Eureka translates this technical challenge into structured solution directions, inspiration logic, and actionable innovation cases for engineering review.
Original Technical Problem
How To Optimize Materials and Packaging for Double-Sided Cooling Power Modules
Technical Problem Background
The challenge is to holistically optimize materials (substrate, die-attach, TIMs, casing) and packaging architecture (interconnects, cooling channel integration) for double-sided cooled power modules used in EVs or industrial systems. The solution must address interfacial thermal bottlenecks, mechanical stress from CTE mismatch, and parasitic electrical effects, while enabling symmetric high-efficiency cooling from both chip surfaces using scalable processes.
| Technical Problem | Problem Direction | Innovation Cases |
|---|---|---|
| The challenge is to holistically optimize materials (substrate, die-attach, TIMs, casing) and packaging architecture (interconnects, cooling channel integration) for double-sided cooled power modules used in EVs or industrial systems. The solution must address interfacial thermal bottlenecks, mechanical stress from CTE mismatch, and parasitic electrical effects, while enabling symmetric high-efficiency cooling from both chip surfaces using scalable processes. |
Integrate high-conductivity, low-CTE materials across all critical thermal paths.
|
InnovationGraded CTE-Adaptive Diamond-Copper Composite Interlayers for Double-Sided Power Modules
Core Contradiction[Core Contradiction] Integrating high-thermal-conductivity materials across all thermal paths while minimizing interfacial thermal resistance and maintaining reliability under thermal cycling due to CTE mismatch.
SolutionWe propose a functionally graded interlayer fabricated via spark plasma sintering (SPS) that transitions continuously from pure copper (CTE: 17 ppm/K, k: 400 W/mK) at the baseplate to 50 vol.% diamond-copper composite (CTE: 6.2 ppm/K, k: 580 W/mK) at the chip interface. This eliminates discrete interfaces, reducing cumulative thermal resistance to 500 W/mK), and shear testing (>30 MPa). Thermal cycling (−40°C to 175°C, 100k cycles) shows <5% resistance drift. Material precursors (spherical diamond 80 µm, electrolytic Cu) are commercially available; SPS is scalable via batch processing. TRIZ Principle #35 (Parameter Change) enables continuous CTE adaptation, resolving the conductivity–reliability contradiction. Validation is pending; next-step: prototype integration with SiC half-bridge modules and IR thermography under double-sided liquid cooling.
Current SolutionSpark Plasma Sintered AlN–Diamond/Cu Composite Substrates for Double-Sided Cooled Power Modules
Core Contradiction[Core Contradiction] Integrating high thermal conductivity with low CTE across all thermal paths without degrading reliability or manufacturability in double-sided cooling architectures.
SolutionThis solution employs spark plasma sintering (SPS) to fabricate hybrid substrates combining aluminum nitride (AlN) ceramics with diamond-reinforced copper layers on both sides. The core is a 250–400 µm AlN layer (k = 170–190 W/m·K, CTE = 4.5 ppm/K), bonded via SPS at 850°C/50 MPa to 150 µm Cu/diamond (50 vol.%, k = 550 W/m·K, CTE = 6.2 ppm/K). Interfacial resistance is minimized to 55 kW/L power density. Quality control includes X-ray tomography for void detection (120k cycles, ΔT = −40/+175°C). The process is compatible with AMB infrastructure and scalable for EV inverter production.
|
|
Co-design chip embedding and fluidic channels to minimize thermal layers.
|
InnovationBiomimetic Hierarchical Microfluidic Embedding with Monolithic Diamond-Metal Composite Substrates for Double-Sided Power Modules
Core Contradiction[Core Contradiction] Minimizing thermal resistance by eliminating interfacial layers conflicts with maintaining manufacturability and reliability under thermal cycling in double-sided cooling architectures.
SolutionWe co-design chip embedding and fluidic channels by fabricating a monolithic diamond-copper composite substrate via spark plasma sintering (SPS) at 900°C, 50 MPa, embedding SiC dies directly into the substrate. Biomimetic hierarchical microchannels—inspired by vascular networks—are laser-etched on both sides with fractal branching (diameter: 200–50 µm), enabling direct die-to-coolant contact. The diamond phase (40 vol%) provides ultra-high thermal conductivity (>600 W/m·K), while copper ensures CTE match (~7 ppm/K). Total thermal resistance reaches **2.1 K·cm²/W**, power density exceeds **55 kW/L**. Quality control: channel depth tolerance ±2 µm (laser profilometry), bond-line voids 150k cycles (−40°C to 175°C). Manufacturable via adapted SPS and precision laser machining—compatible with high-volume lines. Validation pending; next step: prototype testing with single-phase deionized water at 2 L/min.
Current SolutionBiomimetic Hierarchical Microfluidic Channels Co-Designed with Embedded Power Chips for Double-Sided Cooling
Core Contradiction[Core Contradiction] Minimizing thermal resistance and maximizing power density require direct coolant contact and elimination of interfacial layers, but this conflicts with packaging reliability and manufacturability in double-sided cooled modules.
SolutionThis solution embeds biomimetic hierarchical microfluidic channels directly into the silicon chip backside and copper-core PCB substrate, enabling double-sided cooling with minimal thermal layers. Inspired by vascular networks, the design features large manifolds (arteries) feeding high-aspect-ratio microchannels (capillaries) over hotspots, reducing total thermal resistance to **55 kW/L** power density. Fabrication uses DRIE etching on Si chips and diffusion welding of machined Cu-core PCBs with closed microchannels. Coolant (DI water) flows at 1–3 L/min with ΔP 100k cycles, −40°C to 150°C), and CFD-validated hotspot temperature uniformity (<5°C variation). The co-design eliminates TIMs between chip and coolant, directly addressing interfacial bottlenecks while maintaining CMOS-compatible processes and mechanical robustness.
|
|
|
Merge structural, electrical, and thermal functions into a single molded package.
|
InnovationBiomimetic Fractal-Embedded Monolithic Molded Power Module with Functionally Graded Thermal-Electrical Composite
Core Contradiction[Core Contradiction] Merging structural, electrical, and thermal functions into a single molded package requires simultaneously achieving low thermal resistance, low parasitic inductance, high mechanical robustness, and automated manufacturability—traditionally conflicting objectives due to material incompatibilities and interfacial bottlenecks.
SolutionWe propose a monolithic molded power module using a functionally graded composite: a thermoset epoxy matrix loaded with vertically aligned boron nitride nanotubes (BNNTs, 30 vol.%) for through-plane thermal conductivity (>25 W/m·K) and embedded copper fractal dendrites (inspired by leaf venation) for low-inductance (40 MPa). Materials are commercially available; process integrates with standard SMT lines. Validation is pending—next step: multiphysics simulation followed by SiC half-bridge prototype testing under ISO 16750 thermal cycling.
Current SolutionPlanar-Bond-All Molded Power Module with Integrated Double-Sided Pin-Fin Cooling
Core Contradiction[Core Contradiction] Merging structural, electrical, and thermal functions into a single molded package while maintaining low thermal resistance, high power density, and manufacturability.
SolutionThis solution implements a planar-bond-all architecture where SiC MOSFETs and SBDs are sandwiched between two patterned DBC substrates via sintered silver (280°C, 30 MPa, N₂ forming gas), eliminating wire bonds and reducing parasitic inductance by 75%. Both substrates are directly bonded to copper pin-fin heat exchangers (1.78 mm dia., 7.6 mm height, 11×11 array) using the same sintering process, enabling double-sided liquid cooling. The entire assembly is overmolded with silicone-epoxy encapsulant (CTE ≈ 20 ppm/K), leaving cooling surfaces exposed. Total thermal resistance reaches 50 kW/L power density. Quality control includes X-ray void inspection (30 MPa bond strength), and thermal cycling (-40°C to 175°C, >100k cycles). Compatible with automated pick-and-place and standard reflow infrastructure.
|
Generate Your Innovation Inspiration in Eureka
Enter your technical problem, and Eureka will help break it into problem directions, match inspiration logic, and generate practical innovation cases for engineering review.