Thermal Interface Material Thickness and Contact Resistance

The thermal path from a hot component to a heatsink or chassis is only as good as the interface between them. Two flat surfaces that look smooth in the hand touch at a small number of high points, and the air trapped in the gaps is a far better insulator than either material. A thermal interface material fills those gaps, and the thickness it ends up with after assembly is the single most important number in the whole joint. This article covers how to choose a material, how thickness and contact resistance interact, and how to keep the result stable in production.

Why Interface Thickness Decides Performance

Heat flows through the interface by conduction, and conduction through a thin layer improves as the layer gets thinner. A material with a modest bulk conductivity applied as a very thin film can outperform a premium material applied as a thick one. This is why the choice of material cannot be separated from the mechanical design that sets the gap.

The applied thickness is set by the surfaces, the load and the material, not by the specification sheet. A gap that varies across the component footprint produces a bond line that varies with it, and the hottest point of the die sits over the thinnest part of the gap rather than the average. Controlling the geometry is therefore part of thermal design, not a separate assembly detail.

What a Thermal Interface Material Has to Do

A thermal interface material has to displace air, conform to the surface texture without being squeezed completely out of the joint, and remain in place for the life of the product. It also has to survive the temperature range, the thermal cycling and the vibration of the application without pumping, drying or hardening into a brittle layer.

Those requirements pull against each other. A soft material conforms well at low load but pumps out under cycling, while a hard material stays in place but needs more load to conform. The right answer depends on how much clamping force the mechanical design can deliver and how much the joint will move in service.

Pastes, Gap Fillers and Thermal Pads

Thermal grease or paste offers the lowest thermal resistance because it can be applied very thin, but it needs a well controlled dispense and a mechanical design that maintains the load. Gap fillers are dispensed and cured in place, and their advantage is that they accommodate large and variable gaps that a paste cannot bridge without being pushed aside.

Thermal pads are preformed, which makes them consistent and easy to assemble, but their bulk conductivity is usually lower than a paste and the interface resistance is higher because the pad does not wet the surface. A pad is often the right choice for a low power joint with a modest load, while a paste or filler suits a high power joint with tight flatness.

Bond Line Thickness and Contact Resistance

The bond line thickness is the final thickness of the material in the assembled joint, and it is the value that appears in the thermal calculation. Total resistance is the bulk resistance, which scales with thickness, plus the two contact resistances at the surfaces, which do not. Reducing the bond line therefore reduces the part of the resistance that scales, and it also tends to reduce the contact resistance because the material is pressed harder into the surface texture.

Thermal interface material applied between a component and heatsink

There is a limit. Below a certain thickness the material no longer covers the surface completely and dry contact patches appear, so total resistance rises instead of falling. Finding that minimum for a given pair of surfaces is a matter of measurement, and it is the reason a supplier will quote a recommended range rather than a single value.

Compressive Load and Deflection

Load is what turns a compliant material into a thin one. A pad specified to deflect a certain percentage of its thickness at a given pressure will not reach that deflection if the hardware cannot supply the pressure, and the joint will run hotter than the calculation predicted. Springs, shoulder screws and standoffs are the tools that set the gap and the load together.

Uneven load is as damaging as insufficient load. Bolting a heatsink at four corners with a flexible base bows the interface and gives a thin centre and thick edges, which redistributes the heat path away from where it is needed. A stiffer base or a distributed clamping scheme keeps the bond line even.

Pump-Out and Long Term Stability

Thermal cycling makes the two surfaces move relative to each other because they expand at different rates, and a low viscosity material is progressively pushed out of the joint. The result is a bond line that starts thin and ends thinner, with voids at the edges and a slow rise in temperature that is easy to miss in a short test.

A higher viscosity, a filler that increases the yield stress, a controlled bond line and a stable clamping load all resist pump-out. Where the joint will see many thousands of cycles, a short thermal cycling test on a real assembly is a better predictor than a data sheet figure measured on a laboratory press.

Electrical Isolation and Breakdown

Where the interface must also insulate a live tab from a grounded heatsink, the material becomes a dielectric and has to meet a breakdown voltage with margin. Alumina and boron nitride filled materials are common for that role, and the thickness required for the voltage also sets the minimum bond line, which then fixes the thermal resistance.

The dielectric requirement can dominate the design. A joint that needs a thick isolation layer cannot be optimised for the lowest thermal resistance, and the honest approach is to choose the assembly geometry around that constraint rather than to chase a conductivity figure that the thickness will erase.

Dispensing and Placement Control

Paste and gap filler are dispensed by volume, so the dispense pattern and the shot size control the final bond line. A dot pattern that leaves air traps, a shot that is too small to cover the footprint, or an expired material that has thickened in the syringe all produce a joint that looks assembled and performs badly.

Pads are placed by hand or by machine, and the failure mode is different: a pad that is wrinkled, contaminated or stretched during placement will not conform, and a pad that has been repositioned after touching a surface carries debris on both faces. Our component tolerance notes cover the placement discipline that this shares with the rest of the assembly process.

Verifying the Interface in Production

An interface cannot be judged by appearance once the heatsink is fitted, so verification has to rely on measured outcomes. A thermocouple or thermal camera reading on a sample assembly under load is the most direct check, and a bond line measurement on a cross section confirms what was actually achieved rather than what was intended.

Cross section showing bond line thickness of a thermal interface

At gopcb, the underlying board geometry that the interface sits on is controlled through our thermal management design rules, and the fabrication tolerances that affect flatness and thickness are recorded in our fabrication notes. Where the joint also carries current, our current capacity guide explains how the copper and the interface interact.

FAQ

Is a higher thermal conductivity always better? Not by itself, because the interface resistance and the achievable bond line matter as much as the bulk figure. A high conductivity material applied too thickly can be worse than a modest material applied thin.

How thick should the bond line be? As thin as the surfaces and the load allow without leaving uncovered patches, which for many assemblies falls in the range recommended by the material supplier. The value should be confirmed on a cross section rather than assumed.

Can a thermal pad replace thermal grease? In low and medium power joints with an adequate and even clamping load, yes. In a high power joint where resistance must be minimised, a paste or dispensed filler usually performs better.

Leave A Comment