Im seeking an alternative to thermal grease/ pastes. I was researching graphite pads when I stumbled across the soft PGS line from Panasonic.

I need to choose a suitable thermal interface material to connect the CPU die to a heat sink. I do not wish to use thermal paste ( it is greasy). I was wondering which material would be most suitable.

PGS :Excellent thermal conductivity : 700 to 1950 W/(m·K) ( no data for z direction; I found some sources stating that the z direction was about 15W/m-K but I can not verify this).

“GraphiteTIM(Compressible Type) PGS with thermal conductivity : X-Y direction 400W/m∙K, Z direction (28W/m∙K).

The heat sink will be made of aluminium ( roughly 200-250W/m-k) or copper (roughly 350-386W/m-k).

Which is more important? XY thermal conductivity or Z direction thermal conductivity?

Since the heat-sink has a defined thermal conductivity, is there any advantage in choosing a conductivity greater than the heat sink's value? (Example, two materials A and B with a thermal conductivity of 700 and 1000W/m-k respectively for xy plane; assume both have similar z direction conductivity. Does choosing the 1000W/m-k over the 700W/m-k material provide any benefit to the overall system, considering that the heat sink is limited to 386W/m-k? Thanks.


Since the interface material is very thin, its thermal resistance in the z direction will be very low compared to the much thicker heatsink as long as its thermal conductivity is similar to or higher than the heatsink's. Thermal resistance = thickness / thermal conductivity and thermal resistances in series add to give the overall thermal resistance. So I think small differences in z direction thermal conductivity are not important since they're dwarfed by the much higher thermal resistance of the heatsink in series with it.

There would be value in a higher xy thermal conductivity than the heatsink material because that would spread the heat over wider area of the heatsink which would increase the heat flow rate overall as well as reducing the temperature at the interface. I think a small increase in this property would be significant as long as it's already similar to or higher than the heatsink's thermal conductivity.

Here's how I visualize heat flow with a high xy thermal conductivity interface. Without that interface, it would be more concentrated directly over the die, leaving the outer bottom edges of the heatsink relatively unused.

enter image description here

By the way, please only use capital K for kelvin and dot for multiplication in SI units since those are standard. The hyphen is a minus sign, not a multiplication sign! W/(m·K), not W/m-k.

  • $\begingroup$ Thanks, I think I saw a website used that type of notation for the units. So I just copied it. $\endgroup$
    – Void
    Jan 23 '19 at 3:01

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.