SoJi Yong1
YoonYoung Doo2
YoonSang Won1*
-
(Department of Electrical and Computer Engineering, Seoul National University, Seoul
08826, Republic of Korea)
-
(Department of Automotive Engineering, Hanyang University, Seoul 04763, Republic of
Korea)
Copyright © The Institute of Electronics and Information Engineers(IEIE)
Index Terms
Thermal interface material, thermal resistance, indirect cooling, double-sided cooling power module
I. INTRODUCTION
As global interest in sustainable green energy continues to grow, the automotive industry
is undergoing significant transformations, particularly with the rapid expansion of
electric vehicles (EVs) and hybrid electric vehicles (HEVs) [1]. This trend has driven a surge in demand for high-performance power semiconductors,
essential components in power electronic systems. These advancements have led to increased
power density and higher heat generation in power modules [2], which are critical for energy conversion in xEVs. However, excessive heat generation
presents significant challenges, as it can reduce efficiency, compromise reliability,
and shorten the lifespan of power semiconductors [3].
To address these challenges, two advanced cooling strategies are commonly used in
commercial automotive power modules. The first is direct single-sided cooling (Fig. 1(a)), which eliminates the use of thermal interface materials (TIMs) [4]. The second is double-sided cooling (DSC), shown in Fig. 1(b), which provides bidirectional heat dissipation paths. However, most commercial DSC
modules rely on TIMs to fill air gaps between contact surfaces [4,5,6]. While TIMs have higher thermal conductivity than air, their relatively low conductivity
still introduces significant thermal resistance, limiting the overall performance
of indirect DSC structures [7].
Previous studies have shown that even with DSC structures, thermal resistance is only
partially reduced when TIMs are used [6,8,9]. As the thermal conductivity of commercial TIMs remains insufficient, which often
fails to yield proportional reductions in thermal resistance. Direct single-sided
cooling, which avoids TIMs altogether, has been reported to offer comparable thermal
performance to indirect DSC cooling, despite requiring a more complex cooler design
and integration process.
Efforts to enhance indirect DSC performance through structural modifications, such
as adjusting fin dimensions and locations, have shown partial success [8,9]. Some recent studies demonstrated that advanced pin-fin designs, which introduce
a new design named tilted rounded-rectangular fin, can achieve thermal resistance
lower than that of direct cooling designs, even in indirect cooling configurations
[7].
Nevertheless, one of the major challenges in DSC power modules is attributed to the
inherent limitations of commercial TIMs [6,7]. Therefore, this study focuses on analyzing the contribution of TIM characteristics,
such as thermal conductivity and thickness, to the thermal resistance of an indirect
DSC power module structures. Using finite element analysis (FEA), this work investigates
how adjustments to TIM properties can improve thermal performance, attempting to provide
insights into the trends and potential improvements achievable in indirect DSC modules.
Fig. 1. Structure of (a) direct single-sided cooling and (b) indirect double-sided
cooling.
II. THERMAL PERFORMANCE ENHANCEMENT OF INDIRECT DOUBLE-SIDED COOLING STRUCTURE
1. Evaluation Criteria for Thermal Performance of Power Modules
When evaluating the thermal performance of a power module, one of the key factors
is thermal resistance ($R_{th}$). Thermal resistance quantifies the property that
hinders heat transfer and serves as a metric for assessing thermal performance. The
lower the thermal resistance, the more effectively heat can be dissipated, which in
turn enhances the performance and reliability of the power module. $R_{th}$ is defined
by the following equation:
In this equation, $T_{x}$ and $T_{y}$ represent the temperatures at points ${x}$ and
${y}$, respectively, while ${P}$ refers to the power loss.
2. Improving Thermal Resistance of Indirect DSC structure through TIMs Improvement
This section aims to explore methods to improve the thermal performance of the indirect
DSC structure. By examining these material properties, essential input parameters
and boundary conditions are established for FEA simulations. To analyze the thermal
resistance based on TIMs, another equation for thermal resistance is defined as follows:
In this equation, ${L}$ is thickness in this model and refers to the distance along
the path through which heat is transferred. ${A}$ is area and represents the surface
area through which the heat is transferred. ${k}$ is thermal conductivity, which indicates
the material's ability to conduct heat, or how efficiently heat can be transferred
from one point to another.
To reduce the thermal resistance of the TIMs, adjustments to thickness, area, and
thermal conductivity are considered based on the thermal resistance equation (2). However, since the area refers to the surface through which heat is transferred,
simply increasing the area of the structure along the heat path does not necessarily
reduce thermal resistance. Therefore, to better understand the impact of TIM properties,
simulations are conducted by varying thickness and thermal conductivity to analyze
their effects on thermal resistance.
III. POWER MODULE MODELING AND FEM ANALYSIS
1. Power Module Modeling
In Fig. 2, simulation models of the direct single-sided cooling and indirect DSC structures
are shown with the encapsulation removed for better visualization of the internal
structures. The model in Fig. 2(a) stacks the direct bonding copper (DBC) and die on top of a heatsink, covered with
encapsulation. The DBC comprises copper layers bonded to both sides of a ceramic substrate,
offering a stable thermal and electrical interface for power devices.
The model in Fig. 2(b) adds an additional spacer and DBC layer on top. Sintering is used for die attachment,
while solder is used for all other connections. Fig. 3 presents the dimensions of the SiC MOSFET and the bottom DBC [10]. The spacer is designed to match the die's source pad. The DBC copper size is designed
as $16 \times 16 \times 0.3$ mm, and the heatsink is designed with dimensions of $30
\times 30 \times 5.5$ mm. The material properties used in the model are listed in
Table 1 [11,12].
Fig. 2. The simulation models of (a) direct single-sided cooling and (b) indirect
DSC, shown with encapsulation removed for better visualization of internal structures.
Fig. 3. Top view dimensions of SiC MOSFET and bottom DBC with top DBC and spacer omitted.
Table 1. Thermal conductivities of the materials used for the simulation.
|
Material
|
Thermal conductivity
|
|
Silicon carbide
|
370 W/m∙K
|
|
Copper
|
400 W/m∙K
|
|
AlN
|
285 W/m∙K
|
|
Cu60Mo40
|
290 W/m∙K
|
|
Ag
|
240 W/m∙K
|
|
Solder_VACNT
|
59.7 W/m∙K
|
|
EMC
|
0.05 W/m∙K
|
2. Improving TIMs for Enhanced Thermal Performance
In Eq. (2), methods to improve the thermal resistance of the TIMs include adjusting the thickness
and thermal conductivity. Simulations will be conducted using the TIMs listed in Table 2 to analyze their impact. The T670 has a thickness of 0.025 mm, which is thinner than
the conventional TIM, while the 59M39F2P has a thermal conductivity of 64 W/mK, which
is higher than that of the conventional TIM [13,14]. Among the TIMs tested, THERMAL-A-GAP30 served as a conventional TIM and experiments
are conducted with TIMs designed with reduced thickness (T670) and increased thermal
conductivity(59M39F2P) to decrease thermal resistance. Although promising, the 59M39F2P
has not yet been commercialized and remains the subject of ongoing research [14].
Table 2. TIMs properties.
|
Name
|
Types
|
Thickness
(mm)
|
Thermal
Conductivity (W/mK )
|
|
THERMAL
-A-GAP30
|
Gel
|
0.1
|
3.5
|
|
T670
|
Thermal
Grease
|
0.025
|
3
|
|
59M39F2P
|
Advanced TIM
|
0.15
|
64
|
3. Conditions for FEM Analysis
Simulations are conducted using Ansys Mechanical software, incorporating device power
loss as the heat source and heat dissipation through the two heatsinks, with the heat
transfer coefficient applied as the boundary condition. In our indirect DSC model,
SiC chip is assumed to generate a power loss of 313 W.
For the cooling condition, a convective heat transfer coefficient of 24,000 W/m$\cdot$K
is applied to both the top and bottom heatsinks for the double-sided cooling structure.
This coefficient is selected based on the assumption of a fluid-cooled heatsink. Since
this work does not include thermo-fluidic simulations or flowing coolant [7], the junction-to-ambient thermal resistance is used as the performance measure.
IV. SIMULATION RESULTS AND DISCUSSION
The simulation results are shown in Fig. 4. The analysis of the results indicated that the conventional TIM used in the indirect
DSC model exhibits a thermal resistance of 0.489 K/W, which is higher in comparison
to the direct single-sided cooling model, which has a thermal resistance of 0.453
K/W. However, notable improvements are observed with T670 and 59M39F2P.
The model using T670 with reduced thickness shows a thermal resistance of 0.394 K/W,
which is significantly lower than the two models Figs. 4(a) and 4(b). TIM is used
to fill the air gaps between two metal surfaces due to its higher thermal conductivity
compared to air. However, if the TIM is not thick enough, it may not fully fill the
air cavity that forms between the two metals. Since air has a very low thermal conductivity
of 0.026 W/mK, this can significantly increase thermal resistance, thus limiting the
extent to which thickness can be reduced. Therefore, the focus is shifted to analyzing
thermal resistance based on the thermal conductivity of the TIMs.
The model using advanced TIM with high thermal conductivity also shows an improved
thermal resistance of 0.348 K/W. It is observed that high thermal conductivity can
enhance the cooling performance of double-sided cooling. Therefore, more simulations
are conducted to examine the changes in thermal resistance with varying thermal conductivity.
The graph illustrating the relationship between thermal conductivity and thermal resistance
is shown in Fig. 5. A steep initial decrease in thermal resistance is observed, which gradually saturated.
Thermal resistance shows a steep decrease as thermal conductivity increases up to
around 30 W/mK, and beyond that, the impact appears minimal.
Currently, commercially available TIMs typically have a thermal conductivity substantially
lower than 30 W/mK. While some TIMs with superior thermal conductivity, such as the
evaluated 59M39F2P [15], have been reported, they remain at the research stage and are not yet commercially
available. Therefore, from an availability perspective, it would be beneficial to
target a TIM with a thermal conductivity exceeding 30 W/mK for commercialization.
Fig. 4. The simulation results of (a) direct single-sided cooling, (b) indirect DSC
with THERMAL-A-GAP30, (c) T670, and (d) 59M39F2P are shown with encapsulation and
heatsink removed.
Fig. 5. Thermal resistance as a function of TIM thermal conductivity.
V. CONCLUSION
In this study, we investigate the reason why the direct single-sided cooling structure
exhibits better thermal performance compared to the indirect DSC structure, despite
the latter dissipating heat from both sides. The key factor identified is TIMs and
we explore methods to improve its performance by adjusting the thickness and thermal
conductivity. Simulations are conducted based on these parameters, with a particular
focus on how thermal conductivity affects thermal resistance. As shown in Fig. 5, thermal resistance decreases significantly when the thermal conductivity of the
TIM is below 30 W/mK. However, when it exceeds 30 W/mK, the rate of decrease slows
and eventually reaches saturation. Based on these results, it appears reasonable to
target a TIM thermal conductivity of approximately 30 W/mK.
ACKNOWLEDGMENTS
This work was supported by the National Research Foundation of Korea (NRF) through
the Korea Government (MSIT) under Grant 2023R1A2C2006661 and Grant RS-2023-00207865.
The authors appreciate the support from Inter-university Semiconductor Research Center,
Seoul National University, Korea.
References
A. Kumar, M. Moradpour, M. Losito, W.-T. Franke, S. Ramasamy, R. Baccoli, and G. Gatto,
``Wide band gap devices and their application in power electronics,'' Energies, vol.
15, no. 23, 9172, 2022.
J. Lim, D. K. Jang, and S. W. Yoon, ``Novel temperature calibration method of E-mode
and D-mode GaN power modules considering dynamic ON-resistance,'' IEEE Transactions
on Industrial Electronics, vol. 71, no. 5, pp. 5245-5252, 2024.
S. W. Yoon, M. D. Glover, and K. Shiozaki, ``Nickel-tin transient liquid phase bonding
toward high-temperature operational power electronics in electrified vehicles,'' IEEE
Transactions on Power Electronics, vol. 28, no. 5, pp. 2448-2456, May 2013.
S. Jones-Jackson, R. Rodriguez, Y. Yang, L. Lopera, and A. Emadi, ``Overview of current
thermal management of automotive power electronics for traction purposes and future
directions,'' IEEE Transactions on Transportation Electrification, vol. 8, no. 2,
pp. 2412-2428, June 2022.
J. Jeon, J. Seong, J. Lim, M. K. Kim, T. Kim, and S. W. Yoon, ``Finite element and
experimental analysis of spacer designs for reducing the thermomechanical stress in
double-sided cooling power modules,'' IEEE Journal of Emerging and Selected Topics
in Power Electronics, vol. 9, no. 4, pp. 3883-3891, Aug 2021.
M. Liu, A. Coppola, M. Alvi, and M. Anwar, ``Comprehensive review and state of development
of double-sided cooled package technology for automotive power modules,'' IEEE Open
Journal of Power Electronics, vol. 3, pp. 271-289, May 2022.
S. Cho, S. Pyun, and S. W. Yoon, ``Analysis and design optimization of tilted rounded-rectangular
pin-fins heat sink for xEV double-sided cooling power module,'' IEEE Journal of Emerging
and Selected Topics in Power Electronics, vol. 12, no. 2, pp. 1803-1812, April 2024.
S. Pyun, S. Cho, and S.W. Yoon, ``Analysis of perforated pin design use in automotive
SiC power module heatsink,'' Journal of Power Electronics, vol. 23, pp. 1888-1895,
2023.
S. Cho and S. W. Yoon, ``Analysis model for thermal resistance of double-sided cooling
power module with pin-fin heat sink used in xEVs,'' Journal of Power Electronics,
vol. 23, pp. 1880-1887, 2023.
S. Jahdi, O. Alatise, J. Ortiz-Gonzalez, P. Gammon, L. Ran, and P. Mawby, ``Investigation
of parasitic turn-ON in silicon IGBT and silicon carbide MOSFET devices: A technology
evaluation,'' Proc. of European Conference on Power Electronics and Applications (EPE'15),
pp. 1-8, Geneva, Switzerland, September 2015.
Y. Zhang, H.-P. Nee, T. Hammam, I. Belov, P. Ranstad, and M. Bakowski, ``Multiphysics
characterization of a novel SiC power module,'' IEEE Transactions on Components, Packaging
and Manufacturing Technology, vol. 9, no. 3, pp. 489-501, March 2019.
Y. Yang, Y.-C. Tseng, R. Rodriguez, A. D. Callegaro, and A. Emadi, ``Modeling and
analysis of silver-sintered molybdenum packaging for SiC power modules with improved
lifetime and temperature range,'' IEEE Transactions on Components, Packaging and Manufacturing
Technology, vol. 11, no. 12, pp. 2102-2113, Dec. 2021.
J. Broughton, V. Smet, R. R. Tummala, and Y. K. Joshi, ``Review of thermal packaging
technologies for automotive power electronics for traction purposes,'' ASME Journal
of Electronic Packaging, vol. 140, no. 4, December 2018
T. Zhang, B. G. Sammakia, Z. Yang, and H. Wang, ``Hybrid nanocomposite thermal interface
materials: The thermal conductivity and the packing density,'' ASME Journal of Electronic
Packaging, Vol. 140, no. 3, September 2018.
H. Fu, R. Fang, C. Tian, W. Qian, S. Cao, Z, Zhang, X. Xu, C, Yao, Z. Wang, and D.
He, ``High-performance thermal interface materials enabled by vertical alignment of
lightweight and soft graphene foams,'' Nano Res., vol. 17, pp. 9293-9299, 2024.
Ji Yong So received his B.S. degree in automotive engineering from the Hanyang
University, South Korea in 2021. He is currently working toward a unified master's
and doctor's degrees with the Department of Electrical and Computer Engineering, Seoul
National University, Seoul. His current research interests include advanced packaging
and power module packaging.
Young-Doo Yoon was born in South Korea. He received his B.S., M.S., and Ph.D. degrees
in electrical engineering from Seoul National University, Seoul, South Korea, in 2002,
2005, and 2010, respectively. From 2010 to 2013, he was with Samsung Electronics Company,
South Korea, as a Senior Engineer. From 2013 to 2017, he was an Assistant Professor
with the Department of Electrical Engineering, Myongji University, Yongin, South Korea.
Since 2017, he has been a Faculty Member of the Department of Automotive Engineering,
Hanyang University, Seoul, South Korea, where he is currently a Professor. His research
interests include power electronic controls of electric machines, high power converters,
electric vehicles, and electric home appliances.
Sang Won Yoon received his B.S. degree in electrical engineering from Seoul National
University, Seoul, Korea, in 2000 and his M.S. and Ph.D. degrees in electric engineering
and computer science from University of Michigan, Ann Arbor, MI, USA, in 2003 and
2009, respectively. From 2009 to 2013, he was a Senior Scientist and a Staff Researcher
at the Toyota Research Institute of North America, Ann Arbor, MI, USA, where he conducted
research in power electronics and sensor systems for automobiles. From 2013 to 2023,
he was Assistant Professor, Associate Professor, and Professor in the Department of
Automotive Engineering, Hanyang University, Seoul, Korea. Since 2023, he has been
with the Department of Electrical and Computer Engineering at Seoul National University,
Seoul, Korea. His research interests include packaging and reliability of semiconductors,
electronics for mobility, and their applications.