Diamond Compound

wizardmaxx

wizardmaxx

wizardmaxx

Associate
Joined
21 Aug 2006
Posts
1,908
Location
Stafford
Up to my must do it myself thing again and ordered 10g of commercial diamond powder.

Read a couple of writeups off of google and have no cooking clue what Polydimethylsiloxane or Polytetraflouroethylene is so I was just going to use the cheapest silicone based thermal compound (as runny as possible) to make my concoction.

My theory is that due to the 'distance' between the chip and cooler any form of diamond transfer is better than nothing.

If we have any scientists in the building that know what the two ingrdients are and why they were chosen please post.

Others like myself post thoughts on a suitable base material for the prototype compound.

Update on temps once fourmula has been completed. :p
 
PTFE is the chemical name for Teflon. It's a low friction fluoropolymer.
PDMS is a silocone that has good 'flow'characteristics.

Although I am a materials scientist by training, I didn't need any of that to use Google. ;)

Both could be used to enhance a thermal interface material which is what I'm guessing you've going for.

You basically want something that will hold the diamond in place and not be too viscous so as to pour over onto the motherboard, or too stiff and not fill into the micron-sized gaps in the CPU heatspreader/heatsink as required. Which is probably where PDMS comes in.

Respect for the effort you're going through, but unless you spend time optimising your formula, you will probably wnd up with worse results than the stuff that comes with cheap heatinks.

Good luck!
 
I wish you luck :)

I'd be interested in trying IC Diamond 7 if I could find it in the UK. At the moment I'm using the excellent Liquid Metal Pro but it's a chore to remove.

ICD7:

diamond particle loadings @ 92% by weight, 94% after 10 minute dry-out prior to heat sink installation. Material loading above 90% is recommended as the best combination of rheological and thermal properties to minimize interface pump out due to thermal cycling. < 40 µ maximum particle diameter. Silicone free.


Hum, I think my first base would be OCZ Freeze if it doesn't make to thick a paste. I find it very consistent on an application by application basis. So, try pure Freeze, measure, try Wizard Compound, measure, another control application of Freeze. Repeat until happy :)
 
Monkeypuzzle - Cost me £4 with p+p. No particle size, but we will se when it arrives.

Mrthingyx - Thanks for the info. Know where I can buy some or a suitable substitute? Was thinking I would steal the wife's silicone hair stuff, but will have to see what else is in the bottle.

Ecat - Thanks. I will look into that. Is the stuff stiff or will it accept some 'powder'?

The basic idea is to mix the powder with a base compound to achieve as said a not too runny and not too stiff consistency.

Any ideas open to the floor.
 
ecat said:
I wish you luck :)

I'd be interested in trying IC Diamond 7 if I could find it in the UK. At the moment I'm using the excellent Liquid Metal Pro but it's a chore to remove.

ICD7:

diamond particle loadings @ 92% by weight, 94% after 10 minute dry-out prior to heat sink installation. Material loading above 90% is recommended as the best combination of rheological and thermal properties to minimize interface pump out due to thermal cycling. < 40 µ maximum particle diameter. Silicone free.


Hum, I think my first base would be OCZ Freeze if it doesn't make to thick a paste. I find it very consistent on an application by application basis. So, try pure Freeze, measure, try Wizard Compound, measure, another control application of Freeze. Repeat until happy :)
Click to expand...

Im thinking about trying out the Liquid Metal Pads. Im not too hot about the liquid metal pro as it contains gallium. Although the warning deems that it only eats alu like crazy Iv'e heard that it very slowly goes through copper. Def don't want that happening to my HS. If you take off your HS do you notice if the base has changed colour at all?
 
Hey. Materials science here too, but only a student. Just to say thanks for your post, according to wikipedia diamond beats copper by about 5 times over for thermal conductivity. It's wiki, but still.

Interesting question though, I can see it beating as5 or similar. I wonder which part is limiting, heat flow from processor to paste, from paste to heat sink, or from paste to diamond. I think there are too many interfaces between chip and heatsink with the current pastes, rather than the conductivity of the particles within the paste being too low.

At a guess, the best temperature would be achieved by lapping heatsink and processor down to their bare copper and bonding the two together. Two very clean, flat copper surfaces will weld to each other at room temperature. Failing that, welding is not possible. Chemically welding might be, I'm unsure how hot that gets. This would be an effort to achieve, and I think make it impossible to remove the processor from the motherboard.

Definitely a more interesting question than the maths I'm working on :)
Looking forward to test results, could conceivably be quite significant. I'd suggest mx2 as the base, as it tends to rival as5 by itself so mx2 + diamond should be a winner. Might be quite difficult to spread sadly :(
 
JonJ678 said:
Hey. Materials science here too, but only a student. Just to say thanks for your post, according to wikipedia diamond beats copper by about 5 times over for thermal conductivity. It's wiki, but still.
Click to expand...

Wikipedia is just as accurate as the Encyclopedia Britannica on the whole, although EB don't like that being known.

For contentious subjects, one has to form their own view, however.

A couple of things to consider:

CPU heatspreaders look like their electroplated (I think it's a nickel alloy finish) so the size of pores in the surface is going to be extremely variable at best with gaps between the nanometre and tens of microns. As such, you're going to have to be careful with the size of the diamond particles. If they're the consistency of sand, they're going to wreck any continuity between the two surfaces you could have hoped for.

If you're getting stuff which is in the nanometre range, this will do a good job of filling gaps. The bigger the particle size, the worse the effective conductivity will be.

The ultimate surface preparation for thermal interface material/grease would be matched surface finish to particle size, allowing for as continuous an interface (think Lego, but smaller). An alternative would be soldering the heatsink onto the chip, but this is not really a practical consideration.

What you could try (assuming you're brave enough and have access to an electron microscope) is plasma deposition of gold, carbon or platinum onto the heatsink base or (if you're really, really brave) your CPU so that you get a layer of 'loosely' (nanoscopically speaking) packed, highly conductive material on the surfaces required to transfer heat to one another. Clamp them together and you've got an alternative to a continuous interface.

Where would you get PDMS? Try your missus' shampoo - it's what makes her hair feel soft, silky and glossy. If you're metro, try your own.

Try Google for an alternative supplier.
 
About wiki: I generally go with:

If it's about maths/science/computers, it's right 99% of the time (a mistake in an article, rather than a wrong article)

If it's about anything opinion-based eg religion/wars/cults or anything moral-based (especially religion!), take it with a pinch of salt.

Very interested to see how this works out - looks like it should be a fun project, if nothing else :-)
 
Should have it in a day or two. Found el-cheapo silicone conpund as well so may have some results over the weekend.
 
Last edited:
Had a thought. I've lapped things with powdered diamonds in some form of solution. It was blue, and is probably electrically conductive. It might also be very good as a thermal conductor though
 
I will post a picture later of the mysery pixie dust.

All seems not to be lost. I have started to 'seperate' the big bits from the small bits so to say. Something I kinda of learnt making biodiesel.

I should have a decent seperation in about 24 hours. :cool:
 
I am seperating it using the highly scientic method of washing it with water. It was either that or picking out the big bits with a tweezer and that didn't appeal to me so much. No idea on size, but will be really small as only the particles that could be kept in suspension will be used.
 
I can conclusively declare that this did not work.

Oh well I will keep looking for some descent diamond powder as it is not eaily obtainable in small quantities as most supplies are from the USA and postage cost is mentally high.
 
Sorry to here this, even the smallest grains were to large ?

Still, if you mix the stuff with goose grease you can make a +5 damage coating for your sword, or am I getting my realities confused again ?
 
You must log in or register to reply here.
Back
Top Bottom