Author Topic: Thermal conductivity of graphite  (Read 3487 times)

0 Members and 1 Guest are viewing this topic.

Offline vignesh_shanTopic starter

  • Contributor
  • Posts: 14
  • Country: in
Thermal conductivity of graphite
« on: December 01, 2020, 03:02:08 pm »
Hi,
Graphite sheets are better in thermal conductivity. But is it conduct only in surface. I mean if we place graphite sheet between processor and heatsink , will the heat transfer from processor case to graphite sheet and from graphite sheet to heat sink.

Some are saying that graphite sheet conductivity will very low across surfaces(z axis). High in x, y axis..

 Is this true ??

Thanks.
Vignesh.M
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 3163
  • Country: be
    • Freelance Hardware Engineer
Re: Thermal conductivity of graphite
« Reply #1 on: December 01, 2020, 03:18:11 pm »
Well, yes, just look at any datasheet from graphite heatpads:

https://docs.rs-online.com/484d/0900766b8142bbcd.pdf

Awesome (way better than copper) in plane, so great to spread heat around. Mediocre as far as "through thickness" transfer is concerned.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3322
  • Country: gb
Re: Thermal conductivity of graphite
« Reply #2 on: December 01, 2020, 06:54:38 pm »
I mean if we place graphite sheet between processor and heatsink , will the heat transfer from processor case to graphite sheet and from graphite sheet to heat sink.

Perhaps I am misinterpreting this, but how else would conductive heat transfer work?  All practical interface materials have thermal impedance, so you will see some temperature drop across the the interface.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14749
  • Country: de
Re: Thermal conductivity of graphite
« Reply #3 on: December 01, 2020, 07:02:29 pm »
Graphite is pretty anisotropic: the conductivity in the plane of the crystal is pretty good, but out of plane it is pretty horrible. It is no just thermal conductivity, but some other properties too.

The other question is than how the graphite crystals are oriented in the pad - this does not have to kind of match the physical dimensions / symmetry.  For a real life product at an affordable price it would not be a single crystal, so no such extreme values and additional heat resistance from internal interfaces.
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 3888
  • Country: us
Re: Thermal conductivity of graphite
« Reply #4 on: December 01, 2020, 09:28:26 pm »
The pad that Ice-Tea linked to has an anisotropy of around 100:1 which is pretty significant.

But basically the keys are: 1) The pad is very thin (25 micron).  So even poor thermal conductivity in the Z direction won't matter much, and 2) The XY conductivity more than makes up for it.  Remember, heat transport is thermal conductivity times cross section.  By conducting heat laterally very effectively, you get to use more cross section to conduct in the Z direction which is the whole point of heat spreaders.


 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Thermal conductivity of graphite
« Reply #5 on: December 01, 2020, 11:09:46 pm »
All surfaces are not perfectly flat. Whether it's the heat sink or the surface of the component there are always microscopic  pits  and deformations on the surface . Just using a thermal pads will not transfer heat any better than direct contact because of these imperfections of the surfaces . This is why a paste is used to fill those deformations and pits so maximum heat transfer is distributed across the entire contact surface . Thermal pads are used mainly to isolate the component from the heat sink if isolation is required and should not be used without also addressing the surface of the contacting materials. I'm skeptical of claims of graphite pads with a thermal conductivity of 1500 W/m-k   Graphite's thermal conductivity is 168W/m-k . Also graphite mixed with some polymer to make the material pliable would further reduce it's thermal conductivity . Graphite is still excellent when compared to other materials like  Aluminum Oxide at  36W/m-k and Zinc Oxide at 50W/m-k . You must also consider the thermal conductivity of the heat sink as well . Aluminum is around 236W/m-k and Copper is around 400W/m-k . But regardless of graphite's superior thermal conductivity , the surfaces should still be polished or mated with a graphite paste for maximum surface contact.
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 3888
  • Country: us
Re: Thermal conductivity of graphite
« Reply #6 on: December 02, 2020, 07:01:56 am »
Graphite's thermal conductivity is 168W/m-k . Also graphite mixed with some polymer to make the material pliable would further reduce it's thermal conductivity

That is for a randomly oriented mix of graphite.  Graphite is anisotropic, and in plane has similar thermal conductivity to diamond, or in fact even higher.  I have heard of measurements of single layer graphene up to 5000 W/mK.  This isn't a single crystal graphite or graphene, but it is totally possible to have 1500 W/mK.

The pad linked to is not intended to be used for the same purpose as thermal compound or a sil pad which provides thermal contact or electrical isolation.  It is a heat spreader designed to transport the heat laterally so you can use a large heat sink effectively.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf