Gold instead of thermal paste?

Q

Quartz-1

Supreme [H]ardness
Joined
May 20, 2011
Messages
4,257
Has anyone tried using pure gold instead of thermal paste? Is it sufficiently ductile? At ~£35 a gram, and a density of ~20g / cc, a 1mm x 10mm x 10mm layer will cost ~£70. Not that unreasonable for the enthusiast. And it has a thermal conductivity vastly higher than silver paste.
 
If it worked better, I figure we would have seen it being sold by now. I mean shit, we have diamond based compound now, why not gold?
 
Jorona said:
If it worked better, I figure we would have seen it being sold by now. I mean shit, we have diamond based compound now, why not gold?
Click to expand...
The inability to fill micro gaps between the IHS and HSF base?And the tiny price/performance gain over other solutions on the market today?
 
Wiki says that Indium is already used. It has a thermal conductivity of 81 W/m-1. And it's far cheaper.
 
Lattice oriented CVD diamond wins by a longshot.Thermal conductivities in the thousands ...but the 'lattice oriented' part is really important, you can't just grind up diamonds and put it into thermal paste , it doesnt work that way.

Silver is the best 'metal' as far as thermal conductivity goes. Metals like indium and gallium are typically used with silver to produce very soft alloys suitable for TIM use.

You 'could' use gold , but why .

----

Here's my attitude towards fancy TIM's :

Any properly designed thermal solution should operate within spec , without thermal paste.The paste is just a bonus...If it won't , you should be looking at things like mating surfaces (flatness etc) , compression , etc...

I use MG chemicals thermal paste.

:D
 
The operational difference between most thermal pastes is about 5C, if you ignore a couple of outliers that performed particularly badly:

http://www.hardwaresecrets.com/article/Thermal-Compound-Roundup-February-2012/1490/5

The best paste in that particular roundup is aluminum/zinc oxide paste. There are diamond pastes that perform a couple of degrees worse. Mayonnaise is in the middle of the pack.

The only reasonable conclusion you can draw from this is that the thermal conductivity of an element has little relation to its effectiveness in a thermal paste.

But your question is about using a solid between the CPU and the heatsink. How are you planning to perfectly fill the space between them with solid gold? If you can't do that, then the result will be worse than any thermal paste.
 
It would be pointless, what the paste is made out of has been shown to matter little other than with the likes of liquid metal types, because of little to no fillers. What matters are base and IHS type and shape, if the HS base and IHS do not match up well or the base is very rough, thicker TIM's will do better, however when you get into very nice bases or lapping, the thinner TIM's tend to do better as there is more metal to metal contact and the thin TIM's do not hinder this.

I noticed this when lapping many times, using a standard TIM over something like LP, with both unlapped parts the change was 1-5C depending on the setup, however, once lapped the really thin LP jumped ahead with upwards of 10-11C over the normal TIM's. So rather than looking at fancy TIM's, you should be looking at the matting surfaces and try to max out metal to metal contact.
 
evilsofa said:
The only reasonable conclusion you can draw from this is that the thermal conductivity of an element has little relation to its effectiveness in a thermal paste.
Click to expand...

On the contrary..

But your question is about using a solid between the CPU and the heatsink. How are you planning to perfectly fill the space between them with solid gold? If you can't do that, then the result will be worse than any thermal paste.
Click to expand...

I agree, hence my asking if gold were sufficiently ductile.
 
BlueFireIce said:
It would be pointless, what the paste is made out of has been shown to matter little other than with the likes of liquid metal types, because of little to no fillers. What matters are base and IHS type and shape, if the HS base and IHS do not match up well or the base is very rough, thicker TIM's will do better, however when you get into very nice bases or lapping, the thinner TIM's tend to do better as there is more metal to metal contact and the thin TIM's do not hinder this.

I noticed this when lapping many times, using a standard TIM over something like LP, with both unlapped parts the change was 1-5C depending on the setup, however, once lapped the really thin LP jumped ahead with upwards of 10-11C over the normal TIM's. So rather than looking at fancy TIM's, you should be looking at the matting surfaces and try to max out metal to metal contact.
Click to expand...

LP = ???
 
Gold is too pricey. Ultimate material for modern TIM would be a pad or paste flavor of graphene.

Graphene_from_gases_for_new,_bendable_electronics_




This would make the ultimate TIM in form of pad/grease/liquid application.



From: ACS.ORG
nl-2007-031872_0005.gif


We report the measurement of the thermal conductivity of a suspended single-layer graphene. The room temperature values of the thermal conductivity in the range (4.84 ± 0.44) × 103 to (5.30 ± 0.48) × 103 W/mK were extracted for a single-layer graphene from the dependence of the Raman G peak frequency on the excitation laser power and independently measured G peak temperature coefficient. The extremely high value of the thermal conductivity suggests that graphene can outperform carbon nanotubes in heat conduction. The superb thermal conduction property of graphene is beneficial for the proposed electronic applications and establishes graphene as an excellent material for thermal management.
Click to expand...
 
Last edited:
Quartz-1 said:
I agree, hence my asking if gold were sufficiently ductile.
Click to expand...

Ductility doesn't matter. It's still a solid and won't fill the microscopic imperfections necessary for a thermal interface.
 
SonDa5 said:
Gold is too pricey. Ultimate material for modern TIM would be a pad or paste flavor of graphene..
Click to expand...

Same as diamond.

Carbon lattices have amazing conductivities , but , just like diamond , lattice orientation matters , paste no......pads maybe.

Lets not forget the nanotubes......more carbon.Still directional.

-----

I actually priced out a 1" diameter, 0.200 " thickness oriented lattice CVD diamond chip (1800 W·m−1·K−1 ) for a waterblock once.
$8000.

----

To answer the question , No , Gold is not ductile enough.Not at the clamping pressures on a computer.

:D
 
You must log in or register to reply here.
Back
Top