Heatsink Attachment and Thermal Interface

The Thermal Path

Heat leaves a component through a chain of materials, and the weakest link decides the temperature. From the die the heat travels through the package, through the interface between the package and the heatsink, into the heatsink and then to the air. The interface is usually the weakest link, because the two surfaces are not perfectly flat and the air in the gaps is an excellent insulator. The purpose of a thermal interface material is to displace that air with something that conducts better, and the purpose of the mounting is to hold the surfaces close enough that the material layer is thin.

Interface Materials

The common options are thermal grease, a thermal pad, a phase change material and a thermal adhesive. Grease gives the lowest thermal resistance because it can be applied very thin, but it is messy, it can pump out under thermal cycling and it provides no mechanical attachment. A pad is clean, easy to apply and consistent, but it adds a thickness and therefore a resistance, so it suits a rougher interface or a gap that cannot be closed. A phase change material behaves like a pad at room temperature and flows like grease at operating temperature, which combines the handling of a pad with the performance of grease. An adhesive attaches the heatsink and conducts heat, and it is used where a mechanical fixing is not possible.

Mounting Force and Flatness

The thickness of the interface layer is set by the pressure. More pressure thins the layer and lowers the resistance, up to the point where the board or the package is stressed. The force has to be applied evenly, which is why a heatsink with four screws should be tightened in a diagonal sequence and to a specified torque, and why a clip is designed to distribute its load. Surface flatness matters as well: a package with a convex top touches in one place, and the rest of the interface is filled with material that is thicker than intended. Where the surface is irregular, a pad or a phase change material is a more forgiving choice than grease.

mounted heatsink on a PCB assembly with thermal interface material visible

Board Level and Enclosure Level

The heat can also be removed through the board rather than the top of the package. A thermal via array under a device conducts heat into the copper beneath, which spreads it and can transfer it to a chassis or a metal case. That path has the advantage of avoiding a heatsink on the component, which may be difficult to fit, and the disadvantage that the laminate is a poor conductor, so the vias and the copper area have to be substantial. Where the enclosure is metal, a thermal pad between the board and the chassis is often the most efficient solution, and it is one that has to be considered at the design stage rather than added later.

Verification

An attachment should be verified by measurement, not by reasoning. A thermal image of the assembly under the worst case load shows whether the heat is leaving through the intended path, and a thermocouple on the package and on the heatsink shows the temperature difference across the interface, which is a direct measure of the interface resistance. Measuring the same assembly with and without the interface material, or with a different mounting force, quantifies the improvement. Where the result does not match the expectation, the cause is often the interface thickness or the flatness rather than the material.

Reliability of the Attachment

The attachment has to survive the life of the product. Thermal cycling causes the materials to expand and contract at different rates, and a rigid interface can crack or delaminate; grease can migrate or pump out, leaving a dry area; an adhesive can lose its bond; and a clip can relax over time. The verification should therefore include cycling, not only a single measurement, because an interface that performs well when new may not do so after a thousand cycles. Where the product has a long life or a severe environment, the choice of material should be made with that history in mind rather than from the initial thermal resistance alone.

Common Mistakes

The mistakes are consistent. Applying too much grease thickens the layer and increases the resistance rather than reducing it. Reusing a pad that has been removed and replaced leaves it compressed and ineffective. Tightening one screw fully before the others tilts the heatsink and opens a gap at the opposite corner. Placing a component so that the heatsink also loads a nearby part creates a mechanical stress that was not intended. And ignoring the height tolerance of the components means the heatsink does not seat on the tallest device as designed. Each of these is easy to avoid when the process is written down.

PCB manufacturing process

FAQ

What does thermal interface material do? It displaces the air between two surfaces, since air is an insulator, and so lowers the resistance of the joint.

Which material is best? Grease for the lowest resistance, a pad or phase change material for easier handling and tolerance to gaps.

Does more pressure always help? Only until the layer is thin; beyond that the board or the package is stressed and the gain stops.

How is the interface verified? By measuring the temperature difference across it under the worst case load, and by cycling the assembly.

Why does a re-used pad perform worse? It has already been compressed, so it cannot conform and its thickness is no longer as designed.

Conclusion

The thermal interface is usually the weakest link in the path, so choose the material for the gap, control the mounting force and verify with measurement. Design the thermal path with the enclosure in mind. Thermal design belongs to PCB design and layout, the assembly that carries it sits in PCB assembly, and the verification is part of PCBA testing. Thermal attachment for a new product is developed during prototype PCB assembly in 2026.

Leave A Comment