Met a scientist creating graphene heatsinks

[Phryq]

So I met a scientist today who's actually researching graphene as a thermal conductor, and trying to fix the problem of it conducting horizontally but not vertically.

I'm like - yo, let me join your lab so we can make me a good passive-cooling heatsink for my high performance CPU.

Her idea is to layer graphene and copper into sandwiches.

Thoughts?

 

[jeshikat]

If she can solve the problem then that'd be awesome. Graphene is really cool stuff but it's kinda useless as a heatsink due to the lack of thermal transfer in the vertical axis.

But how does sandwiching copper help? Because you still have the problem of transferring heat from the graphene layer up to the copper layer.

 

[Josh | NFC]

Well I do have a box of pencils I no longer use thanks to the invention of word processing--so if you need me to donate to the cause I am willing!

 

[Phryq]

If she can solve the problem then that'd be awesome. Graphene is really cool stuff but it's kinda useless as a heatsink due to the lack of thermal transfer in the vertical axis.

But how does sandwiching copper help? Because you still have the problem of transferring heat from the graphene layer up to the copper layer.

Maybe they do it on a nano scale? e.g. a Nano-layer of graphene and a nano-layer of copper.

I mean, if you can control individual atoms, could you have e.g. carbons bonding on a 2d plane, with certain carbons bonding to specific copper atoms, (but I guess that would make some kind of copper-carbonate). I mean, it's not normal chemistry, so I have no idea how this would work (not that I understand normal chemistry).

I can't find anywhere on the net the rate of thermal radiation from graphene into the air; only its conductivity.

 

[VegetableStu]

Wait, by horizontally you mean by the 2D plane of the nature of graphene? I'm thinking make it a tube and bend it upwards, like heatpipes

 

[Phryq]

Wait, by horizontally you mean by the 2D plane of the nature of graphene? I'm thinking make it a tube and bend it upwards, like heatpipes

That's what I'm trying to understand. I mean, you could simply bend it any way, line your entire passive-cooling case with it, but willl the heat conduct away from the graphene into aluminum/copper.

OR if you explose the graphene to open air, will it radiate (imagine graphene finned heatsink).

 

[Elerek]

if graphite only conducts heat well along the sheet, why not use it as the fins on the heatsink, so the heat can spread along the sheet from standard heatpipes?

 

[Kmpkt]

It wouldn't work for fins as it would not radiate anywhere except the top border. Think of the graphene as being a bunch of small heat tunnels where heat goes in one end and out the other.

 

[Reldey]

It wouldn't work for fins as it would not radiate anywhere except the top border. Think of the graphene as being a bunch of small heat tunnels where heat goes in one end and out the other.

What about if it was porous?

Edit: Ha, with a bit of google-fu: https://www.ncbi.nlm.nih.gov/pubmed/28534904

 

[Kmpkt]

If you had 100 tunnels and you punched holes at various points in all of them I think they would fail. Perhaps my understanding of their function is erroneous though.

 

[EdZ]

What about if it was porous?

Edit: Ha, with a bit of google-fu: https://www.ncbi.nlm.nih.gov/pubmed/28534904

While that paper is paywalled, a similar one is open access.

(green = bulk Copper, black = porous Copper, red = graphene-coated porous Copper)
It appears that regular old bulk Copper is a better choice below ~800K (~525°C) for conducting heat away from a source. However, the graphene-coated porous Copper did outperform bulk Copper for radiating heat (normalised for volume rather than mass) as well as outperforming porous Copper, so if you have a heatsink design currently using porous Copper the graphene-coated variant is a good upgrade.
Current heatsinks are far outperforming bulk copper in conductivity due to the use of heatpipes (phase-change transport). If you want to design a heatsink that outperforms a 'traditional' heatpipe + solid fin heatsink, you first need a design using porous copper that at least matches performance to use as a starting point.

 

[Kmpkt]

I'm leaning towards going the IC Diamond route and doing a diamond/copper 3D printed composite myself : )

 

[ChainedHope]

Very interesting concept and it has potential, but we aren't really there yet. Maybe she will figure it out.

I did some research during an internship with graphene for a small project and we tested using copper VIAs to help transfer the heat up. It ended in "Meh" status and was scrapped but it was a neat idea. In theory it worked, in reality it kinda caught fire (seriously, we melted the first prototype during validation testing lmao but that was human error not because of the design). The idea was to construct a material that used nano sheets of graphene with small hexagon holes filled with a proprietary copper structure that was melded into the nano sheets. The idea was to help get the heat to transfer through the copper to get it to the next layer of graphene. In the end the company decided it wasnt worth the investment tho.

 

[proch55]

Hopefully she gonna crack the problem. I am really looking forward for much more efficient computers enhanced with nano-materials. As an affordable cooling goes, nanodiamond based thermal pastes are pretty interesting. Its more expensive then the alternative but results are better.

 

[dumplinknet]

I don't think anything in nature can NOT absorb heat, thus heat transfer. Maybe greatly reduced it, but not block it.

 

[bledha]

that paper is paywalled

Sci hub