Two-phase jet impingement cooling has emerged as a leading strategy for managing the extreme heat fluxes and local hot spots of modern electronics. Yet, practical application in confined geometries is often hindered by flow maldistribution and instabilities that occur due to outflow of the expanded liquid-vapor mixture. This study proposes a direct-on-chip confined two-phase jet impingement cooling strategy that incorporates a liquid-vapor phase-separation membrane to manage outflow of the two-phase mixture. Implemented through a staggered multi-jet array with distributed inlets and outlets to manage flow distribution in a three-path manifold, the membrane is positioned as the jet confinement wall above the heated surface. Phase separation is thereby achieved by selectively permeating vapor to a dedicated outlet while retaining liquid under controlled transmembrane pressure. An operational regime map is first developed based on the vapor generation rate and vapor venting capacity to identify the theoretical limiting factors. An experimental investigation then systematically characterizes the thermal-hydraulic performance using deionized water as the working fluid, examining the effect of phase separation under varying heat fluxes, flow rates, and transmembrane pressures. The results show that phase separation consistently mitigates both the magnitude and oscillation of pressure drop in two-phase flow regimes compared to non-separation operation where the dedicated vapor outlet is closed. At the maximum steady-state heat fluxes, the implementation of phase separation achieves average pressure drop reductions of 23.0%, 13.0%, and 10.5%, and suppresses the pressure drop oscillations (sample standard deviation) by 51.9%, 57.5%, and 23.5% at flow rates of 0.06, 0.10, and 0.14 L/min, respectively. Under extreme operating conditions near the critical heat flux at low flow rates (0.06 L/min and 175 W/cm2), phase separation effectively suppresses local dryout, reducing the surface temperature by up to 6.7 degrees C and diminishing the surface temperature oscillations from 4.76 degrees C to 0.44 degrees C (90.8% reduction). The membrane enables stable operation by preferentially venting vapor directly from the impingement cavity, which prevents the accumulation of large vapor structures in the liquid path. This research provides key insights into phase-separation-enhanced twophase jet impingement cooling, offering a high-performance solution having stable operation for the thermal management of next-generation electronic devices.

Densely packed arrays of high-power devices are found in various multi-chip electronic systems and packages. Forced liquid cooling, using cold plates having internal finned flow passages to enhance heat transfer, has long provided an effective thermal management solution to mitigate the temperature rise of devices found in these systems. Such extended fin surfaces, while lowering the temperature of devices, also imposes a flow resistance and leads to increased liquid pressure drop across the cold plate. Optimizing the distribution of fins can lead to cold plate designs that achieve lower device temperatures at a given pressure drop. Topology optimization using a homogenization approach to optimize the arrangement and sizing of pin fins in cold plates has been demonstrated to yield high-performance designs for cooling single and multiple heat sources. However, arrays of devices found in telecommunications, solid state illumination, and other systems pose a complex heating boundary condition for which the performance enhancement offered from topology optimized cold plates has not been explored. Specifically, in comparison to cold plates having uniform pin fins throughout, subjected to the same heating and flow boundary conditions. Here, topology optimized cold plate designs are generated for different array configurations of devices, varying the device size, pitch, and number of devices in the array. The performance of these cold plates is assessed via a thermal resistance metric and the pressure drop, both considered in the multi-objective cost function to be minimized, with different weightings considered to explore the Pareto optimality. Cold plates having topology optimized pin fin distributions versus uniform pin fin distributions are compared at equivalent average pressure gradients across the cold plate. It is observed that the relative percentage improvement in the thermal resistance of topology optimized cold plates decreases after exceeding a certain number of devices in the streamline direction. From inspection of the flow and temperature fields, it is observed that topology optimization derives its performance advantage by utilizing the available heat capacity of the coolant via using the pin fin distribution to route low-temperature coolant from the inlet to each device in the array. This work highlights that topology optimization can achieve improved cold plate designs for cooling arrays of devices, with the relative advantage over uniform cold plates depending on the number of devices in streamwise direction.

With increasing heat loads generated by modern electronics and data centers, thermal management systems require compact and high-performance heat exchangers customized to meet application-specific demands. While topology optimization has been extensively applied to design cold plates and heat sinks, its extension to fluid-to-fluid heat exchangers remains limited due to the challenge of ensuring two non-mixing fluid domains within the same design space. This study introduces a novel topology optimization framework that leverages homogenized triply periodic minimal surfaces (TPMS) as tunable building blocks having a geometry that addresses this challenge. Each computational cell in the design domain is represented as a TPMS unit, dividing the space into two interwoven, non-mixing fluid regions suitable for heat exchange. A continuous design variable field controls the local TPMS frequency, which governs the local geometry, and is linked to thermal-hydraulic performance. A two-fluid, local thermal non-equilibrium, steady-state solver is implemented in OpenFOAM to model heat exchange within the TPMS structure, incorporating convective and conductive resistances that determine the local temperature difference between the fluids in each cell. Design sensitivities are computed using the adjoint method, enabling gradient-based optimization while retaining continuous grading of substructures without penalization or projection. Case studies demonstrate the ability of the proposed framework to generate spatially varying TPMS structures that optimize both heat transfer effectiveness and pressure drop. Compared to uniform TPMS designs, the optimized geometries achieve superior performance across the Pareto frontier. The resulting designs are directly manufacturable using additive techniques, with no need for post-processing to enforce binary distributions common to conventional density-based topology optimization approaches. This work provides a scalable and physically interpretable technique for fluid-to-fluid heat exchanger design.

Closed loop thermosyphons (CLTs) are passive heat transfer devices, that when operating in a buoyancy-driven two-phase regime, can transfer heat nearly isothermally and thereby achieve excellent thermal performance. However, CLTs are susceptible to flow instabilities, particularly geysering, which can lead to transient fluctuations in pressure and temperature. This work presents an experimental investigation of geysering instabilities during the powering up of a small-scale CLT. Experiments are conducted for heat loads from 45 W to 260 W and water fill ratios from 48% to 81%. The experimental facility uses transparent tubing, enabling flow visualization. Loop performance is characterized through loop thermal resistance, dominant oscillation frequency of the evaporator outlet temperature, and standard deviation of the transient temperature data. Flow visualization revealed that the geysering instability follows a repeating cycle of subcooled boiling in the evaporator, followed by a buoyancy-driven increase in fluid column height in the riser, which draws subcooled liquid back into the evaporator causing bubble collapse and subsequent flow reversal, before the cycle begins again with subcooled boiling in the evaporator. Two distinct geysering oscillation behavior were observed: low-frequency, high amplitude and high-frequency, low amplitude modes. Lower fill ratios transition from low- to high-frequency oscillations with power, while higher fill ratios maintain low-frequency oscillations. Results show that increasing heat load decreases thermal resistance for all fill ratios, with the 62% fill ratio achieving the lowest values. These findings provide insight into the underlying mechanism of geyser boiling and establish trends in loop behavior with fill ratio and heat load.

Two-phase jet impingement is a thermal management technique that offers the high heat transfer coefficients necessary for cooling of high-heat-flux electronic devices. Use of dielectric fluids for jet impingement further allows direct cooling of electronic devices. While these fluids are attractive owing to their low pressure and saturation temperature, suitable for the operating temperature of electronics, their thermophysical properties are unfavorable for boiling performance. Therefore, boiling enhancement surfaces are of interest to improve boiling heat transfer performance and enable the use of two-phase jet impingement in high-heat-flux thermal management applications. Self-organized laser functionalization (SOLF) offers a scalable approach that creates permanent micro- and nano-scale features on surfaces for boiling enhancement. In this study, two-phase jet impingement experiments are performed using a test chamber and a supporting two-phase flow loop. A single 3 mm-diameter jet orifice is spaced 7.5 mm from a 25.4 mm-diameter circular aluminum heated surface. The boiling performance of SOLF-functionalized heater surfaces is compared to that of a polished surface during pool boiling and two-phase jet impingement. Experiments are conducted over a range of jet inlet subcooling temperatures (5–30 K) as well as inlet velocities (0.5–2 m/s) using the dielectric fluid HFE-7100. Tests under each condition were performed by incrementing the heat input up to the maximum heat flux before a transition to film boiling occurred. SOLF functionalization led to a significant reduction in surface superheat during nucleate boiling, with a reduction of ∼47–74% compared to the baseline polished surface at the same maximum heat flux. However, for the combination of the laser processing parameters used and the range of thermal testing conditions, the effect on critical heat flux (CHF) was minimal. The SOLF surface demonstrated a characteristic boiling curve behavior, for both pool boiling and two-phase jet impingement at low subcooling, in which an increase in the surface heat flux led to a reduction in the surface superheat, known as boiling inversion. Boiling curve hysteresis experiments and high-speed visualization of the boiling process confirm that the mechanism by which SOLF functionalization reduces the surface superheat is activation of additional nucleation sites along the boiling curve.

Two-phase closed loop thermosyphons (CLTs) are passive heat transfer devices that can effectively remove high heat fluxes through phase change, and their scalability makes them applicable to a wide range of industries. Models available to predict the steady behavior of CLTs typically report a range of fill ratios, geometric parameters, and evaporator-condenser height differences for which the system will achieve minimum thermal resistance. However, given that there is a strong coupling between the operating mass flux of a CLT and the heat flux applied, and the refrigerant charge is fixed during operation, these identified optimum points cannot be maintained in cases where the heat flux changes during operation for a fixed design. In this study, we present an operational map for a CLT at saturated conditions across varying heat fluxes and fill ratios. This enables surveying of the fill ratios required for operation in the saturated regime, which is known to offer the lowest thermal resistance, across different heat fluxes. The mass, energy, and momentum equations are solved simultaneously using steady homogeneous two-phase flow assumptions. The loop is divided into evaporator, adiabatic, and condenser sections with appropriate boundary conditions defined for each section, and the governing equations are applied in a one-dimensional lumped model. A parametric study is performed by varying the fill ratio of the system and heat flux applied to the evaporator. The flow regime characteristics are investigated as a function of heat flux, and contour maps of the saturated regime operating envelope that minimizes thermal resistance are presented. The resulting operational map suggests that fill ratios between 0.3 and 0.75 will result in the minimum thermal resistance of the loop. This study provides a means to obtain operational maps for closed loop thermosyphons as a useful design tool to identify the desired operating range of heat fluxes and fill ratios for a given geometry.

Heat sinks having extended surface area can substantially increase the heat dissipation limit of two-phase immersion cooling compared to the critical heat flux limit of flat surfaces. However, the highly nonlinear and boiling-regime-dependent surface boundary condition makes designing an optimal two-phase immersion heat sink a challenge, and no systematic design approach currently exists. In this work, we propose a density-based topology optimization method for designing heat sinks under pool boiling conditions. Boiling on a heat sink is modeled as a conduction problem with convective boundary condition, where the heat transfer coefficient is defined as a function of superheat and spans all boiling regimes (natural convection, nucleate boiling, transition boiling, and film boiling). The heat sink is represented as voxelized density distribution within the design domain, with the thermal conductivity modeled as a function of local density. The boiling boundary conditions are implicitly incorporated at the heat sink interface using an interpolation function of density and superheat. Penalizations are applied to both the thermal conductivity and the boiling boundary conditions to enforce a binary final design, while density filtering and Heaviside projection are applied to control the feature length scale. Optimized heat sink designs are generated iteratively using a multi-objective cost function, with the adjoint method applied to compute sensitivity at each optimization iteration. The resulting coral-like heat sink designs are predicted to effectively dissipate heat at a footprint-area-based power density significantly exceeding the critical heat flux of a flat surface, and are further shown to outperform conventional pin-fin and plate-fin arrays as benchmarks. A detailed analysis of heat sink superheat and heat transfer distributions further reveals that the optimized designs effectively spread the heat load and reduce the conduction resistance by maintaining the surface area with high heat transfer coefficients in the nucleate boiling regime close to the heat input region. The superheat-dependent boundary condition captures the drastic reduction in heat transfer coefficient with transition to film boiling and naturally penalizes operations in such unfavorable regimes, facilitating the generation of heat sinks that avoid thermal runaway. This study demonstrates the effectiveness of proposed topology optimization method in balancing the heat spreading and conduction resistances under a boiling boundary condition to achieve optimal two-phase immersion heat sink performance.

Flow boiling in parallel microchannel heat sinks is attractive for heat dissipation from telecommunications devices, data center server chips, and electrified vehicle power converters and motors. However, there are practical implementation challenges associated with two-phase flow maldistribution through parallel microchannels, which can degrade thermal performance and reliability. Therefore, predicting and understanding the flow distribution mechanics in parallel microchannels based on analysis of the system stability is essential. It has been shown that lateral channel-to-channel thermal coupling has the potential to reduce maldistribution for systems having only two parallel channels. Yet, the effect of thermal coupling on system stability and maldistribution in larger sets of many parallel channels, as would be common in microchannel heat sinks and other components, is unexplored. In the current work, a lumped fluid flow model, which also incorporates the wall thermal capacity effects on the system dynamics, is developed to assess flow distribution with lateral thermal coupling between neighboring parallel microchannels. Linear stability analysis establishes that the effectiveness of lateral thermal coupling in reducing maldistribution weakens with an increasing number of parallel microchannels. And for many parallel microchannels, it is confirmed that lateral thermal coupling alone cannot stabilize the system, and the entire two-phase flow operating region becomes inherently maldistribution-prone, with similar stability to thermally isolated parallel channels. However, with the addition of inlet restrictors to the channels, the impact of lateral thermal coupling becomes significant, regardless of the number of channels. Inlet restrictors are more effective in suppressing maldistribution when there is a higher lateral thermal conductance between neighboring channels. This unintuitive combined influence of lateral thermal coupling and the inlet restrictors in reducing maldistribution is attributed to nonlinear behavior that is explained based on the eigenvalues of the parallel microchannel system.

Microscale oscillating heat pipes (OHPs) are a promising technology for thermal management of electronic devices, offering high effective thermal conductance and scalability for compact integration. Predictive modeling of OHPs at the microscale remains difficult due to extreme aspect ratios, complex two-phase transport, and transitions across multiple operating regimes. In this study, an efficient one-dimensional (1D) homogeneous model is developed to simulate unsteady two-phase flow dynamics of closed-loop microchannel OHPs, governed by conservation of mass, momentum, and energy. The homogeneous assumption treats liquid and vapor phases as uniform mixtures, enabling use of common variables—temperature, pressure, and velocity—to describe the flow. Thermophysical properties are expressed as functions of internal pressure and temperature, thereby capturing the thermomechanical cycles of vapor compression and expansion that drive OHP operation. The model is applied to multiturn microscale OHPs subjected to varying evaporator-to-condenser temperature differences. Simulations capture three operating regimes—prestartup, oscillation, and dryout—and reproduce the strong dependence of thermal performance on regime. In particular, heat transfer enhancement is observed with the onset of oscillatory flow, consistent with prior experimental findings. The model further demonstrates robustness in predicting oscillations driven solely by internal thermomechanical instabilities, without requiring gravity, capillarity, or other external driving forces. Overall, the proposed framework provides a tool for understanding coupled thermal-fluid dynamics of microscale OHPs and interpreting the transitions between the prestartup, oscillation, and dryout operating regimes. In the future, with calibrated heat transfer parameters from experiments, the model may be extended to accurate heat transfer prediction of OHPs and have potential applications in design optimization.

Boiling processes are notoriously difficult to analyze via visual inspection due to the complex interactions between the vapor bubbles and the surface. Unsupervised machine learning (ML) is a powerful tool to uncover physical insights into the bubble dynamics during boiling from image data. In this study, principal component analysis (PCA), an unsupervised dimensionality reduction algorithm, is used to extract new physical descriptors of boiling heat transfer from pool boiling experimental images without any labeling and training. Experiments are conducted with different working fluids and heater surfaces to investigate the effect on the bubble morphology and subsequently on the physical descriptors identified through unsupervised ML. The dominant frequency and amplitude deduced from the Fourier transform of the time series of the first principal component (PC) are compared against physical parameters such as bubble size, bubble count, and vapor area fraction. The new physical descriptors derived from PCA show a positive correlation with conventional parameters related to bubble morphology, as demonstrated by linear regression analysis. Pearson Correlation Coefficients further confirm the strong correlations between dominant amplitude and both bubble size and vapor area fraction, as well as between dominant frequency and bubble count. These strong correlations hold across multiple different working fluids (water and HFE 7100) with different heater surfaces (plain and microstructured surfaces made of copper and silicon materials), demonstrating the potential for these extracted physical descriptors to generalize and act as a surrogate to conventional physical descriptors. This unsupervised learning approach offers a robust alternative to traditional pool boiling analyses or supervised ML approaches that rely on time-consuming manual labeling involving bubble identification and segmentation.

Two-phase jet impingement cooling is a promising solution for high-heat-flux semiconductor thermal management, as it combines strong convective heat transfer with boiling to remove large heat loads at relatively low flow rates and pressure drops. However, practical deployment is hindered by challenges including inconsistent boiling initiation on smooth surfaces, surface dry-out, vapor-induced flow instabilities, and premature critical heat flux (CHF). Excessive vapor generation within confined geometries can disrupt flow uniformity, causing temperature oscillations and unstable operation. To address these challenges, this work presents a confined, direct-on-silicon two-phase jet impingement cooling approach incorporating a porous-wick-assisted phase separation mechanism. The engineered porous wick enhances nucleate boiling and enables in situ phase separation at the boiling surface. Integrated with a custom three-path manifold, the design routes separated liquid and vapor streams, minimizing vapor accumulation within the confined region and suppressing two-phase instabilities. The porous wick is directly printed onto the silicon substrate using advanced additive manufacturing, eliminating the need for a thermal interface material (TIM) and its associated thermal resistance. Thermal-hydraulic characterization using a low-surface-tension dielectric fluid demonstrates that wick-assisted phase separation reduces thermal resistance by 23-29% compared to configurations without phase separation. Extended testing over more than 200 h of continuous operation confirms stable thermal performance and indicates strong potential for long-term reliability. System-level validation is demonstrated in a 1 U server equipped with an NVIDIA V100 GPU (graphics processing unit) incorporating a direct-on-silicon printed wick.

Topology optimization approaches have extensively explored optimizing the fin material distribution in cold plates to minimize device temperature rise given pressure drop. While the fin distribution indeed dictates the pressure drop and thermal performance characteristics of a cold plate, its optimization is typically performed for a pre-specified flow inlet and outlet boundary locations, which also have a significant role on the resulting designs and performance. This study thus presents a novel framework to simultaneously co-optimize flow inlet and outlet locations in combination with the fin material distribution. The developed framework adopts a hybrid optimization approach, combining gradient-based topology optimization of the fin material distribution with a metaheuristic particle swarm optimization algorithm for placement of the inlet and outlet. The topology optimization algorithm uses a homogenization approach to determine an optimized distribution of pin fins with varying sizes throughout the cold plate to route coolant from a specified inlet to outlet. Meanwhile, the particle swarm algorithm is used to simultaneously locate the optimal inlet and outlet positions. The initialization of algorithm consists of defining multiple cold plate instances having a uniform pin fin distribution and different random inlet/outlet locations. Topology optimization is first performed for each instance to generate partially optimized designs while keeping the inlet/outlet locations fixed. The inlets and outlets of all instances are then relocated based on a particle swarm optimization algorithm. The partially optimized pin fin distributions, along with the new inlet/outlet locations, are then used as the initial designs for further iterations in the co-optimization approach. The process is repeated until convergence is observed in the objective function. The objective function chosen for demonstration of the co-optimization approach is a thermal resistance metric that considers the temperature of heated regions given a fixed total inlet pressure above the outlet gauge pressure. The developed framework is applied to two test cases representative of single chip and multi-chip systems. Benchmark cold plates having pre-specified inlet and outlet locations are also topology optimized for comparison. Cold plates having co-optimized inlet/outlet locations and topology are shown to perform better than the benchmark cases based on thermal resistance metric value for both test cases considered. For a test case with a single heated region, there is a ∼19% decrease in the maximum base temperature rise for the co-optimized cold plate when compared with a benchmark case. While relative performance improvements are invariably case-specific, establishing this framework for coupling of gradient-free and gradient-based optimization algorithms to simultaneously locate the inlet and outlet whilst optimizing the fin distribution offers a critical means to speed up the design process when addressing complex cold plate design cases having unintuitive optimal flow inlet and outlet placements.

Recent advancements in power semiconductors devices and electric machines have fueled a growing demand for compact, integrated thermal management approaches. Of particular recent interest is embedded two-phase cooling of electric motors, which have numerous internal heat sources in the form of electrical windings, the power density of which are limited by external cooling approaches and single-phase cooling techniques. However, adequate and uniform cooling with an embedded two-phase cooling approach requires proper flow management via a flow manifold to avoid flow boiling instabilities and maldistribution. This study investigates the flow uniformity and stability of two-phase flow within a parallel microchannel system with a hierarchical manifold. The hierarchical manifold system contains a primary manifold that delivers flow to each parallel tube (array of channels) and sub-manifold that further split the flow to the parallel array of microchannels within each tube. These manifolds are separated by orifices to hydraulically isolate each tube by providing an additional pressure drop. A single-tube and parallel multi-tube test section have been developed to assess the flow distribution and temperature uniformity with R-1233zd(E) as the working fluid. With a test matrix of three inlet subcooling temperatures and varying flow rate, the wall temperature and pressure drop are measured with increasing heat load to characterize the behavior of each test section. The multi-tube test section exhibits decoupled flow boiling incipience between tubes, in which nucleation in each tube can begin at different heat inputs without initiating boiling in the other tubes. Additional test sections were also tested without the hierarchical manifold to test baseline two-phase flow distribution in this design. Without orifices, the temperature uniformity improved at near-incipience heat inputs, as boiling incipience in a single tube triggers boiling across all tubes, as compared to the with-orifice test section.These observations reveal a newly termed incipience-overshoot instability that can occur due to the superheat temperature overshoot that occurs prior to the incipience of boiling, and in particular in systems that have flow restrictions introduced to make the system otherwise stable not considering this overshoot. In such systems, the flow is isolated between tubes, and at incipience, some tubes have two-phase flow at a decreased flow rate, whereas the remaining tubes can maintain a superheated liquid flow at an increased flow rate. This results in slight maldistribution and temperature non-uniformity until further superheating of the liquid in these tubes causes transition to two-phase flow at higher heat loads.

This work demonstrates the use of a transient model to map the operational limits of a heat pipe based on the time-to-dryout, evaporator temperature, and time-to-rewet in response to dynamic high-power pulses and subsequent throttling. The model considers dynamic coupling between the local liquid saturation and capillary pressure, as well as incorporates wetting hysteresis to predict pulse-load-induced dryout and power-throttling-assisted recovery. Traditional heat pipe design and performance metrics consider the ‘steady capillary limit’ (maximum power at which a steady heat load can be maintained without dryout), but this steady-state framework does not capture the transient power excursions common in modern electronic systems. This work therefore puts forth a framework for mapping the ‘dynamic capillary limits’ of a heat pipe as a function of the transient workload of interest. Specifically, the framework produces design-ready operational maps that quantify the time-to-dryout (TTD), peak evaporator temperature during a pulse (Tevap), and time-to-rewet (TTR) during throttling — thereby defining dynamic capillary limits beyond the conventional steady capillary limit. These operational maps are effective as tools for assessing the transient limits of performance and can be generated for various different heat pipe designs.

High heat fluxes in electronic devices must be effectively dissipated to prevent local hotspots, which are critical for long-term device reliability. In particular, advanced semiconductor packaging trends toward thin form factor products increase the need for understanding and improving in-plane conduction heat spreading in anisotropic materials. The 2D laser-based Ångstrom method, an extension of traditional Ångstrom and lock-in thermography techniques, measures in-plane thermal properties of anisotropic sheet-like materials. This method uses non-contact infrared temperature mapping to measure the thermal response to periodic laser heating at the center of a suspended sample. The spatiotemporal temperature data are analyzed via an inverse fitting algorithm to extract thermal conductivities in the in-plane orthotropic directions that best adhere to the governing heat conduction equation. Using this algorithm, we present an approach to simultaneously fit data across multiple heating frequencies, which improves measurement sensitivity because the thermal penetration depth varies with frequency. The accuracy of this technique is assessed by tuning experimental parameters such as sample dimensions and heating frequency. A standardized workflow is proposed for measuring unknown materials and for processing the data, including filtering out regions influenced by laser absorption and heat sink boundary effects. Numerical simulations validate the method across a wide range of thermal conductivities (0.1–1000 W m−1 K−1) and material thicknesses (0.1–10 mm), with accuracy demonstrated for anisotropy ratios up to 1000:1. Experimental measurements on isotropic and anisotropic materials agree well with the benchmark values. Ultimately, standardization of this technique supports the development of engineered anisotropic heat-spreading materials for thermal management and packaging applications.

Electrification of transportation in the aviation industry is challenging in part due to the high power densities necessary for propulsion using electric motors. A commercial, narrow-body aircraft would require electric motor systems having >12 kW/kg power density, over twice the current state-of-the-art. One major limitation to increasing power density are limits on the operating temperature. Electric motor windings, which are wrapped in insulation, produce the majority of the heat in the motor. Positioning the coolant closer to the windings so as to decrease the overall thermal resistance between the heat source and sink is therefore a promising route toward enabling higher power densities. In this study, an additively manufactured stator subsection prototype with embedded microchannels is used to demonstrate two-phase cooling at different mass flow rates of R1233zd(E). Compared to single-phase cooling, utilizing two-phase flow provides higher heat transfer coefficients, which have increasing importance on reducing the overall resistance when the coolant is embedded close to the heat source, as well as offering a nearly isothermal coolant at the saturation temperature independent of mass flux. This prototype test section, which has been demonstrated for continuous operation at 30.4 A/mm2, is experimentally characterized at five flow rates between 0.33 g/s and 0.83 g/s. The average coil temperature is demonstrated to be insensitive to mass flow rate, as is desired for practical operation, owing to the high effective heat capacity rate when operating in the two-phase regime. Instrumentation of the test section with wallembedded thermocouples enables decomposition of the total coil temperature rise into the conductive and convective thermal resistance components. The incorporation of two-phase flow reduced the convective thermal resistance by 77 %. Thermal models for each of these resistance components are developed to validate experimental findings, and further, to allow performance prediction in context of up-scaling to the full motor assembly and higher operating powers.

A topology optimization framework is developed to design embedded microchannel heat sinks that consider multiple workloads and transient thermal management challenges in 3Dstacked semiconductor packages. The framework integrates steady-state optimization with penalties in the design objective that update based on transient performance evaluations to iteratively adjust porosity distributions. The optimization begins with multi-workload steady-state analysis to ensure reliable thermal and hydraulic performance across all power distributions. These designs are modified by imposing localized penalties based on peak temperatures observed during dynamic loads, thus ensuring effective cooling under both continuous and transient conditions. Case studies show that the optimized designs reduce pressure drop and maintain target temperatures while balancing thermal and hydraulic constraints. This approach provides an efficient and scalable framework for the design of embedded cooling solutions for next-generation semiconductor packages.

Outdoor radiative cooling surfaces passively lose heat by reflecting solar irradiation and emitting infrared radiation to cold deep space through the atmospheric sky window (8–13 µm), thereby achieving sub‐ambient temperature. Ultrawhite radiative cooling paints are an emerging technology offering scalable solutions for cooling and passive water harvesting wherein surface wettability plays a key role. This work, examines how radiative cooling paint pigment and binder formulations affect surface morphology, roughness, and dynamic wettability. Samples are prepared with three different nanoparticulate pigments, calcium carbonate (CaCO 3 ), barium sulfate (BaSO 4 ), and hexagonal boron nitride (hBN); two binders, including an acrylic and a waterborne silicone‐modified polyurethane dispersion (SILIKOPUR 8081); and pigment solid volume concentrations from 0% to 80% v/v. The CaCO 3 and BaSO 4 pigments produced paints with rougher textures and higher contact angles due to their pigment particle morphology. While high solar reflectance was achieved across various pigment and binder combinations, wettability exhibited a complex trend with pigment concentration, indicating that maximizing reflectance does not necessarily optimize wetting behavior. This expanded understanding on how pigment type, binder and concentration influence wettability, offering pathways to design coatings with tailored spectral and wetting properties for both self‐cleaning paints and passive water harvesting applications

Jet impingement, enhanced by evaporation and boiling from a porous wick structure, has potential for dissipating high heat fluxes. This study investigates the enhancement of twophase jet impingement cooling using capillary-driven boiling and proposes a concept design to address the challenges associated with two-phase operation, including pressure fluctuations, surface temperature oscillations, and premature critical heat flux. A novel manifold design is proposed, featuring a three-path manifold with distributed liquid inlets, liquid outlets, and vapor outlets. The integration of porous wick structures on the surface increases the availability of nucleation sites that promote boiling and evaporation throughout the evaporator. Embedded tunnels within the wick facilitate liquid bypass to the liquid outlets and act as an integrated liquid reservoir for capillary wicking to the porous regions. Whereas vapor generated during boiling in the wick exits through dedicated vapor outlets to achieve phase-separation. The tunnel spacing is optimized to ensure a sufficient capillary wicking length in the wick structure to avoid dry out at high heat flux operation, even for low-surface-tension dielectric coolants used in electronics cooling. A reduced-order model was developed to assess the thermo-hydraulic performance of the proposed system under various flow rates and heat flux conditions, using R1233zd as the working fluid. The manifold and wick dimensions were designed to maintain a target pressure drop of 40 kPa and heat flux of $500 \mathrm{~W} / \mathrm{cm}^{2}$. This work demonstrates an energy-efficient, low-cost, and package-integrated two-phase jet impingement cooler optimized for minimal pressure drop and maximum thermal performance. By addressing challenges such as phase separation and manufacturing constraints, the proposed design offers a scalable and robust solution for managing extreme thermal loads in high-performance applications.

High heat fluxes generated in electronic devices must be dissipated by conduction through the semiconductor substrate and packaging layers to avoid local high-temperature hotspots that govern device reliability. Deeply buried interfaces within the semiconductor components are challenging to characterize in situ due to their low relative magnitude and location within the chip stack. Here, we develop a novel metrology technique and experimentally demonstrates a non-destructive method for characterization of the thermal interfacial resistances of interfaces buried within stacks of material substrates. This work targets interfaces buried deeper than the thermal penetration depth of available transient measurement techniques, but with thermal resistances below the resolution of most steady-state techniques, on the order of 0.01 cm2 K -1 W-1. This new interfacial conductance measurement technique combines non-contact periodic heating with infrared (IR) thermal sensing to measure the transient temperature response of a multi-layer stack of materials. Specifically, periodic heating of one face and cooling on the opposite phase generates a transient, one-dimensional temperature gradient across the sample stack. The corresponding steady periodic temperature amplitude and phase delay across the thickness of the material are used to extract for the thermal interfacial resistance using an inverse fitting method, assuming the thermal properties of the solid layers are known. Numerical simulations are developed to generate synthetic temperature data, which along with the inverse fitting method, are used to validate the extraction of interfacial thermal resistance in two-layer stacked materials, as well as thermal conductivity of bulk materials without an interface. The data extraction process is shown to accurately extract the interfacial thermal resistances ranging from 0.001 to 1 cm2 K -1 W -1 for interfaces that up to 5 millimeters from the exposed surface. For bulk materials, this technique demonstrates accuracy in extracting the thermal conductivity spanning a thermal conductivity range of 0.1 to 2000 W m -1 K-1. An experimental facility is developed, which includes an IR-transparent heat sink, laser-based heating, and two IR temperature sensors for measurements of thermally opaque samples. Measurements demonstrate the accuracy and sensitivity of the measurement technique for interfacial thermal resistance of two-layer samples, as well as thermal conductivity of bulk materials. The ultimate goal of this work is to develop a standardized technique for measurement of thermal resistances across the range of magnitudes and stack geometries commonly found in modern electronic packages, ranging from near-junction epitaxial semiconductor films to interconnect layers in emerging die-to-die and wafer hybrid bonding technologies.