Thermal grease is used to lower thermal impedance between a heat source and its sink, but do you test how well the various greases perform?
Thermal grease doesn’t get a lot of attention, as it certainly lacks the glamour of many other active and passive components or materials. Still, it’s widely used between a heat source and heat sink to ensure a good thermal interface with as low a thermal impedance as possible between the two, facilitating heat flow from source to sink.
Nonetheless, even a material as humble as thermal grease deserves some attention, yet it is often an underappreciated item in the assembly BOM. That’s why I was fascinated when I came across a 2008 report from the National Renewable Energy Laboratory (NREL), “Thermal Interface Materials for Power Electronics Applications,” which addressed the basic question: “How good is that grease, anyway?” While the report is almost 10 years old, I think it still has validity in both the setup and the results.
Figure 1 shows the role of grease in a typical application with an IGBT. The report is clear: “The silicon die is soldered to the direct bond copper (DBC) layer, which is composed of an aluminum nitride layer sandwiched between two copper layers. This DBC layer is soldered to a copper baseplate, and the grease layer serves as the interface between the baseplate and the heat sink. This thermal grease can be up to 100 microns thick (bond-line thickness, or BLT) and, depending on the formulation, it has a thermal conductivity in the range of 0.4 to 5 W/mK.”
It’s one of several common thermal interface materials (TIMs) which also includes polymer composites (thermal pads) and solders. There are alternatives to grease, yet it is still a preferred solution when the two surfaces to be interfaced have challenging mechanical or physical characteristics in size, shape, finish, or irregularity. While widely used (and seeing an automated dispenser in action is impressive), these greases bring unique issues that impact the overall assembly process, and even longer-term concerns such as drying out.
The interesting thing in the report was the test technique. How do you set up a reliable and consistent test scenario? The NREL test stand shows the overall installation schematically (Figure 2).
What is not shown is how they maintained co-planarity and uniform thickness of the grease layer between the metering block surfaces: they used tiny glass spheres of consistent diameter and a pneumatic press using closed-loop load control via LabView. (These spheres are a standard item in labs, where they are used both as spacers and for linear-measurement calibration.) Similar hot plate/cold plate stacks are widely used in thermal tests, but this one had the added challenge of maintaining the grease thickness layer with a known amount of stack pressure, while not affecting thermal flow.
Going literally from top to bottom, their test setup seems to be consistent in how it measured the thermal performance of the GUT (grease under test). Their table of results for various commercial greases shows a wide range in thermal resistance, which is likely due to their differing compositions (which plays into cost, as well); it also shows the effects of different thicknesses. The pre-experimental modeling they did was also interesting, as it was simultaneously simple and complicated, and the modeling/simulation and experimental results were in close agreement.
Thermal grease is not necessarily the only cost-effective or preferred TIMs solution. NREL is working with Texas A&M University on a Defense Advanced Research Projects Agency (DARPA)-funded project to investigate a new type of of TIMs, which uses the chemical integration of boron nitride nanosheets (BNNS), soft organic linkers, and a copper matrix (see “New Class of Thermal Interface Materials Delivers Ultralow Thermal Resistance for Compact Electronics” and “Synthesis, Production and Characterization of Next Generation Thermal Interface Materials for Electronic Applications.”)
Have you ever used thermal grease as a source-to-sink interface? Did you use a thermal model to determine how well it would work, and were the results of the model in line with your real-world thermal results? Could differences in the grease model and its actual application have affected the test results?
Bill Schweber is an EE who has written three textbooks, hundreds of technical articles, opinion columns, and product features.