Enhancing Thermal Efficiency in Power Electronics: A Review of Advanced Materials and Cooling Methods

Abstract

Over the last several years, a significant advancement in high-voltage electronic packaging techniques has paved the way for next-generation power electronics. However, controlling the thermal properties of these new packaging solutions is still a major challenge. The utilization of wide bandgap semiconductors such as SiC and GaN offers effective methods to minimize thermal inefficiencies caused by conduction losses through high-frequency switching topologies. Nevertheless, the need for high voltage in electrical systems continues to pose significant barriers, as heat generation remains one of the most significant obstacles to widespread implementation. The trend of electronics design miniaturization has driven the development of high-performance cooling concepts to address the needs of high-power-density systems. As a result, the design of effective cooling systems has emerged as a crucial aspect for successful implementation, requiring seamless integration with electronic packaging to achieve optimal performance. This review article explores various thermal management approaches demonstrated in electronic systems. This paper aims to provide a comprehensive overview of heat transfer enhancement techniques employed in electronics thermal management, focusing on core concepts. The review categorizes these techniques into concepts based on fin design, microchannel cooling, jet impingement, phase change materials, nanofluids, and hybrid designs. Recent advancements in high-power density devices, alongside innovative cooling systems such as phase change materials and nanofluids, demonstrate potential for enhanced heat dissipation in power electronics. Improved designs in finned heat sinks, microchannel cooling, and jet impingement techniques have enabled more efficient thermal management in high-density power electronics. By fixing key insights into one reference, this review serves as a valuable resource for researchers and engineers navigating the complex landscape of high-performance cooling for modern electronic systems.

1. Introduction

In recent years, the sales of passenger electric vehicles (EVs) have experienced a significant increase, growing almost 5 times from 450,000 in 2015 to 2.1 million in 2019. Moreover, the recent announcement made by the United States Government to transition almost 650,000 vehicles in its fleet to all-electric vehicles signifies a growing trend toward electrification. In this context, it is necessary to emphasize the significance of power semiconductor technologies as fundamental constituents of traction PE (power electronics) systems [1].

Thermal management in power electronics involves the dissipation of heat generated by active semiconductor devices, such as transistors, diodes, and integrated circuits, as well as passive components like resistors and inductors. These components generate heat due to the inherent power losses they experience during conversion and switching. Maintaining the junction temperatures of semiconductor devices within safe limits is crucial for ensuring optimal performance and reliability [2,3,4,5].

Conventional cooling techniques, such as air-cooling using heat sinks and fans, have been widely employed in power electronic systems [2]. However, the ever-increasing power densities and miniaturization trends have necessitated the exploration of advanced and more efficient thermal management solutions [6,7]. This review paper aims to provide a comprehensive overview of the various techniques and approaches investigated and implemented for effective thermal management in power electronics.

One of the widely studied techniques is the use of fin-based heat sinks, which utilize the principle of extended surfaces to enhance heat dissipation through increased surface area. These heat sinks can be air-cooled or liquid-cooled, depending on the cooling requirements and system constraints [8,9,10]. Micro-channel heat sinks, which incorporate micro-scale channels for coolant flow, have also gained significant attention due to their ability to achieve high heat transfer coefficients and compactness [11,12].

Jet impingement cooling is another promising technique that utilizes the impingement of a high-velocity fluid jet onto the heated surface. This approach offers efficient heat transfer by disrupting the boundary layer and enhancing turbulent mixing, making it suitable for high heat flux applications [13,14]. Phase change materials (PCMs) have also been explored as a thermal management solution, leveraging their ability to absorb and release large amounts of latent heat during phase transitions, effectively acting as thermal buffers [15].

This review addresses the critical role of thermal management in electric vehicle (EV) power electronics, emphasizing the importance of advanced cooling solutions in maintaining optimal operational performance and longevity. Section 2 provides an overview of key components within power modules, including power semiconductor devices, spacers, substrates, thermal interface materials (TIM), and heat sinks, which play a crucial role in heat generation and dissipation. Section 3 then presents a comprehensive analysis of advanced cooling techniques, such as finned surfaces, microchannels, jet impingement cooling, phase change materials (PCMs), and nanofluids. These methods are selected for their potential to meet the high heat flux and efficiency requirements of electric vehicle (EV) power electronics. In Section 4, a comparative assessment of these cooling methods is conducted, highlighting their respective advantages and limitations in terms of thermal performance. This structure allows for a thorough investigation from component-level considerations to advanced cooling strategies, leading to a comprehensive understanding of current and emerging solutions for efficient thermal management in power electronics.

Furthermore, the paper will explore the impact of emerging trends and technologies, such as wide-bandgap semiconductors (e.g., SiC and GaN), on thermal management requirements. These advanced materials offer superior properties, including higher operating temperatures and lower switching losses, but also present new challenges in terms of heat dissipation and reliability [16,17]. Addressing these challenges requires innovative cooling solutions tailored to the unique characteristics of wide-bandgap devices [18].

By providing a comprehensive review of state-of-the-art techniques for effective thermal management in power electronics, this paper aims to serve as a valuable resource for researchers, engineers, and designers working in this field. It will highlight the latest advancements, identify research gaps, and provide insights into future directions for continued innovation and development of efficient and reliable cooling solutions for power electronic systems.

2. Methodology

To ensure a comprehensive and systematic review, the literature search was conducted using Scopus and IEEE Xplore databases to identify peer-reviewed journal articles, conference proceedings, and high-impact review papers relevant to thermal management in power electronics systems. The search queries included the following keywords and combinations: “power electronics thermal management,” “heat sink,” “direct bonded copper (DBC),” “thermal interface material (TIM),” “spacer,” “substrate,” “jet impingement cooling,” “microchannel cooling,” “phase change material,” “nanofluid cooling,” “pool boiling,” and “electronics cooling.”

After initial screening, papers were selected based on their direct relevance to electronics cooling and their contribution to advancing the field, prioritizing studies with high citation counts to ensure quality and impact. The selected papers were categorized into six major sections: finned heat sinks, microchannel cooling, jet impingement, phase change materials (PCMs), nanofluids, and pool boiling.

For each year from 2010 to July 2025, the number of relevant articles was recorded. The percentage values in Figure 1 were calculated as the proportion of publications for each year relative to the total number of selected articles across the study period.

3. Power Module Major Components

Effective thermal management in EV power modules is essential for ensuring performance, reliability, and durability. The thermal path involves several key resistances, including junction-to-case, case-to-heatsink, and heatsink-to-ambient. Junction-to-case thermal resistance is important at the module level. Efficient heat spreading from the case to the heatsink helps to improve cooling. Over time, thermal-related failures like solder joint cracks, delamination, and wire bond lift-off can develop, especially as power cycling increases. These failures reduce heat transfer efficiency, leading to long-term degradation of the module [19].

Inverters and, occasionally, converters utilize power electronics modules in a variety of traction systems. The housing structure of these modules contains power electronics components, including thermistors, power semiconductor dies, resistors, and capacitors [20]. A schematic illustration of a power device packaging module with various components can be found in Figure 2 of [1].

3.1. Power Semiconductor Device

Power electronics play a crucial role in modern electrical and electronic systems, enabling efficient energy conversion, control, and conditioning. As the demand for higher power density and miniaturization continues to rise, effective thermal management has emerged as a critical challenge in power electronic systems. Excessive heat can harm a system’s lifetime and reliability. It may lead to lower performance, lower efficiency, and early failure of electronic components [1,2]. The primary devices used for traction converters are metal-oxide-semiconductor field-effect transistors (MOSFETs), diodes, and insulated-gate bipolar transistors (IGBTs). Automotive applications require devices that can handle a current rating of over 200 A. As a result, heat generation becomes a significant concern, necessitating proper thermal management. Previously, automotive traction systems used silicon-based IGBTs and diodes. However, modern designs increasingly rely on wide bandgap technologies (WBGs), specifically SiC-based devices, to fulfill performance requirements [1]. Figure 2 illustrates a comparative analysis of material properties that are extensively employed in power electronics devices.

In comparison to silicon (Si) devices, silicon carbide devices are renowned for their superior properties. Elevated breakdown voltage levels, greater operating temperatures, rapid switching velocities, and reduced switching loss are some of these characteristics [22]. The comparative analyses between the Si and SiC-based power modules are listed in

Table 1. A comparative overview of Si IGBT and SiC MOSFET modules.

Parameter	Si IGBT	SiC MOSFET
$\mathbf{d}\mathbf{v}/\mathbf{d}\mathbf{t}$	Low	High
Switching Frequency	Slow	Fast
Tail Current	Present	Absent
$\mathbf{d}\mathbf{i}/\mathbf{d}\mathbf{t}$	Low	High
Switching Loss	High	Low
Operating Junction Temperature	Low	High
Power Density	Low	High
Heat Flux	Low	High

concerning multiple parameters. SiC MOSFETs are developed to be significantly more compact than Si IGBTs while retaining identical voltage and current ratings [23]. Switching frequencies for Si-IGBT-based inverters typically do not exceed 20 kHz. However, SiC MOSFETs have the capability to reach a switching frequency of up to 100 kHz when used in a hard switching converter [24]. Additionally, the losses can be further minimized via soft switching, resulting in even greater switching frequencies. The surge in frequency of switching may lead to higher power density as well as the ability to use smaller filters. In contrast to bipolar devices, SiC MOSFETs demonstrate enhanced properties, including the absence of tail current resulting from minor recombination of carriers and a switching rate that is quicker than that of Si IGBTs.

Silicon Carbide (SiC) MOSFETs can work at elevated temperatures, reaching approximately 300 °C. This is almost double the highest temperature that silicon IGBTs can manage. This lowers the need for cooling in addition to a heatsink. However, existing packaging techniques constrain the high-temperature performance of SiC devices. Given the aforementioned advantages, numerous studies have been conducted on SiC MOSFETs in recent years. Silicon carbide (SiC) is a material that exhibits higher stiffness and a greater Young’s modulus compared to silicon (Si). This feature can lead to increased stress levels in the solder connecting SiC devices to the DBC substrate. This stress can create excessive temperatures in certain regions and non-uniform heat distribution in SiC power modules [25]. SiC devices have made it possible to further reduce the size of EV (electric vehicle) inverters, increasing their power density, but this has resulted in thermal management challenges [14].

Figure 3 represents graphically how the packaging trend has changed over the decade. Power electronics must possess an efficient thermal management system to ensure optimal functioning. The dimensional distinctions between Si and SiC-based devices utilized in comparable applications are illustrated in Figure 4.

The properties of Si and SiC devices play a crucial role in the overall performance of power modules. Despite being simple and less expensive, the Si devices provide less switch frequency, increased switching losses, and lower operating temperature than the SiC devices. The lower electrical properties of Si devices restrict them from handling higher power density, dissipating less heat, and thus require high thermal management compared to SiC devices [27].

In contemporary electric vehicles, power semiconductor devices are increasingly integrated within multi-device module packaging, as opposed to the discrete plastic-encapsulated packages used in earlier EVs like the 2012 Tesla Model S and Model X. Discrete packaging keeps each device separate. In contrast, module packaging combines multiple devices into a single unit. This setup contains protective casing, electrical connections, lead frame, and ceramic substrates that insulate electricity effectively. Module packaging offers enhanced heat management, stronger mechanical support, and improved electrical connectivity. Recent advancements have further evolved module packaging from single-sided to double-sided cooling, which significantly improves thermal performance by reducing temperature fluctuations and thermal stress, thereby enabling higher power density and cost-reduction opportunities [28].

3.2. Spacer

The spacer is one of the key components of a double-sided cooling (DSC) module, which plays a crucial role in connecting heat sinks both thermally and electrically while maintaining a safe distance [20].

When designing traction inverter devices, Double Sided Cooling (DSC) power modules provide more design flexibility than single-sided cooling (SSC) power units. The technique is compatible with both IGBT and SiC power modules. However, the double-sided structure of DSC power modules gives rise to severe thermal mismatches, making it easier for attachments to fail. Therefore, thermomechanical stress remains a challenging aspect to ensure the proper functioning of DSC power modules, unless thermal performance is compromised. The spacer is a crucial component in DSC power modules, and its composition and position have a substantial effect on thermomechanical performance. Figure 1 in [20], for example, illustrates the double-sided cooling power module, featuring its key components (including spacer) and their corresponding positions.

Hsieh et al. [29] analyzed the impact of spacer material characteristics on the thermal efficiency and stress distribution of vertically stacked 3D integrated circuit packages. Utilizing response surface methods, a design optimization was carried out to investigate the spacer’s thermal conductivity in connection with its resistance to heat transfer. Furthermore, the associated mechanical characteristics, including the coefficient of thermal expansion and Young’s modulus, concerning the stresses are illustrated. The findings indicate that the thermal performance and stress distributions of the integrated circuit packages are significantly influenced by the material characteristics of the spacers.

Lin et al. [30] suggested a design that utilizes the flip-chip approach and includes a spacer beneath the die.

An inherent benefit of this design lies in the spacer’s dimensions, which surpass the die dimensions. This design differs from the conventional spacer stacking arrangement, in which the spacer lies on the top of the die and is smaller in dimension compared to the power devices. An example of the heat dissipation path in a power module packaging, depending on the spacer positioning, can be found in Figure 4 of [30]. Cao et al. [31] introduced a credible, double-sided package incorporating the bidirectional switch using bridge buffer spacers for enhanced heat dissipation and reduced thermomechanical stresses. Studies have shown that the proposed spacer effectively decreases the peak thermomechanical stress on the connection between the spacer and the IGBT by about 35% and reduces the maximum junction temperature by approximately 9%.

Frequently, spacers are cube-shaped to provide an elevated solder-contact surface when power dies are utilized. Reducing thermal and electrical resistances in this contact region confers an advantageous outcome. However, it has the potential to expedite the degradation of die-bonded solders and increase thermomechanical stress. To address this issue, Jeon et al. [20] proposed spacer designs using an octagonal or etched surface of X-shape and assessed their performance relative to trenched spacers and typical cuboid spacers. Figure 4 of Ref. [20] showcases these three spacer shapes. The octagon and spacers of the X-shape presented in this study provide superior solder reliability compared to the trenched 2 × 2 as well as typical cuboid spacers. While the spacer of x-shape without solder infiltration is the most effective solution, managing the orientation and quantity of redistributed molten solder might be a challenging endeavor. Thus, based on the production process, it can be inferred that the octagonal spacer is an optimal option. The inclusion of an AlN spacer layer has been found to effectively reduce both alloy and interface roughness scattering, resulting in enhanced system mobility. Additionally, it has been observed that by augmentation of the spacer layer’s thickness to 1.5 nm, the cutoff frequency can be enhanced up to 312 GHz [32].

3.3. Substrate

A substrate is an integral part of power module packaging, performing the functions of creating a proper thermal path toward the heat spreaders, electrical connections with low resistance, and necessary electrical insulation. The presence of substrate in power module packaging enables the power module to be operated at high power density [33].

The utilization of wide bandgap (WBG) devices in power modules requires packaging designs that use advanced materials and structures. These designs must fulfill challenges with managing heat and mitigate problems caused by too much electrical parasitic inductance. The direct-bonded copper substrate in the power module consists of a ceramic insulator that is positioned between two copper layers. The uppermost layer comprises critical elements such as wire bond landing, die attachment, tracks, and terminal attachment regions. In contrast, the lowermost layer comprises the contact region between the substrate and the domain of thermal management. The detailed view of a half-bridge power module with each of its components as well as connections based on DBC technology is presented in Figure 1 of [34]. Various insulating ceramics have been examined and documented in research for direct-bonded substrates.

Table 2. Standard ceramic layer thickness and properties [34].

Material	Thickness (µm)	Thermal Conductivity (W/mK)	Coefficient of Thermal Expansion, α (10−6/°C)
Al2O3	381	20	8.1
AlN	635	180	4.5
Si3N4	635	30	3.3
HT-07006	152	24	7.5

lists different types of ceramic layer materials along with their standard thickness, thermal, and electric properties. When power density is not a consideration and mechanical robustness and an elevated coefficient of thermal expansion are desired, aluminum oxide (Al₂O₃) is frequently utilized in substrate design. The manufacturing of Al₂O₃ is relatively inexpensive compared to aluminum nitride (AlN). Silicon nitride (Si₃N₄) is more mechanically robust compared to AlN and is the preferred insulating layer of the DBC substrate. For power electronic applications, the insulated metal substrate (IMS) is a substitute for DBC in which the electrically conducting layers are insulated with a polymer-based coating. IMS consists of conductor wires, an insulating layer, and a metal base. Dielectric polymers possess excellent dielectric strength and can be machined in very thin layers. These are the properties that make IMS more attractive despite having lower thermal performance. Multilayer substrate is also possible using IMS [34]. Due to the advantages such as lower cost, mechanical strength, machineability, and lightweight, IMSs are being used for moderate power applications.

Graphite is emerging as an efficient material for managing heat in electronic systems because it has excellent thermal conductivity as well as a low density, which makes it better than conventional metals such as aluminum and copper. Gurpinar et al. [34] explored the use of insulated metal substrates with incorporated graphite in wide bandgap power modules with SiC MOSFETs. Tang et al. [17] investigated SiC chips that are incorporated into active metal-brazed (AMB) substrates. Badstuebner et al. [35] examined the power cycle durability of SiC MOSFET modules without baseplates, using Al₂O₃ and Si₃N₄ DBC substrates. The simulations conducted indicate that Si3N4 substrates accelerate the rate at which the temperature increases. The study carried out by Fukumoto et al. [33] delved into the impacts of drastic temperature variations ranging from −55 °C to 250 °C on Si₃N₄-AMB substrates. Yu et al. [36] performed research comparing the thermal and electrical performance of GaN transistors installed on ceramic substrates and printed circuit boards (PCBs).

Graphite-embedded substrates possess the potential to decrease the thermal resistance of SiC MOSFETs by as much as 17% and increase the density of current by 10%, regardless of the substrate cooling thermal management strategy employed. The research also validates that the transient thermal impedance of the dies may be lowered by up to 40% as a result of the improved capacity for heat [34]. The thermal performance of the standard power packaging using an SSC (side-side cooling) solution is significantly influenced by the substrate’s material and the thickness of the central layer of an AMB substrate [17]. SiC MOSFET modules with Si₃N₄-DCB substrates offer the advantage of reduced thermal impedance, enabling increased current capability without compromising the bond wires’ reliability [35]. The ceramic substrates offer superior thermal management capabilities, delivering a thermal resistance that is reduced fourfold than that of PCBs while maintaining suitable electrical performance. Direct bonded copper (DBC) substrates are an excellent choice for high-current inverters as they can significantly reduce thermal resistance. However, they may lead to slightly elevated switching losses because of parasitic inductance [36]. The copper layers do not exhibit any signs of delamination from the Si₃N₄ layers even after being subjected to 600 temperature cycles. Upon detailed investigation, it was uncovered that the fundamental reason for the roughening of the surface is the thermal stress resulting from the mismatch in the coefficient of thermal expansion (CTE) between copper and silicon nitride [33].

3.4. Thermal Interface Material

Thermal interface materials (TIMs) refer to special materials that are applied to create a flat and smooth surface between the electronics and the heat sink. TIMs fill the air gap and void areas to enhance the contact area between two surfaces, which ultimately leads to higher heat transfer [37]. Increasing demand for functionality, power density, miniaturization, and thermal dissipation has become a challenging problem for thermal interface applications.

Greases, such as silicon-based or metal-particle-filled greases, are commonly used due to their conformability and ease of application. They exhibit low thermal resistance when applied under high pressure. Burzo et al. [38] investigated three TIMs through transient thermo-reflectance (TTR), and a correlation was observed between the imposed pressures. As the bond line thickness reduces, there is an expected drop in thermal resistance. As the pressure on the contact lowered, the grease created a more effective connection with the surrounding surfaces. Before reaching around 800 kPa, the thermal resistance of the interface dropped in direct correlation with the applied pressure. However, after this threshold was reached, the thermal resistance remained constant.

Phase-Change Materials transition between solid and liquid phases at specific temperatures, filling surface irregularities and reducing thermal resistance. Wen et al. [39] developed a low-melting alloy (comprising Bi, Sn, In, and Ga) with the addition of Cu fillers, resulting in an enhanced thermal interface material (TIM) performance. Various fill ratios of Cu were examined, with a 50% fill ratio yielding the optimal lowest thermal resistance. The composite of phase change alloy has excellent heat conductivity and wettability. Therefore, it is possible to reduce the thermal resistance at the interface between the blocks to a low level of 0.08 cm²K/W.

Solder TIMs, like Indium or Sn-Ag-Cu alloys, are known for their excellent bulk thermal conductivity and durability in thermal cycling. Indium’s low flow stress, which is around 1 MPa, allows it to conform effectively to surfaces and manage thermal expansion stresses. This property helps to protect delicate semiconductor materials from shear forces. Hansson et al. [40] analyzed solder matrix fiber composites (SMFCs) that are Indium or Sn–Ag–Cu alloy-infused networks of fibers comprising varying fiber densities. The study revealed that achieving a thermal interface resistance of less than 4 mm²K/W at reduced fiber densities was possible for composites composed of indium and Sn-Ag-Cu. On the contrary, as fiber densities increased, the number of continuous metallic lines traversing the interface decreased. For 500 thermal cycles, the resistance at the thermal interface remained constant for all sample configurations, and the internal microstructure exhibited no signs of deterioration.

Carbon-based TIMs have gained traction for their high thermal conductivity. CNTs are particularly useful on non-planar surfaces. Idris et al. [41] investigated the impact of carbon nanotube (CNT) input weight percentage, size, and shape on the performance of thermal interface materials (TIMs). To evaluate the efficacy of CNT powder when combined with silicon oil, it was noted that the resistance to thermal contact escalates as the filler loading increases. However, this can be mitigated by applying pressure. The observed rise in resistance to thermal contact was attributed to the reduced conformability of carbon nanotubes in comparison to the spherical configuration. By incorporating carbon nanotubes (CNT), the thermal conductivity of silicone oil increased from 0.165 W/mK to 1.640 W/mK. Moreover, it was observed that the thermal conductivity rose to 4.237 W/mK as the CNT concentration increased to 60% by weight. Prinzi et al. [42] studied carbon nanotube (CNT)–polymer composite thermal interface materials that are intended for cooling solutions for nonplanar devices with minimal thermal resistance as well as elevated compliance. Vertically oriented nanotubes of carbon nanotubes formed on metal foils exhibit minimal thermal resistance, according to the study. Conversely, as the height of the CNTs increases, the TIM exhibits enhanced mechanical flexibility and greater heat resistance. When evaluating the ability of a TIM to establish excellent contact at an interface with non-flat surfaces as a result of manufacturing tolerances, warpage, and other factors, its compressibility is an essential consideration. Mouane et al. [37] aimed to investigate. This study investigates the heat conduction properties of carbon nanomaterials, such as nanospheres (CNS) and nanotubes (CNT), at varying concentrations in comparison to thermal interface materials (TIM). The findings from the measurements indicated that the thermal interface materials composed of 1% carbon nanomaterials exhibited the greatest heat transmission, leading to a peak temperature reduction of 2 °C. Additionally, it was observed that the TIMs comprising carbon nanospheres and 1% carbon nanotubes demonstrated the minimal thermal resistances, measuring 100.9 mm²K/W and 80.14 mm²K/W, respectively. Furthermore, the findings of the study indicated that it was anticipated that the thermal conductivity of the CNTs would surpass that of the CNSs.

Lower thermal resistance ensures efficient heat transfer from the device to the heatsink or cooling system. Wang et al. [43,44] proposed the utilization of liquid metal containing bismuth as a TIM filler between the connecting areas of press-packed IGBT devices. By its elevated thermal and electrical conductivities, the liquid metal assisted in the optimized device’s over 30% reduction in junction-to-case thermal resistance when compared to conventional commercial press-packed IGBT devices. The method for optimizing thermal contact resistance using liquid metal is comparatively economical and easier to implement in comparison to the widely employed nano silver sintering technique.

Bulk thermal conductivity is the ability of the TIM itself to conduct heat. Jiang et al. [44] analyzed various proportions of hybrid fillers consisting of platelet h-BN fillers and mixed spherical fillers in thermal interface materials (TIMs). A specific ratio of combined platelet and spherical h-BN additives can considerably improve the thermal properties of TIMs in comparison to those that only contain platelet or spherical h-BN particles, according to the study. The TIM, consisting of spheres and pure platelets, has thermal conductivities of 0.9 W/mK and 2.3 W/mK, respectively. Nevertheless, the proportion of spherical particles to platelets of 4:1 resulted in a thermal conductivity of 7.1 W/mK, surpassing the conductivity of pure platelets by more than threefold.

Contact resistance occurs at the interface between the TIM and the surfaces it connects. Barako et al. [45] focused on an investigation into the comparative effectiveness of utilizing a reactive metallic bond film instead of indium solder for the attachment and transfer of carbon nanotube bands to metal-coated substrates. By supplying a localized, high-temperature heat source, a reactive metallic adhesive layer facilitates the formation of Sn-Au bonds between metallic CNT bands and substrates. By utilizing reactive metal bonding, it is possible to decrease the thermal boundary resistance to a magnitude less than 27 mm²KW⁻¹, which is over an order of magnitude lower than the resistance of the nonbonded contact. This finding is relevant for applications where maintaining low thermal resistance at the interface is essential, such as the IGBT module and the heatsink.

Many TIMs show a correlation between applied pressure and thermal resistance. Zhang et al. [46,47] analyzed the correlation between thermal resistance and rheological properties. They compared five different silicon gels with varying formulations using Al particles as a filler. The thermal resistance of each specimen decreased as the testing pressure rose.

Emerging research is focusing on novel materials and techniques to further reduce thermal resistance and enhance thermal conductivity. Dan et al. [47,48] developed a network model implemented in Java that depicts the near-percolation movement in particle-filled mechanisms concerning arbitrary distributions of size and interactions between particles. The results of the analysis indicate that for a given volume fraction, there is an optimum level of polydispersity that exhibits the maximum thermal conductivity. The effect of interfacial resistance on the efficient thermal conductivity of TIMs renders the influence of polydispersity negligible. Yi et al. [48] examined the efficacy of oxygen plasma reaction technology in improving the efficiency of heat transfer occurring at the interface of the heat spreader and the TIM. Nickel oxide is utilized as a phonon transfer connect layer between TIM and the heat spreader to improve the transfer of phonons at the interface, according to the study. Furthermore, it was discovered that plasma cleansing enhanced the heat spreader’s wettability, fortified the heat spreader’s connection to the TIM, and decreased the porosity of the interface. Subsequently, the thermal resistance of the package module dropped by 42.50%, while its thermal conductivity climbed to 37.69%. Kemerli et al. [49] evaluated the efficacy of different TIMs in conjunction with metal foams. The results indicate that eGraf demonstrates superior thermal performance due to its effective heat dispersion inside the metal foam ligaments, resulting in a 15–18% enhancement in heat transmission compared to the baseline case. The thermal gap pad demonstrated heat transmission properties that were similar to those of the baseline scenario. Conversely, applying thermal epoxy to the struts caused an improvement in thermal resistance and a 25% reduction in heat transmission. Furthermore, the application of thermal epoxy to the heated surface significantly augmented the thermal resistance, leading to a 50% reduction in the Nusselt number.

Table 3. Thermal conductivity of widely used TIMs analyzed in recent papers [40,44,49,50].

Type	Material	Thermal Conductivity (W/mK)
Gap Filler	h-BN	300
	IC-diamond	2000
	Kryonaut	12.5
	GT-2	9.8
	MX-4	8.5
	Kafuter k-705	2
	Epoxy resin	0.8
	Silicone oil	0.165
	Artic Silver 5	0.96
	Nickel oxide	12
	Thermal epoxy	2.5
Gap Pad	TFX	14.3
	Rocket	7.5
	Kafuter k-5205	2
	Shin-Etsu G751	4.5
	TF 400	2.8
	Thermal gap pad	13.1
	eGraf	10.15
PCM	SY G1000	6
Solder	SAC305	34
Metal Based	Indium	36
	PC alloy with 50% Cu	54
	Bismuth-based liquid metal	70

highlights a variety of TIMs studied from reviewed papers, along with their respective thermal conductivity values.

3.5. Heat Sink

Heat sinks play the role of heat exchangers to dissipate heat effectively from the heat-generating electrical or mechanical equipment. Heat sinks generally utilize three basic principles of heat transfer: conduction, convection, and radiation. These mechanisms are adopted in multifarious ways in practical cases. Figure 5 depicts the mechanisms as well as the techniques that have been adopted for enhanced heat transfer and thermal management of heat sinks.

Effective heat conduction depends on the intimate surface contact along with contact surface resistance. Optimized heat sink design necessitates the study of thermal contact resistance [51]. Convection heat transfer is highly influenced by geometry configuration and optimization. Recent studies focused on heat sink fin shape and configuration optimization [52].

The effective surface area and emissivity of heat sink materials significantly influence radiation heat transfer. Therefore, optimizing the heat sink structure using materials with high emissivity is crucial. Integrated microchannels in heat sinks are being used for high heat flux thermal management. Jet impingement cooling represents the direct cooling method where high-velocity air/water impinges on the surface to be cooled. The application of a jet cooler provides enhanced localized cooling of hotspots.

Recent passive thermal management includes phase change materials (PCMs). PCM absorbs a portion of heat and changes its physical state, and serves as a storage of heat energy.

Integrating PCMs with finned heat sink structures can effectively enhance heat dissipation through conduction and convection mechanisms [53].

Because of the high surface area with respect to size, nanofluids have emerged as an efficient technique to enhance the heat transfer characteristics of the base fluid. A combination of both conduction and convection currents facilitates the heat transfer performance. The selection of appropriate nanofluid and maintaining uniform particle dispersion in the base fluid are the major challenges that arise in nanofluid applications.

The proper design of a heat sink is crucial for ensuring the reliable operation of power devices. In addition to maintaining optimal temperatures for electronic components, the design of the heat sink must also meet criteria for size, weight, durability, and cost.

4. Cooling Techniques

This section provides an overview of advanced cooling techniques utilized to address the thermal management challenges in power electronics. These approaches include finned heat sinks, microchannel cooling, jet impingement, phase change materials, and nanofluids represent diverse strategies aimed at enhancing heat dissipation and ensuring reliable operation across a range of applications.

The flowchart in Figure 5 categorizes cooling techniques for power electronics into passive and active methods. It provides a hierarchical breakdown, illustrating various cooling strategies such as conduction, convection, radiation, phase change, and liquid cooling, further detailing specific techniques under each method.

4.1. Fin

The utilization of fin heat sinks for heat dissipation is a common practice in the industry. This technique is well-suited for cooling lower-power electronic components, such as DC-DC converters, onboard chargers, and other devices that generate moderate heat. Heat is conducted from the electronic device to the fin base, then flows along the fin towards the tip, driven by the temperature gradient. Upon reaching the fin surface, the heat is transferred to the surrounding air through convection. The increased surface area provided by the fins enhances the rate of convective heat transfer, allowing more heat to be removed from the component. The rate of heat transfer enhancement is affected by factors such as the fin surface area, thermal conductivity, shape, and orientation [54].

Heat sinks incorporating pin fins, plate fins, and oblique fins are the three most prevalent types. Among these, the plate-fin heat sink stands out for its simple geometry and cost-effectiveness. Refs. [55,56,57,58,59,60,61] presents some novel analyses of fin configuration and optimization. Here, a brief overview of effective outcomes and a comprehensive analysis of fin design and optimization for thermal improvement, obtained from reviewed papers, is presented in

Table 4. Different analyses on fin design and configuration.

Ref.	Method	Analysis	Outcome
[55]	Numerical	Rectangular, in-line, and staggered fin configurations.	The in-line arrangement provides the lowest junction temperature and highest heat flux. Staggered configuration improves heat transfer compared to normal configuration. Staggered configuration disrupts the boundary layer.
[56]	Experimental and Numerical	Cross-fin heat sink composed of a sequence of long fins and short fins configured perpendicularly.	Better thermal performance considering both the natural convective heat transfer and radiation. Overcome the center cooling problem caused by the minimum air reach at the center of the heat sink. The impinging of air through cross fins enhances local heat transfer. Overall heat transfer coefficient increases up to 11% compared to plate fins.
[57]	Experimental and Numerical	Temperature self-adaptive fin integration.	Shape Memory Alloy is used. Conversion of lattice structure causes actuation fins. Better thermal performance. Dustproof capability.
[58]	Numerical	Tapered fin heat sink.	2° tapered fin exhibits maximum heat transfer. 3° tapered fin shows higher thermal resistance with lower heat transfer due to the stagnation at the fin origin.
[59]	Experimental and Numerical	Two different geometrical perimeter-shaped fin additions.	Rectangular and trapezoidal fins Reduced pressure drop. Interleaved Trapezoidal Fin Module allows extra 10% airflow, enhances the cooling, and reduces the fan power energy consumption.
[60]	Experimental and Numerical	Hybrid heat sink with microchannel and microjet arrays.	The fins are integrated with a microjet called FINJET. Large surface area to volume ratio. Increased fin efficiency.
[61]	Numerical	Hollow hybrid fin optimization.	Comprised of a flat horizontal base plate and a staggered arrangement of hollow hybrid fin array. Aluminum 6063 is used as material. Better thermal performance.
[62]	Numerical	Introduction of Honeycomb structure heat sink.	Aluminum-based honeycomb heat sink structure. For passive cooling by air, a honeycomb sink having three layers of cells provides better heat transfer, along with increased cost.
[63]	Numerical	Integration of metal foam heat sink with pin fins.	More effective in reducing the bottom wall temperature. Higher average Nusselt number. Thermal performance is dependent on porosity and pore density.
[64]	Experimental and Numerical	PCM with honeycomb fin.	Even heat dissipation in the vertical direction. 61% temperature drop compared to heat sink without fins. Lower temperature growth occurs at smaller porosity.
[65]	Experimental and Numerical	Optimization of heat sink size.	Enhanced surface area causes a greater rate of convective heat transfer. Exceeding a threshold in heat sink size does not provide any further benefits in achieving a high heat dissipation rate.

.

4.2. Microchannel

Microchannel heat sinks are an effective way of efficiently eliminating large amounts of heat at moderate temperatures from electronic devices in cooling applications. The heat from the power electronics is first conducted through solid material to the walls of the microchannels. As the coolant flows through the microchannels, it absorbs heat from the channel walls via convection. The small size of the channels increases the surface area-to-volume ratio, leading to a high convective heat transfer coefficient. This significantly improves the cooling efficiency by rapidly removing heat from the component. Multiple microchannel configurations are utilized in heat sinks and studied for efficient thermal management, as presented in [66,67,68,69,70,71,72,73] and later summarized in Table 5. The widespread adoption of hybrid electric vehicles is closely linked to the development of components that possess greater weight-to-volume proportions and are cost-effective to manufacture as well. To accomplish this goal, it is crucial to use novel and enhanced thermal management technologies that decrease both weight and volume while simultaneously improving cooling efficiency.

Conventional heat sink designs use indirect cooling methods, in which a liquid is circulated through a heat sink that has a TIM applied between the heat sink and the semiconductor device. Nevertheless, the heat transfer coefficient and heat transfer area experience a substantial rise as a result of the microchannels’ tiny hydraulic diameter and elevated surface area to volume ratio. The straight square channel is usually a favored option for the development of heat sinks because of its convenient fabrication. However, it typically falls short of achieving the optimum level of heat transfer efficiency owing to the expansion of the thermal boundary layer in the direction of flow and insufficient mixing within the fluid. The significance of those challenges lies in the fact that the heat transfer coefficient is inversely related to the thickness of the thermal boundary layer, which increases in the direction of flow. Furthermore, the mixing of fluids is desirable in improving the occurrence of heat transfer [74]. Incorporating methods such as enhancing flow disturbances, minimizing the thermal boundary layer, and elevating the velocity gradient closest to the wall are effective ways to enhance the heat sink’s thermal management capability. It is feasible to attain a significant rate of heat flux dissipation through the utilization of fluid heating and boiling within microchannels. The two-phase liquid flow, achieved through the interaction of gas and liquid motion within the channel, is commonly employed to accomplish this objective. Flow boiling has greater heat removal capacity and greater heat transfer coefficients than single-phase flow for a specific rate of mass flow [75].

Microchannel cooling strategies can lead to flaws like large pressure drops, increased resistance, manufacturing complexity, etc. In addition, contamination in the cooling fluid can cause clogging issues inside the channel and ultimately cause the cooling system to fail [76].

For applications requiring ultra-high-power density, two-phase cooling offers advantages in terms of cooling efficiency, size, weight, and power consumption. Liquid cooling has gained prominence, particularly for high-power electronics such as insulated gate bipolar transistors (IGBTs) used in inverters and converters [77]. Microchannel cooling is highly effective for compact, high-power-density electronics in electric vehicles (EVs), including inverters and motor controllers.

Table 5 depicts the microchannel cooling strategies along with detailed outcomes that were investigated in the reviewed papers.

Table 5. Different analyses on microchannel cooling.

Ref.	Method	Analysis	Fluid	Outcome
[78]	Numerical	Manifold microchannel heat sink optimization.	50/50 water/ethylene-glycol	Pin fin-type structures disrupt the thermal boundary layer and enhance fluid mixing that lowers the temperature of base plate. Concerning average temperature and pumping power, topology-optimized microchannels performed superior to the typical straight ones.
[67]	Numerical	Optimization of the heat sink with single-layer, double-layer, and double-sided channels with working fluid water and Aluminum oxide.	Water, Water-based Al2O3	The sandwich structure shows better thermal resistance reduction up to 59% whereas it was 15% for the double-layer channels. Uniform temperature distribution in the sandwich structure. Al2O3–water-based nanofluid shows better heat transfer performance. Nanofluids with a higher concentration result in better cooling.
[68]	Experimental	Hierarchical microchannel heat sink.	HFE-7100	Significant pressure drop in the boiling region. Increased pressure with rising flow rate in both single-phase and two-phase regions. The maximum volumetric heat dissipation is 2870 W/cm3 for a 9 × 9 array manifold. The maximum volumetric heat dissipation is 285 W/cm3 for a 3 × 3 array manifold. Increase in pressure drop due to compact manifold design.
[69]	Numerical	Self-adaptive microchannel pin fin heat sink.	Deionized Water	Overcomes the inability to control coolant flow during different heating conditions. Thermal-sensitive PNIPAM Hydrogel is introduced at the microchannel outlet around the pin fins. Intelligent behavior and volume shrinkage with the increase in regional temperature. The maximum temperature can be reduced by 12.2 K with a pressure drop of 34 kPa.
[70]	Numerical	Addition of pillar structure in the microchannel.	Deionized Water	Pillar heights up to 260 µm in the microchannel disturbed the boundary layer. Enhanced mixing of fluid particles with an elevation in heat transfer. A significant pressure drop up to 5.5 kPa is noticed due to the pillars in the channel. Elevated pump power consumption.
[71]	Experimental and Numerical	Topology optimization of spider web heat sink.	Water	Improved flow and thermal performance compared to conventional spider web sinks. The temperature difference in substrate can be lowered by 57.35% compared to conventional spider web heat sink.
[72]	Numerical	Modification of microchannel heat sink incorporating secondary flow channels.	Water	Two secondary channel configurations of regular trapezoidal and parallel with different inlet-outlet flow configurations (I, C, and Z types) are investigated. “I” type configuration provides better flow uniformity. Secondary flow channels enhance heat absorption and temperature uniformity through the heat sink. Regular trapezoidal channels show optimum performance.
[73]	Experimental and Numerical	Microchannel heat sink with rhombus fractal network.	Water	Enlargement of the cross-sectional area results in a decrease in pressure drop. RMCHS (Rhombus Microchannel Heat Sink) covers a larger area, and thus, temperature uniformity is obtained. RMCHS is found to be superior concerning Nusselt number and reduced thermal resistance compared to typical heat sinks with parallel microchannels.
[79]	Experimental	Unidirectional porous heat sink.	Distilled Water	Porous media made of oxygen-free copper is used as a heat sink. Separated passages are used for liquid supply and vapor discharge. Enhanced evaporative heat transfer.
[80]	Numerical	Rectangular microchannel modification in three ways (Front loose back compact, uniformly distributed, and Front compact back loose) with internal spoiler cavities.	Water	Creation of a jetting and throttling effect Reduction in pressure in the three configurations compared to the normal rectangular channel. Enhanced heat transfer. Microchannel with front loose back compact exhibits the lowest maximum temperature for the same inlet velocity.
[81]	Numerical	Interrupted microchannel heat sink with no ribs, rectangle ribs, triangle ribs, and trapezoid ribs.	Water	The rectangular rib is superior in terms of Nusselt number and friction factor increment. The trapezoidal shape exhibits better overall heat transfer performance. The reduced chamfer of the ribs significantly improves the efficiency of heat transfer.
[82]	Experimental and Numerical	Addition of a combined delta winglet composed of three delta winglets in a rectangular channel.	Water	Combined winglets have better heat transfer enhancement. The friction factor shows a higher value for combined delta winglets. Optimum performance is found when the spacing is zero at a rotation angle of 60°.
[83]	Experimental	Additive Manufacturing-based multi-passed microchannel heat sink.	Water	Reduced thermal resistance and enhanced pressure drop compared to conventionally machined and Diffusion-bonded heat sink. Higher surface roughness. Up to 46% of thermal performance increases in the case of an additively manufactured heat sink compared to its traditionally machined counterpart.
[84]	Numerical	Application of side wall staggered ribs in the microchannel.	Water	Significant pressure drop compared to straight microchannel heat sinks. Reduced rib pitch, increased rib dimension, and hydraulic diameter increase heat transfer and elevate the pressure drop.
[85]	Experimental and Numerical	Combining the Si-glass microchannel heat sink.	Deionized Water	High thermal conductivity compared to conventional passive cooling. Dissipate heat flux greater than 150 W/cm2.
[86]	Numerical	Implementation of Y-shaped symmetric and asymmetric bionic fractal networks.	Deionized Water	Asymmetric branching exhibits better temperature uniformity. The symmetric configuration shows a greater pressure drop at the same fractal number. An increase in the fractal number improves temperature uniformity for both networks.
[87]	Experimental and Numerical	Implementation of V-shaped rib with different cross-sections in rectangular low channel.	Water	V-shaped ribs having semicircular, square, trapezoidal, and triangular cross-sections are introduced. Higher heat transfer performance and pump power consumption than the smooth channel and straight rib. Performance enhancement due to the generation of secondary flow behind the rib and boundary layer disruption.
[88]	Numerical	Transverse ribs (rectangular shape) in open microchannel.	Deionized Water	Ribs in the flow path disturb the flow regime. Flow recirculation behind the ribs. Breaking the thermal boundary layer and enhanced flow mixing with higher velocity. Increased heat transfer at the cost of pressure drop compared to open microchannels without ribs. Both the Nusselt number and the heat transfer coefficient increased gradually with the increasing number of ribs.
[89]	Numerical	Circular re-entrant cavity microchannels and sinusoidal wavy microchannels.	Water	The largest pressure drop in the sinusoidal microchannel. Circular re-entrant microchannels outperform the sinusoidal channels in the hybrid performance evaluation criterion.
[90]	Numerical	Addition of different types of ribs and cavities in microchannel heat sink.	Deionized Water	Pressure drops in the flow system due to the cavities and ribs. Reduced pressure drops and increased fluid mixing due to secondary flow channels. Elliptical rectangular ribs exhibit the highest heat dissipation performance with a maximum comprehensive performance evaluation of 1.73.
[91]	Numerical	Initialization of decreasing height bifurcated plate in microchannel.	Water	The heat transfer coefficient increases with the increase in Reynolds number. Generation of the vortices around the plate wall along with the upper surface of the decreasing height bifurcated plate. Decreasing the height of the bifurcated plate heat sink displays the maximum heat transfer coefficient.
[74]	Numerical	Wavy microchannel with varying wavelength.	Water	Horizontal MCHS produces symmetrical vortices. Enhanced heat transfer compared to vertical and straight channels.
[92]	Experimental and Numerical	SiC microchannels with high aspect ratio and decreased hydraulic diameter.	Deionized Water	Low thermal resistance of 0.024 cm2 C°/W is attained at 1 kW/cm2 heat flux.
[93]	Experimental and Numerical	Embedded cooling with Si fabricated 3D manifold microchannel.	Water	A surge in flow rate leads to a decline in the temperature of the heated surface, as well as a reduction in thermal resistance. 34% of the total pressure drop occurs at the inlet because of contraction, and 5% of the total pressure drop exit sections because of expansion.
[66]	Experimental	Integration of straight microchannel and unit cell microchannel with jet impingement.	Water	Larger pressure drop in straight microchannel compared to the unit cells with jet impingement. The maximum pressure drop for the straight channel is 39.8 kPa at a flow rate of 95 mL/min. The maximum pressure drop in the unit cell with the jet is 81.2 kPa at 450 mL/min. The shorter length of the unit cell and arrayed jet orifices contribute to better thermal performance.
[94]	Experimental	CTE-matched two-phase mini-channel heat Sink.	R-245fa	AlN provides lower thermal resistance compared to TIM. Monotonic decrease in thermal resistance due to increased heat load.
[95]	Experimental	Two-pass diverging microchannel and skived microchannel heat sink.	HFC-245fa	R-245fa refrigerant used. 30 to 48% higher pressure drop in the skived channel compared to the diverging channel. Clogging causes increased thermal resistance with increasing heat transfer rate in the skived channel. The diverging channel provides reduced thermal resistance. The partial dry-out effect is minimized in the diverging channel. Total heat transfer in the skived channel is 2.35 times higher than in diverging channel.
[96]	Experimental	Two-phase additively manufactured MCHS with multiple geometric features.	Water	Maximum 900 W heat dissipation with 60 °C inlet secondary coolant temperature. High flexibility. Superior control attributes for thermal hunting and flow avoidance.
[76]	Experimental	Embedded microchannels with 3D manifold.	Water, R-245fa	R245fa as working fluid. Low thermal resistance around 0.07 K/W.

Table 5. Different analyses on microchannel cooling.

Ref.	Method	Analysis	Fluid	Outcome
[78]	Numerical	Manifold microchannel heat sink optimization.	50/50 water/ethylene-glycol	Pin fin-type structures disrupt the thermal boundary layer and enhance fluid mixing that lowers the temperature of base plate. Concerning average temperature and pumping power, topology-optimized microchannels performed superior to the typical straight ones.
[67]	Numerical	Optimization of the heat sink with single-layer, double-layer, and double-sided channels with working fluid water and Aluminum oxide.	Water, Water-based Al2O3	The sandwich structure shows better thermal resistance reduction up to 59% whereas it was 15% for the double-layer channels. Uniform temperature distribution in the sandwich structure. Al2O3–water-based nanofluid shows better heat transfer performance. Nanofluids with a higher concentration result in better cooling.
[68]	Experimental	Hierarchical microchannel heat sink.	HFE-7100	Significant pressure drop in the boiling region. Increased pressure with rising flow rate in both single-phase and two-phase regions. The maximum volumetric heat dissipation is 2870 W/cm3 for a 9 × 9 array manifold. The maximum volumetric heat dissipation is 285 W/cm3 for a 3 × 3 array manifold. Increase in pressure drop due to compact manifold design.
[69]	Numerical	Self-adaptive microchannel pin fin heat sink.	Deionized Water	Overcomes the inability to control coolant flow during different heating conditions. Thermal-sensitive PNIPAM Hydrogel is introduced at the microchannel outlet around the pin fins. Intelligent behavior and volume shrinkage with the increase in regional temperature. The maximum temperature can be reduced by 12.2 K with a pressure drop of 34 kPa.
[70]	Numerical	Addition of pillar structure in the microchannel.	Deionized Water	Pillar heights up to 260 µm in the microchannel disturbed the boundary layer. Enhanced mixing of fluid particles with an elevation in heat transfer. A significant pressure drop up to 5.5 kPa is noticed due to the pillars in the channel. Elevated pump power consumption.
[71]	Experimental and Numerical	Topology optimization of spider web heat sink.	Water	Improved flow and thermal performance compared to conventional spider web sinks. The temperature difference in substrate can be lowered by 57.35% compared to conventional spider web heat sink.
[72]	Numerical	Modification of microchannel heat sink incorporating secondary flow channels.	Water	Two secondary channel configurations of regular trapezoidal and parallel with different inlet-outlet flow configurations (I, C, and Z types) are investigated. “I” type configuration provides better flow uniformity. Secondary flow channels enhance heat absorption and temperature uniformity through the heat sink. Regular trapezoidal channels show optimum performance.
[73]	Experimental and Numerical	Microchannel heat sink with rhombus fractal network.	Water	Enlargement of the cross-sectional area results in a decrease in pressure drop. RMCHS (Rhombus Microchannel Heat Sink) covers a larger area, and thus, temperature uniformity is obtained. RMCHS is found to be superior concerning Nusselt number and reduced thermal resistance compared to typical heat sinks with parallel microchannels.
[79]	Experimental	Unidirectional porous heat sink.	Distilled Water	Porous media made of oxygen-free copper is used as a heat sink. Separated passages are used for liquid supply and vapor discharge. Enhanced evaporative heat transfer.
[80]	Numerical	Rectangular microchannel modification in three ways (Front loose back compact, uniformly distributed, and Front compact back loose) with internal spoiler cavities.	Water	Creation of a jetting and throttling effect Reduction in pressure in the three configurations compared to the normal rectangular channel. Enhanced heat transfer. Microchannel with front loose back compact exhibits the lowest maximum temperature for the same inlet velocity.
[81]	Numerical	Interrupted microchannel heat sink with no ribs, rectangle ribs, triangle ribs, and trapezoid ribs.	Water	The rectangular rib is superior in terms of Nusselt number and friction factor increment. The trapezoidal shape exhibits better overall heat transfer performance. The reduced chamfer of the ribs significantly improves the efficiency of heat transfer.
[82]	Experimental and Numerical	Addition of a combined delta winglet composed of three delta winglets in a rectangular channel.	Water	Combined winglets have better heat transfer enhancement. The friction factor shows a higher value for combined delta winglets. Optimum performance is found when the spacing is zero at a rotation angle of 60°.
[83]	Experimental	Additive Manufacturing-based multi-passed microchannel heat sink.	Water	Reduced thermal resistance and enhanced pressure drop compared to conventionally machined and Diffusion-bonded heat sink. Higher surface roughness. Up to 46% of thermal performance increases in the case of an additively manufactured heat sink compared to its traditionally machined counterpart.
[84]	Numerical	Application of side wall staggered ribs in the microchannel.	Water	Significant pressure drop compared to straight microchannel heat sinks. Reduced rib pitch, increased rib dimension, and hydraulic diameter increase heat transfer and elevate the pressure drop.
[85]	Experimental and Numerical	Combining the Si-glass microchannel heat sink.	Deionized Water	High thermal conductivity compared to conventional passive cooling. Dissipate heat flux greater than 150 W/cm2.
[86]	Numerical	Implementation of Y-shaped symmetric and asymmetric bionic fractal networks.	Deionized Water	Asymmetric branching exhibits better temperature uniformity. The symmetric configuration shows a greater pressure drop at the same fractal number. An increase in the fractal number improves temperature uniformity for both networks.
[87]	Experimental and Numerical	Implementation of V-shaped rib with different cross-sections in rectangular low channel.	Water	V-shaped ribs having semicircular, square, trapezoidal, and triangular cross-sections are introduced. Higher heat transfer performance and pump power consumption than the smooth channel and straight rib. Performance enhancement due to the generation of secondary flow behind the rib and boundary layer disruption.
[88]	Numerical	Transverse ribs (rectangular shape) in open microchannel.	Deionized Water	Ribs in the flow path disturb the flow regime. Flow recirculation behind the ribs. Breaking the thermal boundary layer and enhanced flow mixing with higher velocity. Increased heat transfer at the cost of pressure drop compared to open microchannels without ribs. Both the Nusselt number and the heat transfer coefficient increased gradually with the increasing number of ribs.
[89]	Numerical	Circular re-entrant cavity microchannels and sinusoidal wavy microchannels.	Water	The largest pressure drop in the sinusoidal microchannel. Circular re-entrant microchannels outperform the sinusoidal channels in the hybrid performance evaluation criterion.
[90]	Numerical	Addition of different types of ribs and cavities in microchannel heat sink.	Deionized Water	Pressure drops in the flow system due to the cavities and ribs. Reduced pressure drops and increased fluid mixing due to secondary flow channels. Elliptical rectangular ribs exhibit the highest heat dissipation performance with a maximum comprehensive performance evaluation of 1.73.
[91]	Numerical	Initialization of decreasing height bifurcated plate in microchannel.	Water	The heat transfer coefficient increases with the increase in Reynolds number. Generation of the vortices around the plate wall along with the upper surface of the decreasing height bifurcated plate. Decreasing the height of the bifurcated plate heat sink displays the maximum heat transfer coefficient.
[74]	Numerical	Wavy microchannel with varying wavelength.	Water	Horizontal MCHS produces symmetrical vortices. Enhanced heat transfer compared to vertical and straight channels.
[92]	Experimental and Numerical	SiC microchannels with high aspect ratio and decreased hydraulic diameter.	Deionized Water	Low thermal resistance of 0.024 cm2 C°/W is attained at 1 kW/cm2 heat flux.
[93]	Experimental and Numerical	Embedded cooling with Si fabricated 3D manifold microchannel.	Water	A surge in flow rate leads to a decline in the temperature of the heated surface, as well as a reduction in thermal resistance. 34% of the total pressure drop occurs at the inlet because of contraction, and 5% of the total pressure drop exit sections because of expansion.
[66]	Experimental	Integration of straight microchannel and unit cell microchannel with jet impingement.	Water	Larger pressure drop in straight microchannel compared to the unit cells with jet impingement. The maximum pressure drop for the straight channel is 39.8 kPa at a flow rate of 95 mL/min. The maximum pressure drop in the unit cell with the jet is 81.2 kPa at 450 mL/min. The shorter length of the unit cell and arrayed jet orifices contribute to better thermal performance.
[94]	Experimental	CTE-matched two-phase mini-channel heat Sink.	R-245fa	AlN provides lower thermal resistance compared to TIM. Monotonic decrease in thermal resistance due to increased heat load.
[95]	Experimental	Two-pass diverging microchannel and skived microchannel heat sink.	HFC-245fa	R-245fa refrigerant used. 30 to 48% higher pressure drop in the skived channel compared to the diverging channel. Clogging causes increased thermal resistance with increasing heat transfer rate in the skived channel. The diverging channel provides reduced thermal resistance. The partial dry-out effect is minimized in the diverging channel. Total heat transfer in the skived channel is 2.35 times higher than in diverging channel.
[96]	Experimental	Two-phase additively manufactured MCHS with multiple geometric features.	Water	Maximum 900 W heat dissipation with 60 °C inlet secondary coolant temperature. High flexibility. Superior control attributes for thermal hunting and flow avoidance.
[76]	Experimental	Embedded microchannels with 3D manifold.	Water, R-245fa	R245fa as working fluid. Low thermal resistance around 0.07 K/W.

4.3. Jet Impingement

Active cooling approaches with pumped liquid loops are frequently investigated in micro- or mini-channel flow, spray impingement, and jet impingement. Jet impingement cooling is widely regarded as the most efficient method for electronics that necessitate the removal of a significantly high heat flux. This methodology permits the attainment of heat transfer coefficients exceeding 30,000 W/m²K [97]. An impinging jet’s geometry in a conventional jet impingement cooling solution module generally consists of a nozzle directed at a flat surface at a precise separation [98]. The fluid is expelled from the nozzle at a certain speed, which is directed towards the surface to be cooled. In the stagnation region, the jet first contacts the heated surface, where intense convective heat transfer occurs. The kinetic energy of the jet converts to thermal energy, rapidly cooling the surface. As the jet moves away from the impact point, it begins to spread and slow down, leading to a region where heat transfer efficiency gradually decreases. As the jet interacts with the surface, a boundary layer develops, where the fluid velocity decreases from the jet velocity to zero at the wall. The thickness of this boundary layer affects heat transfer rates, with thinner layers leading to higher heat transfer coefficients [13].

In certain scenarios, conventional convection cooling systems might not suffice as they could result in localized “hotspots” of increased heat flux due to specific configurations of electronics. In such cases, it becomes imperative to employ more focused cooling techniques for effective thermal management. One notable benefit of jet impingement is its capacity to easily modify the target spot. Moreover, the narrow thermal and concentration boundary layers encircling the stagnation region enable it to extract a considerable quantity of mass and heat from the impinging surface. The efficacy of impingement cooling is significantly influenced by geometrical factors, including jet array arrangement, jet configuration, impingement plate elevation, and jet-to-target gap [99]. An illustration of the fluid flow path and different flow interaction regions of jet impingement cooling can be found in Figure 1 of [100].

When exposed to sufficient heat, a jet impingement cooler initiates bubble formation, which allows it to function as a two-phase system. This process leads to the emergence of two fluid zones when the jet impinges on the heated surface. The region where the liquid spreads from the point of impact is known as the flow region, while the impingement zone is where the jet strikes the surface. The disparity in the distribution of cooling potential leads to a significantly greater heat transfer coefficient, as well as following heat elimination in the impingement zone in comparison to other areas.

One of the major limitations of jet impingement cooling is its tendency to cause vigorous boiling that induces premature wall jet separation from the heated surface, due to the flow distribution of impinging jets. In contrast to spray cooling, the liquid layer formed on the heated surface is attached to the surface only at the impingement region, increasing its vulnerability to deterioration and detachment from the remaining impingement surface. As a result, the heated surface experiences dry-out situations, which may lead to burnout of the power component. Jet impingement cooling is a cooling technique that includes high-speed impacts. This might pose a risk of damaging sensitive components, making the application of impingement cooling more complicated [101].

Jet impingement cooling is a method that provides high heat transfer rates by focusing on specific localized hot spots. This technique is ideal for power electronics, such as IGBTs and MOSFETs, that experience concentrated heating in inverters and converters, where conventional cooling may not suffice for high-performance EVs.

A comprehensive description of SiC power module equipped with a liquid jet cooling mechanism is presented in [102], whereas [103] describes single and multiple jet configurations in detail. These two, along with other jet impingement cooling strategies investigated by different authors are listed in

Table 6. Different analyses on jet impingement cooling.

Ref.	Method	Analysis	Outcome
[99]	Experimental	Bare die 3D printed thermal packaging with jet impingement.	Achieves a chip temperature reduction by a factor of 4 and 5 concerning the lidded package and bare die in comparison to standard air cooling. The bare die package outperforms the lidded package by 44%. Three-dimensional-printed microfluidic heat sink reduces cooler size and weight.
[100]	Experimental and Numerical	Submerged staggered liquid jet arrays featuring various discharge manifolds.	Decreasing pitch causes jet degradation to occur more rapidly. Heat transfer increases with increasing manifold angle and decreasing pitch values.
[102]	Experimental and Numerical	Jet impingement cooled heat exchanger designed by subjecting the module base plate to an array of nozzles with a diameter of 200 µm.	The SiC JFET device’s junction temperature drops to 169 °C with optimum jet impingement cooling. In comparison, cold plate cooling yields a junction temperature of 290 °C, whereas micro-channel cooling has 215 °C. Jet impingement cooling increases power dissipation to 167 W from 60 W with a cold plate and 99 W with a micro-channel cooler at 195 cc/min.
[103]	Numerical	Utilization of porous inserts (metal foam).	Porous inserts improve the cooling effectiveness and temperature control. The porous substrate provides a large surface area.
[104]	Experimental	Heat transfer and fluid flow characteristics of water jet impingement on a flat plate.	Nusselt number and hydraulic jump diameter follow the increasing trend when the H/D is below 0.4, named as the jet deflection region. Negligible effect on Nu when the H/D is within the region 0.4 to 1, termed an inertia-dominant region.
[105]	Experimental	Synthetic jet impingement with four different aperture cases.	Four aperture configurations: open, closed, 20 mm intermediate diameter, and contracting diameter, are investigated with varying jet ratio H/D for each case, which proved itself as an influential factor on overall thermal performance. Due to the re-entrainment of warmer fluid back into the synthetic jet orifice, the variable diameter jet exhibits a lower heat transfer performance at H/D = 0.5. A reliable increase in heat transfer is obtained through variable jet diameter within H/D = 1 to 3.
[106]	Experimental	Confined two-phase jet impingement.	HFE 7100 used as coolant. In the fully boiling regime, nucleate boiling is the primary method of heat transfer. Critical heat flux increases significantly with the increased jet velocity.
[107]	Numerical	Combination of air jet impingement and cross-coolant flow.	Three air jet cross-sections (square, streamwise rectangular, spanwise rectangular) with in total of eight different configurations with hotpot are investigated. The streamwise cross-section provides better cooling at the hotspot. The square cross-section provides better temperature uniformity.
[108]	Numerical	Direct contact jet impingement with ATF (Automatic Transmission Fluid).	Eliminates heat sink and interface materials. Better cooling scheme with a reduced volume of coolant and increased power density.
[16]	Numerical	Implementation of single-side and double-side jet impingement cooling with manifold design.	The increased distance between the jet and the impinge plate hurt thermal performance. A decrease in nozzle diameter led to reduced temperature due to the generation of high jet velocity. A large number of nozzles helped in pressure drop reduction. Circular configuration is more effective. The double-sided liquid jet is superior in temperature reduction.
[109]	Numerical	Jet in cross flow for hotspot cooling.	The case of pure jet impingement (PJI) provides the lowest central stream-wise temperature distribution, but at the maximum pressure drop. JICF (jet in crossflow) configuration results in a lower temperature distribution compared to PCF (Pure crossflow) at a lower pressure drop. An inclined jet with crossflow enables a smooth mixing of fluid, which helps to reduce the pressure drop to a large extent.
[110]	Experimental and Numerical	Surface augmentation in jet impingement fountain region.	The streamwise ribs having triangular cross-sections are superior in terms of surface temperature reduction. Offset cones are also beneficial compared to transverse ribs by promoting secondary fountain development and shielding the downstream jet. Transverse ribs with angled walls become detrimental due to the crossflow effect and large pressure drop.
[13]	Experimental and Numerical	Investigation of different outlet arrangements and pin-fin layouts.	Reduced flow distance across all exterior jets after impacting the target plate reduces flow resistance in a two-way opening setup. Removing the pin fins right below the jets can improve heff by 4.5%. Staggered pin-fin causes minor improvement of heff; however, p can be reduced by 11.1%. heff can be further elevated by 8.8% incorporating a shorter jet-to-target distance. Applying a four-way opening configuration and a greater central jet concurrently reduces the local Nusselt number non-uniformity to 1.8%.
[111]	Experimental and Numerical	Implementation of single-phase jet impingement cooling directly at the power electronic substrate layer to enhance thermal management.	Up to 25% lower thermal resistance than conventional packaging. Reduced number of jets and increased jet spacing enhance the heat transfer coefficient.
[112]	Numerical	Effect of flow pulsations on local, time-averaged Nusselt number of an impinging air jet.	Pulsated impinging jet enhances the performance of impinging jets.
[98]	Experimental and Numerical	Influence of geometric and flow factors on the heat transfer properties of both steady and unsteady jets.	Optimal cooling is achieved for both steady and unsteady jets when the jet spacing to diameter ratio (Z/D) is 5. The stagnation heat transfer coefficient increases 25% when an unstable jet is utilized instead of a steady jet at Reynolds number of 20,000.

with effective outcomes.

4.4. Phase Change Material

Efficient thermal management is a substantial challenge in the implementation and reduction in the size of electrical devices with a high concentration of power. A high-power density electronic device that is used intermittently necessitates the adoption of an effective passive cooling method, such as a finned heat sink based on phase change material (PCM). PCMs act as thermal energy storage devices by changing their physical state from solid to liquid depending on the melting temperature. PCMs absorb a large amount of latent heat energy and thus perform the cooling task when the temperature exceeds the melting point of PCMs.

Table 7. Melting temperature of different phase change materials for EV power electronics cooling [113,114].

PCM	Melting Temperature (°C)	Thermal Conductivity (W/mK)	Heat of Fusion (KJ/kg)
RT 11	10–12	0.2	160
Capric acid + Lauric acid	18	0.143	120
Paraffin C16–C18	20–22	0.2	152
RT 27	25–28	0.2	149
Paraffin (RT44HC)	41–44	0.2	250
Sodium Acetate Trihydrate	58–62	0.82	269
RT 70	70	0.2	230

lists some available phase change materials for power electronics cooling with their melting temperature [113].

Designing a thermal management system using PCMs is crucial, as it requires proper PCM selection. Factors like high latent heat, high thermal conductivity, chemical stability, a small change in volume, small vapor pressure, congruent melting, corrosiveness, etc., have a profound impact on PCM selection [113].

Deng et al. [15] experimented to analyze the transient thermal performance of finned heat sinks using PCMs to manage the heat generated by electronic devices. The research highlights the significance of heat load and the volumetric proportion of material PCM.

The findings of this study can be highly beneficial for the advancement of thermal management techniques in electronics. Integrating fins into heat sink designs that use phase change materials has an effective impact on the overall thermal efficiency of electronics. It has been observed that an augmentation in the quantity of fins leads to a reduction in the operational temperature, which translates to longer-lasting devices that operate within acceptable temperature limits. The addition of fins improved the melting rate and temperature distribution in comparison to heat sinks without fins. Furthermore, the number of fins has been found to have a direct correlation with the enhancement ratio and thermal capacity of PCM-based heat sinks. Higher input heat fluxes correspond to higher operating temperatures and shorter device life at acceptable temperatures. Increasing the volumetric fraction of PCM has been shown to enhance thermal storage capacity and reduce temperature levels, thereby resulting in enhanced ratios and extended operating times are the characteristics of PCM-based heat sinks with more fins.

Hasan et al. [115] investigated the micro pin finned heat sink, which consists of various fin shapes, including circular, triangular, and square, in addition to an unfinned micro heat sink. The study further explores the use of phase change materials (PCMs) with different types and configurations as a cooling medium, after the initial use of air in the heat sink. The results indicate that the use of PCMs in heat sinks leads to a reduction in temperature compared to traditional air cooling. Moreover, the circular fins proved to be more effective in reducing temperature compared to triangular and square fins.

Marri et al. [116] conducted a study on a heat sink that utilizes a phase change material, complemented by a compact heat pipe to improve its thermal performance. The study found that by placing the evaporator of the heat pipe in symbiosis with the PCM in the heat sink, a substantial improvement in efficacy is attainable. The study revealed that while performing the discharging cycle, a PCM-filled heat sink featuring a heat pipe demonstrated superior performance compared to a heat sink featuring radial fins and a central stem. At power levels of 6, 8, and 10 W, the time required to return the heat sink to near-ambient temperature differed by 28%, 126%, and 175%, respectively.

To improve the thermal conductivity of PCMs utilized in thermal energy storage (TESs) and thermal management systems (TMSs), metal foams have been implemented extensively. The advent of additive manufacturing technology has enabled the fabrication of intricate and complex architectural structures with ease, presenting an opportunity to utilize numerous other foam and structure candidates in TESs and TMSs. Qureshi et al. [117] utilized three triply periodic minimal surfaces (TPMS)-based foams, namely Gyroid, IWP, and Primitive, in a finned metal foam-PCM (FMF-PCM) system. Figure 6 illustrates the 3D view of the finned metal foam PCM-based heat sink. In pure conduction, the IWP foam outperformed the Gyroid, Primitive, and Kelvin cells. Primitive performed best under natural convection, followed by Gyroid, IWP, and Kelvin. Pure conduction performance was better for TPMS foams than Kelvin cells when compared to natural convection.

A significant constraint on the implementation of phase change materials (PCMs) in passive thermal management systems is their inadequate thermal conductivity. To address this challenge, Liu et al. [64] proposed a novel honeycomb fin that can effectively retain the temperature. A cooling system utilizing a PCM-based honeycomb structure is illustrated in Figure 7. The implementation of honeycomb fins facilitates uniform heat distribution of phase change materials (PCMs) in the vertical direction. This design enables PCMs with varying thicknesses to commence melting at nearly the same time, resulting in optimized heat absorption of the PCM. The cooling plate thickness plays a significant role in reducing the cell temperature, and the melting liquid phase fraction of PCM decreases with increasing thickness. Overall, this innovative honeycomb fin design has the potential to significantly enhance the performance and safety of passive thermal management systems.

Phase change materials (PCMs) can manage transient heat loads by absorbing large amounts of thermal energy. This method is especially useful for components that experience intermittent high heat loads, such as those involved in regenerative braking systems or in conditions where power demand fluctuates rapidly.

4.5. Nanofluid

Nanofluids have emerged as a promising cooling technology owing to their exceptional heat transfer properties that enable efficient thermal management of devices with high power density and heat flow. Nanofluids are the dispersion of nanomaterials in the base fluid. The suspended nanoparticles enhance the thermal conductivity of the base fluid with more stability. The random motion of nanoparticles (Brownian motion) contributes to the increased effective thermal conductivity. The thermal performance of nanofluid is also affected by the higher surface-to-volume ratio [118].

Tan et al. [119] examined the characteristics of the flow and thermal conductivity of nanofluids consisting of water and Al₂O₃ particles in microchannels. The analysis is conducted using the Eulerian–Lagrange model. The findings indicated that the application of nanofluids may greatly improve the efficiency of microchannel heat sinks in facilitating heat transfer. An increase in the volume percentage of nanofluids leads to an improvement in the performance of microchannel heat sinks. Nanofluids have enhanced heat transmission capabilities compared to water, but with a higher pressure drop. The dispersion of nanoparticles in the three-dimensional flow field is heterogeneous, with a higher concentration of particles near the wall.

Zhang et al. [120] suggested a hybrid system, comprising a passive ferrofluid (FF) cooling system and a housing water jacket. The system utilizes an oil-based liquid infused with nano-sized ferromagnetic particles.

Olmo et al. [121] investigated the effect of magnetic nanoparticles on the cooling capacity of natural ester through prepared samples of ferrofluid by adding maghemite to the natural ester at varying concentrations. The results showed that the nanofluid exhibited superior behavior to the base fluid, with lower temperature gradients relative to the ambient temperature. The largest improvement was observed at the maximum load regime, with a reduction in temperature gradient of up to 11.2%. Moreover, nanoparticles increased the breakdown voltage of natural ester, leading to an improved insulating capacity of the base fluid. The findings suggest that the incorporation of maghemite nanoparticles boosts the cooling and insulating characteristics of natural ester, hence enhancing its overall performance.

Vaishnav et al. [122] investigated the impact of ZrO₂ concentration (g/L) on the electrical, thermal, and rheological properties of nanofluids. The findings revealed that nanofluids with ZrO₂ demonstrated improved cooling, insulation, and breakdown voltage (BDV) compared to pure transformer oil. The optimal performance of the nanofluid was observed at a concentration of 0.2 g/L of ZrO₂ nanoparticles in the transformer oil. The BDV of the nanofluid was found to increase substantially as compared to pure oil, with a maximum BDV of 21.53 kV achieved at a 0.2 g/L ZrO₂ nanoparticle concentration. The study also showed that the thermal conductivity and viscosity of the samples increased with increasing ZrO₂ nanoparticle concentration. Specifically, the thermal conductivity rose from 0.167 to 0.259 W/mK.

Sakanova et al. [123] investigated multiple nanofluids, including diamond–water, SiO₂–water, and CuO–water in wavy and traditional microchannels. The results indicate that wavy microchannel heat sinks (MCHSs) perform better in cooling compared to traditional rectangular MCHSs when using nanofluids and pure water as coolants. Shorter wavelengths and higher amplitudes lead to lower thermal resistance. Conversely, the thermal resistance increases with decreasing wavelength and increasing amplitude. Increasing the amplitude and flow rate causes a significant surge in pressure drop. Diamond-water nanofluid exhibits the best performance compared to other nanofluids. Interestingly, the study results indicate that a higher concentration of nanofluids in a wavy channel does not necessarily result in superior performance compared to a traditional channel.

Tharayil et al. [124] aimed to investigate the impact of graphene nanofluid on the thermal performance of a miniature loop heat pipe (mLHP) featuring a nanoparticle-coated evaporator. The study compared the experimental results of mLHPs with plain and nanoparticle-coated evaporators using distilled water as the working fluid. The findings indicate that the combination of nanoparticle coating and nanofluid yields the highest heat transfer as compared to the mLHP. The utilization of nanofluid amplifies the heat transfer performance of mLHP by reducing the thermal resistance by an average of 45.2% and enhancing the evaporator heat transfer coefficient by an average of 113.4% for the optimum nanofluid volume concentration of 0.006%. Moreover, the research revealed enhancements in the thermal effectiveness of the heat pipe while using an evaporator coated with nanoparticles and nanofluid.

Kumar et al. [125] examined the effects of a hybrid nanofluid comprising Alumina and MWCNT (Multi-walled Carbon Nanotube) at different mixture volume ratios while maintaining a constant concentration. The study focused on heat transfer coefficient and pressure drop ratio, with a particular emphasis on determining the optimal performance of the alumina and MWCNT mixture at a 3:2 ratio. The study revealed that the concentration of MWCNT significantly impacts the heat transfer coefficient, with a 44% increase observed at the MWCNT (5:0) ratio. The evaluation of performance criteria (PEC) revealed that all the nanofluids had a PEC greater than 1, indicating that nanofluids are a better option than conventional fluid (DI water).

Patel et al. [126] investigated the preparation and performance analysis of two distinct nanofluid mixtures, SiO₂ with water and SiO₂ with EG. Their study focused on analyzing the density, viscosity, thermal conductivity, and specific heat of different mixtures. The results indicated an increase in density, viscosity, and thermal conductivity with the particle concentration, except for specific heat. The stability of the nanofluid was found to be a function of the particle size. The results showed that the thermal conductivity of nanofluids improved with increasing nanoparticle volume fraction. Furthermore, the analysis revealed that EG had better thermal conductivity and performance than deionized H₂O.

Mansour et al. [127] studied the thermal characteristics of oil-based nanofluids, utilizing nanoparticles of Al₂O₃, TiO₂, and SiO₂ at four different weight fractions. Results showed that SiO₂ at 0.01 g/L had the highest heat transfer coefficient. However, increased weight fraction led to reduced heat transfer coefficient, attributed to Brownian motion caused by reduced velocity of larger particle agglomerations. The study also found that the viscosity of nanofluids was less dependent on shear rate, behaving as a Newtonian fluid.

Shill et al. [128] analyzed the insulation and cooling performance of transformer oil-based nanofluids. The nanofluids were prepared using SiO₂ nanoparticles with different concentration ratios of 0.1, 0.3, 0.5, and 0.7 to the base oil. The research revealed that the thermal conductivity of nanofluid increased in correlation with the augmentation of nanoparticle concentration within the base oil. More precisely, the thermal conductivity of nanofluid showed a 5% increase when the percentage of its concentration was 0.7% in mineral oil and a 10% rise when it was present in MIDEL-7131.

Turgut et al. [129] studied the thermal efficiency of an electronic cooling system. The study was carried out utilizing a nanofluid composed of water and Alumina (Al₂O₃) particles embedded in water. The findings indicate that including nanofluids with a modest volume concentration (1% vol.) of Alumina nanoparticles may significantly decrease the maximum temperature of the system by around 2.7 °C, in comparison to using water alone.

Nanofluids, which enhance the heat transfer properties of liquid cooling systems, are beneficial for high-heat-flux applications, making them suitable for critical power electronics like high-power inverters and battery thermal management systems.

The choice of nanofluid for a practical application relies on factors such as thermal conductivity, convective heat transfer efficiency, lower viscosity, nanoparticle dispersion stability, lower pumping power, pressure drop, corrosivity, and toxicity. Nanofluid with aluminum oxide (Al₂O₃) provides higher thermal conductivity with a moderate increase in viscosity in electronic chip cooling. TiO₂ nanofluid shows higher thermal conductivity and stability, but its higher price limits its practical implementation. SiO₂ nanofluids offer good thermal conductivity with low viscosity, making them suitable for applications needing efficient heat transfer with minimal pumping power. ZrO₂ nanofluids provide stable dispersion and moderate thermal conductivity, making them ideal for environments where stability and low corrosivity are essential. CuO nanofluids are known for their high thermal conductivity, significantly improving heat transfer [130].

Aksoy et al. [131] investigated the effect of nanofluid concentrations on spray cooling performance. Under both boiling and non-boiling conditions, nanofluids at concentrations of 0.05 wt.%, 0.1 wt.%, and 0.2 wt.% enhanced the cooling performance of an aluminum block by approximately 6%, 12%, and 25%, respectively, compared to deionized water. In another study, Aksoy et al. [132] explored the effects of deposition layers and nanoparticle-coated surfaces on cooling performance. Water sprayed onto a TiO₂-coated surface exhibited superior thermal performance compared to uncoated surfaces, with a 0.2 wt.% rutile TiO₂–water nanofluid achieving a cooling rate nearly three times greater than that of a water spray. Repetitive spraying without removing deposited nanoparticles further enhanced heat transfer by approximately 13%.

Conversely, another study [133] on convective heat transfer in silicon microchannel heat sinks revealed no significant enhancement with nanofluids compared to deionized water, attributed to nanoparticle clustering, surface erosion, and material degradation. These practical challenges, along with increased pressure drops and nanofluid discoloration, highlight the need for further research to optimize nanofluid applications in microchannel systems.

4.6. Pool Boiling

Pool boiling is a phase-change heat transfer process where liquid in contact with a heated surface boils, forming vapor bubbles that absorb significant latent heat, effectively cooling the surface. Surface modifications play a critical role in enhancing pool boiling performance by increasing nucleation sites and improving heat transfer efficiency.

Hadzic et al. [134] investigated laser-functionalized copper superbiphilic surfaces with monolayer coatings, achieving a heat transfer coefficient (HTC) of up to 299 kW/m²·K, a 434% enhancement over untreated surfaces. Sun et al. [135] demonstrated that three-dimensional complex structures (3D-CS) significantly improve pool boiling performance, with boiling curves showing a sudden increase in HTC just before reaching critical heat flux (CHF) due to reduced wall superheat. Shi et al. [136] explored advanced surface engineering, including semi-layered microporous copper structures and multi-level micro cubic fin silicon substrates, which enhanced heat transfer by providing additional nucleation sites and extended surface areas. This approach achieved a CHF of 90 W/cm² over a 2.25 cm² heated area.

Zhang et al. [137] developed copper powder-mesh composite surfaces using screen printing technology, comparing mesh surface (MS), fully printed mesh surface (FPMS), and locally printed mesh surface (LPMS). The LPMS configuration exhibited the best pool boiling performance, with a CHF of 214.1 W/cm² and an HTC of 38.4 W/(cm²·K) at a heat flux of 90.7 W/cm². Additionally, a study by [138] applied porous Cu with R-141b refrigerant to enhance pool boiling through nano- and microporous structures formed via two-stage electrodeposition, achieving a 53% increase in HTC compared to uncoated surfaces.

Ma et al. [139] investigated homogeneous and fractal array micro pin finned surfaces with FC-72 as the working fluid. Fractal arrays impeded bubble coalescence in subcooled boiling, maintaining liquid supply channels to prevent dry-out, resulting in a CHF of 120.3 W/cm² at 25 K subcooling, 4.72 times higher than that of a smooth surface.

5. Performance of Cooling Solutions

Coolants are substances utilized in power electronics cooling systems to absorb and remove heat from electronic components, including power modules and semiconductors, to prevent overheating and ensure optimal functionality. The pie chart in Figure 8 visualizes the statistical breakdown, shown in percentages, of the coolants utilized in the reviewed papers.

Power electronics liquid coolants should be non-flammable, non-toxic, cost-effective, and have optimal thermophysical properties—high thermal conductivity, specific heat, and heat transfer coefficient, along with low viscosity. The coolant must remain effective over the expected operational temperature range of the system. For instance, water and water–ethylene glycol mixtures have broad temperature ranges and are widely used in power electronics thermal management. Coolants such as HFE 7100 and refrigerants like R134a and R245fa are non-flammable and chemically inert, reducing corrosion risks. However, these may pose environmental concerns, such as high global warming potential (GWP) and potential ozone depletion effects. Their dielectric properties make them ideal for applications where electrical isolation is essential. Deionized water, nanofluids, and water-ethylene glycol mixtures are commonly applied due to their low viscosity and strong heat transfer properties. Deionized water has low electrical conductivity, which helps reduce short-circuit risks, but it can cause corrosion over time if not treated with corrosion inhibitors [6].

Several factors, including the power density of the electronic components, the operating environment, the design of the cooling system, and economic considerations, influence the choice of the appropriate coolant. To ensure optimal performance in a vast array of applications, including industrial power systems and consumer electronics, effective cooling is vital for prolonging the life and dependability of power electronic equipment.

Table 8. Performance of reviewed microchannel systems.

Ref.	Fluid	Feature	heff (W/cm2K)
[140]	Water	Nano- and micro-technology-based packaging optimization.	1.05
[141]	Ethylene glycol water	Multi-pass branching microchannel.	1.25
[142]	CuO	Rectangular microchannel with varying CuO nanoparticle concentration.	1.85
[143]	Water	Submillimeter channels directly manufactured in the ceramic substrate’s back-metallization layer.	1.96
[144]	R134a	Heat transmission during flow boiling in micro-channels with a rectangular cross-section.	1.97
[145]	HFE 7100	Critical heat flux (CHF) limit in sub-cooled flow boiling in the microchannel.	2.16
[146]	Water	Cooling system with hot water coolant.	2.39
[147]	R134a	Heat transmission during boiling in microchannels with two-phase flow.	2.48
[148]	Water	A chip connected to a single-phase Si microchannel cooler.	3.57
[149]	Water	Microchannel equipped with oblique fins.	3.61
[150]	HFE 7100	Investigation of microchannel heat sink aspect ratio regarding flow boiling heat transfer.	4.40
[151]	Water	Chip-attached Si microchannel cooler using Ag epoxy.	4.63
[152]	R236fa	Cooling of a high heat flux in a Si microchannel via two phases.	6.07

and

Table 9. Performance of reviewed jet impingement systems.

Ref.	Fluid	Feature	heff (W/cm2K)
[11]	HFE 7100	Two-phase hybrid microchannel jet impingement.	1.67
[153]	FC-72	Two-phase jet cooling with a smooth surface.	1.82
[154]	FC 40	Single-phase heat transfer under circular microjet arrays.	2.39
[155]	HFE 7100	Two-phase free-faller circular jet.	2.5
[156]	HFE 7100	Single-phase hybrid microchannel jet impingement.	2.82
[157]	Water	Single-phase heat transfer using droplet sprays.	3.94
[153]	FC-72	Two-phase jet cooling with finned surface.	3.95
[14]	Water-Ethylene glycol	Pin fins in the DBC substrate using laser powder bed fusion additive manufacturing.	4.1
[153]	FC-72	Two-phase jet cooling with pin-finned surface.	5
[158]	Water	Microjet array cooling with distributed return network.	5.2
[159]	Water	Single-phase distributed return submerged direct liquid jet impingement cold plate having nozzle pitch of 100 µm.	5.87
[154]	Water	Single-phase heat transfer under circular microjet arrays.	6.1
[160]	Water	4 × 4 array jet impingement PVC cooler with direct cooling.	6.25
[161]	R-245fa	Two-phase jet impingement using closed tunnel porous coated surface.	6.93
[161]	R-245fa	Two-phase jet impingement using open tunnel porous coated surface.	6.93
[161]	R-245fa	Two-phase jet impingement using flat porous coated surface.	8.6
[157]	Water	Two-phase heat transfer using arrays of microjets.	9.68
[161]	R-245fa	Two-phase jet impingement using pin fins with porous coating.	9.78
[162]	Water	Ultrathin manifold microchannel heat sink with impinging liquid slot-jets.	10.8
[163]	Water	Polymer nozzles with nine staggered orifices.	12
[164]	Water	Si hybrid heat sink and diamond heat spreader with microchannel flow, microjet array impingement.	18.9

illustrate effective heat transfer coefficients of heat sinks with microchannels and jet impingement systems, in which it is evident that convective heat transfer is optimized by jet impingement studies. The main reason for the increased performance of this technology is largely due to the removal of thermally ineffective layers, such as TIMs, that are located between the device and the cooling system. Additionally, the two-phase heat transfer phenomena also contribute to this improved performance.

6. Qualitative Comparison of Cooling Techniques

Selecting the cooling method depends heavily on the specific nature of the heat load being addressed. Figure 9 provides a concise overview of the effective application scenarios for thermal management techniques.

7. Challenges in Quantitative Analysis

The findings of the reviewed papers, summarized in Table 4, Table 5 and Table 6, indicate the availability of extensive research data on each cooling method, including finned heat sinks, microchannel cooling, jet impingement, phase change materials, and nanofluids. However, quantitative comparisons are challenging due to variations in testing conditions, such as heat sources, coolants, flow rates, and boundary conditions.

To elaborate, these papers use different parameters to evaluate cooling performance. Secondly, the variation in testing conditions, such as heat sources, coolants, flow rates, and boundary conditions, makes the task of quantitative comparison difficult. Also, some studies investigate cooling methods in the context of specific applications. So, it becomes difficult to generalize the findings and compare across different applications.

To enable a comprehensive quantitative comparison, future research should prioritize standardized testing methodologies, consistent performance metrics, and comprehensive data reporting across different cooling methods. This would allow for a more accurate and meaningful assessment of the relative performance of these cooling techniques and facilitate informed decision-making in thermal management applications.

8. Conclusions

This review categorizes a range of thermal management techniques, including finned heat sinks, microchannel cooling, jet impingement, phase change materials (PCMs), and nanofluids. A critical comparison reveals distinct advantages and limitations for each method, necessitating tailored approaches for specific applications.

Nanofluids offer high convective heat transfer due to enhanced thermal conductivity from nanoparticle suspensions [118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133]. However, challenges such as nanoparticle clustering, increased pressure drops, and stability issues limit their practical implementation, particularly in microchannel systems [133]. Future research should focus on optimizing nanoparticle types, concentrations, and surface treatments to mitigate these challenges.

PCMs provide effective transient thermal buffering by absorbing latent heat during phase transitions [64,113,114,115,116,117]. However, their low thermal conductivity and degradation during thermal cycling restrict their long-term efficacy. Innovations such as metal foam integration or biometric fin designs can enhance conductivity and heat distribution, warranting further investigation [64,117].

Finned heat sinks remain prevalent due to their simplicity and cost-effectiveness [54,55,56,57,58,59,60,61]. Configurations such as cross-fin or tapered-fin designs enhance heat dissipation, but their performance is limited by air-cooling constraints. Microchannel cooling offers high heat transfer coefficients due to increased surface area-to-volume ratios [66,67,68,69,70,71,72,73,74,75,76,77], yet faces challenges like pressure drops and potential clogging [76]. Jet impingement cooling excels in addressing localized hotspots with high heat transfer rates [97,98,99,100,101,102,103], but its complexity and risk of component damage require careful design optimization.

Hybrid cooling systems, combining techniques such as PCMs with finned heat sinks or microchannels with jet impingement, show promise for achieving optimal thermal performance in high-density power electronics [15,60]. Future studies should explore these hybrid approaches, focusing on cost-performance trade-offs and long-term reliability under cyclic thermal loading. Additive manufacturing offers opportunities for designing complex heat sink geometries, potentially improving thermal performance while reducing weight and cost [117].

The advent of wide-bandgap (WBG) semiconductors, such as SiC and GaN, has driven advancements in power electronics, enabling smaller, high-frequency devices with increased thermal challenges [16,17]. Optimizing packaging components, such as thermal interface materials (TIMs) and baseplates, is critical for efficient heat transfer. Double-sided cooling techniques are increasingly preferred over single-sided cooling due to enhanced heat dissipation [28]. Two-phase cooling, including pool boiling and flow boiling, offers high convective heat transfer coefficients, with supercritical CO₂ emerging as a promising coolant despite challenges in maintaining its supercritical state [6].

For accurate comparisons of cooling performance, standardized testing conditions are essential, including consistent coolant flow rates, input power density, and ambient temperatures. Future research should prioritize establishing key performance metrics, such as pressure drop, Nusselt number, and power density, to enable robust comparisons across cooling systems [54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139]. These efforts will guide the development of efficient, reliable thermal management solutions for next-generation power electronics in automotive and industrial applications.