| Metadata | Details |
|---|
| Publication Date | 2023-05-01 |
| Authors | Henry A. Martin, Marcia Reintjes, Dave Reijs, Sander Dorrestein, Martien Kengen |
| Institutions | Delft University of Technology |
| Citations | 7 |
| Analysis | Full AI Review Included |
- Core Achievement: Successful heterogeneous integration of CVD diamond heat spreaders into Power Quad-Flat No-Lead (PQFN) packages to enhance thermal dissipation in high-power electronics.
- Thermal Performance: The thermally enhanced PQFN demonstrated a 9.6% reduction in maximum junction temperature (Tj) compared to a standard PQFN when subjected to a 6.6W input power (~10.5 W/mm2 heat flux).
- Heat Spreading Effect: The diamond spreader effectively redistributed heat, resulting in a significantly lower thermal gradient and more uniform temperature distribution across the active device surface.
- Material Optimization: Optimal layer thicknesses were determined via transient thermal simulation: a 50 ”m thinned Silicon die integrated with a 100 ”m CVD diamond heat spreader.
- Integration Method: A robust, double pressureless sintering process using nano-silver paste was employed for bonding the die-to-diamond and the diamond-to-leadframe stack.
- Reliability: The enhanced package exhibited excellent thermo-mechanical reliability, showing less than 2% thermal degradation after 200 thermal cycles ranging from -55 °C to 150 °C.
- Efficiency Scaling: The efficiency of the diamond heat spreader was found to increase with increasing applied input power, making this solution highly beneficial for high-power applications (greater than 6.6W).
| Parameter | Value | Unit | Context |
|---|
| Diamond Thermal Conductivity (Grade 1) | 1800 | W/mK | CVD Diamond used as heat spreader material |
| Optimal Silicon Die Thickness | 50 | ”m | Thinned from 400 ”m for optimal performance |
| Optimal Diamond Spreader Thickness | 100 | ”m | Thickness chosen for experimental study |
| Maximum Input Power (Test) | 6.6 | W | Applied at Resistor 1 location for 1 second |
| Maximum Heat Flux (Test) | ~10.5 | W/mm2 | Equivalent flux density on the active area |
| Tj Reduction (Max) | 9.6 | % | Reduction achieved at 6.6W input power |
| Standard PQFN Max. ÎTj | 43.28 | °C | Maximum temperature rise recorded at 6.6W |
| Enhanced PQFN Max. ÎTj | 39.13 | °C | Maximum temperature rise recorded at 6.6W |
| Thermal Degradation (Reliability) | < 2 | % | After 200 thermal cycles (-55 °C to 150 °C) |
| Sintered Interface Thickness | ~35 | ”m | Thickness of the nano-Ag sinter layer |
| Resistor 1-10 Resistance Sensitivity | 0.462 | Ω/°C | Used for temperature sensing (TCR) |
| Spiral Resistor Resistance Sensitivity | 72 | Ω/°C | Used for temperature sensing (TCR) |
| Back-side Metallization Stack | 100/200/600 | nm | Ti/Pt/Au layer structure |
- Substrate Thinning: Commercial Thermal Test Chips (TTCs) were mechanically thinned from 400 ”m down to 50 ”m via a back-side polishing process to minimize the intrinsic thermal resistance of the silicon.
- Metallization: The polished back-side of the TTCs and the 110 ”m CVD diamond slabs were sputter coated with a Ti/Pt/Au stack (100/200/600 nm) to ensure robust adhesion during subsequent sintering.
- Die-to-Diamond Sintering: Nano-Silver sinter paste was dispensed onto the metalized diamond slab. The thinned TTC was wet-mounted, and the stack was sintered using a pressureless sintering process under a Nitrogen atmosphere.
- Diamond-to-Leadframe Assembly: Nano-Ag paste was screen-printed onto the copper lead frame die pad. The pre-sintered Die-Diamond stack was placed onto the lead frame and sintered again (second pressureless sintering step).
- Electrical Interconnection: Electrical connections were established using 99.99% pure gold wire bonds (25 ”m wire, 50 ”m bump) to connect the TTC resistors (Resistors 1-3, 5-6, 8-10, and the spiral resistor) to the lead frame bond pads.
- Final Packaging: The wire-bonded stack was encapsulated via transfer molding using an epoxy molding compound, followed by singulation into the final PQFN package form factor.
- Thermal Testing: Packages were soldered onto a custom Printed Circuit Board (PCB) with 4-point Kelvin connections. Transient thermal measurements were performed inside a temperature-cycling oven (TMCL) at 25 °C ambient, applying 6.6W input power for 1 second.
- Reliability Assessment: Thermo-mechanical reliability was tested using 200 thermal cycles, ranging from -55 °C to 150 °C, with in-situ measurement of junction temperature at 25 °C intervals.
- Electric Vehicle (EV) and Hybrid Electric Vehicle (HEV) Systems: Essential for managing the increasing power density and heat flux in automotive electronics, which are critical for vehicle performance and reliability.
- High-Power Radio Frequency (RF) Devices: Applicable to Gallium Nitride (GaN) High Electron Mobility Transistors (HEMTs) and other high-frequency devices where concentrated thermal regions (hot spots) limit performance and lifetime.
- Industrial Power Conversion: Used in high-efficiency power modules and inverters where device failure due to thermal breakdown is a primary concern.
- Advanced Heterogeneous Integration: Provides a viable back-end packaging solution for integrating materials with vastly different thermal properties, enabling multi-chip high-power packaging applications.
- Aerospace and Defense: Suitable for electronics operating in harsh environments requiring high thermal stability and proven reliability under extreme temperature cycling (-55 °C to 150 °C).
View Original Abstract
<p>Integrated Circuits and Electronic Modules experience concentrated thermal hot spots, which require advanced thermal solutions for effective distribution and dissipation of heat. The superior thermal properties of diamonds are long known, and it is an ideal material for heat-spreading applications. However, growing diamond films to the electronic substrate require complex processing at high temperatures. This research investigates a heterogeneous method of integrating diamond heat spreaders during the back-end packaging process. The semiconductor substrate and the heat spreader thicknesses were optimized based on simulations to realize a thermally enhanced Power Quad-Flat No-Lead package. The performance of the thermally enhanced PQFN was assessed by monitoring the temperature distribution across the active device surface and compared to a standard PQFN (without a heat spreader). Firstly, the thermally enhanced PQFN indicated a 9.6% reduction in junction temperature for an input power of 6.6W with a reduced thermal gradient on the active device surface. Furthermore, the diamond heat spreaderâs efficiency was observed to increase with increasing power input. Besides, the reliability of the thermally enhanced PQFN was tested by thermal cycling from -55°C to 150°C, which resulted in less than 2% thermal degradation over two-hundred cycles. Such choreographed thermal solutions are proven to enhance the packaged deviceâs performance, and the superior thermal properties of the diamond are beneficial to suffice the increasing demand for high power. </p>
- 2013 - Diamond-based heat spreaders for power electronic packaging applications