I can’t spare you a little bragging knowledge now, but if you didn’t sleep completely in school, you’ll be able to understand the following easily, because I’m trying to simplify it as much as possible.
Thermal expansion of materials
As the temperature rises, all materials, including the thermal compound, expand thermally. This is well known, whereby the thermal paste can expand both in its surface area and in its thickness. And because such a paste is a viscoelastic material, it has both flowing (viscous) and restoring (elastic) properties, which are collectively referred to as viscoelastic properties. Nevertheless, we need to carefully separate the two at this point.
Viscous reaction
At higher temperatures, the material usually becomes softer. This means that it tends to flow more, especially under the pressure that exists between the two contact surfaces. So as the temperature rises, the paste may start to flow sideways, but it can still increase in thickness as the materials move apart due to thermal expansion.
![](https://www.igorslab.de/wp-content/uploads/2024/04/GPU-and-Grease-980x696.jpg)
Elastic reaction
The elastic component of the paste can cause it to attempt to retain its original shape when stretched by thermal expansion. As the temperature rises and the adjacent materials expand, the paste can develop a restoring force that favors an increase in thickness.
Reduced support function, pump-out and bleeding
At high temperatures (and pressures), there is another important factor for the development of the so-called pump-out effect: the loss of the support function. This is because at high temperatures the paste can become softer and less resistant to pressure loads. However, if the paste is less viscous, it offers less resistance to “bleeding” or lateral displacement. At the same time, the paste can expand upwards (i.e. in thickness) due to the reduced viscosity and thermal expansion of the material.
![](https://www.igorslab.de/wp-content/uploads/2024/04/Grease-980x735.jpg)
The “bleeding” of thermal pastes refers to the leakage of liquid components from the paste over time. The liquid that bleeds is usually the oil or carrier that holds the solid particles (such as metallic or ceramic fillers) in suspension in the paste. The heat can cause the liquid components of the paste to become less viscous and separate from the solid fillers, causing the liquid to bleed. This may be due to poor mixing of the components or the use of inferior oils and fillers. Caution: If too much paste is applied, there is a higher chance that the excess will separate and bleed. A thin, even layer is normally sufficient to ensure optimum heat transfer.
In some cases, and with lower quality pastes such as Thermal Hero Quantum, the release of gases trapped in the paste can also play a role at high temperatures, most notably in a change in behavior after several cycles have been completed. When these gases are released, they can form bubbles within the paste and lead to an increase in BLT.
Change in material properties
The material properties of the thermal paste, such as cohesion and adhesion to the contact surfaces, can change with increasing temperature, which in turn has a very negative effect on the interface resistance. At higher temperatures, degradation or chemical changes can also occur, causing the paste to swell or change its structure. Some materials that make up the thermal paste can start to decompose or react chemically at higher temperatures. This can also cause the paste to increase in volume and the BLT to increase. Like many materials, thermal pastes can also age over time. The chemical structure of some components of the paste can change over time, which can lead to separation of the components. This is particularly common with pastes that use silicone oil as a carrier, as silicone oils tend to oxidize and become volatile.
Development of shear forces
An increase in temperature often results in shear forces within the paste, especially if the surfaces (heatsink, die/IHS) have different coefficients of thermal expansion. This can lead to thermal stresses developing and the paste also being stretched in thickness while the surfaces move relative to each other.
![](https://www.igorslab.de/wp-content/uploads/2024/08/08-LIBS-7-980x735.jpg)
Summary and conclusion
The bondline thickness (BLT) increases with increasing temperature mainly due to the thermal expansion of the materials, the decrease in viscosity and the associated softening of the paste, the release of gases and bubble formation as well as the change in material properties and the effect of shear forces. All of these effects can reduce thermal efficiency, as a thicker heat conducting material has a higher thermal resistance (thermal resistance Rth).
It is precisely this behavior of the heat-conducting paste that also leads to constant degradation, which is faster or slower depending on the composition. Nobody except the manufacturers can be expected to carry out tests with 1000 or more cycles, but if you are testing thermal pastes as a reviewer, you should also carry out at least three longer cycles of warming up and cooling down in the “real-world tests” before determining a measurement result. This is because poor pastes in particular will quickly fall through the cracks and won’t confuse the reader with performance explosions that are only available for a short time.
If you measure with test setups whose tolerance limit is already at least 1 to 2 Kelvin, then you should at least include the degradation factor. Although this increases the time required enormously, it also ensures that posers such as the Thermal Hero Quantum can be reliably identified and then correctly classified. I don’t really want to write any more about this, because it’s already enough. Just to give you food for thought as to why some pastes can no longer deliver what they promise on closer inspection and why my results may differ from others at first glance.
66 Antworten
Kommentar
Lade neue Kommentare
Veteran
Veteran
Veteran
Urgestein
1
Urgestein
1
Urgestein
Urgestein
1
Veteran
Urgestein
1
Veteran
Urgestein
Urgestein
Urgestein
Mitglied
1
Alle Kommentare lesen unter igor´sLAB Community →