plots for Chips 1 and 2. As depth increased to
400 lm, there is an increase of 150% in HTC. A similar trend is
observed in Chips 10, 11 and 12. It can therefore be concluded that
the deeper channels have greater enhancements as compared to
shallow channels over the range of tests conducted. The heat transfer area is increased with deeper fins, and also the liquid comes in
contact with higher temperature surfaces near the channel bottom.
HTC of Chips 3 and 12 were comparable.
4.2. Effect of channel width
Fig. 7. Heat transfer coefficient of the test chips using projected area during
saturated distilled pool boiling of distilled water at atmospheric pressure.
Fig. 9(a) and (b) shows the effects of variation in channel width
on HTC and CHF. From these plots it can be observed that as the
channel width increases, HTC decreases up to a point, beyond
which it increases again. From the plot in Fig. 9(a), it can be
observed that HTC reduced for Chips 6 and 9 as compared to Chip
3, but increased for Chip 12. CHF on the other hand increased as
channel width increased, but there is a certain value beyond which
it started to decrease. A record high HTC of 995 kW/m2 °C was
obtained. For Chip 3 with narrow channels, an HTC of 867
kW/m2 °C was obtained for Chip 12. CHF increased from 2420 kW/
m2 for Chip 3 to 3250 kW/m2 for Chip 9, and then reduced again
to 2410 kW/m2 for Chip 12. From these plots, it can be concluded
that to increase CHF, wider channels could be employed, but up
to a critical limit. For increased HTC, wider (greater than 700 lm)
or narrower channels (less than 300 lm) could be used. The width
of the channel affects the heat transfer surface as well as the crosssectional area available for liquid circulation.
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
821
Fig. 9. Plots showing effects of channel width on (a) CHF, (b) HTC.
Fig. 10. Plots showing effect of fin widths on (a) CHF, and (b) HTC.
4.3. Effect of fin width
Fig. 10(a) and (b) shows the effect of fin width on CHF and HTC.
From the pool boiling curve, it is evident that as the fin width
increased, its CHF decreased, and wall superheat increased. Chip
13 had a CHF of 2400 kW/m2. While Chip 14 had a CHF of
1850 kW/m2. Chip 13 showed an enhancement of 100% over the
plain surface, while Chip 14 showed an enhancement of 55% over
the plain surface.
Fig. 10(b) shows a comparison of HTC for Chips 12, 13 and 14. It
can be seen that as the fin width increased, its heat transfer performance dropped. For Chip 13, a highest HTC was 238 kW/m2 °C,
which represents an enhancement of 500% over the plain surface,
while Chip 14 had a highest HTC of only 121 kW/m2 °C, which
indicates an enhancement of 200% over the plain surface. From this
study, it is evident that as the fin width increases, the performance
of the test surface drops. It is noted that as the fin width increases,
the total available surface area decreases, while the area of the
microporous surface increases.
Based on pool boiling tests conducted on Chips 1–14, it can be
concluded that for these chips, in order to attain higher HTC and
higher CHF, either wider (excess of 700 lm) or narrower (less than
400 lm) channels, with depths around 400 lm and fin widths less
than 250 lm are preferred for water as the working fluid at atmospheric pressure.
is 30% lower than that for Chip 12. There was a reduction of 600% in
wall superheat for Chip 12 as compared to the uncoated microchannel chip. Chip 12UC had a maximum HTC of only 190
kW/m2 °C. The HTC of coated microchannel chip (Chip 12) at CHF
was 425% more than the uncoated microchannel chip and 2300%
more than plain chip as seen in Fig. 12.
The heat transfer performances of all chips tested in this study
are tabulated in Table 2. All microchannel chips performed better
than the plain chip. The wall superheat for all chips was less than
15 °C. Chip 14 had the lowest CHF of 1855 kW/m2 among the
coated chips while Chip 9 had the highest CHF of 3250 kW/m2.
4.4. Comparison of the highest performing chip with uncoated
microchannel chip
Fig. 11 shows a comparison of pool boiling performance for Chip
12 with a plain chip and Chip 12UC, which was an uncoated chip of
the same dimensions but without any coating on it. It can be seen
from this plot that the CHF of Chip 12UC was 1920 kW/m2, which
Fig. 11. Pool boiling curves showing comparison of Chip 12, Chip 12UC and a plain
chip.
822
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Fig. 12. Pool boiling curves showing comparison of HTC for Chip 12, Chip 12UC and
a plain chip.
Fig. 14. Comparison of HTC of Chips 3, 9 and Chip 12 with results available in
literature [4,5,7,14–16].
4.5. Comparison with literature
Pool boiling curves for the three highest performing chips, Chips
3, 9 and 12 from the test matrix, were compared with performance
plots available in the literature for other enhancement techniques
[4,5,7,14–16] and are presented in Fig. 13. The CHF of Chip 9 is
comparable to Li and Peterson [5], Mori and Okuyama [4] and Kandlikar [14]. Cooke and Kandlikar [7] and Chen et al. [10] have CHF
comparable to Chips 3 and 12. However, the wall superheats from
other studies are considerably higher. For example, the wall superheats for Mori and Okuyama, and Li and Peterson are in excess of
50 °C, while they are consistently below 5–10 °C for the chips studied in this investigation. Chip 12 had the lowest wall superheat of
2.4 °C at its CHF of 2410 kW/m2, while Chip 9 had a wall superheat
of 7.3 °C at its CHF of 3250 kW/m2. Fig. 14 shows a comparison of
HTC of these chips.
Fig. 13. Comparison of pool boiling curves of Chips 3, 9 and Chip 12 with results
available in literature [4,5,7,14–16].
5. Heat transfer mechanism
5.1. Comparison of pool boiling curves using normalized heat flux
CHF of the plain chip was under 1200 kW/m2. A maximum HTC of
995 kW/m2 °C was obtained for Chip 12. The plain chip had an HTC
of 44 kW/m2 °C. Thus, it could be concluded from the result table
that the combination of microchannel with microporous coating
on fin tops provided significant enhancement in pool boiling heat
transfer performance.
The heat flux for Chips 3, 6, 9 and 12 were divided by their area
enhancement factor to normalize the heat flux to the plain surface
neglecting the fin efficiency effects. The pool boiling curve of the
normalized heat flux plotted against wall superheat considering
base temperature of the fin is shown in Fig. 15. Chip 9 had a
Table 2
Critical heat flux and heat transfer coefficient of the test matrix.
Sr. No.
Fin width (lm)
Channel width (lm)
Channel depth (lm)
Number of channels
Critical heat flux (kW/m2)
Heat transfer coefficient (kW/m2 °C)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
200
200
200
200
200
200
200
200
200
200
200
200
500
1000
300
300
300
400
400
400
500
500
500
762
762
762
762
762
200
300
400
200
300
400
200
300
400
200
300
400
400
400
20
20
20
17
17
17
15
15
15
10
10
10
8
5
2120
2063
2426
2088
2719
2819
2269
2995
3250
2319
2275
2420
2405
1855
356
384
867
291
720
509
279
326
461
152
183
995
238
121
823
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
respective wall superheats at the base and the top surfaces of the
microchannels.
When the pool boiling curves of the microchannel chips were
normalized to the surface area, it could be represented by strips
of alternate porous coatings (200 lm wide) and plain surfaces
(300–762 lm wide). Although the pool boiling heat transfer performance of the microchannel chips with heat flux normalized to
the surface area was similar to the pool boiling curve of Chip 6P
[2], the normalized surface had alternate strips of microporous
and plain surfaces. This showed that microchannels had a significant role to play in the enhancement of pool boiling heat transfer.
5.2. Effect of coating
Fig. 15. Pool boiling curve of Chips 3, 6, 9, and 12 normalized to the surface area
and Chip 6P [2] with wall superheat calculated using temperature at base of
microchannels.
The effect of coating on microchannel tops can be clearly
described through Figs. 11 and 12. Fig. 11 was described earlier
to show a comparison between an uncoated microchannel chip
with a coated microchannel chip with the same dimensions. The
enhancement in CHF can be attributed to the change in enhancement mechanism introduced by the microporous coating. Fig. 12
is even more illustrative as the base microchannel surface area is
the same between these two chips, but the heat transfer coefficient
is dramatically improved due to coating on microchannel fin tops.
The simultaneous increase in CHF makes this mechanism very
attractive.
5.3. High speed imaging
Fig. 16. Plot of Chips 3, 6, 9, 12 with heat flux normalized to the surface area and
Chip 6P [2] versus HTC calculated using temperature at base of microchannels.
normalized CHF of 1444 kW/m2 at a wall superheat of 19 °C. The
pool boiling curves were compared to the plot of Chip 6P from part
[2]. Chip 6P is a plain chip with porous coating on the entire boiling
surface with the same morphology as that of the porous surface on
top of the microchannel fins. Fig. 16 shows the comparison of the
HTC plotted against wall superheat. Chip 3 had the highest HTC
of 120 kW/m2 °C while Chip 9 had the lowest HTC of 70 kW/m2 °C.
Table 3 shows CHF values for different chips over their
projected and normalized areas. Also shown in the table are the
To understand the bubble dynamics involved, high speed
images of boiling over coated microchannels were captured using
a Photron high speed camera. Images were captured at
2000 frames per second (fps); hence, each subsequent image is
0.5 milliseconds (ms) apart. Fig. 17(a)–(h) shows imaging sequence
for Chip 9 at a heat flux of 189 kW/m2. The microchannel fin tops
appear darker as compared to the base of the channels as porous
coatings do not reflect as much light as polished sides and base
of the microchannels. The channels were machined using a climb
cutting technique, giving it better polish, and reduced number of
potential nucleation sites on the sidewalls and the base of the
channels. Similar images were taken at higher heat flux of
613 kW/m2, as shown in Fig. 18(a)–(h).
From Figs. 17(a)–(g) and 18(a)–(g), it is seen that the bubbles
grow and depart from fin tops of the microchannels. In Fig. 17, a
bubble could be seen growing and departing on top of the microchannel fin. In Fig. 18(a) vapor column of coalescing bubbles is
seen. This showed that at higher heat fluxes, frequency of nucleation of bubbles was higher. The channels were observed to be
flooded with water.
5.4. Proposed model for bubble dynamics and heat transfer
Based on the images and the heat transfer results, a heat transfer mechanism is postulated. Bubbles grow and depart from the
top of the microchannel fins. Porous surfaces enhanced this bubble
Table 3
Comparison of pool boiling results using heat flux over projected and normalized areas for different test chips.
Chip
Area enhancement
factor
CHF over projected
area (kW/m2)
CHF over normalized
area (kW/m2)
Wall superheat at top
surface of the fins (°C)
Wall superheat at the base
of microchannels (°C)
Plain
Chip 6P [2]
Chip 3
Chip 6
Chip 9
Chip 12
0
0
2.26
2.38
2.25
1.76
1043
1400
2426
2819
3249
2408
1043
1400
933
1184
1444
1368
27.5
7.1
2.8
5.5
7.0
2.42
27.5
7.1
10.1
17.2
19.5
6.9
824
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Fig. 17. (a)–(g) Successive images (0.5 ms apart) of bubble dynamics of Chip 9 at a heat flux of 189 kW/m2.
dynamics further. The microporous network provided additional
nucleation sites of different sizes, resulting in initiation of nucleate
boiling at lower wall superheats. As the bubbles departed, the
vacated volume of the bubble was replaced by water. The microchannels act as passages for water supply from all sides, creating
a convective flow. This flow enhanced the convective heat transfer
through the channels, augmenting the heat transfer further. The
channels were flooded with water and ensured that water was
constantly fed to the heated surface enhancing the CHF. Cooke
and Kandlikar [7] proposed that channels with a lower width have
higher heat transfer coefficient enhancement due to their smaller
hydraulic diameters. This could be clearly observed in pool boiling
curves for Chips 3, 6 and 9. The CHF for these chips was attained as
a result of coalescence of the bubbles due to increased frequency of
nucleation at higher heat fluxes, preventing water from filling up
the microchannels. For wider channels, at higher heat fluxes,
nucleation might occur inside the channels. Limited nucleation
and vapor formation within the channels enhanced the heat transfer performance. However, at higher vapor generation rates within
the channels, the liquid circulation mechanism would be hindered.
This would result in a lower CHF as compared to the narrower
channels. These trends could be observed from pool boiling curves
shown in Fig. 9. A schematic of bubble dynamics and fluid flow
through the microchannels is shown in Fig. 19.
Based on above description, it is evident that for wider channels, there might be nucleation occurring at the base. Since, the
microchannel surfaces are not polished; there is a possibility of
nucleation from the base of the microchannels at higher heat
fluxes. As the possibility of nucleation from the base of the microchannels cannot be ruled out, pool boiling curves for the test chips
were plotted using temperature at the base of the microchannels
for calculating wall superheats. Fig. 20 shows pool boiling curves
of Chips 1–12 with wall superheat calculated using temperature
at base of the channels. The plots were compared to the original
pool boiling curves and it was observed that all the curves had
shifted towards right as expected. This meant that there was a
reduction in the overall HTC due to increased wall superheats. Chip
12, which reached CHF at a wall superheat 2.4 °C using the fin top
temperature, had a wall a superheat of 6.9 °C using the base temperature. This reduced the overall HTC to 234 kW/m2 °C indicating
that there was a significant reduction in HTC. Similarly for Chip 3,
there was a reduction in HTC from 867 kW/m2 °C to 242 kW/m2 °C.
5.5. Theoretical analysis of the proposed model
An analytical study was then conducted to evaluate the HTC of
the microchannel fins in Chip 12. The microchannels were
assumed to be smooth and flow of the fluid through them was
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
825
Fig. 18. (a)–(g) Successive images (0.5 ms apart) of bubble dynamics of Chip 9 at a heat flux of 613 kW/m2.
assumed to be laminar and having a uniform velocity. The flow
path and flow boundaries (dotted lines) of water assumed in the
analysis are shown in Fig. 19. The water enters the microchannels
from the central region, and is fed to the vacated areas from the
sides. Since it is symmetric, only half of the microchannel and fin
top is under consideration. A bubble may nucleate and start growing even at a different location close to the departing bubble, even
before it departs. Thus, to account for this simultaneous growth,
we assume that the half channel under consideration provides
water for growth and departure of an entire bubble on its adjacent
fin top. This flow also promotes convective heat transfer. Once the
bubble departs, water is fed to the evacuated region from the adjacent microchannel. This region is shown in the box in Fig. 19. This
is because when the bubble departs, the water is fed from adjacent
microchannels, and only one microchannel is considered for this
analysis. The analysis could be divided in three parts. In the first
part of the analysis, HTC from the side walls and quarter of the base
of microchannel adjacent to the side walls was evaluated. In the
second part, HTC of the top of the microchannel fins with micro
porous coatings was evaluated. In the third part, the HTC due to
entrance of the fluid in the microchannels due to jet action of the
flow was evaluated.
The flow in the passage between the microchannel fins was
assumed to be similar to the flow inside two infinitely long parallel
plates, as depth of the microchannels (400 lm) was very small as
compared to the length of the microchannels (10 mm). The length
of the microchannel under consideration was equivalent to the
average diameter of the bubble. The HTC with this liquid flow
was solved as an entry region problem.
The fluid flow inside the microchannels is represented by a
schematic as shown in Fig. 21. The flow is assumed to be a flow
between two parallel plates. The figure shows two parallel plates,
spaced 190.5 lm. Width of the plates is 587 lm. The liquid was
assumed to flow with a uniform velocity Vm along the direction
as shown in Fig. 21. The central region of the microchannel was
assumed to be the region from where the fluid entered the microchannels. The microchannel was divided in four equal parts as
shown in Fig. 19. Each part had a flow cross-section height of
190.5 lm. The region adjacent to the microchannel wall underneath the departing bubble was considered for this analysis. The
826
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
Fig. 22. Local HTC at the distance x measured from the entrance.
Fig. 19. Schematic of the micro convective flow in the microchannel due to
dynamics over the porous microstructures on fin tops.
Fig. 20. Pool boiling curve of the test chips with wall superheat calculated using
temperature at the base of the microchannels.
Table 4
Bubble diameters and frame number obtained from high speed images.
Bubble number
Bubble diameter (lm)
Frame number
1
2
3
4
5
622
542
604
648
516
149
168
176
189
550
images captured at 2000 fps. Data from the video are tabulated in
Table 4. Diameters of the bubbles were measured at their departure. Frame numbers were obtained from the software controlling
the videos. All the images were captured at a heat flux of 580 kW/
m2.
From the above data, mean diameter of the bubbles was calculated to be 587 lm. Also, from the frame numbers for the first four
bubbles, the frequency of departure was calculated to be 200 bubbles/s. Once a bubble departs, it leaves behind a vacated volume
equivalent to its volume at departure. Thus, from knowledge of
the average diameter of bubble at departure and frequency of the
bubbles per second, the volumetric flow rate of water was calculated to be 2.11 108 m3/s. Thus, using the volumetric flow rate,
the velocity of the water was calculated as 0.36 m/s.
Assuming the flow to be between two parallel plates, the
hydraulic diameter of the liquid flow channel was calculated as
follows:
Dh ¼ 2 spacing between the parallel plates
ð1Þ
From the known velocity, viscosity and hydraulic diameter, the
Reynolds number was calculated as 467. Prandtl number for water
is estimated to be 1.75. From the Reynolds and Prandtl numbers,
the Peclet number could be evaluated using the following equation:
Pe ¼ Re Pr
ð2Þ
Now, if x is the distance from the entrance, the non-dimensional
distance x⁄ is defined as,
Fig. 21. Schematic showing the dimensions and spacing of parallel plates, and
direction of fluid flow velocity (not to scale).
effect of the 90° bend was neglected for simplifying the calculations, as the aim of the calculations was to find a first order approximation for the HTC of recirculated liquid flowing over the
microchannel bottom and side walls.
The liquid flow was determined from the volumetric flow rate
of bubbles leaving the porous fin tops. The bubble departure diameter and the bubble frequency were obtained from the high speed
x ¼
x
Dh Pe
ð3Þ
Thus, x⁄ was calculated for different values of x in the entrance
region. With knowledge of x⁄, local Nusselt number (Nux,h) under
constant heat flux boundary condition was calculated at a distance
of x from the entrance using Eq. (4) from [17]:
Nux;h
"
#1
1
1 X
1 expð4n2 p2 x Þ
¼
6 n¼1
n2 p2
ð4Þ
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
827
Fig. 23. (a) Schematic showing a bubble nucleating from top of the microchannel fin, and liquid flow through the central region similar to an impinging jet, and (b) equivalent
diagram of a submerged slot jet with dimensions (not to scale).
Using the value of Nux,h obtained from the plot, HTC was evaluated using the definition of Nusselt number:
Nux;h
hDh
¼
k
ð5Þ
From Eq. (5), local HTC was calculated for different locations. The
variation of HTC with the distance is shown in Fig. 22. The average
HTC over the entire length was estimated to be 57 kW/m2 °C.
To evaluate HTC for the top of the microchannel fins, results
from the pool boiling curve of Chip 6P from [2] were used.
Morphology of the porous surface at the top of the microchannel
fins was similar to that of Chip 6P. Since pool boiling performance
depends on the surface morphology, the results for the plain surface with microporous coatings could be used for microporous
fin tops. At the heat flux of 580 kW/m2, an HTC of 84 kW/m2 °C
was obtained for plain chip.
Further analysis was conducted to calculate the heat transfer
rate from the microchannel surface. Water was assumed to enter
the channels through the central region and hit the base of the
microchannel. Although water flow in the axial direction could
be occurring, it was neglected in this preliminary analysis. This
flow configuration could be treated as a slot nozzle jet submerged
in water. The velocity of the water in the central jet was estimated
to be 0.36 m/s. The width of the nozzle was considered as half of
the channel width, which was 381 lm. Schematic of the water flow
in the system and its equivalent submerged slot jet nozzle is
shown in Fig. 23(a) and (b).
The heat transfer in the stagnation region under the jet has been
studied extensively in literature. Narayanan et al. [18] performed
an experimental study on submerged slot jets impinging normally
on the surface and provided a detailed analysis on its flow fields,
surface pressures and HTC. Various researchers [18–21] have given
correlations for Nusselt number as a function of Reynolds number
and Prandtl number. Chen et al. [19] presented a correlation applicable for Reynolds number between 55–407 for a slot width of
234 lm. Reynolds number calculated was 467, which is slightly
higher than the range for the correlation. Since no correlation for
the calculated Reynolds number was found in literature, the
following Chen et al. correlation was used:
Nuo ¼ 0:408Re0:596 Pr 1=3
ð6Þ
From the above correlation, the value of Nusselt number was
calculated to be 19.2. Using a hydraulic diameter of 0.000381 m,
the HTC was estimated to be 33.7 kW/m2 °C.
Thus, the overall HTC on the microchannel chip would be the
average of the top, sidewalls and the bottom surfaces of the microchannels. An area weighted value of the overall HTC of 58 kW/
m2 °C was obtained. This value is in reasonable agreement (within
less than 15%) with the HTC of 69 kW/m2 °C at a normalized heat
flux of 330 kW/m2 from Fig. 15.
Although there are several approximations involved in this
analysis, it is seen that the predictions are within less than 15%
of the experimental value. Although a more exhaustive analysis
could be performed for the proposed mechanism, the agreement
with the preliminary analysis presented here suggests that the
proposed microconvective liquid flow is a possible enhancement
mechanism in this configuration. This will enable us to further
improve the performance by considering factors that can be
adjusted to improve heat transfer under this mode.
6. Conclusions
The electrodeposition coating technique developed in [2]
was applied over fin tops of copper microchannel test chips.
The effect of microchannel dimensions – fin width, channel
width and channel height – was systematically studied on pool
boiling performance of saturated water over copper substrates
at atmospheric pressure. The performance of these chips is
seen to be superior to any other techniques reported in literature. The CHF improved, and a significant reduction in wall
superheat was noted. The performance highlights are given
below.
Highest CHF – 3250 kW/m2 at a wall superheat of 7.3 °C for Chip
9.
Highest HTC – 995 kW/m2 °C at a CHF of 2420 kW/m2 for Chip
12 based on temperature at the fin top surface.
828
C.M. Patil, S.G. Kandlikar / International Journal of Heat and Mass Transfer 79 (2014) 816–828
The effect of microchannel geometry was also studied. Based on
this study, the following observations can be made among the
chips tested:
Thinner fins performed better than thicker ones.
Deeper channels performed better than shallow ones.
A channel width of 762 lm (Chip 12) gave the highest HTC. A
channel width of 300 lm (Chip 3) yielded an HTC comparable
to Chip 12. Thus, chips with wider channels performed better
with respect to HTC. Narrow channels (less than 300 lm) gave
a comparable result.
A channel width of 500 lm (Chip 9) gave the highest value of
CHF. It was observed that as channel width increased from
300 to 500 lm, CHF increased. As channel width increased
beyond this value, the CHF dropped, suggesting that there is a
critical value of channel width for enhancing CHF.
High speed images were captured at a rate of 2000 frames per
second. Based on the images obtained on different chips, it was
observed that nucleation occurred on the top of the microchannel
fins, with porous surfaces providing nucleation sites and channels
acting as water supply conduits. Departing bubbles created a current of liquid flow in the microchannels enhancing the microconvective heat transfer. A theoretical model was proposed based on
the microconvective flow in the microchannels and the results
are in reasonable agreement (within 15%) of the experimental values. CHF for narrow channels was reached when the coalescence of
bubbles formed a vapor film on the top and prevented water from
reaching the microchannels. For wide channels, the heat transfer
performance improved as it could sustain the micro convective
flow at higher heat fluxes. However, further increase in the width
of the channels resulted in nucleation at the base between the
channels, hindering the replenishment of water in the channels
and reaching an early CHF.
It is believed that a similar performance enhancement can be
obtained with other fluids and under different operating conditions, although the geometrical parameters and electrodeposition
process parameters may need to be optimized for those conditions.
Conflict of interest
None declared.
Acknowledgments
The work was performed in the Thermal Analysis, Microfluidics
and Fuel Cell Laboratory in the Mechanical Engineering Department at the Rochester Institute of Technology, Rochester, NY. The
authors gratefully acknowledge the financial support provided by
the National Science Foundation under CBET Award No. 1335927.
References
[1] C.M. Patil, S.G. Kandlikar, Review of the manufacturing techniques for porous
surfaces used in enhanced pool boiling, Heat Transfer Eng. 35 (10) (2014) 887–
902.
[2] C.M. Patil, K.S.V. Santhanam, S.G. Kandlikar, Development of a two-step
electrodeposition process for enhancing pool boiling, Int. J. Heat Mass Transfer
79 (2014) 989–1001.
[3] X.S. Wang, Z.B. Wang, Q.Z. Chen, Research on manufacturing technology and
heat transfer characteristics of sintered porous surface tubes, Adv. Mater. Res.
97–101 (2010) 1161–1165.
[4] S. Mori, K. Okuyama, Enhancement of the critical heat flux in saturated pool
boiling using honeycomb porous media, Int. J. Multiphase Flow 35 (10) (2009)
946–951.
[5] C. Li, G.P. Peterson, Parametric study of pool boiling on horizontal highly
conductive microporous coated surfaces, J. Heat Transfer 129 (11) (2007)
1465.
[6] D. Cooke, S.G. Kandlikar, Effect of open microchannel geometry on pool boiling
enhancement, Int. J. Heat Mass Transfer 55 (4) (2012) 1004–1013.
[7] D. Cooke, S.G. Kandlikar, Pool boiling heat transfer and bubble dynamics over
plain and enhanced microchannels, J. Heat Transfer 133 (5) (2011) 052902.
[8] P.S. O’Neill, C.F. Gottzmann, J.W. Terbot, Novel heat exchanger increases
cascade cycle efficiency for natural gas liquefaction, in: K.D. Timmerhaus (Ed.),
Advances in Cryogenic Engineering, Springer, US, 1972, pp. 420–437.
[9] A.E. Bergles, M.C. Chyu, Characteristics of nucleate pool boiling from porous
metallic coatings, J. Heat Transfer 104 (1982) 279–285.
[10] Y. Chen, C. Li, C. Wang, D. Wu, J. Zuo, X. You, Synthesis, structure and physical
properties of the one-dimensional chain complex of tetrathiafulvalene
carboxylate, Sci. China, Ser. B: Chem. 52 (10) (2009) 1596–1601.
[11] O.N. Mankovskij, O.B. Ioffe, L.G. Fibgant, A.R. Tolczinskij, About boiling
mechanism on flooded surface with capillary-porous coating, Ing. Fiz. J. 30
(2) (1976) 975–982.
[12] H.F. Smirnov, An approximate theory of heat exchange during boiling at
surfaces coated with capillary-porous structures, Teploenergetika 2 (9) (1977)
77–80.
[13] J.T. Cieśliński, Nucleate pool boiling on porous metallic coatings, Exp. Therm.
Fluid Sci. 25 (7) (2002) 557–564.
[14] S.G. Kandlikar, Controlling bubble motion over heated surface through
evaporation momentum force to enhance pool boiling heat transfer, Appl.
Phys. Lett. 102 (5) (2013) 051611.
[15] R. Chen, M.-C. Lu, V. Srinivasan, Z. Wang, H.H. Cho, A. Majumdar, Nanowires for
enhanced boiling heat transfer, Nano Lett. 9 (2) (2009) 548–553.
[16] E. Forrest, E. Williamson, J. Buongiorno, L.-W. Hu, M. Rubner, R. Cohen,
Augmentation of nucleate boiling heat transfer and critical heat flux using
nanoparticle thin-film coatings, Int. J. Heat Mass Transfer 53 (1–3) (2010) 58–
67.
[17] S. Kakac, R.K. Shah, W. Aung, Handbook of Single-Phase Convective Heat
Transfer, John Wiley and Sons, New York, NY, 1987.
[18] V. Narayanan, J. Sayed-Yagoobi, R.H. Page, An experimental study of fluid
mechanics and heat transfer in an impinging slot jet flow, Int. J. Heat Mass
Transfer 47 (8) (2004) 1827–1845.
[19] Y.C. Chen, C.F. Ma, Y.X. Li, Forced convective heat transfer with impinging slot
jets of meso-scale, Int. J. Heat Mass Transfer 49 (1) (2006) 406–410.
[20] M. Qin, Local heat transfer and laminar-turbulent transition with impinging
submerged and free-surface slot jets of dielectric liquid, Master’s Thesis,
Beijing University of Technology, Beijing, China, 1996.
[21] D.T. Vader, F.P. Incropera, R. Viskanta, Local convective heat transfer from a
heated surface to a planar jet of water with nonuniform velocity profile, J. Heat
Transfer 34 (3) (1990) 611–623.