Satish G. Kandlikar1
Fellow ASME
e-mail: [email protected]
Theodore Widger
Ankit Kalani
Valentina Mejia
Mechanical Engineering Department,
Rochester Institute of Technology,
Rochester, NY 14623
Enhanced Flow Boiling
Over Open Microchannels
With Uniform and Tapered
Gap Manifolds
Flow boiling in microchannels has been extensively studied in the past decade. Instabilities, low critical heat flux (CHF) values, and low heat transfer coefficients have been
identified as the major shortcomings preventing its implementation in practical high heat
flux removal systems. A novel open microchannel design with uniform and tapered manifolds (OMM) is presented to provide stable and highly enhanced heat transfer performance. The effects of the gap height and flow rate on the heat transfer performance have
been experimentally studied with water. The critical heat fluxes (CHFs) and heat transfer
coefficients obtained with the OMM are significantly higher than the values reported by
previous researchers for flow boiling with water in microchannels. A record heat flux of
506 W/cm2 with a wall superheat of 26.2 C was obtained for a gap size of 0.127 mm. The
CHF was not reached due to heater power limitation in the current design. A maximum
effective heat transfer coefficient of 290,000 W/m2 C was obtained at an intermediate
heat flux of 319 W/cm2 with a gap of 0.254 mm at 225 mL/min. The flow boiling heat
transfer was found to be insensitive to flow rates between 40–333 mL/min and gap sizes
between 0.127–1.016 mm, indicating the dominance of nucleate boiling. The OMM geometry is promising to provide exceptional performance that is particularly attractive in
meeting the challenges of high heat flux removal in electronics cooling applications.
[DOI: 10.1115/1.4023574]
Keywords: flow boiling, stable boiling, electronics cooling, high heat flux, boiling heat
transfer, open microchannels, uniform and tapered manifold
1
Introduction
For the last several decades, air has been the preferred fluid for
cooling electronics due to its availability, low cost (cooling fans),
and reliable system operation. As chip power densities increased,
liquid cooling systems have been introduced. The pioneering
work of Tuckerman and Pease [1] revealed the potential for electronics cooling with microchannels. Colgan et al. [2] illustrated
how liquid cooling in microchannels can be used to dissipate heat
fluxes of approximately 1 kW/cm2. However, the temperature variation of the chip surface along the coolant stream and large
pumping power required for their system are of concern. Flow
boiling in microchannels was expected to provide effective cooling without these concerns; however, the available research indicates that such high heat fluxes are not projected with current
microchannel designs.
A number of researchers have investigated the effects of operating conditions on flow boiling heat transfer in microchannels. Qu
and Mudawar [3] reported that the heat transfer coefficient for
microchannel flow boiling is highly dependent on the mass velocity. Steinke and Kandlikar [4] observed that an increase in vapor
quality caused a sharp decrease in the heat transfer coefficient
(with a peak value between 100–200 kW/m2 C near the onset of
nucleate boiling (ONB) close to x ¼ 0), and significantly lower
values of less than 50 kW/m2 C at higher vapor qualities
downstream.
Wu and Cheng [5] conducted a series of experiments using
trapezoidal microchannels over a range of heat and mass fluxes.
They observed three kinds of unstable modes: (1) liquid/two1
Corresponding author.
Manuscript received October 4, 2012; final manuscript received January 15,
2013; published online May 16, 2013. Assoc. Editor: Leslie Phinney.
Journal of Heat Transfer
phase alternating flow at low heat flux and high mass flux, (2) continuous two-phase flow at medium heat and mass fluxes, and (3)
liquid/two-phase/vapor alternating flow at high heat flux and low
mass flux. Kandlikar [6] studied the flow instabilities due to nucleation in microchannels using high speed visual observations. Hetsroni et al. [7] also reported similar instabilities. Later, Kandlikar
et al. [8] and Kuo and Peles [9] employed nucleation cavities and
inlet restrictors to initiate nucleation and reduce instabilities.
Zhang et al. [10] studied the static Ledinegg instability in horizontal microchannels. The authors concluded that increasing the system pressure and channel diameter, reducing the channel length,
and adding inlet flow restrictors stabilized the flow. Mukherjee
and Kandlikar [11] numerically studied bubble growth in a microchannel and proposed tapered microchannels in order to reduce
instabilities. Lu and Pan [12] obtained stable flow through microchannel heat sinks using diverging parallel microchannels with
laser etched artificial nucleation sites. The authors confirmed that
significant improvements were observed in suppressing flow reversal, pressure oscillations, and inlet temperature fluctuations.
Heat transfer performance data during flow boiling in microchannels was obtained by several researchers. Qu and Mudawar
[13] used a microchannel heat sink containing 21 parallel microchannels 231 lm 713 lm and obtained a maximum heat flux of
130 W/cm2. Kosar et al. [14] used microchannels with a hydraulic
diameter of 227 lm and 7.5 lm wide reentrant cavities on the
sidewalls. High boiling and Reynolds numbers were reported to
increase convective boiling. Hetsroni et al. [15] used triangular
microchannels and obtained heat fluxes ranging from 8–33 W/cm2
with mass fluxes between 95–340 kg/m2 s. Liu and Garimella [16]
experimentally investigated flow boiling in microchannels with
inlet water temperatures of 67–95 C, and mass fluxes of
221–1283 kg/m2 s. The authors obtained a maximum heat flux of
129 W/cm2 and a maximum exit quality of 0.2. Kuo and Peles
C 2013 by ASME
Copyright V
JUNE 2013, Vol. 135 / 061401-1
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
[17] obtained a maximum critical heat flux (CHF) of 643 W/cm2
at a mass flux of 520 kg/m2 s using microchannels with reentrant
cavities. They used a set of five microchannels on a
2 mm 10 mm heater substrate with inlet restrictors for stabilization. The wall superheat was as high as 80 C with a heat transfer
coefficient of around 50,000 W/m2 C. Such high wall superheats
may not be suitable in electronics cooling application.
Khanikar et al. [18] experimentally investigated the effect of
carbon nanotubes on heat transfer using de-ionized water. They
concluded that at low mass fluxes, bubble nucleation was promoted, however, the effect was severely compromised at higher
mass fluxes. Bhide et al. [19] used multiwalled carbon nanotubes
on a silicon wafer as their enhanced surface. Liquid inlet temperatures of 40 C and 60 C and mass fluxes of 18–25 kg/m2s were
employed. The flow boiling heat flux was enhanced by 30–80% in
the fully developed nucleate boiling regime. Krishnamurthy and
Peles [20] used circular micro-pin fins with mass fluxes from 346
to 794 kg/m2s. Vapor slugs and annular flow patterns were
observed through high speed flow visualization. The authors concluded that the two-phase heat transfer coefficient was moderately
dependent on the mass flux and independent of the heat flux.
Morshed et al. [21] experimentally investigated the heat transfer
performance of a microchannel with a copper nanowire coating.
The authors used a single microchannel of 672 lm hydraulic diameter with de-ionized water at atmospheric pressure. Nanowires
were grown using an electrochemical deposition technique on the
bottom surface of the microchannel. They obtained a maximum
heat flux of 58 W/cm2 at a wall superheat of 12 C, with an inlet
fluid temperature of 40 C and a mass flux of 251 kg/m2 s.
It is seen from the available experimental data on flow boiling
in microchannels that the maximum reported heat flux is around
130 W/cm2 by Qu and Mudawar [13] and Liu and Garimella [16]
for a 10 mm 10 mm chip with wall superheats of around
20–30 C. This maximum reported heat flux is considerably below
the projected cooling requirement of 1 kW/cm2 for advanced computer chips [22] and is also lower than the values reported in the
literature for some of the enhanced pool boiling surfaces, which
are briefly reviewed in the next section.
2 Current State of Enhanced Pool Boiling Heat
Transfer
Pool boiling heat transfer is at the core of any flow boiling system. A brief review of the current status of pool boiling is presented first to guide us in setting goals for flow boiling in
microchannels. The main advantages of pool boiling for high heat
flux removal are a uniform surface temperature and an efficient
heat removal system. Recent advances have made this boiling process quite efficient using nanowires, open microchannels, and
combinations of different hierarchical surface structures consisting of nanoscale and microscale features. Figure 1 shows a comparison of some of the recent high performing pool boiling
Fig. 1 Comparison of enhanced pool boiling performance
reported in the literature
structures reported in the literature. Mori and Okuyama [23] used
a porous plate, while Chen et al. [24] and Yao et al. [25,26] used
copper nanowires. Yang et al. [27] employed an open foam cover
and Li and Peterson [28] employed a sintered wire mesh. Cooke
and Kandlikar [29,30] used open microchannels that resulted in a
heat flux of 244 W/cm2 with a heat transfer coefficient of
269,000 W/m2 C. A microstructure design based on utilizing the
evaporation momentum force, proposed by Kandlikar [31,32],
was able to dissipate a heat flux of over 300 W/cm2 with a record
heat transfer coefficient of 629,000 W/m2 C. In comparison,
microchannel flow boiling systems have a significantly lower performance, as seen from the literature review. In order to justify
the additional complexities over pool boiling, it would be reasonable to expect a flow boiling system to provide a stable operation
and superior heat transfer coefficient and CHF values.
The objectives of the current study are set to explore the novel
open microchannel with manifold (OMM) flow boiling configuration and present the experimental results validating its potential to
stabilize the flow and enhance heat transfer performance with
water. The goal is to obtain heat fluxes and heat transfer coefficients that are at least comparable to the highest pool boiling systems reported in the literature. In addition, the relevant parameters
governing the heat transfer and critical heat flux will be identified.
3 Open Microchannels With Uniform and Tapered
Manifolds (OMMs) in Flow Boiling
The two main issues that need to be addressed before implementing microchannels in flow boiling applications are: (1) flow
stability, and (2) improved heat transfer and CHF. Flow boiling
instability has been identified as a limiting factor in microchannels. Reducing the downstream flow resistance effectively mitigates the backflow associated with such instabilities. Mukherjee
and Kandlikar [11,33] suggested several tapered microchannel
designs to overcome this problem. Since tapered microchannels
are difficult to implement, a new tapered manifold design with
open microchannels has been developed in the present study, as
shown in Fig. 2. The microchannels have an open area above
them with a gap, which provides increased flow area for the fluid.
The gap increases downstream, thereby decreasing the slope of
the pressure gradient versus the flow rate curve for the flow channel and increasing the flow stability, as illustrated by Zhang et al.
[10].
The tapered manifold and open microchannel design also
addresses the two factors responsible for low CHF—removal of
the vapor from the heat transfer surfaces and the supply of liquid
to the nucleation sites. The extra flow area provided by the manifold above the microchannels is helpful in removing the generated
vapor without an excessive pressure drop penalty. Since vapor
generation increases in the flow direction, a uniform or a tapered
manifold provides the extra space downstream for the vapor to
flow away from the heat transfer surface. Liquid flow is favored
inside the microchannel region due to capillary forces. Finally, the
Fig. 2 Schematic of the open microchannels with tapered
manifolds for stable enhanced flow boiling
061401-2 / Vol. 135, JUNE 2013
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
Transactions of the ASME
open microchannel configuration has been shown to be very effective in enhancing pool boiling heat transfer [29,30] and is
expected to provide similar enhancement under flow boiling conditions in the present OMM configuration shown in Fig. 2.
The techniques available in the literature to reduce the instabilities can be summarized as artificial nucleation sites, inlet restrictors [8,9] and tapered channels [11,12]. Artificial nucleation sites
alone are not able to effectively suppress the instabilities. The
main disadvantage of the inlet restrictors and the tapered microchannels is the added pressure drop. The OMM design offers a
lower pressure drop due to the availability of the additional flow
area. One of the key aspects of this design is that the microchannels promote nucleation and boiling heat transfer, while the open
manifold region provides a path for the vapor flow without
adversely affecting the liquid flow to the nucleating cavities and
the heat transfer regions inside the open microchannels. However,
the OMM design may require an additional space above the microchannels and may need careful consideration before implementing
it in the 1-D and 3-D electronic chip cooling applications.
3.1 Test Fixture. Flow boiling experiments were performed
using the test section configuration shown in Fig. 2. It shows open
microchannels with a tapered manifold that provides a larger flow
area toward the exit. Figure 3 shows a schematic of the test assembly, which consists of a copper block with eight 200 W cartridge
heaters to serve as the main heating unit. Some of the early tests
were performed with a heater which could deliver a maximum
heat flux of only 300 W/cm2. The tip of the square heating block
measures 10 mm 10 mm and has three equally spaced K-type
thermocouples that were used to determine the heat flux. A copper
test chip contacts the tip of the heating block and a fourth thermocouple in the chip measures its temperature. The chip was supported by a ceramic plate and a manifold block delivers and
removes water from the surface of the chip. A silicone gasket
between the chip and the manifold block was used to seal the system and provide the manifold gap that can be readily varied by
changing the gasket thickness. The gaskets used during testing
had thicknesses ranging from 0.127 to 1.524 mm. A Keyence
VW-6000 camera with recording speeds of up to 24,000 frames
per second (fps) was used for recording and observing flow boiling patterns, although most of the videos were recorded at
1000–2000 fps. A MicropumpV pump provided desired flow rates
of distilled and degassed water to the test setup. A reservoir containing the water (not shown) was heated to saturation temperature
with an electric hot plate. To overcome the heat losses in the supply line between the reservoir and the test section, an inline heater
was installed and a subcooling of 2–5 C was maintained at the
inlet of the test section. The supply tubing, pump, and flow meter
were wrapped in fiberglass insulation to minimize heat losses.
Prior to conducting the tests, the gasket, test chip, manifold
block, and water flow rate were selected and the heater was initialized at 35 V input (corresponding to a minimum heat flux of about
24 W/cm2). Once the system achieved steady state (characterized
by a negligible change in heat flux over a 10 min period) the thermocouple temperatures and pressure drop readings were recorded.
The heater voltage was then increased by 5 or 10 V, and the system was again allowed to reach a new steady state. For safety purposes, this testing cycle continued only until the chip surface
temperature approached 125 C, or the cartridge heater temperature reached 600 C. High-speed videos of the chip surface were
taken at each set point for visualization of the water flow and bubble dynamics. Following the completion of a test, the power supply to the heater was shut off so that the setup can cool down. The
water flow rate was then adjusted to the next set point and the tests
were repeated. A maximum of five flow rates were tested during
the experiments.
The heat flux to the chip was calculated by Fourier’s law for
one dimensional heat conduction in the heater block
q00 ¼ kCu
dT
dx
(1)
The temperature gradient dT/dx was calculated from a three-point
backward difference Taylor series approximation using the heating block temperatures T1, T2, and T3, each at a distance Dx of
5 mm apart
dT 3T1 4T2 þ T3
¼
dx
2Dx
(2)
Finally, the chip surface temperature was obtained from the measured chip temperature T4, the heat flux, and the distance
L ¼ 1.5 mm between the chip thermocouple and the chip surface
Ts ¼ T4 q00
L
kCu
(3)
An uncertainty analysis was conducted in a manner similar to
that described by Cooke and Kandlikar [29,30]. At the highest
flow rate (333 mL/min) the rotameter had an uncertainty of 3%
and at the low flow rate the uncertainty was 5%. The individually
calibrated thermocouples have an accuracy of 0.1 C. The resulting uncertainty in the heat flux and surface temperature at high
heat fluxes were 4% and 0.24 C, respectively. All heat fluxes are
reported on the basis of the projected area of the boiling surface,
which is 100 mm2.
R
Fig. 3
Schematic of the test section and the heater assembly
Journal of Heat Transfer
3.2 Test Chips. The experiments were performed with two
copper chips: one with a plain surface and the other with microchannels in the central 10 mm 10 mm heated region. The front
and back views of the microchannel chip are shown in Fig. 4. The
microchannels were computer numerical control machine milled
and had a channel width of 217 lm, a fin width of 160 lm, a channel depth of 162 lm, and a length of 10 mm, giving a hydraulic diameter of 185 lm.
The chip and the heater block designs were similar to the design
reported by Cooke and Kandlikar [29]. Specifically, a 2 mm wide
and 2 mm deep groove was machined on the underside of the chip
to reduce the heat losses and the heat spreading effect from the
chip.
Fig. 4 Schematic of the 3 mm thick copper microchannel chip
with a 217 lm wide, 162 lm deep, and 160 lm fin width in the
central 10 mm 3 10 mm region (left image), and a 2 mm
wide 3 2 mm deep groove on the underside (right image)
JUNE 2013, Vol. 135 / 061401-3
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
evaluated for a plain chip. Second, the heat dissipation performance of the plain and microchannel chips were compared to identify the performance improvements due to microchannels. Finally,
the effects of the manifold depth and volumetric flow rate (V) on
the heat dissipation performance were investigated. The tests were
conducted at five flow rates of 333, 225, 152, 80, and 40 mL/min.
The vapor quality changed along the flow direction over the chip
surface due to the heat addition. Most of the tests were carried out
with a water inlet temperature of 95–98 C.
Fig. 5 Pressure drop fluctuations over a 20 sec period at different heat fluxes for uniform (UM) and tapered (TM) manifolds,
S 5 0.127 mm and V 5 80 mL/min
3.3 Manifold Blocks. Manifold blocks were constructed out
of LexanV and polysulfone and were polished to improve transparency for flow visualization with the high speed camera. Two
types of manifolds were employed: the tapered manifold, which
provided a tapered gap above the microchannels and increased the
manifold depth by 0.180 mm from the inlet to the outlet and the
uniform manifold, which had a flat surface without any recess in
the block. In both cases the manifold depth was controlled by
inserting a gasket of the desired thickness between the chip and
the manifold block, as shown in Fig. 3. Five gaskets of thicknesses
of 0.127, 0.254, 0.508, 1.016, and 1.524 mm were used to provide
the desired manifold depth over the microchannels. The resulting
hydraulic diameters of the open microchannel and manifold
region are 0.305, 0.476, 0.809, 1.442, and 2.035 mm, respectively.
The gaskets served another purpose of limiting the heat transfer
to the microchannel region. A 10 mm 10 mm opening in each
gasket was aligned over the 10 mm 10 mm microchannel region
in order to inhibit contact between the working fluid and the outer
edges of the test chip. The heat losses and the heat spreading
effect from the test section were minimized due to the gasket and
the groove on the underside, as shown in Fig. 4. The errors in heat
flux due to heat losses and heat spreading effects were estimated
to be less than 1%. The active area of the chip was thus defined by
the square opening in the gasket over the chip region with microchannel features.
R
4
Results
There were three primary objectives addressed in this study.
First, the effect of a tapered manifold block on flow stability was
4.1 Effect of Tapered Manifold. The effect of the manifold
taper on flow stability was evaluated by comparing the inlet pressure fluctuations at steady state conditions for both uniform and
tapered manifold blocks, with a plain chip and an entry spacing of
S ¼ 0.127 mm, as shown in Fig. 5. For each heat flux, the pressure
drop readings taken over a 20 sec period at 5 Hz frequency are
shown. The uniform and tapered manifold tests were run with a
flow rate of V ¼ 80 mL/min. Uniform manifold shows pressure
fluctuations between 80–160 kPa as seen in Fig. 5 for heat fluxes
in the range of 50–250 W/cm2. Tapered manifold in the same heat
flux range shows pressure fluctuations below 20 kPa. This indicates that the tapered manifold mitigates the flow boiling instability experienced with the uniform manifold.
The improvement in stability is also demonstrated in video
footage of the bubble formation for each test. Figure 6 shows a
sequence of three high speed images 8 ms apart, with a uniform
manifold block and a plain test chip at V ¼ 103 mL/min and
S ¼ 1.524 mm. A nucleating bubble forms on the right edge of the
frame and then expands both upstream and downstream, thereby
inducing a back flow and pressure fluctuations in the test section.
The image sequence shown in Fig. 7 corresponds to a tapered
manifold block with a plain test chip at V ¼ 74 mL/min and
S ¼ 1.524 mm. In this case, the nucleating bubble forms and
detaches from the chip surface without expanding against the flow
direction. This results in a reduction in pressure fluctuations and
improved flow boiling stability.
4.2 Comparison of the Tapered Manifold and the Uniform
Manifold. Figure 8 illustrates the comparison of the results for
the tapered manifold and the uniform manifold on the heat transfer performance. The tests were conducted with a microchannel
chip with S ¼ 0.127 mm and V ¼ 225 and 40 mL/min. The depth
of the tapered manifold increased by 0.180 mm over the length of
the microchannel region. At V ¼ 40 mL/min, the tapered manifold
achieved 315 W/cm2 at DTsat ¼ 17.2 C, while the uniform manifold reached 308 W/cm2 at DTsat ¼ 20.5 C. Both manifolds did
not reach the CHF limits.
The tapered manifold seems promising, as seen from its
improved performance in Fig. 8. This is believed to be due to the
flow stabilization effect of the tapered manifold. By mitigating
the back pressure caused by the expansion of large bubbles on the
chip surface, the working fluid has an uninterrupted flow passage
and a more continuous contact with the heated region. Reducing
Fig. 6 Successive images taken at 8 ms time intervals for unstable flow boiling on
a plain chip surface with the uniform manifold V 5 103 mL/min and S 5 1.524 mm
061401-4 / Vol. 135, JUNE 2013
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
Transactions of the ASME
Fig. 7 Successive images taken at a 3 ms time interval for stable flow boiling on a
plain chip surface with the tapered manifold V 5 74 mL/min and S 5 1.524 mm
Fig. 8 Comparison of boiling performance for the tapered (TM)
and uniform (UM) manifold blocks with V 5 225 mL/min,
V 5 40 mL/min, and S 5 0.127 mm
Fig. 9 Effect of the manifold depth on the boiling performance
for the tapered manifold block, the microchannel test chip with
V 5 40 and 225 mL/min and S 5 0.127 and 0.254 mm
the manifold depth may provide additional heat transfer performance enhancement by increasing flow velocity and by further preventing back flow and bubble expansion.
4.3 Effect of Manifold Depth With the Tapered
Manifold. The effect of the manifold depth was studied for the
tapered manifold block with S ¼ 0.254 and 0.127 mm, at both
V ¼ 40 and 225 mL/min. The results are shown in Fig. 9. At a heat
flux of approximately 270 W/cm2 and a flow rate of 40 mL/min,
the thinner gasket had a DTsat ¼ 15.5 C, while the thicker gasket
had a DTsat ¼ 37.1 C. When a flow rate of 225 mL/min is applied,
the performance difference is evident, but less dramatic; the same
heat flux of 270 W/cm2 shows a DTsat difference of 10.4 C.
The performance with a 0.127 mm gap is better than that with a
0.254 mm gap for both flow rates. In addition, the performance
with 40 mL/min is better than that with 225 mL/min for both gaps.
Thus, the smaller gap with the lower flow rate is seen to perform
better than the other configurations with the tapered manifold.
4.4 Performance Enhancement With Microchannels With
the Uniform Manifold. The uniform manifold configuration
forms the baseline for the OMM design. It is investigated in more
detail in the following sections. The heat dissipation performance
was evaluated for the plain and microchannel test chips using a
uniform manifold with a spacing of S ¼ 0.254 mm. Tests were
conducted at two flow rates of V ¼ 225 and 333 mL/min. The
results are shown in Fig. 10. The microchannel chip performed
significantly better than the plain chip for both flow rates. The
plain chip dissipated 100 W/cm2 between DTsat ¼ 20 and 25 C
Journal of Heat Transfer
Fig. 10 Comparison of boiling performance for the microchannel and plain chips with the uniform manifold block with
V 5 333 and 225 mL/min and S 5 0.254 mm
and the CHF was reached at this heat flux. The microchannel chip
dissipated over 255 W/cm2 at DTsat ¼ 14.1 C for V ¼ 225 mL/min
without reaching the CHF. As the flow rate was increased to
333 mL/min, the performance slightly deteriorated and the test
was conducted only up to 140 W/cm2. By comparison, at a
DTsat ¼ 14.1 C and V ¼ 225 mL/min, the plain test chip dissipated
JUNE 2013, Vol. 135 / 061401-5
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
Fig. 11 Dry surfaces (outlined regions are shown inside the bubble region) of a
plain (left image), and the microchannel test chip (right image) under a bubble
only 54.8 W/cm2. Boiling curves for the plain test chip were unaffected by changes in the flow rate.
The difference in the performance between the plain and microchannel test chips is partly the result of the differences in the chip
surface structure and boiling behavior. When nucleate boiling
occurs on a plain chip, the area surrounding a nucleation site is
obstructed by the vapor bubble and the effective wetted area of
the chip is reduced. This is shown in Fig. 11, where portions of
the plain chip under the nucleating bubble have dried out, as seen
from the lighter color in the outlined region. The microchannel
test chip avoids this problem by preventing large sections of the
surface to be covered by a nucleating bubble. The land regions of the
microchannels under a bubble are dry (indicated by the light color),
but the microchannel cavity coloration is unchanged, indicating that
the channels are still wet. The outline of the bubble surrounding the
dry area is also visible over the microchannels. The microchannel
chip also has an inherent surface augmentation factor of 1.8 and,
therefore, has 80% more area than the plain chip. The heat flux
improvement beyond the area augmentation factor is due to the
enhanced boiling mechanism in the present OMM configuration.
4.5 Effect of Manifold Depth With the Uniform
Manifold. Five different gasket thicknesses (0.127, 0.254, 0.508,
1.016, and 1.524 mm) were used to study the effect of the manifold depth on the flow boiling performance. The boiling data
shown in Fig. 12 was obtained with a constant flow rate of
333 mL/min, the microchannel test chip, and the uniform manifold
block. A maximum heat flux of 360 W/cm2 at DTsat ¼ 28.4 C was
achieved using a manifold depth of 0.127 mm. The boiling curves
for all depths were similar, except for the 1.524 mm gasket. The
configurations with depths of 0.254, 0.508, and 1.016 mm were
not tested in the high heat flux region due to the heater limitation
in the earlier version of the test setup. When the manifold depth
was increased to 1.524 mm, the heat transfer performance dropped
significantly to a maximum heat flux of only 122 W/cm2 at
DTsat ¼ 22.0 C.
The boiling data in Fig. 12 suggests that there is a critical manifold depth, while the flow rate also has a small effect, as seen
from Fig. 10. The performance increases with a decrease in the
manifold depth from 1.524; however, the performance with a
depth of 0.127 mm seems to be slightly lower than the intermediate three depths.
The manifold depth also has an important role on the pressure
drop that must be considered in a test setup. Figure 13 shows a
plot of the pressure drop versus the heat flux for both 0.127 and
0.508 mm manifold depths using the uniform manifold
V ¼ 225 mL/min and the microchannel chip. At the highest heat
flux of 250 W/cm2, the 0.127 mm manifold had a pressure drop of
Fig. 12 Effect of the manifold depth on the boiling performance with the uniform manifold block and the microchannel test
chip with V 5 333 mL/min and S 5 0.127 mm, 0.254, 0.508, 1.016,
and 1.524 mm
Fig. 13 Effect of the manifold depth on the pressure drop with
the uniform manifold block and the microchannel test chip with
V 5 225 mL/min for S 5 0.127 and 0.508 mm
061401-6 / Vol. 135, JUNE 2013
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
Transactions of the ASME
Fig. 14 Effect of the flow rate on the boiling curves with the
uniform manifold block and the microchannel test chip with
V 5 333, 225, 152, 80, and 40 mL/min and S 5 0.127 mm
Fig. 15 Effect of the flow rate on the boiling curves with the
uniform manifold block and the microchannel test chip with
V 5 333, 225, and 152 mL/min and S 5 0.254 mm
17.0 kPa and the 0.508 mm manifold had a pressure drop of
9.5 kPa. Similarly, at a heat flux of 66 W/cm2, the smaller manifold depth had a pressure drop of 13.0 kPa and the larger manifold
depth had a pressure drop of 4.4 kPa.
4.6 Effect of Flow Rate With the Uniform Manifold. Three
different tests were conducted with the uniform manifold in order
to investigate the effect of the flow rate and manifold depth on the
flow boiling performance. The first test had a set manifold depth
of 0.127 mm with five different flow rates (V ¼ 333, 225, 152, 80,
and 40 mL/min) and the microchannel chip. The highest heat flux
obtained for this manifold depth was 506 W/cm2 at
DTsat ¼ 26.2 C with a flow rate of 152 mL/min, as seen in Fig. 14.
Tests using flow rates of 80 and 40 mL/min performed very similarly with no more than one or two degrees of wall superheat separating each test at any given heat flux. Higher flow rates of 333
and 225 mL/min showed reduced performance, since they typically had a DTsat that was 5–10 C higher than the other tests for
any given heat flux. The CHF was not reached in any of the tests.
The second set of tests used a manifold depth of 0.254 mm, the
uniform manifold, and the microchannel chip. The results are
shown in Fig. 15. A maximum heat flux of 258 W/cm2 at
DTsat ¼ 14.6 C was obtained with a flow rate of 152 mL/min for
this configuration. The boiling data for tests with V ¼ 225 and
152 mL/min have very comparable profiles, while the data for
V ¼ 333 mL/min has a slightly lower heat flux for a given wall
superheat in the low heat flux region. Higher heat fluxes could not
be obtained since these tests were performed with an earlier version of the test setup with a lower power rating.
The final flow rate tests used a manifold depth of 1.524 mm, the
uniform manifold, and the microchannel chip (see Fig. 16). The
flow rate of 333 mL/min produced a maximum heat flux of
127 W/cm2 and the boiling performance for each test was comparable in profile and performance. Only the lowest flow rate of
152 mL/min showed a slightly better performance. The performance with the manifold depth of 1.524 mm was substantially lower
than those with the thinner manifold spacing.
Figures 14 through 16 illustrate that the flow rate has a minimal
effect on the heat dissipation performance in comparison to the
manifold depth. The boiling performance was consistent across a
wide range of flow rates for each depth, however, manifold depths
of 1.016 mm and thinner dramatically outperformed the largest
depth of 1.524 mm throughout the range of heat fluxes tested.
4.7 Comparison to Pool Boiling. The best pool boiling performance in terms of heat transfer coefficient currently exceeds
Journal of Heat Transfer
Fig. 16 Effect of the flow rate on the boiling curves with the
uniform manifold block and the microchannel test chip with
V 5 333, 225 and 152 mL/min and S 5 1.524 mm
the OMM configurations investigated in this paper. Cooke and
Kandlikar [29,30] achieved a heat flux of 250 W/cm2 at
DTsat ¼ 10 C using the best performing microchannel chip tested
in pool boiling, while in the initial tests reported here, a maximum
heat flux of 506 W/cm2 was obtained with DTsat ¼ 26.2 C. It may
be noted that the CHF was not reached during any of the tests due
to the temperature limitation of the heater and only one microchannel configuration was tested to validate the concept. Thus, the
CHF was significantly enhanced in the OMM design.
Figure 17 shows a comparison between the heat transfer coefficients for pool boiling and flow boiling. The three tests which
reached the maximum heat flux of above 500 W/cm2 are included
in this plot. The three cases yield very similar heat transfer coefficients. The pool boiling setup with open microchannels yielded a
maximum heat transfer coefficient of 270 kW/m2 C at a heat flux
of 250 W/cm2 [30]. While pool boiling with a novel microstructure design by Kandlikar [30,32] yielded a heat transfer coefficient
of 629 kW/m2 C. The current OMM flow boiling setup attained a
maximum heat transfer coefficient of 193 kW/m2 C at a heat flux
of 506 W/cm2. Pool boiling outperformed flow boiling in terms of
the heat transfer coefficient at all heat fluxes. These results indicate that flow boiling in the OMM does not improve the heat
transfer efficiency in the nucleate boiling mode within the
JUNE 2013, Vol. 135 / 061401-7
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
of achieving very high heat fluxes with high heat transfer coefficients. The following specific conclusions are drawn from this
study:
Fig. 17 A heat transfer coefficient versus heat flux comparison
between pool boiling and flow boiling with the uniform manifold
block, the microchannel test chip, and V 5 152 mL/min and
S 5 0.127 mm
microchannels and that its utility primarily lies in enhancing the
CHF through efficient vapor removal and liquid supply to the
nucleation sites.
It is seen that the OMM design exceeds the pool boiling performance of the advanced open microchannel design in terms of
the CHF. Further improvements are expected as the manifold gap,
taper, and microchannel geometry undergo refinements.
5
Discussion and Future Work
The OMM design was able to enhance the CHF beyond 500 W/
m2 with a wall superheat of only 26.2 C. The CHF was not
reached in any of the OMM configurations tested. Initial testing
was limited to heat fluxes below 250–350 W/cm2 since such high
performance was not anticipated. Even the new heater design
delivering 500 W/cm2 heat flux over a 10 mm 10 mm area was
exceeded by the OMM design without reaching the CHF. This is
obviously an area for future experimental work in establishing the
CHF limits of different OMM configurations.
This study shows that the manifold depth, manifold taper, and
flow rate influence the heat transfer performance. The effect of the
flow rate seems to be relatively small over the ranges of the heat
and mass fluxes tested. The inlet subcooling is also an important
parameter, which was not studied in this investigation since the
tests were carried out with a near-saturation inlet with a subcooling of 2–5 C.
The work presented here confirms that the OMM design is able
to provide high CHF and heat transfer coefficients during flow
boiling. It opens up a new field where a number of parameters,
such as microchannel geometry, manifold gap depth, and taper
configuration, could be varied and optimized to achieve even
higher values of heat fluxes and heat transfer coefficients. Other
techniques recently employed in pool boiling enhancement, such
as nanowires, nanofluids, and hydrophobic/hydrophilic hierarchical surface patterns, are expected to provide further performance
enhancements. Experiments are also planned to study the effect of
a higher vapor quality on the heat transfer performance.
6
Conclusions
A novel flow boiling configuration with open microchannels
and a uniform or tapered manifold (OMM) is presented. The
experiments conducted indicate that this configuration is capable
1. A tapered manifold profile reduces the back flow caused by
the bubble nucleation and stabilizes flow boiling when compared to a uniform manifold at low flow rates. The pressure
fluctuations exhibited during flow boiling with a uniform
manifold of spacing 1.524 mm are mitigated by the use of a
tapered manifold.
2. The open microchannel surface shows a dramatic improvement in the heat transfer performance over a plain surface.
High-speed video footage shows that liquid flow in the
microchannels prevents the nucleating bubbles from drying
out large areas of the heated surface.
3. The manifold depth has a significant effect on the heat transfer performance. A maximum heat flux of 506 W/cm2 at a
wall superheat of 26.2 C was achieved using a uniform
manifold with a spacing of 0.127 mm, a microchannel test
chip, and a volumetric flow rate of 152 mL/min. The highest
heat transfer coefficient obtained during this test was
193,000 W/m2 C with a 0.254 mm manifold depth. Manifold depths of 0.254 mm to 1.016 mm had similar boiling
curves, which reached a maximum heat flux of 160 W/cm2
at a wall superheat of 14.1 C. The largest manifold depth of
1.524 mm had the worst overall performance.
4. Changes in the volumetric flow rate for a given manifold
depth have only a small impact on the heat transfer performance. Varying the flow rates for manifold depths of 0.127,
0.254, and 1.524 mm produced similar boiling curves for
each manifold depth.
5. The flow boiling heat transfer coefficient in the OMM configuration is currently exceeded by the most recent pool boiling performance reported in the literature with open
microchannels [29,30] and novel microstructures [31,32].
6. The CHF was not reached in any of the OMM testing. The
maximum heat flux reached was 506 W/cm2. This indicates
the tremendous potential offered by the OMM configuration.
7. Future work is planned in order to conduct an in-depth parametric study to evaluate and extend the performance limits
of the OMM.
Acknowledgment
The authors are grateful to the undergraduate co-op students
Camila Gomez Serrano and Noah Barker-Eldon for their valuable
contributions in the design and operation of the experimental
setup. Some of the work reported in this paper was supported by
the National Science Foundation under Award No. CBET123602.
Nomenclature
h¼
kCu ¼
L¼
P¼
q00 ¼
S¼
T¼
V¼
x¼
heat transfer coefficient, W/m2 C
thermal conductivity of copper block, W/m C
length from the thermocouple to the chip surface, mm
pressure, Pa
applied heat flux, W/cm2
manifold depth (gap), mm
temperature, C
volumetric flow rate, mL/min
vapor quality
Greek Symbols
DTsat ¼ wall superheat, C
Dx ¼ distance between thermocouples, mm
Superscript
s ¼ surface
061401-8 / Vol. 135, JUNE 2013
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms
Transactions of the ASME
References
[1] Tuckerman, D. B., and Pease, R. F. W., 1981, “High-Performance Heat Sinking
for VLSI,” IEEE Electron Device Lett., 2(5), pp. 126–129.
[2] Colgan, E. G., Furman, B., Gaynes, M., LaBianca, N., Magerlein, J. H.,
Polastre, R., Bezama, R., Marston, K., and Schmidt, R., 2007, “High Performance and Subambient Silicon Microchannel Cooling,” ASME J. Heat Trans.,
129(8), pp. 1046–1051.
[3] Qu, W., and Mudawar, I., 2003, “Flow Boiling Heat Transfer in Two-Phase
Micro-Channel Heat Sinks—II. Annular Two-Phase Flow Model,” Int. J. Heat
Mass Transfer, 46(15), pp. 2773–2784.
[4] Steinke, M. E., and Kandlikar, S. G., 2004, “An Experimental Investigation of
Flow Boiling Characteristics of Water in Parallel Microchannels,” ASME J.
Heat Trans., 126(4), pp. 518–526.
[5] Wu, H. Y., and Cheng, P., 2004, “Boiling Instability in Parallel Silicon Microchannels
at Different Heat Flux,” Int. J. Heat Mass Transfer, 47(17–18), pp. 3631–3641.
[6] Kandlikar, S. G., 2002, “Fundamental Issues Related to Flow Boiling in Minichannels and Microchannels,” Exp. Therm. Fluid Sci., 26(2–4), pp. 389–407.
[7] Hetsroni, G., Mosyak, A., Segal, Z., and Ziskind, G., 2002, “A Uniform Temperature Heat Sink for Cooling of Electronic Devices,” Int. J. Heat Mass Transfer, 45(16), pp. 3275–3286.
[8] Kandlikar, S. G., Kuan, W. K., Willistein, D. A., and Borrelli, J., 2006,
“Stabilization of Flow Boiling in Microchannels Using Pressure Drop Elements
and Fabricated Nucleation Sites,” ASME J. Heat Trans., 128(4), pp. 389–396.
[9] Kuo, C. J., and Peles, Y., 2008, “Flow Boiling Instabilities in Microchannels
and Means for Mitigation by Reentrant Cavities,” ASME J. Heat Trans.,
130(7), p. 072402.
[10] Zhang, T., Tong, T., Chang, J.-Y., Peles, Y., Prasher, R., Jensen, M. K., Wen, J.
T., and Phelan, P., 2009, “Ledinegg Instability in Microchannels,” Int. J. Heat
Mass Transfer, 52(25–26), pp. 5661–5674.
[11] Mukherjee, A., and Kandlikar, S. G., 2009, “The Effect of Inlet Constriction on
Bubble Growth During Flow Boiling in Microchannels,” Int. J. Heat Mass
Transfer, 52(21–22), pp. 5204–5212.
[12] Lu, C. T., and Pan, C., 2009, “A Highly Stable Microchannel Heat Sink for
Convective Boiling,” J. Micromech. Microeng., 19(5), p. 055013.
[13] Qu, W., and Mudawar, I., 2003, “Measurement and Prediction of Pressure Drop
in Two-Phase Micro-Channel Heat Sinks,” Int. J. Heat Mass Transfer, 46(15),
pp. 2737–2753.
[14] Kosar, A., Kuo, C.-J., and Peles, Y., 2005, “Reduced Pressure Boiling Heat
Transfer in Rectangular Microchannels With Interconnected Reentrant Cavities,” ASME J. Heat Trans., 127(10), pp. 1106–1114.
[15] Hetsroni, G., Mosyak, A., Pogrebnyak, E., and Segal, Z., 2005, “Explosive
Boiling of Water in Parallel Micro-Channels,” Int. J. Multiphase Flow, 31(4),
pp. 371–392.
[16] Liu, D., and Garimella, S. V., 2007, “Flow Boiling Heat Transfer in Microchannels,” ASME J. Heat Trans., 129(10), pp. 1321–1332.
[17] Kuo, C.-J., and Peles, Y., 2007, “Local Measurement of Flow Boiling in Structured
Surface Microchannels,” Int. J. Heat Mass Transfer, 50(23–24), pp. 4513–4526.
Journal of Heat Transfer
[18] Khanikar, V., Mudawar, I., and Fisher, T. S., 2009, “Flow Boiling in a MicroChannel Coated With Carbon Nanotubes,” IEEE Trans. Compon. Packag.
Technol., 32(3), pp. 639–649.
[19] Bhide, R. R., Singh, S. G., Sridharan, A., and Agrawal, A., 2011, “An Active
Control Strategy for Reduction of Pressure Instabilities During Flow Boiling in
a Microchannel,” J. Micromech. Microeng., 21(3), p. 035021.
[20] Krishnamurthy, S., and Peles, Y., 2008, “Flow Boiling of Water in a Circular
Staggered Micro-Pin Fin Heat Sink,” Int. J. Heat Mass Transfer, 51(5–6), pp.
1349–1364.
[21] Morshed, A. K. M. M., Yang, F., Yakut Ali, M., Khan, J. A., and Li, C., 2012,
“Enhanced Flow Boiling in a Microchannel With Integration of Nanowires,”
Appl. Therm. Eng., 32, pp. 68–75.
[22] Kandlikar, S. G., and Bapat, A. V., 2007, “Evaluation of Jet Impingement,
Spray and Microchannel Chip Cooling Options for High Heat Flux Removal,”
Heat Transfer Eng., 28(11), pp. 911–923.
[23] Mori, S., and Okuyama, K., 2009, “Enhancement of the Critical Heat Flux in
Saturated Pool Boiling Using Honeycomb Porous Media,” Int. J. Multiphase
Flow, 35(10), pp. 946–951.
[24] Chen, R., Lu, M.-C., Srinivasan, V., Wang, Z., Cho, H. H., and Majumdar, A.,
2009, “Nanowires for Enhanced Boiling Heat Transfer,” Nano Lett., 9(2), pp.
548–553.
[25] Yao, Z., Lu, Y.-W., and Kandlikar, S. G., 2011, “Direct Growth of Copper
Nanowires on a Substrate for Boiling Applications,” IET Micro Nano Lett.,
6(7), pp. 563–566.
[26] Yao, Z., Lu, Y.-W., and Kandlikar, S. G., 2011, “Effects of Nanowire Height
on Pool Boiling Performance of Water on Silicon Chips,” Int. J. Therm. Sci.,
50(11), pp. 2084–2090.
[27] Yang, Y., Ji, X., and Xu, J., 2010, “Pool Boiling Heat Transfer on Copper Foam
Covers With Water as Working Fluid,” Int. J. Therm. Sci., 49(7), pp.
1227–1237.
[28] Li, C., and Peterson, G. P., 2007, “Parametric Study of Pool Boiling on Horizontal Highly Conductive Microporous Coated Surfaces,” ASME J. Heat
Trans., 129(11), pp. 1465–1475.
[29] Cooke, D., and Kandlikar, S. G., 2011, “Pool Boiling Heat Transfer and Bubble
Dynamics Over Plain and Enhanced Microchannels,” ASME J. Heat Trans.,
133(5), p. 052902.
[30] Cooke, D., and Kandlikar, S. G., 2012, “Effect of Open Microchannel Geometry on Pool Boiling Enhancement,” Int. J. Heat Mass Transfer, 55(4), pp.
1004–1013.
[31] Kandlikar, S. G., 2012, “Devices With An Enhanced Boiling Surface With Features Directing Bubble and Liquid Flow and Methods Thereof,” U.S. Patent
Application No. 13/471,043.
[32] Kandlikar, S. G., 2013, “Controlling Bubble Motion Over Heated Surface
Through Evaporation Momentum Force to Enhance Pool Boiling Heat Transfer,” Appl. Phys. Lett., 102, p. 051611.
[33] Mukherjee, A., and Kandlikar, S. G., 2007, “Numerical Study of Single Bubbles
With Dynamic Contact Angle During Nucleate Pool Boiling,” Int. J. Heat Mass
Transfer, 50(1–2), pp. 127–138.
JUNE 2013, Vol. 135 / 061401-9
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 08/07/2013 Terms of Use: http://asme.org/terms