See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245351182

Merits of Employing Foam Encapsulated Phase Change Materials for Pulsed

Power Electronics Cooling Applications

Article  in  Journal of Electronic Packaging · June 2008

DOI: 10.1115/1.2912185

CITATIONS READS

26 154

3 authors:

Khalid Lafdi Osama Mesalhy

University of Dayton University of Central Florida
215 PUBLICATIONS   3,284 CITATIONS    21 PUBLICATIONS   660 CITATIONS   

SEE PROFILE SEE PROFILE

Ahmed Elgafy
University of Cincinnati
22 PUBLICATIONS   766 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Élaboration de capteurs électrochimiques et de membranes de nano-filtration à base de nanocomposite polyméres-graphène ou d'oxyde métallique-graphène: Application
à la détection et la rétention des cations métalliques View project

Nanomaterials - nanoparticles analysis and behaviours for different applications such as medical, biotechnology and environmental ones View project

All content following this page was uploaded by Khalid Lafdi on 16 June 2015.

The user has requested enhancement of the downloaded file.

 Merits of Employing Foam
Encapsulated Phase Change
K. Lafdi1 Materials for Pulsed Power
e-mail: lafdi@udri.udayton.edu

O. Mesalhy
Electronics Cooling Applications
In the present work, the potential of using foam structures impregnated with phase
University of Dayton,
change materials (PCMs) as heat sinks for cooling of electronic devices has been nu-
300 College Park,
merically studied. Different design parameters have been investigated such as foam prop-
Dayton, OH 45469
erties (porosity, pore size, and thermal conductivity), heat sink shape, orientation, and
use of internal fins inside the foam-PCM composite. Due to huge difference in thermal
properties between the PCM and the solid matrix, two energy equation model has been
A. Elgafy adopted to solve the energy conservation equations. This model can handle local thermal
MET Department, College of Applied Science,
nonequilibrium condition between the PCM and the solid matrix. The numerical model is
University of Cincinnati,
based on volume averaging technique, and the finite volume method is used to discretize
Cincinnati, OH, 45206
the heat diffusion equation. The findings show that, for steady heat generation, the shape
and orientation of the composite heat sink have significant impact on the system perfor-
mance. Conversely, in the case of power spike input, use of a PCM with low melting point
and high latent heat is more efficient. 关DOI: 10.1115/1.2912185兴

Keywords: heat sinks, PCM, foams, electronic devices

1 Introduction the generated power could be converted into oscillations in the

In new electronic applications, the electronic chips are sub-
jected to high heat fluxes due to the increasing demand to reduce
their sizes and to increase their functions. The generated heat from
these electronic chips may be steady or transient in multichip
modules and sensors that irregularly operate or have variable
power. Dissipating heat from these devices and suppressing tem-
perature overshoots during the transient power spikes has become
an important challenge in electronic packaging design. As a con-
sequence, the cooling system must be designed for the peak power
operation; otherwise, another system should be incorporated to
accommodate the power variation during the transient operation
conditions. Different designs of active cooling have been used
utilizing microchannels and finned high thermal conductivity
foams 关1,2兴. Nevertheless, these kinds of designs for active cool-
ing will effectively work for steady state operations, but for tran-
sient operations, the temperature may jump above the maximum
allowed temperature, which is assigned to achieve efficient per-
formance. Additionally, using active cooling by circulating a cool-
ant fluid to absorb the heat from these kinds of electronic chips is
known to consume too much power and to add complexity to the
system design. Therefore, developing a new generation of elec-
tronic devices, which have high capabilities and high mobility,
requires using passive cooling systems, which can be achieved
using phase change materials 共PCMs兲. These systems do not need
any power for cooling. In some cases, just a dc fan represents the
best-compromised solution. PCMs seem to be a good choice for
use in passive cooling of electronic devices. The latent heat of
fusion, which is usually higher than the sensible heat, adds an
extra heat capacity to the cooling system. Also, the fluctuations of
1
Corresponding author.
Contributed by the Electrical and Electronic Packaging Division of ASME for
publication in the JOURNAL OF ELECTRONIC PACKAGING. Manuscript received February
18, 2007; final manuscript received November 12, 2007; published online April 25,
2008. Assoc. Editor: Dereje Agonafer.
solid-liquid interface without significant fluctuations in the elec-
tronic device temperature.
Some analytical studies have been carried out to investigate the
potential use of PCM in the cooling applications of high power
electronics 关3,4兴. Yet, the disadvantage of using such analytical
techniques is that they are limited to simple configurations and
boundary conditions. Siva et al. 关5兴 numerically and experimen-
tally studied the feasibility of using metallic solid-liquid PCM in
periodic power dissipating devices. They noticed that PCM per-
forms well at lower power levels for silicon carbide semiconduc-
tor devices, but the use of high thermal conductivity spreaders
such as diamond becomes mandatory at higher power levels. A
thermal analysis was performed for a package design relevant to
power electronics by Evans et al. 关6兴. The role of PCM in sup-
pressing temperature elevations during pulses was evident.
For mobile electronic devices, PCMs appear to be a good
choice to extend the operational periods. Marc et al. 关7兴 experi-
mentally studied the feasibility of transient thermal management
of handsets using PCMs. The results showed that using PCMs
could substantially increase the time of service for the handset.
Tan and Tso 关8兴 experimentally investigated cooling of electronic
devices using a heat storage unit filled with PCM inside the de-
vice. They noticed that using PCM maintains the chip temperature
below the allowable service temperature for 2 h of transient op-
eration.
PCMs have also been used to reduce the temperature of photo-
voltaic devices to increase their efficiency. Huang et al. 关9兴 studied
the use of PCMs to moderate building integrated photovoltaic
temperature rises experimentally and numerically. The results
showed that the moderation of temperature achieved could lead to
significant improvements in the operational efficiency of photo-
voltaic facades.
On the other hand, different types of PCMs have been utilized
such as metallic and organic PCMs. Metallic PCMs have the ad-
vantage of having high thermal conductivity, but at the same time
suffer from having higher densities, which add more weight to the
system. On the other hand, organic PCMs have high latent heat

Journal of Electronic Packaging Copyright © 2008 by ASME JUNE 2008, Vol. 130 / 021004-1

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 and lower densities but suffer from having very low thermal con- Table 1 Source term definitions
ductivity 关10兴. Shankar and Suresh 关11兴 concluded that the use of
organic PCMs without thermal conductivity enhancers is a poor Su Sv STf STs
choice due to its low thermal conductivity. As a consequence,
many studies have been conducted to enhance the thermal perfor- ␮ ␳C兩u兩u ␮ ␳C兩u兩v hsf asf 共Ts − T f 兲 hsf asf 共T f − Ts兲
− u− − v−
mance of PCMs. Bugaje 关12兴 introduced methods of enhancing K 冑K K 冑K + −␦␳L
df l
the thermal response of organic PCM by incorporation of alumi- ␳g␤ f 共T f − Tref兲 dt
num thermal conductivity promoters of various designs. On the
contrary, Elgafy and Lafdi 关13兴 dispersed two types of carbon
nanofibers with different surface characteristics and different mass
J ⳵␳u 1 ⳵␳Uu 1 ⳵␳Vu
ratios into organic PCM to enhance its thermal conductivity. They + 2 + 2
reported that the thermal performance of the composite was sig- ␧ ⳵t ␧ ⳵␰ ␧ ⳵␩
nificantly enhanced by increasing the mass ratio and by activating
the nanofiber surface. Wirtz et al. 关14兴 developed a multifunc-
tional lamination, which consists of a paraffin impregnated porous
= − y␩
⳵p
⳵␰
⳵p ⳵ ␮
+ y␰ +
⳵␩ ⳵␰ J␧
⳵u
a11 − a12
⳵␰
⳵u
⳵␩
冋冉 冊册
graphite core encapsulated between rigid graphite/epoxy compos-
ite skins. The study showed that the graphite foam is an effective
thermal conductivity enhancer to the imbedded paraffin and it has
+
⳵ ␮
冋冉
⳵␩ J␧
⳵u
− a12 + a22
⳵␰
⳵u
⳵␩
冊册
effectively immobilized the liquid paraffin. On the other hand, + Su共␰, ␩兲J 共2兲
Ishizuka 关15兴 studied the performance of metallic alloy PCM to
control the operational time of high density electronic packaging In v-momentum,
by using a thermal network method. The thermal experiments us- J ⳵␳v 1 ⳵␳Uv 1 ⳵␳Vv
ing this alloy confirmed that the substrate backsurface temperature + 2 + 2
␧ ⳵t ␧ ⳵␰ ␧ ⳵␩

冋冉 冊册
could be kept constant at the melting temperature of the PCM for
several minutes by thermal absorption, while the PCM phase ⳵p ⳵p ⳵ ␮ ⳵v ⳵v
= − x␰ + x␩ + a11 − a12
changed from solid to liquid state. Also, in his study, he confirmed ⳵␩ ⳵␰ ⳵␰ J␧ ⳵␰ ⳵␩

冋冉 冊册
that the thermal network method is practical for the thermal de-
sign of packages with PCM. ⳵ ␮ ⳵v ⳵v
+ − a12 + a22
In the present work, the potential of using different foam struc- ⳵␩ J␧ ⳵␰ ⳵␩
tures impregnated with PCMs as heat sinks for cooling of elec-
tronic devices is introduced. Different design parameters have + Sv共␰, ␩兲J 共3兲
been investigated such as foam properties 共porosity, pore size, and In PCM phase energy equation,

冋 冉 冊册
thermal conductivity兲, heat sink shape, heat sink orientation, and
use of internal fins inside the foam-PCM composite. Two energy ⳵␳c p f T f ⳵␳c p f UT f ⳵␳c p f VT f ⳵ k fe ⳵T f ⳵T f
J␦ + + = a11 − a12
equation model has been adopted to solve the energy conservation ⳵t ⳵␰ ⳵␩ ⳵␰ J ⳵␰ ⳵␩

冋 冉 冊册
equations. The numerical model is based on volume averaging
technique, and the finite volume method is used to discretize the ⳵ k fe ⳵T f ⳵T f
+ − a12 + a22 + ST f 共␰, ␩兲J 共4兲
heat diffusion equation. A line-by-line solver based on tridiagonal ⳵␩ J ⳵␰ ⳵␩
matrix algorithm 共TDMA兲 is used to iteratively solve the alge- In solid matrix energy equation,

冋 冉 冊册
braic discretization equations.
⳵␳c psTs ⳵ kse ⳵Ts ⳵Ts
J共1 − ␦兲 = a11 − a12
⳵t ⳵␰ J ⳵␰ ⳵␩
2 Numerical Model
2.1 Governing Equations and Assumptions. The governing
+ 冋 冉
⳵ kse
⳵␩ J
− a12
⳵Ts
⳵␰
+ a22
⳵Ts
⳵␩
冊册 + STs共␰, ␩兲J

partial differential equations in 2D describing the melting of PCM 共5兲

inside the porous matrix are obtained from volume averaging of
the main conservation equations of mass, momentum, and energy. where ␦ is the solid matrix porosity and ␧ is the liquid phase
Due to the large difference in thermal properties between the PCM fraction. The value of ␧ is defined as the product of the PCM
and the solid matrix in most of the cases, two energy equation liquid fraction and the foam porosity. The geometrical diffusion
model is adopted to solve the energy conservation equations. This coefficients a11, a12, a21, a22 and Jacobian J are given by
model can handle local thermal nonequilibrium condition between
the PCM and the solid matrix. Through formulating the problem,
the PCM and the porous matrix are assumed to have constant
a11 = 冉 冊 冉 冊 冉 冊冉 冊 冉 冊冉 冊
⳵x
⳵␩
2
+
⳵y
⳵␩
2
, a12 =
⳵x
⳵␰
⳵x
⳵␩
+
⳵y
⳵␰
⳵y
⳵␩
properties, and the volume change due to the melting process is
neglected. Finite volume technique in conjunction with numerical
grid generation based on body fitted coordinate transformation has
= − a21,冉 冊 冉 冊 a22 =
⳵x
⳵␰
2
+
⳵y
⳵␰
2

been adopted. The governing equations are transformed into 2D

general coordinates by making coordinate transformations from
the regular 共x-y兲 coordinate to the general coordinate system
冉 冊冉 冊 冉 冊冉 冊
J=
⳵x
⳵␰
⳵y
⳵␩
−
⳵x
⳵␩
⳵y
⳵␰
共␰ , ␩兲. The contravariant velocities U and V are defined as
The mathematical formulation of the governing equations on ⳵y ⳵x ⳵x ⳵y
the nonorthogonal curvilinear system can be written as follows. U=u −v and V = v −u
In continuity, ⳵␩ ⳵␩ ⳵␰ ⳵␰
The source terms that appear in the equations can be expressed in
⳵ ⳵ ⳵ Table 1. The first term that appears in the momentum source terms
共␳J兲 + 共␳U兲 + 共␳V兲 = 0 共1兲 in the u and v directions is the Darcy effect where K is the per-
⳵t ⳵␰ ⳵␩ meability of the porous matrix. The second term represents the
In u-momentum, Forchiemer effect and C is an empirical constant. ␤ f is the fluid

021004-2 / Vol. 130, JUNE 2008 Transactions of the ASME

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 Fig. 2 Physical domain of the heat sink

thermal expansion coefficient and g is the acceleration of gravity.

The first term that appears in energy source terms STf and STs
represents the heat exchange between the solid matrix and PCM
phase. The second terms in STf represent the melting source terms
where f l and L are the liquid fraction and the latent heat of fusion,
respectively.
2.2 Grid Generation. The grid generation is carried out by
solving the differential equations, which describe the coordinate
transformation between 共x , y兲 and 共␰ , ␩兲 coordinates.
a11x␰␰ − 2a12x␰␩ + a22x␩␩ + J2共Px␰ + Qx␩兲 = 0

a11y ␰␰ − 2a12y ␰␩ + a22y ␩␩ + J2共Py ␰ + Qy ␩兲 = 0

where P and Q are controlling functions that can be used to con-
centrate the grid near the walls where large gradient can be antici-
pated. These parameters are specified as mentioned by Warsi et al.
关16兴.
2.3 Numerical Solution. The governing equations have been
discretized by using a finite volume approach. A nonstaggered
grid is used where all the variables 共velocities and all the other
scalar variables兲 are stored at the control volume center. The stan-
dard semi-implicit method for pressure linked equation 共SIMPLE兲
algorithm has been used to solve the coupled continuity and mo-
mentum equations. Rhie and Chow’s interpolation technique 关17兴
is used to define the convection flux at the control volume sur-
faces to eliminate the pressure oscillations. A fully implicit formu-
lation has been used for the time-dependent terms, and the com-
bined convection diffusion coefficients are evaluated by using a
hybrid difference scheme. A line-by-line solver based on TDMA
is used to iteratively solve the algebraic discretized equations 关18兴.
The convergence is assumed to be achieved when the dependent
variable residuals and the mass imbalance, which appears in pres-
sure correction equation, agreed to four significant digits at each
nodal point. In porous regions, the solid phase of the porous foam
is assumed to exchange heat with the PCM through the foam
interfacial surface area, hsf asf 共TPCM-Ts兲. The porous matrix ge-
ometry is assumed to be an intersected mesh of fibers with a
circular cross section area and with diameter d fi 关19兴. The estima-

Fig. 1 „a… Solid-liquid interface after 10 min: comparison be- Table 2 PCM properties for the heat sink
tween present model and predicted results given in Ref. †21‡;
„b… snapshots for liquid-solid interface at time= 75 min for dif- Property Property
ferent porosity aluminum foams infiltrated with PCM, Ref. †22‡;
and „c… comparison between present numerical findings and Tm 共°C兲 36 ␳ 共kg/ m3兲 769
experimental findings presented in Ref. †22‡ for solid-liquid in- k 共W/m K兲 0.15–0.2 L 共kJ/kg兲 247.3
terface for aluminum foam 10 PPI, 93.4% porosity at different ␮ 共kg/m s兲 0.00415 ␤ 共1/K兲 0.00085
time levels c p 共J/kg K兲 2460

Journal of Electronic Packaging JUNE 2008, Vol. 130 / 021004-3

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 tion procedures followed to define the interfacial heat transfer
coefficient h fs between the PCM and the solid matrix is based on
quasisteady heat conduction between the porous matrix and the
PCM with neglecting the liquid phase motion. In this case, the
heat exchange between the matrix fibers and the PCM can be
characterized by heat conduction 关19兴. The interfacial surface area
asf and permeability of the porous matrix are assumed, as pro-
posed by Fourie and Du Plessis 关20兴. While the effective thermal
conductivity of the solid matrix and the PCM has been consid-
ered, as given by Mesalhy et al. 关19兴.
2.4 Model Verification. The present numerical code has been
validated by comparing its findings with both numerical and ex-
perimental results performed by previous works. First, the numeri-
cal code has been applied on the case of Khillarkar et al. 关21兴. The
solid-liquid interface for the two works is illustrated in Fig. 1共a兲,
which shows a good agreement between them. Second, the find-
ings of the present numerical code have been compared to the
experimental results performed by Mesalhy 关22兴. Figure 1共b兲
shows snapshots of the liquid-solid interface for different porosity
aluminum foams infiltrated with PCM 关22兴.
Figure 1共c兲 shows a comparison between the present predicted Fig. 3 PCM and solid phase temperature contours: continu-
results and the experimental results performed by Mesalhy 关22兴 ous „solid… and dotted „PCM…
for solid-liquid interface for an aluminum foam of 93.4% porosity
and 10 PPI. The thick lines in this figure represent the predicted
mushy zones between the totally liquid PCM on the left side and
the totally solid PCM on the right side, while the dotted lines heat generation condition are shown in Fig. 3. This case corre-
represent the extracted points from the pictures. It can be observed sponded to the use of aluminum foam of porosity 94% and 10
from the figure that the numerical model has accurately predicted PPI. It can be observed from the figure that the PCM temperature
the liquid-solid interface. deviated from the foam temperature, which means that the heat
2.5 Physical Domain. A heat sink consisting of high thermal transfer was far away from the local thermal equilibrium. The heat
conductivity foam 共aluminum or carbon兲 saturated with a PCM diffused faster inside the foam material while the solid-liquid in-
represents the physical domain, as shown in Fig. 2. The dimen- terface acted as a thermal barrier. This prevented the heat from
sions of the heat sink are assumed to be 10 cm in height and 5 cm diffusing inside the solid PCM region. At the same time, the hot
in width. The heating source is an electronic chip brazed to an PCM liquid phase beside the heating source moved to the top
aluminum substrate. A thin aluminum enclosure is attached to the region due to the buoyancy forces. The PCM temperature is
substrate to contain the foam with the PCM. To enhance the con- higher beside the hot substrate and at the top region of the heat
vective heat transfer from the surface, extended fins are attached sink, while the solid temperature is higher beside the right wall
to the heat sink enclosure. The heat transfer coefficient from the and at the bottom region of the heat sink.
enclosure walls depends on the cooling system and the number of The heat exchange between the two phases made the solid tem-
the attached fins. A suitable value is assumed 共200 W / m2 K兲, al- perature to follow the same pattern as the PCM phase. The result-
ing HSR for steady heat generation ranged from 0.34 to 0.404.
though the actual magnitude is something arbitrary. Changing this
The temperature-time history at three different locations at the
value will not affect the outcome of the present study, yet it would
substrate is presented in Fig. 4, which shows that the temperatures
vary final results without varying trends. Due to the small thick-
at the top and the bottom levels inside the foam were close to each
ness of the real electronic chip, its heat capacity is neglected, and
other at the beginning of the melting process. After elapsing
the internal heat generation is replaced by a heat flux at the alu-
500 s, they started to deviate from each other due to the convec-
minum substrate surface. The foam material is assumed to be in
tion motion of the liquid phase.
good contact with the enclosure walls and with the substrate. In
the case of applying steady heat generation, a 100 W heat source
is assumed, which is comparable to the power generated from a
high-energy power electronic chip. Ecosane, which has a melting
temperature of 36° C, is used as the PCM, Table 2.
The substrate and the thin aluminum enclosure that contained
the foam-PCM composite have been included in the computa-
tional domain. In the case of applying pulsed heat generation, the
heat source power is assumed to fluctuate like a step function
between a minimum and a maximum value. Different design pa-
rameters have been studied such as foam properties 共porosity, pore
size, and thermal conductivity兲, heat sink shape, heat sink orien-
tation, and use of internal fins inside the foam-PCM composite.

3 Results and Discussion

3.1 Steady Heat Generation. For steady heat generation, the
results of the heat sink design are represented in terms of the
percentage of improving the heat sink resistance 共HSR兲, which is
defined as the maximum temperature rise per unit input power.
The lower value means better performance. The temperature con- Fig. 4 Temperature-time history at different points in the junc-
tours for both of the PCM and the foam materials under steady tion for Al foam „porosity= 94% and pore size= 10 PPI…

021004-4 / Vol. 130, JUNE 2008 Transactions of the ASME

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 Fig. 5 Maximum junction temperature for different foam Fig. 7 Maximum junction temperature for different interfacial
porosities heat transfer coefficients

3.2 Changing Foam Porosity, Pore Size, and Interfacial
Heat Transfer Coefficient. It has been depicted from the litera-
ture review that most important foam parameters, which can affect
the heat sink performance, are the foam thermal conductivity, pore
size, and the interfacial heat transfer coefficient between the foam
and the PCM. The foam effective thermal conductivity depends
on the foam porosity while permeability depends on both the po-
rosity and the pore size. The porosity has been changed from 88%
to 97%, which is the maximum value that aluminum foam can
attain. The pore size has been changed from 0.0012 m 共corre-
sponding to 40 PPI兲 to 0.00434 m 共corresponding to 5 PPI兲. Only
one parameter is allowed to change; the pore size is kept constant
while the porosity changed and vice versa. Maximum junction
temperature-time histories for using different foam porosities,
pore sizes, and interfacial heat transfer coefficients are shown in
Figs. 5–7, while Table 3 summarizes the HSR for the heat sink
and the percentage of change of the HSR. For each figure, one
parameter varies, and the other two parameters are kept constant.
For instance, as in Fig. 5, the foam porosity varies, while the pore
size and interfacial heat transfer coefficient are kept constant.
From Figs. 5–7 and Table 3, one can see that decreasing the po-
rosity from 97% to 90% decreased the HSR but decreasing poros-
ity from 90% to 88% did not make a significant difference. This
occurred due to the reduction in liquid motion in low porosity
foam. On the other hand, for the same foam porosity, changing the
pore size created a significant change in the HSR. Changing the
pore size from 5 PPI to 40 PPI showed more than 7.5% change in
HSR, which can only be achieved by changing the porosity from
97% to 88%. This means that by using high pore size foam, we
can save about 11% of the foam material weight. Moreover,
changing the interfacial heat transfer coefficient by ten times can
improve the performance by 5%. However, after increasing hsf
over 500, the change in HSR became negligible. This is because at
this high value of hsf , the PCM and solid foam reach local thermal
equilibrium 共Ts = TPCM兲, which means that any further increase in
hsf cannot make any difference.
3.3 Adding Internal Fins. Using internal fins can alter the
spreading of the heat inside the heat sink. In this design study, we
checked the effect of adding one or two internal fins inside the
porous matrix. In the case of adding one fin, a 3 mm thickness fin
was attached to the substrate. In the case of using two fins, each
fin thickness was assumed to be 1.5 mm to keep the same extra
weight to the system constant. The presence of the fins signifi-
cantly influenced the temperature field inside the heat sink, as
shown from the PCM temperature contours in Fig. 8. It is found
that inserting these supporting fins especially at the early times of
the melting process during conduction heat transfer mode has en-
hanced the heat diffusion. However, increasing the number of the
fins from 1 to 2 did not lead to a significant difference on the
steady state junction temperature. As shown in Fig. 9, adding two
fins may slightly enhance the heat sink performance. However,
adding more than two fins may deteriorate the performance by
suppressing the liquid phase convection motion.
3.4 Changing Shape and Orientation. To investigate the ef-
fect of the shape and orientation of the heat sink, the shape of the
heat sink has been changed by increasing the width of the top
edge and decreasing the bottom edge. The volume of the heat sink
is kept constant, as the rectangular shape. Figure 10 shows the
PCM temperature contours after 8 min from applying the heat
source for different angles of the right side of the heat sink enclo-
sure while Figs. 11 and 12 show the temperature contours for
different heat sink orientations, heating from bottom and top. The
temperature contours took the same pattern in the case of chang-

Table 3 Changing of HSR for different foam properties

% of
change
Change HSR of HSR

共97%, 10 PPI, hsf = 200兲 0.425 −5.2

共94%, 10 PPI, hsf = 200兲, 0.404 0
base case
共90%, 10 PPI, hsf = 200兲 0.398 1.485
共88%, 10 PPI, hsf = 200兲 0.397 1.7326
共94%, 5 PPI, hsf = 200兲 0.398 1.485
共94%, 20 PPI, hsf = 200兲 0.411 −2.475
共94%, 40 PPI, hsf = 200兲 0.429 −6.188
共94%, 10 PPI, hsf = 50兲 0.418 −3.465
共94%, 10 PPI, hsf = 500兲 0.4 1.0
Fig. 6 Maximum junction temperature for different foam pore 共94%, 10 PPI, hsf = 1000兲 0.398 1.485
sizes

Journal of Electronic Packaging JUNE 2008, Vol. 130 / 021004-5

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 Fig. 8 PCM temperature contours for a heat sink supported with one and
two fins
ing the inclination angle of the right side like the rectangular case
but by changing the heat sink orientation, the temperature con-
tours became completely different. In the case of heating from the
bottom side, small Benard convection cells appeared. These cells
merged with each other to make bigger ones as the melting pro-
cess proceeded. At the end of melting, after all the PCM converted
to liquid, only one big cell appeared. For heating from the top
side, the temperature contours were straight, exactly like the pure
conduction case except that the temperature showed higher values
Fig. 9 Maximum junction temperature for using fins
Fig. 10 PCM phase temperature contours for different angles „80 deg and
70 deg…
021004-6 / Vol. 130, JUNE 2008
Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
beside the metal enclosure wall. Figure 13 shows the effect of
changing the heat sink shape and orientation on the maximum
junction temperature. The calculated HSR values for the case of
adding fins, changing shape, and changing the orientation are
listed in Table 4. It is obvious that changing orientation created
the biggest effect on the heat sink performance. The difference in
the HSR for top and bottom heating was greater than 27%.
3.5 Performance for Fluctuating Power Generation. The
PCM played the role of a temperature damper in the heat sink,
which is subjected to fluctuating heat generation. The key param-
eters that define the performance of the temperature damping ef-
fect of the PCM are the melting point and the latent heat. The
melting point defines the starting point of energy absorption as a
latent heat, while the latent heat defines the capacity of the energy
storage. In this section, the effect of changing the PCM melting
temperature and latent heat on a heat sink performance subjected
to periodic spikes of heat generation has been investigated. The
generated heat ranged from a minimum value of 30 W to gener-
ated energy spikes of 150 W, which lasted for 10 s and repeated
every 300 s, as shown in Fig. 14.
The latent heat has been changed from
150 kJ/ kg to 250 kJ/ kg, and the melting temperature was
changed from 36° C, which was just above the ambient tempera-
ture by 0.5– 45° C. As shown from Figs. 15 and 16, the junction
temperature for both cases followed the same frequency as the
Transactions of the ASME
 Table 4 HSR for adding fins and changing shape and
orientation

Case HSR % of improvement

One fin 0.394 2.475

Two fins 0.391 3.217
Angle 80 0.397 1.733
Angle 70 0.389 3.713
Bottom heating 0.359 11.14
Top heating 0.471 −16.58

Fig. 11 PCM phase temperature contours for bottom heating
applied heat flux. Using the higher latent heat and the lower melt-
ing temperature PCM resulted in lower junction temperature. This
result was expected because the lower the melting temperature,
the sooner the latent heat will participate in damping the tempera-
ture, and higher latent heat will produce a higher damping effect.
It seems that using PCMs in heat sink designs is a very effective
solution especially when the system experiences various levels of
energy spikes. As shown in Table 5, the calculated HSR based on
the maximum power ranged from 0.108° C / W to 0.136° C / W
while it was about 0.404° C / W for the steady heat generation
case. This was due to the heat absorption inside the PCM as a
latent heat during the energy spike period and the release of it
again during the minimum heat generation period.
4 Conclusions
A parametric study has been carried out to investigate the key
parameters that affect the performance of foam-PCM composite
as a heat sink for electronic devices. Some of these parameters are
dependent on the properties of the foam 共such as porosity, thermal
conductivity, and pore size兲 and PCM 共melting point, latent heat,
and viscosity of the liquid phase兲. It is found that decreasing foam
porosity improves the performance of the heat sink while using
foam with large pore size improves its performance. On the other
hand, adding a limited number of fins inside the foam structure
improves heat spreading, which may enhance the performance of
the heat sink. Conversely, changing the orientation and shape of
the heat sink are the most important parameters that can affect the
heat sink performance for steady heat generation. On the contrary,
use of PCM could be a very effective solution for designing heat
sinks for electronic chips subjected to heat generation spikes.
Lower melting point and higher latent heat resulted in better per-
formance.

Fig. 14 Heat flux at the left side for fluctuating energy

Fig. 12 PCM phase temperature contours for top heating

Fig. 13 Maximum junction temperature for different shapes Fig. 15 Maximum junction temperature for different latent
and orientations heats

Journal of Electronic Packaging JUNE 2008, Vol. 130 / 021004-7

Downloaded 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 Subscripts
f ⫽ PCM phase
fe ⫽ PCM effective
s ⫽ solid matrix phase
se ⫽ solid effective
fi ⫽ fiber
i ⫽ inner
o ⫽ outer
References
关1兴 Li, J., Peterson, G. P., and Cheng, P., 2004, “Three-Dimensional Analysis of
Heat Transfer in a Micro-Heat Sink With Single Phase Flow,” Int. J. Heat
Mass Transfer, 47, pp. 4215–4231.
关2兴 Bhattacharya, A., and Mahajan, R. L., 2002, “Finned Metal Foam Heat Sinks
for Electronics Cooling in Forced Convection,” ASME J. Electron. Packag.,
124, pp. 155–163.
关3兴 Lu, T. J., 2000, “Thermal Management of High Power Electronics With Phase
Fig. 16 Maximum junction temperature for different melting
Change Cooling,” Int. J. Heat Mass Transfer, 43, pp. 2245–2256.
temperatures 关4兴 Chakraborty, S., and Dutta, P., 2003, “Analytical Solution for Heat Transfer
During Cyclic Melting and Freezing of Phase Change Material Used in Elec-
tronic or Electrical Packaging,” ASME J. Electron. Packag., 125, pp. 126–
133.
Nomenclature 关5兴 Siva, P., Yogendra, K., and Jungho, K., 2002, “Thermal Management of High
cp ⫽ specific heat 共J/kg K兲 Temperature Pulsed Electronics Using Metallic Phase Change Materials,” Nu-
mer. Heat Transfer, Part A, 42, pp. 777–790.
dfi ⫽ fiber diameter 共m兲 关6兴 Evans, A. G., He, M. Y., Hutchinson, J. W., and Shaw, M., 2001, “Temperature
dp ⫽ pore diameter 共m兲 Distribution in Advanced Power Electronics Systems and the Effect of Phase
g ⫽ gravity acceleration 共m / s2兲 Change Materials on Temperature Suppression During Power Pulses,” ASME
J. Electron. Packag., 123, pp. 211–217.
hsf ⫽ interfacial heat transfer coefficient 共W / m2 K兲 关7兴 Marc, H., Randy, D. W., Stephen, J. P., Jason, M. P., Lou, M., and Calvin, C.,
J ⫽ transformation Jacobian 2002, “Transient Thermal Management of a Handset Using Phase Change
K ⫽ permeability 共m2兲 Material 共PCM兲,” ASME J. Electron. Packag., 124, pp. 419–426.
关8兴 Tan, F. L., and Tso, C. P., 2004, “Cooling of Mobile Electronic Devices Using
k ⫽ thermal conductivity 共W/m K兲 Phase Change Materials,” Appl. Therm. Eng., 24, pp. 159–169.
L ⫽ latent heat of fusion 共J/kg兲 关9兴 Huang, M. J., Eames, P. C., and Norton, B., 2004, “Thermal Regulation of
p ⫽ static pressure 共Pa兲 Building-Integrated Photovoltaics Using Phase Change Materials,” Int. J. Heat
Mass Transfer, 47, pp. 2715–2733.
T ⫽ temperature 共K兲 关10兴 Pal, D., and Joshi, Y. K., 1997, “Application of Phase Change Materials to
t ⫽ time 共s兲 Thermal Control of Electronic Modules: A Computational Study,” ASME J.
u ⫽ velocity in the x-direction 共m/s兲 Electron. Packag., 119, pp. 40–50.
关11兴 Shankar, H., and Suresh, V. G., 2004, “Thermal Management of Transient
v ⫽ velocity in the y-direction 共m/s兲 Power Spikes in Electronics—Phase Change Energy Storage or Copper Heat
x ⫽ distance in the x-direction Sinks,” ASME J. Electron. Packag., 126, pp. 308–316.
y ⫽ distance in the y-direction 关12兴 Bugaje, M. I., 1997, “Enhancing the Thermal Response of Latent Heat Storage
Systems,” Int. J. Energy Res., 21, pp. 759–766.
Greek Letters 关13兴 Elgafy, A., and Lafdi, K., 2005, “Effect of Carbon Nanofiber Additives on
␣ ⫽ thermal diffusivity 共m2 / s兲 Thermal Behavior of Phase Change Materials,” Carbon, 43, pp. 3067–3074.
关14兴 Wirtz, R., Fuchs, A., Narla, V., Shen, Y., Zhao, T., and Jiang, Y., 2003, “A
␤f ⫽ thermal expansion coefficient 共1 / K兲 Multi-Functional Graphite/Epoxy-Based Thermal Energy Storage Composite
␦ ⫽ porosity of solid matrix for Temperature Control of Sensors and Electronics,” AIAA Paper No. 0513.
关15兴 Ishizuka, M., 2004, “Operation Time Control of a High Density Packaging
␧ ⫽ liquid phase fraction Using a Low Melting Point Alloy,” IEEE Trans. Compon. Packag. Technol.,
␰,␩ ⫽ coordinates in computational plane 27共2兲, pp. 239–243.
␳ ⫽ density 共kg/ m3兲 关16兴 Warsi, Z. U., Thompson, J. F., and Mastin, C. W., 1985, Numerical Grid
⌬t ⫽ time step 共s兲 Generation Foundations and Applications, North-Holland, Amsterdam.
关17兴 Rhie, C. M., and Chow, W. L., 1983, “Numerical Study of the Turbulent Flow
␮ ⫽ dynamic viscosity 共kg/共m s兲兲 Past an Airfoil With Trailing Edge Separation,” AIAA J., 21, pp. 1525–1533.
关18兴 Patankar, S. V., 1980, Numerical Heat Transfer and Fluid Flow, Hemisphere,
Washington, DC.
关19兴 Mesalhy, O., Lafdi, K., Elgafy, A., and Bowman, K., 2005, “Numerical Study
Table 5 HSR for fluctuating heat generation for Enhancing the Thermal Conductivity of Phase Change Material 共PCM兲
Storage Using High Thermal Conductivity Porous Matrix,” Energy Convers.
% of improvement Manage., 46, pp. 847–867.
HSR based on compared to steady 关20兴 Fourie, J. G., and Du Plessis, J. P., 2002, “Pressure Drop Modeling in Cellular
Case the max power heat generation Metallic Foams,” Chem. Eng. Sci., 57, pp. 2781–2789.
关21兴 Khillarkar, D. B., Going, Z. X., and Mujumdar, A. S., 2000, “Melting of a
Phase Change Material in Concentric Horizontal Annuli of Arbitrary Cross-
Tm = 36, L = 250 kJ/ kg 0.1086 73
Section,” Appl. Therm. Eng., 20, pp. 893–912.
Tm = 46, L = 250 kJ/ kg 0.136 66.3 关22兴 Mesalhy, O., 2005, “Heat Transfer Phenomena in Foams Infiltrated With Phase
Tm = 36, L = 150 kJ/ kg 0.1375 65.9 Change Materials: Applications to Cooling for Electronics and Energy Storage
Devices,” Ph.D. thesis, University of Dayton, Dayton, OH.

021004-8 / Vol. 130, JUNE 2008 Transactions of the ASME

Downloaded
View publication stats 05 Jun 2012 to 131.238.91.168. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm