I have been reading a fair bit about Thermal grease & adhesive comparisons. There is a lot of information out there - even including tests made with tooth paste(!), which is mind blowing for me.

I have access to a limited variety of thermal grease & adhesive products, however. And I would like to understand what to look for when buying these products.

My question:

Given two similarly priced thermal grease, paste, or adhesives:

  • Do I go for the one with higher conductivity level AND lower impedance, or vice-versa (ie., lower the conductivity & higher the impedance), or other?

Higher thermal conductivity (lower thermal resistance/impedance) means less temperature drop, all other things being equal. That is what you want as it will minimize the heating of the semiconductor die by minimizing the temperature difference between the die and ambient.

The purpose of thermal grease is to fill the small (air) gaps between surfaces that are not perfectly smooth. If the surfaces were ground smooth and flat to a mirror finish, thermal grease would be of limited value as it would space the surfaces apart. Very smooth metal blocks such as gauge blocks used by machinists and in metrology labs can be stuck together simply by twisting (wringing) them together, and the gap is negligible.

I have doubts about materials that may contain solid particles such as toothpaste- even if the thermal conductivity seems acceptable, the minimum layer thickness may be too high. If they're water-based, the water will dry out leaving air gaps. Grease has high viscosity but can be squeezed quite thin, and a good grease will not dry out.

  • \$\begingroup\$ thank you for explaining the question! can I just ask if higher thermal conductivity always equals lower resistance, ie, these two values are always related to each other in the opposite sense? also, for example, is it possible that an aluminium heat sink, which looks and feels very smooth and flat to the human eye & touch, has gaps that also need filling with these heat transferring components? \$\endgroup\$ – Phil Dec 26 '14 at 20:15
  • 2
    \$\begingroup\$ Yes the thermal conductivity and thermal impedance numbers will always be related and inversely proportional. Note that even super polished aluminium will have a somewhat microscopic grainy surface. It is not just feeling flat and smooth that counts. It could feel flat but still be out of flat by multiple thousandths of an inch. \$\endgroup\$ – Michael Karas Dec 26 '14 at 20:24
  • 1
    \$\begingroup\$ If the units are the same, one is just the reciprocal of the other. Unless both surfaces are both incredibly flat and look just like a perfect mirror (not just pretty flat and smooth) you can benefit from thermal grease, which is to say just about anything that doesn't cost a fortune to manufacture. \$\endgroup\$ – Spehro Pefhany Dec 26 '14 at 20:30
  • 1
    \$\begingroup\$ @MichaelKaras: You are right about apparent flat surfaces which are in fact not flat. On the other hand, that there is residual roughness resulting in an air gap does not mean you will need thermal grease to reach your needed thermal impedances in your system. Air has a conductivity, too. And if the gap is small enough it can be neglected. \$\endgroup\$ – Ariser Dec 26 '14 at 20:44
  • 1
    \$\begingroup\$ @Phil: Whether you need a thermal grease or not depends on several parameters. If you can spread the heat inside your primary heat generating module to a high surface, you can pass on using a thermal grease. But you better formulate a new question with exact parameters, because this is way to complex to elaborate in comments. \$\endgroup\$ – Ariser Dec 26 '14 at 20:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.