International Journal of Heat and Mass Transfer 81 (2015) 325–336

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Fluid flow and heat transfer investigations on enhanced microchannel

heat sink using oblique fins with parametric study
Yong Jiun Lee, Pawan K. Singh ⇑, Poh Seng Lee
Department of Mechanical Engineering, Faculty of Engineering, National University of Singapore, 9 Engineering Drive 1, EA-07-08, Singapore 117575, Singapore

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

Article history: Enhanced microchannel heat sink with sectional oblique fin is used to modulate the flow in contrast to
Received 15 April 2014 continuous straight fin. The re-initialization of thermal boundary layer at the leading edge of each oblique
Received in revised form 28 August 2014 due to breakage of continuous fin into oblique sections and the secondary flow due to these oblique cuts
Accepted 10 October 2014
resulted in better heat transfer and a comparable pressure drop. Extensive experimental investigations
Available online 6 November 2014
are carried out with silicon test vehicle with hydraulic diameter of 100 lm and 200 lm and de-ionized
water as flowing fluid. A parametric study involving the oblique angle, fin pitch is also carried out. Appre-
Keywords:
ciable heat transfer augmentation is also achieved with maximum heat transfer performance enhance-
Enhanced microchannel
Oblique fins
ment at 47% when Re = 680. Comparable pressure drop to conventional microchannel is maintained up
Thermal management to Re = 500. Parametric study suggests that smaller oblique angle and smaller fin pitch are beneficial
Electronic cooling for heat transfer enhancement. The performance of the microchannel with 100 lm channel width and
Fin pitch 27° oblique angle is found to be optimum.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction 1250 W/cm2, still holds great promise of microchannels as a viable

The advancement in packaging technology has led to smaller
chip size associated with higher and more concentrated heat flux.
With smaller chip size and high heat flux comes the generation of
high temperature with addition of hot spots which can accelerate
the meantime to failure (MTTF) and reduce the lifespan of elec-
tronic devices as described by Black0 s equation [1]. Therefore, cool-
ing of electronic devices is a persistent challenge in package design.
Heat sinks with mini/microchannel are currently most widely used
for the cooling of small but highly heated electronic devices, due to
their advantages such as compactness, light weight and higher
heat transfer surface area to fluid volume ratio compared with
other macro-scale systems. Since the concept of microchannel heat
sink was first proposed by Tuckerman and Pease [2] in 1981,
numerous studies has been conducted to investigate its flow and
heat transfer characteristic in microchannels. The study of Cope-
land et al. [3] revealed the application difficulties for the conven-
tional microchannel, which were associated with high pressure
drop and significant lateral temperature gradient. However, the
report of Prasher and Chang [4] that single phase microchannel
liquid cooling was capable of cooling heat flux as high as
⇑ Corresponding author. Tel.: +65 6516 4657; fax: +65 6779 1459.
E-mail address: mpepks@nus.edu.sg (P.K. Singh).
cooling option for microelectronics cooling.
Extensive research on techniques to enhance the heat transfer
performance of microchannel heat sinks has been conducted over
the past decades. Based on the minimization of the flow resistance
between a volume (volumetric heater) and a point (cooling liquid
stream), a tree-shaped channel network was proposed by Bejan
and Errera [5]. Chen and Cheng [6] proposed a right-angled bifur-
cation in a rectangular shaped heat sink while Pence [7] preferred a
smaller bifurcation angle in a disk shaped heat sink. Both designs
deployed self-similar fractal-like branches in a heat sink at a fixed
ratio between the upstream and downstream channel width and
channel length, leading to identical bifurcating pattern at each
level. On the other hand, the concept of deploying re-entrant space
in microchannel heat sink was explored experimentally and
numerically by Xu et al. [8–9]. Their studies indicated that the
thinning of boundary layers after every re-entrant space resulted
local heat transfer enhancement and similar or reduced pressure
drop across the enhanced heat sink compared to that of the con-
ventional configuration. Lee et al. [10] proposed a hot spot mitiga-
tion scheme by creating a recess at the cover lid just before hot
spot region.
As an effective method to promote better fluid mixing, secondary
flow has been listed as one of the efficient heat transfer augmenta-
tion techniques. Steinke and Kandlikar [11] suggested two potential
methods in generating secondary flow for microchannel application.

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.10.018
0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
326 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336

Nomenclature

A convection heat transfer area, m2 q00 heat flux, W/m2

Ac fin area, m2 u velocity, m/s
Dh hydraulic diameter, m w width, lm
H channel height, lm
K loss coefficient Greek symbols
L heat sink length, mm a aspect ratio
L0 characteristic length h oblique angle, °
N number of fin D gradient
Nu Nusselt number l dynamic viscosity, Ns/m2
P fin perimeter, m q mass density, kg/m3
PP pump power, W
DP pressure drop, Pa Subscripts
Pr Prandtl number ave average
Q volumetric flow rate, m3/s b base (unfinned)
R thermal resistance, °C/W
c contraction
Rs summation of spreading resistance and 1D conduction ch channel
resistance, °C/W CM conventional microchannel
Re Reynolds number
conv convective
T temperature, °C EM enhanced microchannel
W heat sink width, mm e expansion
X, Y, Z global coordinate system f fluid
X0 dimensionless axial distance, X0 = X/L
fin fin
Y0 dimensionless channel height, Y0 = Y/H HS heat sink
Z0 dimensionless heat sink width, Z0 = Z/W in inlet
cp specific heat capacity, kJ/kgK max maximum
f friction factor
ob oblique channel
h heat transfer coefficient, W/m2K out outlet
k thermal conductivity, W/mK p plenum
l fin length, lm
si silicon
m _ mass flow rate, kg/s sp spreading
n number of fin row tot total
p fin pitch, lm w wall
q heat transfer rate, W
q⁄ heat transfer rate per unit heater, W

The first suggestion was to add smaller channels at a certain angle
between two main liquid channels. Alternatively, secondary flow
can also be generated by a venturi effect. Both methods can lead to
heat transfer enhancement phenomenon without or with slight
additional pumping power consumption. This technique can be
applied by various kinds of fins and plates arrangement. The study
of Tatsumi et al. [12] on parallel plate fin arrays with oblique notches
suggested that the presence of oblique notches resulted in heat
transfer enhancement through interrupting thermal boundary lay-
ers and promoting the generation of spanwise flow (secondary
flow).
The concept of oblique fins in microchannel heat sinks was pro-
posed by Lee et al. [13] By breaking the continuous fins into obli-
que sections, significant local and global heat transfer enhanced
was achievable with little or negligible pressure drop penalty.
Fan et al. [14] introduced the concept of oblique fins into the
design of cylindrical heat sinks and obtained a much larger average
Nusselt number compared with that of conventional straight fin
heat sink. Detailed numerical and experimental studies were sub-
sequently performed by Lee et al. [15] to further look into the
hydrodynamics and thermal development along the oblique finned
microchannel heat sink, providing some insights to the fundamen-
tal mechanism of this technology. These all works focus on overall
flow and heat transfer in oblique fin channel. Since the change in
flow and heat transfer behavior in oblique fin can be attributed
to local geometry change such as oblique fin and oblique channel,
it will be interesting to see the local flow and heat transfer behav-
ior. Dimension is another important aspect of the oblique fin chan-
nel which can affect the flow and heat transfer. In a recent review
article, it has been observed by Kandlikar et al. [16] that the heat
transfer coefficients of these oblique finned copper microchannels
were relatively low due to large hydraulic diameters and suggested
it will be interesting to see performance at smaller level. In addi-
tion, there are a few key design parameters that greatly influence
the heat transfer and pressure drop performance in oblique finned
microchannels. Therefore, a parametric study is essential to
explore the outcome of varying these parameters.
The present work tries to address these challenges by experi-
mentally investigating the oblique fin microchannels of 100 micron
and 200 micron nominal width. The local flow and heat transfer
behavior is observed and analyzed. A parametric study looking into
the effect of oblique angles and fin pitch is carried out. At last, an
optimum geometry is suggested based on the results.
2. Experimental setup
A schematic of the experimental flow loop is shown in Fig. 1.
Deionized water from a reservoir tank is driven through the flow
loop using a micro-annular gear pump. This pump forces the cool-
ant through a 15 lm filter and a flow meter before entering the
microchannel test section. A heat exchanger (connected to water
bath) is used to regulate the water temperature before coolant
enters the test section. Pressure transmitters are attached to the
 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
Fig. 1. Schematic for experimental flow loop.
manifolds immediately before and after the test piece to measure
the pressure drop across the test section. The water temperatures
immediately before, Tf,in and after, Tf,out the microchannel heat sink
in the manifolds are also measured with two T-type thermocouples
(copper-constantan). Heated water that exits the test section is
cooled by passing it through a liquid-to-air heat exchanger before
it returns to the reservoir tank.
After the test section is assembled and properly sealed, the gear
pump is switched on and the desired flow rate through the flow
loop is controlled with the digital control on the gear pump. As
the flow rate is stabilized, the power supply to the heaters is
switched on, and steady state is usually reached in 30–45 min.
The power input to the heaters in the test piece is controlled by
a DC power supply unit. The voltage across the heaters is measured
directly while the high current through the heaters is calculated
from Ohm0 s law based on the voltage measured across a shunt
resistor that is connected in series with the heater. Input power
to the heaters is calculated as the product of measured voltage
and calculated current. Steady-state readings from the thermocou-
ples, differential pressure transmitter and others are recorded by
data acquisition system and stored in a computer throughout the
duration of the experiment. Each steady state value was calculated
based on the average of 100 readings sampled at 0.1 Hz.
2.1. Test sections
Silicon-based microchannel heat sinks are created on flip chip
packages for experimental investigation. Channels are cut on the
backside of the silicon thermal test chips using wafer dicing saw
blades of different width. Measurement is performed with a 3-axis
measurement microscope (Mitutoyo MF-B1010C) with 200 mag-
nification at 9 different points (3  3 grid) on the microchannel
heat sink. The average surface roughness (Ra) measured at the bot-
tom wall of the channel is 178 nm.
Fig. 2(a) displays the test section of microchannel heat sinks in
this experiment. It consists of a polycarbonate manifold that is
mounted onto a flip chip package, where the microchannels are
laid. Each thermal test die is 0.100  0.100 (2.54 mm  2.54 mm) in
size and when diced in 5  5 grid array, this results in an overall
footprint of 0.500  0.500 (12.7 mm  12.7 mm) with 0.65 mm thick-
327
ness for the current test vehicles. Each thermal test die has a heater
(doped silicon well) at the bottom wall of the chip at 2 mm  2 mm
and a series of thermal sensors (5 diodes connected in series), which
enable independent heater control and local temperature measure-
ment, providing greater insight into local heat transfer behavior and
the overall temperature mapping. The resistive heating of the sili-
con thermal test dies is accomplished by driving the current
through the doped silicon well, where the input power is controlled
by a DC power supply unit. Fig. 2(b), on the other hand, shows the
test piece and the enlarged view of the 5  5 arrays of thermal test
dies. A coordinate numbering scheme (X, Z) for thermal test dies
identification is also indicated in the figure.
Temperatures on the thermal test chip are indicated as voltage
drop across the thermal diode sensors. Prior to the experiment, the
voltage-temperature response of these thermal diode sensors are
established through calibration. The calibration is performed in a
convection oven from 30 °C to 90 °C, in steps of 10 °C. Temperatures
and voltage drop are then recorded when both temperatures of oven
and thermal sensors reach steady state, typically in an hour time.
2.2. Data reduction
Fig. 3 shows the representative image of oblique fin channel. In
the schematic, l is fin length, p is fin pitch, h is oblique angle, wch is
channel width, ww is fin width, wob is oblique channel width.
The sensible heat gained by the coolant is determined from an
energy balance:
q ¼ qcp Q ðT f ;out  T f ;in Þ ð1Þ
It is found that more than 95% of the heat input is transferred to the
coolant across all the experimental runs, where the heat loss is
deemed minimum or negligible.
As for material properties, silicon heat sink are assumed to have
constant thermal conductivity, ksi = 148 W/mK. The density, spe-
cific heat capacity, thermal conductivity and dynamic viscosity of
water are evaluated at the mean fluid temperature (average of
the fluid inlet and outlet temperatures) based on the formulas sta-
ted in [17].
As the pressure transmitters are located at the manifolds, the
pressure drop measurement represents the combined losses due
328 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336

Fig. 2. (a) Microchannel test section, (b) test piece with 5  5 array of thermal test dies.

DPch ¼ DP  ðDP c1 þ DP c2 þ DP e2 þ DP e1 Þ
where DPc1 and DPc2 are the contraction pressure losses from the
deep plenum to the shallow plenum, and from the shallow plenum
to the microchannel. These minor losses can be expressed as [19]:
1   K
c1
DPc1 ¼ qf u2p2;in  u2p1;in þ q u2 ð3Þ
2 2 f p2;in
1   K c2
DPc2 ¼ qf u2in  u2p2;in þ q u2 ð4Þ
2 2 f in
where p1 and p2 denote the deep plenum and shallow plenum,
respectively, and Kc1 and Kc2 are the loss coefficients for the abrupt
contractions. On the other hand, DPe2 and DPe1 express the pressure
losses from the microchannel to the shallow plenum and from the
shallow plenum to the deep plenum, which can be written as
follow:
1   K
e2
DPe2 ¼ qf u2p2;out  u2out þ q u2 ð5Þ
2 2 f out
Fig. 3. Schematic of the oblique fin and its nomenclature. 1   K
e1
DPe1 ¼ qf u2p1;out  u2p2;out þ q u2 ð6Þ
to the frictional loss in microchannels and minor losses due to 2 2 f p2;out
abrupt contraction and expansion at the inlet and outlet [18], where Ke1 and Ke2 represent the losses coefficients due to the abrupt
which can be written as follow: expansion. For the present heat sink test section geometry, the
value of Kc1, Kc2, Ke1 and Ke2 are close to unity [20].
DP ¼ DPc1 þ DPc2 þ DPch þ DPe2 þ DPe1 ð2Þ
In addition to the above data reduction, thermal spreading resis-
The pressure drop across microchannel can then be calculated as tance is calculated to better assess the channel wall temperature of
 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
silicon-based microchannel heat sinks. Thermal spreading resistance
occurs whenever heat leaves a heat source of finite dimensions and
enters into a larger region. The thermal test dies used in the
experiment has 64% (80% length  80% width) of its bottom
surface area covered by heater. Thus, heat spreading takes place
when heat is transported from the bottom of the dies to the base of
the microchannel. A 3D thermal spreading resistance model proposed
by Yovanovich et al. [21] is adopted to account for the effect of
thermal spreading. This model assumes a planar rectangular heat
source situated on one end of a flux channel, where it is cooled on
the other end through an average heat transfer coefficient, h while
the flux channel is laterally adiabatic.
The summation of the thermal spreading and 1D conduction
resistances across the thermal test chip is then found as Rs,(x,z)
while the temperature at the base of microchannel can be deter-
mined as the heat supplied by each individual heater, q⁄ and tem-
perature at the bottom surface of thermal test dies, Th are known.
Total resistance = Thermal spreading resistance + 1D conduc-
tion resistance
Rs;ðx;zÞ ¼ Rsp;ðx;zÞ þ R1D

q Rs;ðx;zÞ ¼ T h;ðx;zÞ  T w ; ðx; zÞ
where Thhðx;zÞ is the measured temperature at bottom of thermal
test dies and Thhðx;zÞ is calculated channel wall temperature.
As the local temperatures of the silicon-based microchannel
heat sink can be measured by individual thermal diode sensor in
a 5  5 grid array across the heat sink, local heat transfer coeffi-
cient of the heat sink section on top of each thermal test chip is
then determined using the correlation:
q
hðx;zÞ ¼
Atot ðT wðx;zÞ  T f ;x Þ
where q⁄ is the local heat transfer rate evaluated as the average sen-
sible heat gain by coolant per unit thermal test dies. A⁄tot on the
other hand, is the average convective heat transfer area per unit
thermal test dies.
Total thermal resistance of the heat sink is defined as
T max  T f ;in
Rtot ¼
q
where Tmax is the maximum measured temperature of the heat sink,
Tf is the inlet coolant temperature and q is the heat supplied into the
heat sink.
The streamwise local heat transfer coefficient, hx is defined as
the average heat transfer coefficient of thermal test chips in the
spanwise direction at the identical streamwise distance from the
inlet.
hðx;1Þ þ hðx;2Þ þ hðx;3Þ þ hðx;4Þ þ hðx;5Þ
hx ¼
5
Subsequently, the average heat transfer coefficient of the heat sink
is calculated as the average value from all the twenty-five thermal
test dies.
The corresponding Nusselt number is then calculated as
Nu ¼ hDkf
h
Standard error analysis adopting the principles by Taylor [22] is
performed with the uncertainties in the reported Nusselt numbers
falling between 6.6% and 15.6%. The highest percentage of error
occurs at the maximum flow rate, mainly due to the lowest sensi-
ble water temperature rise from the inlet to the outlet of micro-
channel. The main contributor of uncertainties in this experiment
is identified as the measurement of water temperature (±0.3 °C).
Sequential perturbation method [23] is used to address the uncer-
tainty of spreading resistance, due to the complexity of the
equation. On the other hand, uncertainties for pressure drop
329
measured across microchannel heat sinks are at 1.8–12.6%. The
nominal uncertainty in each parameter is shown in Table 1.
3. Results and discussion
3.1. Silicon-based microchannel heat sink with 100 lm nominal
channel width
The experiments are conducted over the flow rates ranging
from 100 mL/min to 500 mL/min, which corresponds to Reynolds
numbers of 180–680. By employing silicon thermal test chip in
these experiments, the local behavior (local temperature and local
heat transfer coefficient profiles) of the enhanced heat sink can be
investigated. In this case, the performance of the enhanced micro-
channel heat sink with 100 lm nominal width under the experi-
mental conditions of 160 mL/min (Re = 252) total coolant flow
rate and total 273 W heater power of will be presented in detail.
The measured temperatures of the thermal test chips are plot-
ted in Fig. 4 based on the spanwise and streamwise position on
the heat sink. The figure is plotted for different Z where Z is Dimen-
ð7Þ
sionless heat sink width. The wall temperature is lowest at Z0 = 0.5
ð8Þ followed Z0 = 0.3 and 0.7. The heat transfer is almost same in this
region and the plots overlap each other. Higher temperatures are
observed beyond Z0 = 0.3 and 0.7. Wall temperature of heaters at
Z0 = 0.9 is the highest followed by heaters at Z0 = 0.1. The difference
is clearly due to the lateral heat conduction from discrete heat
sources towards the peripheral of thermal test chips.
The enhanced microchannel shows significant reduction in total
thermal resistance compared to the conventional microchannel.
The total thermal resistance comprises of conductive, spreading,
convective and caloric thermal resistances. Generally, the conduc-
ð9Þ
tive thermal resistance remains constant while spreading, convec-
tive and caloric thermal resistances reduce with the increasing
Reynolds number, resulting in lower total thermal resistance. The
improved heat transfer performance of the enhanced microchannel
is demonstrated in Fig. 5, where the total thermal resistance of the
enhanced microchannel heat sink is consistently lower than the
conventional configuration in the range of Reynolds numbers stud-
ied. At low Reynolds number 180, the difference in thermal resis-
ð10Þ tances of both the channels cannot be claimed. As Reynolds
number rises, the effectiveness of enhanced microchannel
increases and the percentage of reduction in total thermal resis-
tance quickly increase. The maximum total thermal resistance
reduction achieved at Reynolds number 690 is as much as 25%
(Rtot,EM = 0.089 °C/W versus Rtot,CM = 0.119 °C/W). In the context of
current experiment, the highest percentage of total thermal
Table 1
Uncertainties for various critical parameters for silicon-based microchannel heat
ð11Þ sinks.
Variable Absolute uncertainties/fractional
uncertainties
Channel width, wc, fin width, ww and 2 lm
fin length, l
Channel height, H 5 lm
Oblique angle, h 0.15°
Volumetric flow rate, Q 1.9–9.8%
Temperature, Tf, Tcu 0.3 °C
Reynolds number, Re 3.2–10.3%
Wall temperature, Tw 0.9–1.2%
Hydraulic diameter, Dh 1.4%
Fin efficiency, g 3.7–15.8%
Total heat transfer area, Atot 3.4–13.7%
Heat transfer coefficient, h 6.4–15.6%
Nusselt number, Nu 6.6–15.6%
Pressure, DP 1.8–12.6%
Friction factor, f 2.7–13.6%
330 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
Fig. 4. Measured heat sink wall temperature of the enhanced microchannel heat
sink.
Fig. 5. Comparison of total thermal resistance between the microchannel heat
sinks.
resistance reduction translates to an 8.1 °C lower in maximum
chip temperature for the enhanced microchannel heat sink
(Tmax,EM = 50.5 °C versus Tmax,CM = 58.6 °C).
As discussed, the introduction of oblique fins and channels
leads to a much uniform heat removal capability across the micro-
channel heat sink. Fig. 6(a) and (b) illustrate the comparison of
local heat transfer performance of microchannel heat sinks for
two of the experimental runs at Re  250 and Re  680. The mark-
ers denote the location of thermal diode sensors on the silicon test
chip in the streamwise direction, where the temperature measure-
ment is made. It is noticed that both conventional and enhanced
microchannel heat sinks display a relatively close Nusselt number
at the upstream of the heat sink, X0 = 0.1 when the boundary layers
are thin. The distinguishing point for these two heat sinks is the
development of the flow regime and boundary layers as coolant
travels downstream. As boundary layers continue to thicken with
the flow distance, heat transfer performance of the conventional
microchannel deteriorates rapidly, as showed in the figures for
both Reynolds number cases. A maximum 48% drop in Nusselt
number is observed between X0 = 0.1 and X0 = 0.9, resulting in a
highly non-uniform heat transfer performance within the heat
sink. Instead of declining, the enhanced microchannel keeps the
Nusselt number at a much elevated level and consistent across
the enhanced microchannel heat sink. The profile of the stream-
wise local Nusselt number is very uniform, confined to a narrow
range of 7.0–8.2 and 12.4–14.4 for both Reynolds number cases
respectively from the inlet of the microchannel to its outlet. This
phenomenon is highly due to the frequent re-initialization of ther-
mal boundary layers and generation of secondary flows. This com-
bination ensures that the flow is in a constant state of development
thus having a sustainable performance close to that of the flow
upstream.
The average heat transfer performance of the microchannel
heat sinks is plotted in Fig. 7. Generally, the average Nusselt num-
ber, Nuave increases with Reynolds number as the thermal bound-
ary layer thickness decreases with the increased fluid velocity.
However, the heat transfer for the enhanced microchannel with
oblique fins is highly augmented in comparison with conventional
microchannel heat sink. At Reynolds number 180, the average
Nusselt number of the enhanced microchannel is 15% higher than
that of conventional microchannel. As Reynolds number rises, the
Nusselt number increases by almost 47%, from 9.0 to 13.2. The
appreciable enhancement in heat transfer is due to the combined
effects of thermal boundary layer re-development at the leading
edge of each oblique fin and the secondary flows generated by flow
diversion through the oblique channels. As the flow rate induced
through the heat sink is increased, more fluid is diverted from
the main channel into the oblique channel. This secondary flow
thus carries higher momentum and further disrupts the boundary
layers development and augments the heat transfer.
Nevertheless, the most interesting feature of the enhanced
microchannel heat sink is that the significant heat transfer aug-
mentation is achieved with small pressure drop penalty. The
enhanced microchannel heat sink, which employs secondary flow
to enhance its heat transfer, still managed a comparable pressure
drop with the conventional microchannel heat sink for Reynolds
number lower than 500, as displayed in Fig. 8. This distinguishes
the proposed scheme from the conventional heat transfer enhance-
ment scheme, where trade-off in terms of pressure drop penalty is
inevitable. As the Reynolds number increases, a higher percentage
of coolant will be diverted into oblique channels. This creates a sec-
ondary flow with stronger momentum, which further augments
the heat transfer but incurs additional pressure drop penalty.
Therefore, the pressure drop for enhanced microchannel starts to
deviate and increases more than the conventional configuration.
However, the magnitude of increment of pressure drop is consid-
ered manageable for the same micropump.
3.2. Parametric study
As discussed in introduction, there are few key design parame-
ters. As these parameters vary, the overall fin layout will change.
This change has an enormous effect on the flow field and conse-
quently it affects the performance of the heat sink. Therefore, a
parametric study is essential to explore the outcome of varying
these parameters. The two critical design parameters identified
for the parametric study are; oblique angle and oblique fin pitch.
A total of seven configurations with three oblique angles (15°,
27° and 45°) coupling with three fin pitches (400 lm, 800 lm
and 1500 lm) are fabricated on silicon-based microchannel heat
sink for the performance evaluation. The benchmarking is based
on two conventional microchannel configurations, which have
100 lm and 200 lm nominal channel width respectively. Detailed
dimensions of various microchannel heat sinks are tabulated in
Tables 2 and 3. Microchannel heat sinks with the similar channel
width, fin pitch but different oblique angles are grouped together
 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
Fig. 6. Comparison of local heat transfer performance of microchannel heat sinks for (a) Re  250 (b) Re  680.
Fig. 7. Comparison of heat transfer performance for microchannel heat sinks.
Fig. 8. Comparison of pressure drop for microchannel heat sinks.
to compare the effect of oblique angle variation and it is vice versa
when evaluating the effect of fin pitch variation.
331
While the combination of a different oblique angle and pitch,
can result in a wholly new enhanced microchannel heat sink lay-
out, it also re-distributes the fin and un-finned surface areas avail-
able for convective heat transfer. Table 4 compares the fin, un-
finned and total surface areas available for convective heat transfer
of different microchannel configurations. From Table 4, it is clear
that the introduction of oblique fins and channels in microchannel
heat sink increases the total heat transfer area by 5–30% with
smaller fin pitch and oblique angle leading to larger surface areas.
3.2.1. The effect of oblique angle variation
The effect of a change in oblique angle to heat transfer, and
pressure drop performance, is diagnosed with three groups of
microchannel heat sinks. The first group are three microchannel
heat sinks with 100 lm nominal channel width, which consists
of a conventional straight channel heat sink (Con-100) as bench-
mark and two enhanced heat sink (400 lm nominal fin pitch)
with 27° and 45° oblique angle respectively (27° – NP0 and 45° –
NP0 ). The experimental results of both heat transfer and pressure
drop are presented in Fig. 9. Fig. 9(a) compares the average Nusselt
number of three microchannel heat sinks. It is observed that the
average Nusselt numbers of Con-100 and 45° – NP0 almost overlap
each other, while that of 27° – NP0 is significantly higher. This sug-
gests that 45° – NP0 is not as effective as 27° – NP0 in augmenting
heat transfer performance. In contrast, the heat transfer augmenta-
tion for 27° – NP0 rises with Reynolds number. As a result, this con-
figuration records the lowest total thermal resistance among the
threes, as shown in On the other hand, the pressure drop across
Con-100 and 45° – NP0 overlap each other as they did in the Nusselt
number plot, as shown in Fig. 9(b). The 27° – NP0 also displays a
pressure drop that is comparable to the other two for Reynolds
numbers lower than 500, signaling that the heat transfer enhance-
ment can be achieved with negligible pressure drop penalty.
The second group of microchannel heat sinks for comparison
has the same features as the first group, with exception that they
are 200 lm in nominal channel width. In this group, conventional
microchannel (Con-200) is adopted as the baseline for comparison
with two enhanced microchannel heat sinks (800 lm nominal fin
pitch) with 27° and 45° oblique angle respectively (27° – NP and
45° – NP). A similar trend is observed from the heat transfer and
pressure drop performance of this group, as displayed in Fig. 10,
in comparison to the first group, which has 100 lm nominal
332 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336

Table 2
Dimensional details of microchannel heat sinks with 100 lm channel width.

Characteristic Conventional microchannel #1 (Con-100) Enhanced microchannel #1 (27° – NP0 ) Enhanced microchannel #2 (45° – NP0 )
Material Silicon
Footprint, width  length (mm2) 12.7  12.7
Number of fin row, n 61
Main channel width, wc (lm) 115 113 110
Fin width, ww (lm) 85 87 91
Channel depth, H (lm) 387 379 351
Aspect ratio, a 3.37 3.35 3.19
Oblique channel width, wob (lm) – 49 48
Fin pitch, p (lm) – 405 400
Fin length, l (lm) – 292 332
Oblique angel, h (°) – 26.3 45.0

– NP0 denotes nominal pitch (400 lm) for enhanced microchannel heat sink with 100 lm channel width.

Table 3
Dimensional details of microchannel heat sinks with 200 lm channel width.

Characteristic Conventional Enhanced Enhanced Enhanced Enhanced Enhanced

microchannel #2 (Con- microchannel #3 (27° – microchannel #4 (45° – microchannel #5 (15° microchannel #6 (27° microchannel #7 (45°
200) NP) NP) – LP) - LP) - SP)
Material Silicon
Footprint, 12.7  12.7
width  length
(mm2)
Number of fin row, 30
n
Main channel 205 205 203 205 206 204
width, wc (lm)
Fin width, ww (lm) 195 195 197 194 194 196
Channel depth, H 450 418 445 455 405 449
(lm)
Aspect ratio, a 2.20 2.04 2.19 2.17 1.97 2.20
Oblique channel – 104 103 101 103 103
width, wob (lm)
Fin pitch, p (lm) – 783 800 1502 1536 400
Fin length, l (lm) – 550 653 1104 1301 253
Oblique angel, h (°) – 27.1 45.0 14.9 26.3 45.0

– NP denotes nominal pitch (800 lm) for enhanced microchannel heat sink with 200 lm channel width.
– LP denotes large pitch (1500 lm) for enhanced microchannel heat sink with 200 lm channel width.
– SP denotes small pitch (400 lm) for enhanced microchannel heat sink with 200 lm channel width.

Table 4
Comparison of convective heat transfer areas for microchannel heat sinks.

Configuration Oblique angle (°) Fin pitch (lm) Oblique fin perimeter (lm) Fin area (mm2) Un-finned area (mm2) Total heat transfer area (mm2)
Con-100 – – – 609.5 90.6 700.1
27° – NP0 26.3 405 976.6 718.7 107.5 826.2
45° – NP0 45.0 400 919.8 635.6 98.1 733.7
Con-200 – – – 354.0 80.8 434.8
27° – NP 27.1 783 1956.9 409.1 102.8 511.9
45° – NP 45.0 800 1864.8 406.4 93.6 500.0
15° – LP 14.9 1502 3718.2 441.6 100.5 542.1
27° – LP 26.3 1536 3476.4 359.1 92.4 451.5
45° – SP 45.0 400 1059.3 464.5 107.8 572.3

channel width. Both Con-200 and 45° – NP result in comparable aver-
age Nusselt numbers and pressure drop for Reynolds number up to
1000, while that of 27° – NP is consistently higher. This again
indicates that smaller oblique angle is better for heat transfer
enhancement. The gradient for 45° – NP in the average Nusselt num-
ber increases beyond Re 1000, suggesting that this configuration can
be more effective in heat transfer for higher Reynolds number.
The third microchannel group is three microchannel heat sinks
with 200 lm nominal channel width, with the conventional micro-
channel heat sink (Con-200) up against two enhanced microchannel
(1500 lm large fin pitch) with 15° and 27° oblique angle (15° – LP
and 27° – LP). Fig. 11(a) again demonstrates that smaller oblique
angle can lead to better heat transfer performance with higher aver-
age Nusselt number. 27° – LP also stands out from the comparison
with its comparable pressure drop with Con-200, within the range
of Reynolds number examined, even though the heat transfer
enhancement is not as high as 15° – LP.
From the assessment employing multiple microchannel groups,
it is consistently demonstrated that a smaller oblique angle con-
tributes to higher heat transfer performance, with a generally
higher pressure drop. Even though the smaller oblique angle
results in lesser flow resistance in secondary oblique channel than
the larger oblique angle, the longer oblique length contribute to
higher pressure drop for smaller oblique angle fin. Similar to liquid
 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336 333

Fig. 9. Comparison of (a) average Nusselt number, (b) pressure drop for microchannel heat sinks with different oblique angles (100 lm nominal channel width and 400 lm
nominal fin pitch).

Fig. 10. Comparison of (a) average Nusselt number, (b) pressure drop for microchannel heat sinks with different oblique angles (200 lm nominal channel width and 800 lm
nominal fin pitch).

flow through a bend, loss coefficient increases for sharper bend as
area of flow separation becomes extensive [24]. Thus, a smaller
oblique angle helps to translate the diffusive oblique channel into
a smoother and smaller flow area expansion from the main chan-
nel. This can help to reduce the boundary layer separation that
occurs in the oblique channel and increase the secondary flow gen-
eration. However, one should be aware that for a smaller oblique
angle with similar fin pitch, oblique fins will become thinner and
this might compromise the structural integrity. In fact, the author
has struggled to fabricate the enhanced microchannel heat sinks
with 15° oblique angle through mechanical wafer cutting.
3.2.2. The effect of oblique fin pitch variation
The impact of oblique fin pitch variation to the enhanced micro-
channel performance is evaluated in two groups of microchannel
heat sinks, with 200 lm nominal channel width, by varying the
oblique fin pitch while other parameters including oblique angle
are fixed. The first group of microchannel heat sinks includes a
conventional microchannel (Con-200) and two enhanced micro-
channel heat sinks (both have 27° oblique angle) with 800 lm
and 1500 lm fin pitch respectively (27° – NP and 27° – LP). The
second group of microchannel heat sinks for comparison comprises
of conventional microchannel (Con-200) and two enhanced micro-
channel heat sink (both have 45° oblique angle), with 400 lm and
800 lm fin pitch (45° – SP and 45° – NP). The experimental heat
transfer and pressure drop findings of the first group is presented
in Fig. 12. The average Nusselt number is the highest for 27° –
NP followed by 27° – LP while that of Con-200 is the lowest. It also
suggests that the maximum temperature is lower than conven-
tional microchannel. On the other hand, the pressure drop for
27° – NP is the highest while both 27° – LP and Con-200 display
comparable pressure drop. This indicates that shorter fin pitch is
beneficial for the heat transfer but incurs some pressure drop
penalty.
On the other hand, Fig. 13 displays the comparison of heat trans-
fer and pressure drop between Con-200, 45° – SP and 45° – NP
microchannel heat sinks. It is found that 45° – SP prevails with the
highest heat transfer performance followed by 45° – NP and Con-
200. Both 45° – SP and 45° – NP heat sinks demonstrate an incre-
ment in the heat transfer enhancement as Reynolds number rises
334 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336

Fig. 11. Comparison of (a) average Nusselt number, (b) pressure drop for microchannel heat sinks with different oblique angles (200 lm nominal channel width and 1500 lm
nominal fin pitch).

Fig. 12. Comparison of (a) average Nusselt number, (b) pressure drop for microchannel heat sinks (200 lm nominal channel width and 27° oblique angle) with different fin
pitches.

above 1000. In addition, the pressure drop for these two heat sinks is
higher than Con-200. Based on the evaluation with two different
microchannel groups, it can be concluded that smaller fin pitch is
beneficial for higher heat transfer performance with generally
higher pressure drop. The introduction of oblique fins and channels
aims to promote thermal boundary layer re-development and sec-
ondary flow generation, which are the critical elements in augment-
ing heat transfer performance. With the smaller fin pitch, more
oblique fins and channels can be inserted in a heat sink and this
enables the higher occurrence of boundary layer re-development
and secondary flow generation.
3.2.3. Overall performance comparison
Besides evaluating the heat transfer of microchannel heat sinks
to the Reynolds number, the total thermal resistance of the indi-
vidual heat sink is always compared between each other under
the constant pressure drop, and, also, pump power constraint, in
the selection process to determine the best performing heat sink
that meets the operating constraints. Fig. 14 demonstrates the
comparison of total thermal resistances of various microchannel
heat sinks under the constraint of pressure drop. Among all the
heat sinks tested, the enhanced microchannel with 100 lm chan-
nel width and 27° oblique channel (27° – NP0 ), achieves the lowest
total thermal resistance at the lowest pressure drop. This observa-
tion demonstrates the effectiveness of the enhanced microchannel
through the integration of smaller channel size (hydraulic diame-
ter) that has large surface area-to-volume ratio with short sec-
tional oblique fins. For instance, the pressure drop that is
required to achieve 0.119 °C/W employing conventional micro-
channel (Con-100) can be reduced by 55% (from 45 kPa to
25 kPa) by adopting the enhanced microchannel (27° – NP0 ).
A similar trend is observed for microchannel heat sinks with
200 lm nominal channel width. 45° – SP and 15° – LP are the
two microchannel configurations that record the lowest total ther-
mal resistance, in comparison to other microchannel heat sinks
with 200 lm nominal channel width. This finding also points out
the effectiveness of a smaller fin pitch and a smaller oblique angle
in enhancing heat transfer performance. More importantly, these
configurations perform better than conventional microchannel
heat sink with 100 lm nominal channel width at the similar
 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
Fig. 13. Comparison of (a) average Nusselt number, (b) pressure drop for microchannel heat sinks (200 lm nominal channel width and 45° oblique angle) with different
oblique fin pitches.
Fig. 14. Comparison of total thermal resistance to the pressure drop across
microchannel heat sinks.
pressure drop. By incorporating sectional oblique fins, both chan-
nel width and fin width of a microchannel heat sink can be relaxed,
allowing a much simpler, and economical, fabrication process.
In summary, the heat transfer of microchannel heat sinks can be
ranked as follow under constant pressure drop or pump power
constraint:
27° – NP0 > 45° – NP0 > Con-100, 45° – SP and 15° – LP > 27° –
NP > 27° – LP > 45° – NP > Con-200
3.3. Conclusion
Experimental validation is performed on the proposed heat
transfer augmentation concept, using silicon-based test vehicles.
Both local and overall characteristics in heat transfer are examined,
as well as pressure drop across the heat sinks. A parametric study
is then conducted to study the effect of variation in two critical
design parameters; oblique angle and oblique fin pitch. The follow-
ing key conclusions are drawn from this study:
335
1. Appreciable heat transfer augmentation is achieved with sili-
con-based enhanced microchannel with maximum heat trans-
fer performance enhancement at 47% when Re = 680.
Comparable pressure drop to conventional microchannel is
maintained up to Re = 500.
2. The parametric study shows that a smaller oblique angle is ben-
eficial for heat transfer enhancement. The smaller oblique angle
results in smaller flow resistance that could generate larger sec-
ondary flow rate through it.
3. Smaller fin pitch enables higher occurrence of thermal bound-
ary layer re-development and secondary flow generation, thus
further improve the heat transfer performance.
4. Among all the configurations tested, configuration 27° – NP0 is
the most effective microchannel heat sink through the combi-
nation of small hydraulic diameter (104 lm), small oblique
angle (27°) and small oblique fin pitch (400 lm). On the other
hand, enhanced microchannel heat sinks with 200 lm nominal
channel width demonstrate the capability to either equal or
enhance the thermal resistance of conventional microchannel
with 100 lm nominal channel width. This provides an opportu-
nity for channel width to be relaxed in the micro-fabrication
without compromising the heat transfer performance.
Conflicts of interest
None declared.
Acknowledgement
The authors gratefully acknowledge the MOE Academic
Research Fund (AcRF) Tier 2 research project Singapore for their
financial support. The Project title is Investigations on Micro scale
Transport Phenomena in Novel Oblique Structures with Project No.
MOE2011-T2-2-126 (WBS No: R265-000-423-112).
References
[1] J.R. Black, Electromigration – a brief survey and some recent results, IEEE
Trans. Electron Devices 16 (4) (1969) 338–347.
[2] D.B. Tuckerman, R.F.W. Pease, High-performance heat sinking for VLSI, IEEE
Electron Device Lett. 2 (5) (1981) 126–129.
[3] D. Copeland, H. Takahira, W. Nakayama, Manifold Microchannel Heat Sinks
Theory and Experiment, ASME – PUBLICATIONS – EEP, 1995.
336 Y.J. Lee et al. / International Journal of Heat and Mass Transfer 81 (2015) 325–336
[4] R. Prasher, J. Chang, Cooling of electronic chips using microchannel and micro-
pin fin heat exchangers, in: Proceedings of the 6th International Conference on
Nanochannels, Microchannels, and Minichannels, ICNMM, 2008, pp. 1881–
1887.
[5] A. Bejan, M.R. Errera, Convective trees of fluid channels for volumetric cooling,
Int. J. Heat Mass Transfer 43 (17) (2000) 3105–3118.
[6] Y. Chen, P. Cheng, Heat transfer and pressure drop in fractal tree-like
microchannel nets, Int. J. Heat Mass Transfer 45 (13) (2002) 2643–2648.
[7] D. Pence, Reduced pumping power and wall temperature in microchannel heat
sinks with fractal-like branching channel networks, Microscale Therm. Eng. 6
(4) (2003) 319–330.
[8] J.L. Xu, Y.H. Gan, D.C. Zhang, X.H. Li, Microscale heat transfer enhancement
using thermal boundary layer redeveloping concept, Int. J. Heat Mass Transfer
48 (9) (2005) 1662–1674.
[9] J. Xu, Y. Song, W. Zhang, H. Zhang, Y. Gan, Numerical simulations of interrupted
and conventional microchannel heat sinks, Int. J. Heat Mass Transfer 51 (25–
26) (2008) 5906–5917.
[10] Y.J. Lee, P.S. Lee, S.K. Chou, Hotspot mitigating with obliquely finned
microchannel heat sink – an experimental study, IEEE Trans. Compon.
Packag. Manuf. Technol. 3 (8) (2013) 1332–1341.
[11] M.E. Steinke, S.G. Kandlikar, Single-phase heat transfer enhancement
techniques in microchannel and minichannel flows, in: Second International
Conference on Microchannels and Minichannels, Rochester, NY, June, 2004, pp.
17–19.
[12] K.Tatsumi, M.Yamaguchi, Y. Nishio, K. Nakabe, Heat transfer and pressure loss
characteristics of obliquely-arranged cut-fins, in: Proceedings of the 13th
International Heat Transfer Conference 2006, CD ROM (No. THE-15), http://
dx.doi.org/10.1615/IHTC13.p17.150.
[13] Y.J. Lee, P.S. Lee, S.K. Chou, Enhanced microchannel heat sinks using oblique
fins, ASME (2009).
[14] Y. Fan, P.S. Lee, L. Jin, B.W. Chua, A simulation and experimental study of fluid
flow and heat transfer on cylindrical oblique-finned heat sink, Int. J. Heat Mass
Transfer 61 (2013) 62–72.
[15] Y.J. Lee, P.S. Lee, S.K. Chou, Enhanced thermal transport in microchannel using
oblique fins, J. Heat Transfer 134 (10) (2012) 101901.
[16] S.G. Kandlikar, S. Colin, Y. Peles, S. Garimella, R.F. Pease, J.J. Brandner, D.B.
Tuckerman, Heat transfer in microchannels – 2012 status and research needs,
J. Heat Transfer 135 (9) (2013) 091001.
[17] Y.J. Lee, P.S. Lee, S.K. Chou, Numerical study of fluid flow and heat transfer in
the enhanced microchannel with oblique fins, J. Heat Transfer 135 (4) (2013)
41901.
[18] W. Qu, Transport phenomena in single-phase and two-phase micro-channel
heat sinks, J. Electron. Packag. 126 (2004) 213–226.
[19] N.E. Todreas, M.S. Kazimi, Nuclear Systems I – Thermal Hydraulic
Fundamentals, Hemisphere Publishing Corporation, USA, 1989.
[20] W.M. Rohsenow, J.P. Hartnett, E.N. Ganić, Handbook of Heat Transfer
Fundamentals, second ed., McGraw-Hill publication, 1985.
[21] M.M. Yovanovich, Y.S. Muzychka, J.R. Culham, Spreading resistance of isoflux
rectangles and strips on compound flux channels, J. Thermophys. Heat
Transfer 13 (4) (1999) 495–500.
[22] J.R. Taylor, An Introduction to Error Analysis: The Study of Uncertainties in
Physical Measurements, University Science Books, 1997.
[23] R.J. Moffat, Describing the uncertainties in experimental results, Exp. Therm
Fluid Sci. 1 (1) (1988) 3–17.
[24] J.F. Douglas, J.M. Gasiorek, J.A. Swaffield, L.B. Jack, Fluid Mechanics, fifth ed.,
Pearson, 2005. p. 389.