Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Thermal Science and Engineering Progress

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

New Journal and we have not received input yet 20 (2020) 100742

Contents lists available at ScienceDirect

Thermal Science and Engineering Progress


journal homepage: www.sciencedirect.com/journal/thermal-science-and-engineering-progress

Heat transfer enhancement of conventional aluminum heat sinks with an


innovative, cost-effective, and simple chemical roughening method
Mohammad Reza Attar a, Majid Mohammadi b, Amin Taheri b, Saman Hosseinpour c,
Mohammad Passandideh-Fard b, Mohsen Haddad Sabzevar a, Ali Davoodi a, *, 1
a
Department of Materials and Metallurgical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
b
Department of Mechanical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
c
Institute of Particle Technology (LFG), Friedrich-Alexander-Universität-Erlangen-Nürnberg (FAU), Cauerstraße 4, 91058 Erlangen, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: Increased temperature of electronic devices during their performance results in the decline of their performance
Cooling efficiency and reduces their lifetime due to the increased thermal damage, a problem which is very substantial in minia­
Chemical roughening turized electronic components. Here, we present an innovative, cost-efficient, simple, and environmentally
Heat sink
benign chemical method for creating micro/nano roughened structure on the conventional aluminum heat sinks
Thermal management
to increase their cooling capability. The surface structures of the roughened heat sinks were examined using
Thermal resistance
scanning electron microscopy and atomic force microscopy. The performance of conventional and micro/nano
roughened aluminum heat sinks was analyzed for thermal management of a printed circuit board (PCB) pro­
ducing heat fluxes equivalent to those of a Core i7-1068NG7 processor. At forced air convection condition and
under heat fluxes of 8000, 10,000 and 12,000 W/m2, the temperature of PCBs cooled with roughened heat sinks
were found to be lower than those cooled with the conventional heat sinks by impressive values of 9.2 ◦ C,
11.0 ◦ C, and 12.8 ◦ C, respectively. The thermal resistance of the novel heat sink decreased by 38.01%, 39.14%,
and 39.96%, respectively, compared to that of the conventional heat sink. In free convection, employing the
roughened heat sink improved the PCB cooling effectiveness by roughly 14% at heat fluxes of 8000 W/m2 at
cutoff temperature. The proposed simple method in this study opens up the possibility of reducing the size (or
mass) of aluminum heat sinks or employing a heat sink of a certain size for the thermal management of a system
with higher heat flux.

1. Introduction continuing miniaturization (i.e size and weight reduction) and an in­
crease in output power density of electronic devices have led to
According to the atlas of economic complexity [1], the current gross increased demand for more efficient thermal management [8,9]. Proper
export of the USA in the area of electronics alone exceeds 170 billion US thermal management of electronics also affects their lifetime and per­
$. Apart from the direct investment costs of the electronic devices, their formance which are effectively reliant on the operating temperature of
power consumption for operation and thermal management adds to the electronic devices [10,11]. Based on the above details, thermal man­
total costs of the electronics. Therefore, it is not a big surprise that the agement of electronic components is of priority concern [12].
thermal management of electronics, for maintaining them in ideal Several innovative mechanical designs and new fabrication strate­
operating conditions, with a market share of 11 billion US$ has attracted gies, nowadays, particularly by 3D printing are used to improve the
huge attention [2,3]. Energy saving of cooling systems not only reduces cooling performance of electronic devices, but less attention has been
the costs but also contributes to the reduction of carbon emission from made on engineering intrinsic surface properties of cooling devices
electronic devices in the environment [4]. For instance, 1.2 TWh energy using industrially scalable technologies [13–15]. With respect to the
consumption by personal computer (PC) units can be significantly materials and methods which are utilized for the thermal management
reduced by using more efficient cooling systems [5–7]. Moreover, of electronic devices, cooling systems can be categorized into different

* Corresponding author.
E-mail addresses: a.davodi@um.ac.ir, adavoodi@seas.harvard.edu (A. Davoodi).
1
John A. Paulson School of Engineering and Applied Sciences, Harvard University, 29 Oxford Street, Cambridge, Massachusetts 02138, USA.

https://doi.org/10.1016/j.tsep.2020.100742
Received 31 May 2020; Received in revised form 23 September 2020; Accepted 28 September 2020
Available online 5 October 2020
2451-9049/© 2020 Elsevier Ltd. All rights reserved.
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

types such as heat pipes [16,17], liquid-based cooling systems techniques to fabricate microscale structure on surface for heat transfer
[11,18,19], spray cooling [20,21], phase changing materials [22,23], improvement. Based on their experimental results, at a subjected heat
thermoelectric cooling [24], and air-based cooling [25–27]. These flux of 500 W/cm2, in comparison to the conventional smooth surface,
cooling techniques are broadly classified into passive and active [26]. the modified microscale surface exhibited improved heat transfer co­
The passive cooling techniques operate without any external source for efficients up to 76%. The performance of micro-, nano-, and micro/
heat removal and are, therefore, rather inefficient for the thermal nano-structured surfaces on heat transfer characteristics of copper sur­
management of electrical components with a high output heat flux. In faces was studied by Sadaghiani et al. [35]. In their study, the me­
contrast, the active cooling techniques, which benefit from an external chanical sanding method was utilized for surface roughening. In
thermoelectric cooling compartment, can be easily adopted for systems comparison to the smooth reference sample, the authors found the
with high heat flux. Nevertheless, often the thermoelectric coolers suffer maximum heat transfer improvement of about 16%, 25%, and 36%, at a
from relatively low efficiency and high capital cost. When liquid-based mass flux of 50 kg/m2 and heat flux of 17 W/cm2 for micro-, nano-, and
cooling modules are used as the external cooling compartment, any micro/nano-structured surfaces, respectively.
leakage in the module may lead to an irreversible damage to chips and It is thus clear that the surface roughening results in an improved
electronic devices. In addition, liquid coolants, either aqueous or heat transfer efficiency in heat sinks, which in turn, leads to better
nonaqueous, usually are not utilized as they adds more weight to the thermal management of electronic devices. Nonetheless, surface struc­
system [28] and often exhibit freezing problems in low-temperature turing and fabrication of roughened surfaces using additive
conditions [29]. Therefore, cooling by air-forced circulation deems a manufacturing are tedious, expensive, and require specialized equip­
viable option for active cooling systems. ment. Chemical etching is another versatile method for roughening the
Although air has low thermal properties (i.e. heat capacity and surface of metals. Nevertheless, despite the application of chemical
thermal conductivity), it should not be overlooked as a viable coolant, etching as a way for the fabrication of roughened surfaces, such strate­
because of its accessibility, ease of control for pressure drop, and low gies are often implemented for changing the physicochemical properties
leakage threats [25]. In order to overcome the low thermal properties of of the surface such as its hydrophobicity, water or ice repellency, as well
air and for enhancement of heat transfer efficiency of the air-cooled as the paint adhesion to a limited extent. Thus, in this study, we propose
systems several methods have been offered, including the application a simple and innovative chemical etching method for fabrication of a
of plain and pin fins, metal foams, porous media, and dimpled and micro/nano roughened structures on a conventional aluminum heat sink
roughened surfaces [30]. The main disadvantage of porous media and for the purpose of enhancing its heat transfer efficiency. To the best of
metal foams originates from their low mechanical resistance. Addi­ our knowledge, such chemically induced surface roughening method
tionally, creating pin, fins, and dimples on porous surfaces requires has not yet been utilized for improving the heat transfer efficiency of a
utilization of costly equipment including but not limited to the computer surface. Furthermore, in most of the studies in which chemical etching
numerical control (CNC) machining. The mentioned methods are often has been implemented for surface roughening (for other purposes), often
not appropriate for scale up in large-scale industrial production. a combination of multiple etchants in a relatively complex protocol is
Therefore, developing simple and low-cost strategies for improving the used which involves the application of sonication bath and multiple
thermal properties of the air-cooled systems is a necessity. washing steps [42–44]. In contrast, in this study, only one type of
It is proven that surface roughening profoundly enhances the heat chemical (i.e. HCl) is used for surface roughening without further
transfer rate in various applications including in solar air heaters application of ultrasonic treatment (ethanol is also used to clean the
[31,32], boilers [33,34], electronic devices [35], and heat exchangers surface from oil and grease, a step which can be omitted for as received
[36]. The improved heat dissipation in roughened heat sinks has been heat sinks). This simplification allows for upscaling the process and
experimentally and numerically studied before. For example, Ceylan enables the utilization of our proposed methodology in the industrial
et al. [37] proposed a numerical correlation between the surface relative applications. In many studies of this kind, extremely hard to handle
roughness and the convective heat transfer coefficient in the pipelines, chemicals (such as HF) are used for surface structuring. Apart from being
based on the limited number of available experimental data. Chakroun difficult to implement, these methods also leave behind negative envi­
et al. [38] found that the surface roughness has a strong effect on the ronmental impacts. Also, extra steps such as neutralization of the waste
flow characteristics of an air jet impinged on a rough surface. They are required for the safe collection of the solutions after application. In
showed that the surface roughness affects the mean velocity as well as our proposed method, however, less environmentally dangerous etchant
the turbulence intensity of the flow. Ventola et al. [30] experimentally (i.e. HCl) is used and the waste treatment is extremely simplified. This is
investigated, for the first time, the artificial roughness manufacturing by an important aspect simplifying the industrialization of the proposed
direct metal laser sintering (DMLS) to augment convective heat transfer method.
in flat and finned heat sinks. They observed that on rough flat and rough Since up to now, most of the experimental works in this field are
fin surfaces (Ra = 43 µm) the average convective heat transfer was performed using expensive and complicated surface roughening meth­
enhanced by 63% and 35% as compared to smooth surfaces (Ra≈1 µm), odologies, the substantial enhancement in heat transfer efficiency of the
respectively. Oguntala et al. [39] numerically assessed the effects of heat sinks using the simple/innovative proposed chemical method for
cylindrical micro-fins with artificial roughness on heat transfer the particular application of cooling electronic devices is rather unique
improvement for the thermal management of microprocessors. Their in the field of thermal design and management.
results revealed that the surface roughness as well as the geometric ratio In the following, the procedure for the preparation and character­
of the micro fins enhance the thermal performance and cooling effi­ ization of roughened surfaces is discussed. The thermal properties of
ciency of the device compared to smooth fins. The effect of micro- roughened heat sinks are then compared with those of the conventional
roughness shapes of concentric, cubic, and cylindrical produced by ad­ systems under different conditions. The proposed method in this work
ditive manufacturing on heat transfer and fin effectiveness was later for surface roughening utilizes nontoxic chemical and can be easily used
assessed both numerically and experimentally by Singh et al. [40]. They in manufacturing scale-up. Additionally, currently available conven­
found that the concentric shaped roughness elements have a higher heat tional aluminum heat sinks at any size and geometry can be treated
transfer improvement and fin effectiveness as compared to their cylin­ using this method, which makes it industrially attractive.
drical and cubic roughened shapes counterparts. Overall, all these
roughened structures, which were made via additive manufacturing,
exhibited substantial improved heat transfer efficiency compared to a
smooth target in jet impingement cooling systems.
Bostanci et al. [41] utilized plasma spraying and particle blasting

2
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

2. Experimental details spectrometer (OES, Foundry-Master Smart) as: Si 0.40%, Fe 0.32%, Cu


0.04%, Mn 0.02%, Mg 0.57%, Zn 0.05%, and Al balance (in wt%),
2.1. Roughening process matching AA6063 (ASM standard).

The simple and facile chemical etching treatment depicted in Fig. 1 is


utilized in accordance with our earlier publication [45] to develop 2.3. Experimental setup
micro/nano-scale roughness on the surface of conventional heat sinks.
This method is utilized as a viable alternative to the more expensive and An experimental setup including two main sub-systems and mea­
demanding bottom-up fabrication approaches such as DMLS. As indi­ surement equipment is developed to scrutinize and compare the heat
cated in Fig. 1, the conventional heat sink is initially washed in alcohol/ transfer performance of the conventional and roughened heat sinks for
distilled water mixture and dried at room temperature to remove the the cooling of a PCB as the heat generating electronic component. In
excess water/ethanol from the surface. In the etching step, the Fig. 2a different units of the setup including, a conventional and two
aluminum heat sink is immersed in an aqueous solution of hydrochloric roughened aluminum heat sinks, DC CPU fans, K-type thermocouples,
acid (37 wt%) (1:1 vol ratio) at room temperature for 4 min, inducing PCBs operating under similar conditions, and insulation layer can be
the surfaces roughness. The etching time was optimized to ensure the observed. The role and details of each component in the experimental
formation of micro/nano-roughness on the surface of the conventional setup schematically shown in Fig. 2b are described as follows:
sink, as will be discussed later. The sample is then washed and dried for
5 min in a flow of dry air. All steps of the presented method take just • A DC power source (30 V-5 A ATTEN) is connected to the PCB, which
about 20 min. As a result, the rapid process of application of non-toxic regulates the open-circuit voltage and short circuit current to provide
chemical allows for upscaling this method for industrial use. a desired heat flux between 8000 to 12,000 W/m2.
• The conventional (or micro/nano roughened) aluminum heat sink is
2.2. Surface characterization used for the thermal management of the PCBs. The total number of
fins in each heat sink is 17 and the dimensions of each fin are: length
Examination of the surface structure before and after the roughening = 3.0 cm, width = 9.0 cm, thickness = 0.1 cm. The distance between
process is performed using scanning electron microscopy (SEM, Zeiss, the two fins = 0.3 cm with total apparent interior surface of ~886.8
LEO 1450VP, Voltage: 20 KV) equipped with the secondary electrons cm2. The mass of the conventional heat sink is 210 gr. After the
detector (SE-SEM). For surface topography observation and quantifica­ roughening process, due to the aluminum dissolution and metal
tion of surface roughness, the atomic force microscopy (AFM, Ara release, the mass of the heat sink is reduced to 195 gr. The rough­
Research) is used (Tip: Silicon, Al backside coating, Rtip = 8 nm). The ening process and surface heat sink characterization are discussed in
specific surface area of the conventional, as well as the micro/nano the previous sections, vide supra.
roughened heat sinks, is determined using Brunauer-Emmett-Teller • The K-type thermocouples for measuring temperatures are placed
(BET) surface analysis method (Quantachrome NovaWin). A high- between the PCB and copper plate, at the center of cooling units. In
speed camera (960 frames per second) is used to determine the wetta­ order to accurately determine the transient temperature-time
bility of the heat sink surface after roughening. The chemical composi­ response of the PCB at constant input heat fluxes, thermocouples
tion of the conventional heat sink is determined using optical emission are fixed 1.5 mm deep in the slot. For practical purposes, a cutoff

Conventional heat sink Distilled water Alcohol

Washing process Drying process


3 min min, 25°C
Etching
process
4 min
Distilled water

Hydrochloric
acid/water

Drying process Washing process


min, 25°C 3 min

Micro/nano-roughened heat sink

Fig. 1. Details of the roughening process employed on heat sink.

3
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

(a)

Surface SEM image Surface SEM image


(b)

9 cm 9 cm

3 cm 3 cm

Thermal paste Thermal paste


Copper plate Copper plate

PCB PCB

Conventional Micro/Nano roughened


heat sink heat sink

Fig. 2. a) A photograph of the experimental setup including the heat sinks and measurement equipment. Two heat sinks, conventional and roughened, with the same
apparent surface area (i.e 886.80 cm2) as well as one roughened heat sink with reduced surface area (i.e. 554.25 cm2) are depicted in this picture. b) Schematic
representation of different elements in the experimental setup.

temperature of 75 ◦ C is considered, to stop the measurements before Czech Republic) to experimentally determine the air velocity close to
the device failure. the heat sinks. The Reynolds number (Re) based on the velocity of the air
• A digital data-logger (Lutron TM-947SD) is utilized to record the and the geometry of the heat exchange unit was estimated to be 722,
PCBs and ambient temperatures simultaneously. The recorded tem­ which guarantees a laminar air flow close to the fins.
peratures are saved on an SD card every 10 s.
• Three similar DC CPU fans (PBT-GF30-FR, 12 V, and 0.12 A) are used
over the heat sink to provide forced air convection. The free con­ 2.4. Experimental uncertainty investigation
vection is assessed by switching off the DC CPU fan.
• To ensure a uniform heat flux distribution on the PCB, a thin copper A systematic uncertainty analysis is used to ensure the reliability of
plate (4.4 cm × 6 cm × 0.1 cm) is placed between the cooling unit the experiments. The uncertainty is assessed based on the accuracy of
and the PCB. the measuring equipment and repeatability of the experiments, in
• Two layers of thermal paste (high voltage silicon rubber, Hong-Fa accordance with the procedure previously described in literature [46].
Technology Co. Model Number: 7099) are used and spread uni­ In Eq. (1), for R as a function of n independent linear parameters (R =
formly to fill the air gaps between the PCB, heat sink, and copper R(ν1 , ν2 , ⋯, νn )), the uncertainty is given as:
plate. The thermal paste reduces the thermal resistance and increases √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ( )2 ( )2̅
the heat transfer rate between the PCB, copper plate, and heat sink ∂R ∂R ∂R
δR = δν1 + δν2 + ⋯ + δνn (1)
(Fig. 2b). Using thermal paste can also improve the contact of two ∂ν1 ∂ν2 ∂νn
surfaces. In this study, the thickness of each layer of thermal paste is
Herein, ∂Rand δνi are the uncertainty of function R and the uncer­
1.5 mm. Based on calculations, this thickness causes a maximum
tainty of parameter Vi, respectively. In addition, ∂ν
∂R
is the partial deriv­
thermal resistance of 0.0002 K/W; that is 0.01 ◦ C difference between i

each two sides. ative of R with respect to νi . The total uncertainty (Eq. (2)) can be
• An elastomeric insulation layer (K-FLEX®) with a thickness of 2 cm calculated using uncertainty of repetition (Eq. (3)) and uncertainty of
and a thermal conductivity of 0.0034 W.m− 1.K− 1 is used at the the equipment
bottom of the PCBs. Thus, the heat transfer only takes place between (Eq. (4)):
the PCBs and aluminum heat sinks. √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∂v = (∂vrep )2 + (∂veqp )2 (2)
We utilized a hand held Mini Anemometer (UT363, from UNI-T,

4
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

S performance in the presence of an external fan (i.e. forced air convec­


∂vrep = √̅̅̅ (3)
n tion) and in the absence of the external fan (i.e. free convection). For a
conventional heat sink, as shown in Fig. 5, when no external fan is used,
a the PCB temperature rises monotonically until the device temperature
∂veqp = √̅̅̅ (4)
3 reaches the cutoff temperature (after almost 370 sec. at ~75 ◦ C).
Herein, S is standard deviation [47],n is the number of the repeated However, when a fan is used to force the air convection, the system
measured data, and a is half of the measuring equipment accuracy. The reaches a steady-state temperature of 49.8 ◦ C within the first 300 sec.
uncertainties related to the measuring equipment are provided in and the device maintains its function throughout the whole measure­
Table 1. For all considered cases in this study and based upon Table 1, ment. For a roughened heat sink, the steady-state temperature of 40.6 ◦ C
the calculated uncertainty is less than 3% which shows a high accuracy is reached also after 300 sec. when an external fan is used. In the case of
of experimental data. free convection, the cutoff temperature is reached after 520 sec. The
reduction of steady-state temperature or prolongation of the time to
3. Results and discussion reach the cutoff temperature in free convection after roughening the
heat sink surface will be discussed in detail in the following (vide infra).
3.1. Surface properties
3.2.1. Forced air convection
Fig. 3 displays the SEM images of as-received conventional and
roughened aluminum heat sink surfaces with different magnifications. 3.2.1.1. Effect of surface roughness on cooling performance of aluminum
The surface of the conventional heat sink is rinsed prior to imaging to heat sinks under different heat fluxes. We first consider the forced air
remove the possible contaminations and dust from the surface. The convection cases and compare the steady-state temperatures of systems
surface of the conventional heat sink (Fig. 3a and b) is relatively smooth when the conventional and roughened heat sinks operate under
and some local pits are observed at its grain boundaries. As clearly different heat fluxes. Fig. 6a-c depict the transient time-temperature
shown in Fig. 3c and d, the relatively smooth surface of the conventional response of conventional and roughened heat sinks under heat flux of
aluminum heat sink is decorated extensively with micro/nano scale 8000, 10,000, and 12,000 W/m2, respectively. As is observed in these
features after the roughing process. The AFM results in Fig. 3e clearly curves, in all cases roughening the heat sink surfaces results in a sub­
exhibit the three dimensional (3D) structure of the surface features stantial decrease in the temperature of the system, from the beginning to
(~200 nm size and only few nm in height) that are formed, thanks to the end. Similar to the trend that was discussed earlier for the heat flux
selective dissolution of aluminum at certain crystallographic directions of 8000 W/m2, the steady-state temperatures in a conventional heat sink
at the dislocation defects that are common in crystalline aluminum at the heat flux of 10,000 and 12,000 W/m2 reach 53.7 ◦ C and 58.6 ◦ C,
[48,49]. The conventional aluminum substrate presented an average respectively, whereas the corresponding steady-state temperatures in
roughness of 41.1 nm (Ra = 41.1 nm) whereas the roughened aluminum the roughened heat sink are 42.7 ◦ C and 45.8 ◦ C, respectively. By
substrate produced during etching process showed a notable increase of creating roughness on the surface of the heat sink, the effective heat
the average roughness (Ra = 169.4 nm) due to the formation of disso­ transfer surface increases and according to Newton’s law of cooling, the
lution patterns. Similar dissolution patterns have also been observed on heat output will be higher. In addition, the main mechanism of the
monocrystalline aluminum surfaces [50]. In Fig. 4, the contact angle reduction of the system’s temperature with forced air convection when
evolution of 10 μl distilled water droplet after impact on the roughened the surface is roughened is the formation of additional air vortices near
aluminum surface is presented. As shown in this figure, the surface of the the nano and micron size features in the roughed sink surface. As can be
roughened aluminum sample is superhydrophilic (i.e. water droplet seen, the absolute effect of surface roughness in reducing the PCB
contact angle ~0◦ ). The water droplet wets the surface completely temperature rise (as compared with the ambient temperature, ΔT1 −
within 20 ms after the impact and probably penetrates the micro/nano ΔT2) is dependent on the system’s heat flux and increases when higher
scale features of the roughened surface. As comparison, the contact heat fluxes are employed. Nevertheless, the ratio of ΔTs (i.e. ΔT1/ ΔT2) in
angle of the water droplet on a conventional aluminum heat sink was all cases remains nearly constant (~1.6), which will be discussed further
~87◦ in details. The cooling effectiveness (ε) of the roughened heat sink can be
defined as:

3.2. Thermal assessments ΔT1 − ΔT2


ε= (5)
ΔT1
To investigate the effectiveness of the presented surface roughening The calculated cooling effectiveness is therefore 38.01%, 39.14%,
method in improving the cooling capacities of conventional heat sinks, and 36.96% for the heat fluxes of 8000 W/m2, 10,000 W/m2, and
the thermal properties of conventional and roughened heat sinks are 12,000 W/m2, respectively. This means that replacing a conventional
compared for the thermal management of PCB at heat fluxes equivalent heat sink with the roughened heat sink of the same size results in more
to those of a Core i7-1068NG7 processor. The initial experiments with a than 35% enhancement of cooling effectiveness under forced air
heat flux of 8000 W/m2 are carried out to compare the heat sinks

Table 1
Accuracy and uncertainty of equipment.
Equipment Measurement section Equipment Uncertainty of Standard Accuracy Total
uncertainty repetition deviation uncertainty

Thermocouple PCB temperature (conventional heat 0.144 ◦ C 0.057 ◦ C 0.099 ±0.5 ◦ C 0.154 ◦ C
sink)
Thermocouple PCB temperature (roughened heat 0.144 ◦ C 0.076 ◦ C 0.132 ±0.5 ◦ C 0.163 ◦ C
sink)
Thermocouple Ambient temperature 0.144 ◦ C 0.025 ◦ C 0.121 ±0.5 ◦ C 0.146 ◦ C
Digital multimeter (VC9805, Power source voltage 0.144 V – – ±0.5 V 0.144 V
China)
Digital multimeter (VC9805, Power source current 0.231 A – – ±0.8 A 0.231 A
China)

5
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

(a) (b)

(c) (d)

(e)
24.24 nm
24.24 nm

1 μm

0 nm

Fig. 3. SEM images of (a-b) conventional and (c-d) micro/nano roughened aluminum heat sinks. e) AFM image of the micro/nano roughened aluminum heat
sink surface.

0 ms 4 ms 8 ms 12 ms 16 ms 20 ms

Fig. 4. Consecutive contact angle of 10 μl water droplet on the roughened heat sink surface.

convection, irrespective of the system’s heat flux. As a result, theoreti­ 3.2.1.2. Roughened surfaces for more efficient heat sinks. To quantify the
cally, systems generating higher heat fluxes may be thermally managed improved performance of a conventional heat sink after surface rough­
when roughened heat sinks of the same sizes are employed under forced ening treatment, a simple energy balance can be executed as [51]:
air convection condition.
Q∝hA(TPCB − T∞ ) (6)

6
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

(hA)2 ΔT1
= (9)
(hA)1 ΔT2

The measured values for temperature differences at a constant heat


flux of 10,000 W/m2 for the conventional and roughened heat sinks are
28.1 ◦ C and 17.7 ◦ C, respectively. Therefore, the ratio of total heat
transfer coefficient (hA)
(hA)
2
for the two heat sinks is ≃ 1.6. This means that
1

using a micro/nano roughened heat sink instead of a conventional heat


sink results in almost 60% enhancement in total heat transfer coefficient
(hA) under forced air convection condition.
To experimentally verify the possibility of replacing a conventional
heat sink with a similar size roughened heat sink for the thermal man­
agement of systems with higher heat fluxes, we compared the perfor­
mance of a conventional heat sink at a heat flux of 12,000 W/m2 with
that of a roughened heat sink at 1.6 times higher heat flux (i.e. 19,200
W/m2). Alternatively, the performance of a roughened heat sink at a
heat flux of 12,000 W/m2 is compared with that of a conventional heat
sink at nearly 1.6 times lower heat flux (i.e. 7500 W/m2). The transient
Fig. 5. Evolution of PCB temperature in free and forced air convection at the
time-temperature responses of these two systems are presented in Fig. 7a
heat flux of 8000 W/m2 when conventional or roughened heat sinks are used
and b, respectively. As seen in these curves, the PCB temperature for the
for thermal management.
two cases is very similar. Therefore, at a constant PCB temperature, the
60% enhancement of the heat transfer by roughening the surface of the
herein, Q is the rate of generated heat of the PCB (i.e. the heat flux times heat sink (previous section) is further verified.
the PCB area), h the convection heat transfer coefficient, A the heat sink The inverse of the total heat transfer coefficient can be regarded as
total surface area, TPCB the PCB temperature, and T∞ the ambient tem­ the average thermal resistance of the system and is considered as an
perature. For the conventional heat sink [52]: important parameter in assessing the performance of cooling units [53].
Q The average thermal resistance (Rth ) of cooling modules can be written
(hA)1 ∝ (7) as [12]:
ΔT1
TPCB − T∞ 1
and for the roughened heat sink: Rth = = (10)
Q hA
Q In Fig. 8, the average thermal resistance for the conventional and
(hA)2 ∝ (8)
ΔT2 roughened heat sinks under three heat fluxes are provided. Accordingly,
To obtain the ratio of total heat transfer coefficient (hA) for the two the average thermal resistance is reduced by 38.01%, 39.14%, and
heat sinks, considering the fact that the heat flux and area of the PCB are 36.96% for heat fluxes of 8000, 10,000, and 12,000 W/m2, respectively,
the same for both heat sinks (i.e. Q is the same), we can have [52]: when the roughened heat sink is utilized instead of the conventional
heat sink.

Fig. 6. The ambient temperature and transient time-temperature responses of conventional and micro/nano roughened heat sinks for forced air convection under a
heat flux of a) 8000, b) 10,000, and c) 12,000 W/m2. The ratio ΔT1/ ΔT2 is also plotted on the right axis where ΔT1 and ΔT2 are the differences between the PCB and
the ambient temperatures for the conventional and roughened heat sinks, respectively.

7
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

Fig. 7. The time response of the measured PCB temperature for conventional heat sink and its roughened counterpart at a) 1.6 times higher flux for roughened sink
and b) 1.6 times lower flux for the conventional heat sink.

Fig. 8. The thermal resistance of conventional and micro/nano roughened heat Fig. 9. The time response of the measured PCB temperature for conventional
sinks at different heat fluxes. heat sink and its roughened counterpart with 37% smaller surface area for a
system with 12,000 W/m2 heat flux.
3.2.1.3. Roughened surfaces for reducing the size of heat sinks. From
these results it can be inferred that for maintaining a PCB with a certain 3.2.1.4. Roughened surfaces for lighter heat sinks. The heat sink mass
heat flux at a target temperature value, a conventional heat sink can be before and after the etching process was 210 gr and 195 gr, respectively,
replaced by a smaller heat sink with roughened surface. This is an reflecting 7.1% reduction of the total mass. Assuming that the redepo­
important feature from the economical point of view and can aid with sition of dissolved metal during the chemical etching process is negli­
manufacturing smaller electronic components. This size reduction pos­ gible, this reduction of the mass of heat sinks can be attributed solely to
sibility is successfully demonstrated by comparing the PCB temperature the dissolution of metal, especially at the dislocations and defects (see
that is cooled by a conventional heat sink with that cooled by a rough­ Section 3.1). This is another advantage of the present method for
ened heat sink with an almost 1.6 times smaller (37% reduction) surface improvement of the cooling efficiency of heat sinks, as compared to
area, as shown in Fig. 9. In other words, the effective heat exchange other methods such as DMLS which often result in an increased mass of
surface ration between the roughened and conventional heat sinks (A*/ the sinks after modification.
A) is 63%.
The BET surface analysis, however, showed that the specific surface 3.2.1.5. The heat sinks performance under an intermittent load. Thermal
area of the conventional heat sink (0.028 m2/g) increases after rough­ hysteresis and thermal fatigue are important characteristics of cooling
ening process (0.119 m2/g) by a factor of 4.25. From the BET surface modules. To further investigate the performance of conventional and
analysis, it becomes obvious that not all the micro/nano rough features roughened heat sinks under intermittent loads (i.e. on/off switching),
that are formed on the heat sink after roughening process contribute to four cycles of heating/cooling (30 sec./15 sec.) are selected. The
the heat transfer enhancement. In fact, it is expected that the very small behavior of transient PCB temperature at a heat flux of 10,000 W/m2 for
air pocket circulation and additional air vortices near the nano and the two systems is shown in Fig. 10. As is depicted in this figure, the
micron size features are redundant in the improvement of the heat maximum and minimum temperatures (peaks and valleys) for conven­
transfer of the sinks. Nevertheless, as will be discussed later, the for­ tional heat sink are 48.7 ◦ C and 28.6 ◦ C, whereas for roughened heat
mation of these deep and small features results in the reduction of the sink these values are nearly 41 ◦ C and 27.7 ◦ C, respectively. In neither
overall mass of the roughened sinks. case, the temperature–time hysteresis is detected. Consequently, using a

8
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

Overall, the innovative approach that is presented in this work en­


ables industrial production of more efficient, smaller, and lighter heat
sinks through a very cheap, simple, and environmentally procedure and
would open up the possibility of reducing the overall costs of the thermal
management of the electronic devices in large scale and contributes to
reducing the carbon emission to the environment from electronics.

Declaration of Competing Interest

The authors declare that they have no known competing financial


interests or personal relationships that could have appeared to influence
the work reported in this paper.

Acknowledgements

MRA, MM, AT, MPF, MHS, and AD acknowledge FUM for financial
supporting. SH thanks Prof W Peukert, BMBF-MSRT (CAlSAB,
01DK20097), and Emerging Talents Initiative (ETI) 2018/2_Tech_06,
Fig. 10. The transient time-temperature response of conventional and rough­ FAU (Grant No. 2018/2), Germany, grants for supporting his research.
ened heat sinks under intermittent loads for a system with the heat flux of Sandra Wittpahl (FAU) is acknowledged for the BET surface area
10,000 W/m2. The on and off cycles were 30 s and 15 s, respectively. measurement.

heat sink with a roughened surface can decrease the PCB temperature Authors contribution
under intermittent loads without influencing the service performance of
the cooling device during the on/off cycles. MRA conceptualization, data collection, experimental design, partial
writing. MM and AT data collections, experimental design, partial
3.2.2. Free convection writing. SH data collection, partial writing, review and editing. MPF,
When the conventional and roughened heat sinks are compared MHS, and AD funding acquisition, leading the project, review and
under free convection (i.e. without external fan), the beneficial effect of editing.
surface roughening is less pronounced compared to the forced air con­
vection scenario (discussed in Section 3.2.1). This phenomenon may be References
attributed to a very small air velocity near the heat sink surfaces in free
convection. According to Eq. (5), the cooling effectiveness ratio upon [1] A. Alam, K.-Y.J.S. Kim, A.B. Chemical, Mixing performance of a planar micromixer
with circular chambers and crossing constriction channels, 176 (2013) 639-652.
roughening treatment of the heat sink surface in free convection case is [2] C.J. Ho, J.-C. Liao, C.-H. Li, W.-M. Yan, M. Amani, Experimental study of cooling
0.14 at cutoff temperature. In another word, even in the absence of an performance of water-based alumina nanofluid in a minichannel heat sink with
external fan, the cooling effectiveness of a roughened heat sink is almost MEPCM layer embedded in its ceiling, Int. Commun. Heat Mass Transfer 103
(2019) 1–6.
14% improved compared to its conventional counterpart. As mentioned
[3] X. Chen, T. Li, H. Zeng, Z. Hu, B. Fu, Numerical and experimental investigation on
earlier, the micro/nano features on the surface of roughened sinks, micromixers with serpentine microchannels, Int. J. Heat Mass Transfer 98 (2016)
which in turn increase the effective surface area of the sinks, are 131–140.
responsible for this improved cooling efficiency. [4] M. Masataka, T. Nguyen, M. Koichi, S. Yuji, W. Xiao Ping, T. Nguyen, W. Vijit,
Thermal management in high performance computers by use of heat Pipes and
vapor chambers, and the challenges of global warming and environment, in: 2009
4. Conclusions 4th International Microsystems Packaging, Assembly and Circuits Technology
Conference, 2009, pp. 191–194.
[5] N. Mills, E. Mills, Taming the energy use of gaming computers, Energ. Effi. 9 (2)
In this study, to enhance the cooling capability of conventional (2016) 321–338.
aluminum heat sinks, an economical, innovative, and environmentally [6] J.L. Reyna, M.V. Chester, Energy efficiency to reduce residential electricity and
friendly method was proposed by creating micro/nano-roughness on the natural gas use under climate change, Nat. Commun. 8 (2017) 14916.
[7] E. Gelenbe, Y. Caseau, The impact of information technology on energy
heat sink surface with the aid of a chemical etching process. Two consumption and carbon emissions, Ubiquity 2015 (2015) 1–15.
different cooling modules, conventional and micro/nano roughened [8] M.L. Elsayed, O. Mesalhy, Studying the performance of solid/perforated pin-fin
heat sinks, were tested for the thermal management of a printed circuit heat sinks using entropy generation minimization, Heat Mass Transfer 51 (2015)
691–702.
board (PCB) at heat fluxes of 8000–12,000 W/m2. The performance of [9] M.L. Elsayed, O. Mesalhy, R.H. Mohammed, J.P. Kizito, Q.H. Leland, L.C. Chow,
conventional and micro/nano heat sinks was compared under forced air Enhancement of a heat sink performance using a partial shield and/or a guide
convection and free convection conditions and the obtained results can plate, Int. J. Heat Mass Transfer 134 (2019) 948–958.
[10] M. Saeed, M.-H. Kim, Heat transfer enhancement using nanofluids (Al2O3-H2O) in
be summarized as:
mini-channel heatsinks, Int. J. Heat Mass Transfer 120 (2018) 671–682.
[11] A. Taheri, M. Ghasemian Moghadam, M. Mohammadi, M. Passandideh-Fard,
• The cooling effectiveness of the conventional heat sink under forced M. Sardarabadi, A new design of liquid-cooled heat sink by altering the heat sink
air convection was improved by 38.01%, 39.14%, and 36.96% for heat pipe application: experimental approach and prediction via artificial neural
network, Energy Convers. Manage. 206 (2020), 112485.
the heat fluxes of 8000 W/m2, 10,000 W/m2, and 12,000 W/m2, [12] M.R. Shaalan, M.A. Saleh, O. Mesalhy, M.L. Elsayed, Thermo/fluid performance of
respectively after surface roughening. a shielded heat sink, Int. J. Therm. Sci. 60 (2012) 171–181.
• At free convection, the improved cooling efficiency of the roughened [13] A.L. Moore, L. Shi, Emerging challenges and materials for thermal management of
electronics, Mater. Today 17 (2014) 163–174.
heat sink was 14% at heat fluxes of 8000 W/m2. [14] Thermal Management of Telecommunications Equipment, ASME Press, 2013.
• Replacing a conventional heat sink with its micro/nano roughened [15] L.T. Yeh, R.C. Chu, Thermal Management of Microelectronic Equipment, ASME
counterpart allows for the reduction of the size of the heat sink by Press, 2002.
[16] G. Zhou, J. Li, Z. Jia, Power-saving exploration for high-end ultra-slim laptop
37% for the thermal management of PCBs of similar heat fluxes. computers with miniature loop heat pipe cooling module, Appl. Energy 239 (2019)
• The roughening process results in the mass reduction of the heat sink 859–875.
by 7.1%, which allows for production of lighter electronic devices. [17] G. Xiahou, J. Zhang, R. Ma, Y. Liu, Novel heat pipe radiator for vertical CPU
cooling and its experimental study, Int. J. Heat Mass Transfer 130 (2019) 912–922.

9
M.R. Attar et al. Thermal Science and Engineering Progress 20 (2020) 100742

[18] S.S. Khaleduzzaman, I.M. Mahbubul, M.R. Sohel, R. Saidur, J. Selvaraj, T.A. Ward, surfaces for high heat flux electronics cooling systems, Appl. Therm. Eng. 127
M.E. Niza, Experimental analysis of energy and friction factor for titanium dioxide (2017) 484–498.
nanofluid in a water block heat sink, Int. J. Heat Mass Transfer 115 (2017) 77–85. [36] S.-S. Hsieh, C.-T. Liauh, A.C. Ku, Thermal analysis of the performances of helical-
[19] S.S. Khaleduzzaman, M.R. Sohel, R. Saidur, I.M. Mahbubul, I.M. Shahrul, B. type, roughened, double-pipe heat exchangers, Appl. Energy 26 (1987) 67–73.
A. Akash, J. Selvaraj, Energy and exergy analysis of alumina–water nanofluid for [37] K. Ceylan, G. Kelbaliyev, The roughness effects on friction and heat transfer in the
an electronic liquid cooling system, Int. Commun. Heat Mass Transfer 57 (2014) fully developed turbulent flow in pipes, Appl. Therm. Eng. 23 (2003) 557–570.
118–127. [38] W.M. Chakroun, A.A. Abdel-Rahman, S.F. Al-Fahed, Heat transfer augmentation
[20] J.-X. Wang, Y.-Z. Li, Y.-F. Mao, E.-H. Li, X. Ning, X.-Y. Ji, Comparative study of the for air jet impinged on a rough surface, Appl. Therm. Eng. 18 (1998) 1225–1241.
heating surface impact on porous-material-involved spray system for electronic [39] G. Oguntala, R. Abd-Alhameed, G. Sobamowo, H.-S. Abdullahi, Improved thermal
cooling – an experimental approach, Appl. Therm. Eng. 135 (2018) 537–548. management of computer microprocessors using cylindrical-coordinate micro-fin
[21] Q. Gao, Y. Liu, G. Wang, F. Deng, J. Zhu, An experimental investigation of heat sink with artificial surface roughness, Eng. Sci. Technol., Int. J. 21 (2018)
refrigerant emergency spray on cooling and oxygen suppression for overheating 736–744.
power battery, J. Power Sources 415 (2019) 33–43. [40] P. Singh, M. Zhang, S. Ahmed, K.R. Ramakrishnan, S. Ekkad, Effect of micro-
[22] D. Zou, X. Liu, R. He, S. Zhu, J. Bao, J. Guo, Z. Hu, B. Wang, Preparation of a novel roughness shapes on jet impingement heat transfer and fin-effectiveness, Int. J.
composite phase change material (PCM) and its locally enhanced heat transfer for Heat Mass Transfer 132 (2019) 80–95.
power battery module, Energy Convers. Manage. 180 (2019) 1196–1202. [41] H. Bostanci, V. Singh, J.P. Kizito, D.P. Rini, S. Seal, L.C. Chow, Microscale surface
[23] A. Farzanehnia, M. Khatibi, M. Sardarabadi, M. Passandideh-Fard, Experimental modifications for heat transfer enhancement, ACS Appl. Mater. Interfaces 5 (2013)
investigation of multiwall carbon nanotube/paraffin based heat sink for electronic 9572–9578.
device thermal management, Energy Convers. Manage. 179 (2019) 314–325. [42] X. Li, Q. Zhang, Z. Guo, T. Shi, J. Yu, M. Tang, X. Huang, Fabrication of
[24] D. Liu, F.-Y. Zhao, H.-X. Yang, G.-F. Tang, Thermoelectric mini cooler coupled with superhydrophobic surface with improved corrosion inhibition on 6061 aluminum
micro thermosiphon for CPU cooling system, Energy 83 (2015) 29–36. alloy substrate, Appl. Surf. Sci. 342 (2015) 76–83.
[25] S. Kumar, M. Sarkar, P.K. Singh, P.S. Lee, Study of thermal and hydraulic [43] Z. Huang, Y.F. Li, P. Jin, M.S. Hu, B. He, N.K. Gao, H. Zhang, H.Y. Jin, Fabrication
performance of air cooled minichannel heatsink with novel geometries, Int. of aluminum superhydrophobic surface with facile chemical etching method,
Commun. Heat Mass Transfer 103 (2019) 31–42. Mater. Sci. Forum 804 (2015) 103–106.
[26] Z. Khattak, H.M. Ali, Air cooled heat sink geometries subjected to forced flow: a [44] R. Wu, S. Liang, A. Pan, Z. Yuan, Y. Tang, X. Tan, D. Guan, Y. Yu, Fabrication of
critical review, Int. J. Heat Mass Transfer 130 (2019) 141–161. nano-structured super-hydrophobic film on aluminum by controllable immersing
[27] M.L. Elsayed, O. Mesalhy, Effect of a slotted shield on thermal and hydraulic method, Appl. Surf. Sci. 258 (2012) 5933–5937.
performance of a heat sink, J. Electronic Packaging 137 (2014). [45] M.R. Attar, E. Khajavian, S. Hosseinpour, A. Davoodi, Fabrication of micro–nano-
[28] O. Mesalhy, M.L. Elsayed, J.J. Corona Jr, A.A. Kwarteng, J.P. Kizito, Q.H. Leland, roughened surface with superhydrophobic character on an aluminium alloy surface
L.C. Chow, Study of a high-reliability dual-fan system for cooling aerospace by a facile chemical etching process, Bull. Mater. Sci. 43 (2019) 31.
electromechanical actuators, Thermal Sci. Eng. Prog. 18 (2020), 100540. [46] S.J.J.M.E. Kline, Describing uncertainty in single sample experiments, 75 (1953) 3-
[29] M.L. Elsayed, O. Mesalhy, J.P. Kizito, Q.H. Leland, L.C. Chow, Performance of a 8.
guided plate heat sink at high altitude, Int. J. Heat Mass Transfer 147 (2020), [47] R. Nasrin, N.A. Rahim, H. Fayaz, M. Hasanuzzaman, Water/MWCNT nanofluid
118926. based cooling system of PVT: experimental and numerical research, Renew. Energy
[30] L. Ventola, F. Robotti, M. Dialameh, F. Calignano, D. Manfredi, E. Chiavazzo, 121 (2018) 286–300.
P. Asinari, Rough surfaces with enhanced heat transfer for electronics cooling by [48] B. Qian, Z. Shen, Fabrication of superhydrophobic surfaces by dislocation-selective
direct metal laser sintering, Int. J. Heat Mass Transfer 75 (2014) 58–74. chemical etching on aluminum, copper, and zinc substrates, Langmuir 21 (2005)
[31] S.B. Bopche, M.S. Tandale, Experimental investigations on heat transfer and 9007–9009.
frictional characteristics of a turbulator roughened solar air heater duct, Int. J. [49] X. Xiao, W. Xie, Z. Ye, Preparation of corrosion-resisting superhydrophobic surface
Heat Mass Transfer 52 (2009) 2834–2848. on aluminium substrate, Surf. Eng. 35 (2019) 411–417.
[32] N.M. Phu, V. Tuyen, T.T. Ngo, Augmented heat transfer and friction investigations [50] Y.B. Xu, K.F. Ha, Z.G. Wang, X.H. Wang, J. Li, Dislocation distribution near the
in solar air heater artificially roughened with metal shavings, J. Mech. Sci. crack tip of I and II modes in bulk aluminum single crystal, 69 (1991) 203-206.
Technol. 33 (2019) 3521–3529. [51] A.H. Beitelmal, M.A. Saad, C.D. Patel, Effects of surface roughness on the average
[33] G. Liang, I. Mudawar, Review of pool boiling enhancement by surface heat transfer of an impinging air jet, Int. Commun. Heat Mass Transfer 27 (2000)
modification, Int. J. Heat Mass Transfer 128 (2019) 892–933. 1–12.
[34] D.S. Wen, B.X. Wang, Effects of surface wettability on nucleate pool boiling heat [52] B.J. Jones, J.P. McHale, S.V. Garimella, The influence of surface roughness on
transfer for surfactant solutions, Int. J. Heat Mass Transfer 45 (2002) 1739–1747. nucleate pool boiling heat transfer, Journal of Heat Transfer 131 (2009).
[35] A.K. Sadaghiani, N.S. Saadi, S.S. Parapari, T. Karabacak, M. Keskinoz, A. Koşar, [53] K.-T. Chiang, Optimization of the design parameters of Parallel-Plain Fin heat sink
Boiling heat transfer performance enhancement using micro and nano structured module cooling phenomenon based on the Taguchi method, Int. Commun. Heat
Mass Transfer 32 (2005) 1193–1201.

10

You might also like