Thermal conductive epoxy: powdered diamond, only $1200 for 5g

Pressure is important, but can also be a problem. What you're doing by applying pressure are:

1. Compressing the TIM (thermal interface material) so that it fill the cracks and gaps.
2. Exude any excess TIM goo out of the gap, to produce a thinner TIM.
3. Straightening the base metals if they are not perfectly flat.
4. Compressing the metals so that they mix (alloy) at the interface. This probably takes more pressure than the typical heat sink can provide, but it might be possible if the metals are plated with a softer metal (gold or indium).

The idea behind a TIM is to *NOT* require that the heat sink be applied under pressure. Instead, the imperfections in the metal parts are filled by a thermal conductive paste, which requires only modest pressure. Applying too much pressure to a ceramic case RF power transistor or ceramic package CPU, will probably crack the case. In laptops and some desktops, the heat sinks use coil springs and shoulder screws to control the maximum pressure.

Your brass standoffs changed in contact resistance because the ends were not flat or polished, which reduced the electrical contact area. Half the bulk(?) resistance would imply half the contact area.

Like thermal resistance optimization, the objective is to maximize the contact area. For thermal, I can use either conductive or non-conductive paste or filler. For electrical, it has to be conductive. For brass, I guess(tm) you could smear the end of the standoff with solder paste, do the same to the other side, add flux, shove a large DC current (AC causes spatter) through the joint, and produce a soldered connection. If that's not good enough, think about submerged arc welding. If that's too complexicated, think about pounding the brass standoff into the base metal or heat sink using an interference fit. If a PCB or sheet metal, use a swaged standoff:

On Tue, 22 Dec 2015 15:51:18 -0800, Jeff Liebermann Gave us:

They were talking about making prophylactics from it too.. So they

It is also a much bigger molecular lattice than the Helium atom is, yet it will not all them to pass. They are talking about lining helium airships with it to reduce the inevitable leakage which currently occurs.

So I think it would make great ESD mats or one could impregnate carbon nanotubes into thin Teflon sheets for that.

But yeah, I think there can be applications for thermal abatement here.

You don't?

Is that elaborate enough (or at a high enough rate)?

Well, I'm sure you could conjure some application that might involve moving heat from here to there via graphene. However, for the immediate problem of finding the best TIM (thermal interface material) for RF power transistors and aluminum heat sinks, graphene is not a candidate.

Hardly. What's missing are numbers, calcs, references, URL's, buzzwords, acronyms, and other devices that allow me to distinguish between a likely fact and your opinion. You are of course entitled to pass judgement, offer an opinion, and inflict both upon this newsgroup. However, without substantiation, methinks opinions, pontifications, and judgements are generally worthless. In this case, how would you use graphene as a TIM and show some examples or prior art. If the idea is totally new, produce some numbers that might suggest that it's possible.

However, you got me thinking again about using graphic powder (not flakes). No need for graphene here. If I smear some graphite powder on both the power transistor and the heat sink, the cracks will mostly fill with the powder. I can then scrape off the excess and rub the power transistor onto the heatsink in a figure 8 (which someone suggested previously), which should remove any excess graphite (if the surfaces are reasonably flat). It will also make a huge mess, but that can be cleaned later.

On Tue, 22 Dec 2015 18:25:53 -0800, Jeff Liebermann Gave us:

Referencing that someone mentioned making a prophylactic from it was not an opinion.

Also., here is a top research facility talking about using it.

Cosmetic grade BN powder isn't as messy. And you can get it in different colors.

Hi Jeff, sorry my posts suffer from not enough information sometimes. I was using brass standoffs as "weak" thermal links from a ln2 bath to a sample stage. You can calculate what the thermal resistance should be. And what I found was that the measured thermal resistance was something like twice the calculated value to (maybe) four times bigger. Depending on how hard the screws were tightened. I assumed the extra resistance was the contact/ surface resistance... (the thing you are trying to reduce with the TIM) And since there were two surfaces I apportioned the extra resistance equally to each side... hence my 1/2 to twice value. I should have tried some goop at the interface and measured the change.. but I didn't... Maybe next time.

George H.

Finally remembered. The paste I was thinking of was some special Shin Etsu stuff, around 4.5-5 W/m/K.

Cheers

Phil Hobbs

Dow has some grease that's around 4.

They claim anti-pump-out too.

Thanks. Looks like they have taken it to market: Claims 4.5 W/m-K so it might be the same or similar stuff.

I don't have a clue what's in it, except that it looks, smells, and works much like Arctic Silver, which contains a little of everything (silver, boron, alumina, zinc oxide, etc). However, Artic Silver claims 8.7 W/m-K while NREL tests show only 0.89 W/m-K. Interesting reading and quite a list of different magic formulas. The winner was Dow Corning TC-5022 at 4.0 W/m-K. Second best was Shinetsu X-23-7762-S at 3.7 W/m-K (claimed 4.0).

None of the greases even come close to the dry flake, powder, and foil thermal conductivity numbers. Obviously, pastes are easier to handle, but with a potential 100:1 performance benefit, it would seem that the dry TIMs would be worth investigating.

Bah Humbug (T'is the season, so I can say that).

I didn't see that, but congratulations on having read what I posted. Unfortunately, there's no mention in your comments on how a graphene based TIM is going to solve the problem of filling cracks, and reducing gaps in my RF power device heat sinks. I can see some possibilities for moving heat from the transistor die, but not between the package and the aluminum heat sink. Keep trying. It's a topic of interest to me.

Ummm... I get a list of URL's. Any particular URL that I should be reading? If you read the article, did it give a number for thermal conductivity achievable in the lab?

Carbon nanotubes have long been known to have amazing thermal conductivity, but bonding them to thermal interfaces has been problematic. That's exactly the problem we're discussing. The carbon nanotube may be a great thermal conductor, but if it can't connect to the device, package, or heat sink, it's useless. Besides, if finding the material with the best thermal conductivity were the only parameter of importance, we would be using thermal paste made with diamond powder or maybe diamond films.

Thanks. I didn't know that there was a cosmetic flavor. The h-BN type forms flakes, which are the best for thermal conductivity. I'll try not to think about the minimum purchase quantity and price:

List of candidates and typical thermal conductivity: W/m*K Diamond 1000 c-BN 740 (Cubic Boron Nitride) h-BN 600 (Hexagonal Boron Nitride) Silver 406 Copper 385 Gold 314 Aluminum 205 Graphite 200 Carbon 150 SiC 120 Brass 109 Indium 86 Pastes 4.0 SilPad 2000 3.5 Dow Corning 340 0.54

On Wed, 23 Dec 2015 10:34:03 -0800, Jeff Liebermann Gave us:

The DARPA links. Remember what I wrote? "Top researchers". How many of those do you think there are?

This is likely the right stuff:

Then the real problem becomes: "What does one do with the other 127 oz?"

Yeah, it happens, but as you mention, the contact area problems are the same. By tightening the screw, you're increasing the area of the contact patch, eliminating any air gaps, and crushing any rounded granules into something flat. You could do better with a hammer or a hydraulic press.

LN2 would be 77K, which is above the temperature where alloys and impurities dominate the resistance of brass. Still, it might be worth checking to see if the resistance/temp is still linear at 77K.

I think conventional thermal goo would have made it worse. You didn't disclose the mechanics of your (microscope?) (aluminum?) sample stage. If it's been machined reasonably flat, and isn't coated, plated, or painted with something thermally disgusting, then simply machining and polishing the end of the brass standoff should have produced good results without super-tightening the screw. If my guess(tm) is correct, adding some dust to fill in the imperfections, from my list of TIM candidates, might improve things.

Disclaimer: My very limited cryogenics experience was in about 1980 making ice cream with liquid nitrogen, so please assume that I don't have a clue what I'm doing or am seriously out of date.

I guess I'm among the "top researchers". First, I search. Then, I do it again which is called re-search. I usually start at the top and work my way down. That makes me a "top re-searcher" (after the 2nd search). I don't know how many "top re-searchers" there might be. Probably everyone that refine their searches starting from the top.

However, you're right about one thing. I forgot what you wrote. If you want to prevent that from happening in the future, I suggest you write something worth remembering. Try to reverse our positions. If I wrote the above, would you even want to read it?

I'll look at the DARPA links after I recover from having three Christmas lunches today. Too much of a good thing and far too much chocolate. To recover, I'll be exercising more and fasting for at least a week or three.

Drivel: Santa (almost) breaks the net: I think I wrote that in about 1986, when the world population was still around 5 billion. The technobabble is dated, but still applicable. More:

Merry Christmas and Bah Humbug (T'is the season).

Nope. The real problem is that it's boron nitride based PAINT. Methinks it's fairly apparent that the superior thermal conductivity of any of the previously mentioned powders, it thoroughly ruined by the addition of some kind of fluid, paste, goo, grease, oil, slime, or whatever. No thanks.

I think it's just packaged wet to avoid dusting with a little bentonite clay pushed to alkaline with a little ammonia to form a colloidal suspension. After the water goes away it'd be pretty fluffy - no non-volatile organic binders, oil etc.

Nope. It seems rather odd when they can just bag the h-BN along with an inert gas and not need the water carrier. The MSDS call it a "paint": Water 7732-18-5 60 - 100 % Amine Coated Bentonite 68953-58-2 1 - 5 % Boron nitride 10043-11-5 10 - 30 % That's an awful lot of water for just a carrier. 100% water? How does that work when there's nothing left for the h-BN?

However, you're probably right about suspending the h-BM particles in solution since this suppose to be some kind of ceramic high temperature lubricant. No data sheet, no clue on the grain size, and no OEM part number.

Going to the source and digging through a web site obviously designed as a work of art instead of information: I find this under "paint": BN sprays feature high-purity, fine BN crystals suspended within a fast-drying, solvent-based carrier formulation. Available in 13-ounce aerosol cans and made available in 12-can carton shipments, each BN spray can is able to provide about 100 square feet of coverage and is designed to produce a smooth coating with very low binder content. Ok, it really is a paint, which the data sheets has morphed into a "coating": Well, the applications mix are rather interesting, but nothing involving high thermal conductivity. Also, no grain size, or how to get rid of the residual bentonite clay. No thanks.

I think it is a mold release.

Dan