View Full Version : Gold Leaf for Cooling?
Bad Maniac
06-20-01, 06:28 AM
What do you guys think?
Gold has a VERY high heat transfer coefficent, has it not?
What about using common gold leaf/gold foil from an art supplier instead of Arctic silver or anything similar. Gold foil is VERY thin, must hav better heat transfer rate, and since it is very thin and gold is already a very soft metal, putting a gold leaf betwwen the cpu and the heatsink would squeeze the gold leaf to very exact fit against the two surfaces, giving a very good contact, and a very good transfer.
This is just an idea I have been playing with, and I dont have equippment to test it at the moment. Please give me input on this topic.
Actually gold has a lower heat transfer rate than silver AND copper. Interestingly enough it also has a lower electrical conductivity than both silver and copper. However, I believe it is used in many electircal components due to it's low tendency to corrode, and it's malleability(it's very smushy, so they can make pins out of it/mold it onto pins very easily). Find a periodic table that lists heat transfer properties, you can find out information like that there. Also, the "gold leaf" that you buy from an arts supply store probably isn't actually gold, or at least nothing like pure gold. About an ounce of gold can be flattened into a sheet the size of a tennis court, THAT is gold leaf, and costs you both your testicles, and one of your ass cheeks to buy, simply because it's incredibly difficult and time consuming to make gold leaf(if you tear it even a tiny bit, unless you're at the very end, you have to start allll over again).
Are you sure, Spewn? The properties table in my heat transfer textbook lists pure gold as having a heat transfer COEFFICIENT of 317 W/(K.m). Pure copper's is at 401 W/(K.m). and pure aluminium's at 237 W/(K.m). Stainless steel is around 13 to 15 W/(K.m) -and the surface of a boiling kettle can still scald you. Heat transfer RATE depends on geometry, area of contact, and temperature difference, as well as the heat transfer coefficient of a metal. Any how, gold is expensive, and yes, it's very chemically inert (which includes oxidation/corrosion).
a thermal properties list....its quite extensive...
http://www.ai.mit.edu/people/tk/tks/tcon.html
dew.
Yeah he's right
It goes a little something like this...
cvd diamond
Silver
Copper
Gold
Aluminium
Wood :D
cvd diamond is far too costly though, silver is the best practical solution but still costly, copper is cheap and good but is hard to work with and heavy and aluminium is great at everything but heat transfer which it's just ok at.
I've read about someone who tryed gold leaf and though silver is a better condicter than gold, Artic silver isn't so gold leaf may be quite good as long as air doesn't get trapped during instalation
I have gold leaf and silver foil if anyone wants to send me an ass check I will mail some to you!!! LOL
seriously if you want to trade some 80mm fans I will send some to you to test out.
I think there is some fancy restaurant somewhere which serves dishes with thin, gold foil, and gold dust. You supposed to EAT it along with the other stuff. Can't remember where tho, saw it a long time ago on TV.
I recomend leaving it with some pressure (Either from the heatsink clip or using a clamp) for a while to crush the gold leaf into place as well.
Bad Maniac
06-20-01, 11:24 AM
How close to Pure silver is the arctic silver paste then? If you could get (and leaf gold IS rather pure) pure gold then it might have good properties. Anyone know how good the silver paste is compared to the real thing?
it's about 80% silver but it's not the content that does it. It's it being micronised to form a paste and to help against electrical conduction.
I think Arctic Silver 2 is rated at 8 W/(K.m).
And I think silver is rated at about 60
Magistrate
06-20-01, 02:12 PM
What about Silver Leaf? Does such a thing exist?
cjtune (Jun 20, 2001 11:01 a.m.):
I think there is some fancy restaurant somewhere which serves dishes with thin, gold foil, and gold dust. You supposed to EAT it along with the other stuff. Can't remember where tho, saw it a long time ago on TV.
I had some pie at a Russian restaurant with gold dust on it. The gold dust didn't have a whole lot of taste, but it was pretty neat! I like pie.
cjtune (Jun 20, 2001 09:40 a.m.):
Are you sure, Spewn? The properties table in my heat transfer textbook lists pure gold as having a heat transfer COEFFICIENT of 317 W/(K.m). Pure copper's is at 401 W/(K.m). and pure aluminium's at 237 W/(K.m). Stainless steel is around 13 to 15 W/(K.m) -and the surface of a boiling kettle can still scald you. Heat transfer RATE depends on geometry, area of contact, and temperature difference, as well as the heat transfer coefficient of a metal. Any how, gold is expensive, and yes, it's very chemically inert (which includes oxidation/corrosion).
what do you mean am I sure? If stainless steel is lower than copper, and aluminum is lower than copper, then obviously higher = better, right? Seeing as the area of contact and temperature difference won't change depending on whether you use gold or silver or copper, you basically just said what I said, that gold wouldn't be as good as silver OR copper at transferring heat :)
William
06-20-01, 08:00 PM
gold is an excellent conductor, but copper and especially silver are better. I am chemistry nut, so I am quite sure. Whoever posted about Gold being used because it doesn't corrode is quite right, thats golds weirdish trait is that it does not corrode at all. I can actually explain why(it transfers an electron from a different shell) but as far as pure conductivity goes, its lower, and would still be pricier. The best metal to use to keep from corrosion with a good thermal conductivity would have to Platinum. My chemistry teachers wedding ring is platinum becuase it is extremely unreactive. But of course, one can only really try it before we can say whether we are right. This very well may be better than Artic Silver.
CrystalMethod
06-20-01, 11:17 PM
Hell! Grab that puppy and beat it into some platinum leaf to share with all of us! I'm sure your teacher wouldn't mind. After all, it's all in the name of science, right?
...well sort of...
William
06-20-01, 11:37 PM
i don't think i could beat her up. She has a black belt and regularly does Karate. Plus a friend of hers trains the attack dogs for the miami police and he trained her dog. So in fear of my life, i don't think i willl do that.
CpuWaterCool.com
06-21-01, 08:46 AM
I read this a while back. Go out and search for a silver dollar or other pure but cheap coin (silver). It was $3 USD or so. Take it and lap it to death. Place VERY VERY thin layer of grease and sandwich between slug and HSF. If I remember correctly, that netted the reviewer about 3-5c.
Warning to Advertizers !
If you have an article or review that you feel would be of interest to our readers, contact Ed or Joe on our Front Page. If they agree, they will post a mention and a link.
We do not allow even our Seniors to advertise their sites. On occasion I will mention/recommend one or two of their sites at a time. This is in recognition of their contributions to this Forum and the writing they have done for our Front Page.
Only Seniors of this forum are permitted to use their own sites' URL in their signature.
We feel no obligation to provide space anywhere on our site for the free use of other sites.
We deal with encroachment by issuing a warning. If it goes unheeded, we ban the offender.
Kingslayer
06-21-01, 09:44 AM
This is actually very interesting. But wouldnt gold leaf totally defeat the purpose that paste was created for.
The whole idea with the paste, is to fill the nooks and crannies of both the heatsink and the core, improving the efficiency of the cooling process, by increasing surface area. Silver is just added to the paste to increase the coefficient of the paste. Yes, gold is very malleable, but, there isn't a heatsink that creates enough pressure to force it into the nooks and crannies of a heatsink. You will actually lose surface area by using the gold. Even with the Global-Win gorilla clips.
Just my $.02.
coldmop
06-21-01, 03:54 PM
the neat thing about gold leaf is when it is burnished onto say the bottom of a heatsink it sticks like crazy.
How come? It has to do with the sharing of valence electrons, I'm pretty sure about this. But been wrong before. Anyways the theory here would be that those two materials are then very close together. Now If the two mating surfaces (cpu, heatsink, ) were really true, and you could somehow gaurentee that the gold would be allowed to fill the gap, with sufficient pressure and extreme cleanliness I think the gold would be a better themal bridge than the paste.
William
06-21-01, 07:24 PM
cold mop (Jun 21, 2001 03:54 p.m.):
the neat thing about gold leaf is when it is burnished onto say the bottom of a heatsink it sticks like crazy.
How come? It has to do with the sharing of valence electrons, I'm pretty sure about this. But been wrong before. Anyways the theory here would be that those two materials are then very close together. Now If the two mating surfaces (cpu, heatsink, ) were really true, and you could somehow gaurentee that the gold would be allowed to fill the gap, with sufficient pressure and extreme cleanliness I think the gold would be a better themal bridge than the paste.
gold doesn't share valence electrons. it electron configuration is as follows
[Xe] 6s2 5d1 4f14 5d8
What Gold does it shift it so that its D orbital is full and its S is half full which is much more stable
[Xe] 6s1 5d1 4f14 5d9
Now i don't know quite why gold leaf would stick like crazy. electro static properties would do that as well as an attraction to the electron rich gold. But no sharing is trully going on(well maybe, we still aren't too sure how metallic bonds work). The interesting question here is If you use gold leaf instead of goop, will the increase in the thermal properties equal out or improve upon a lesser material which contacts more surface area. This would be an excellent experiment.
coldmop
06-21-01, 07:51 PM
thanks for the clarification, no sharing huh, may not be any good then. Could the gold adopt (see it as one of it's own) one of the electrons from the other material so the S would be full as well. I can't remember this stuff. "Of all the things I've lost..I miss my mind the most"
William
06-21-01, 08:50 PM
no it won't. All metals have very low electron affinities which is the desire of an atom to gain an electron. They all want to loose their electrons to achieve a stable orbital. What gold will do is loose that one electron in the outer S shell.
If my calcs are correct (which I doubt...), then wouldn't the gold foil only have to cover 2.5% the surface area as the Arctic Silver II to be just as effective? I'm not doing any of the Physics heat flow calculations (I forgot how too...) but just simple division. (8/317)*100. Would this not be a correct way to determine the percentage area it would have to cover to be just as effective?
If someone here remembers the physics (or is willing to help me out with them :) ) I would like a double check on my figure. Seems OK to me.. And at ony 2.5% the area, not much of the chip has to be directly touched... Of course I am only guessing at the math since I forgot the physics, so I'm not sure.
JigPu
Forgetter of physics...
William
06-22-01, 12:27 AM
probably something like that. This is a great idea, oh won't somebody try it out!
JigPu (Jun 21, 2001 10:26 p.m.):
If my calcs are correct (which I doubt...), then wouldn't the gold foil only have to cover 2.5% the surface area as the Arctic Silver II to be just as effective?
You HAVE TO COVER THE ENTIRE CPU SURFACE with somekind of interface material to the heatsink. Your CPU heats up all over and if you cover only 2.5% of it than you only have a small cold spot while the other parts get burnt on the core! I think the silicon and ceramic of the core is pretty poor at conducting heat so it'll be hard for the other parts of a CPU core to pass on their waste heat to the only area that has a 'direct link' to the heatsink -unless they are in direct contact with the HS themselves. Think of this as a bottleneck problem.
William
06-22-01, 03:04 AM
cjtune (Jun 22, 2001 02:42 a.m.):
JigPu (Jun 21, 2001 10:26 p.m.):
If my calcs are correct (which I doubt...), then wouldn't the gold foil only have to cover 2.5% the surface area as the Arctic Silver II to be just as effective?
You HAVE TO COVER THE ENTIRE CPU SURFACE with somekind of interface material to the heatsink. Your CPU heats up all over and if you cover only 2.5% of it than you only have a small cold spot while the other parts get burnt on the core! I think the silicon and ceramic of the core is pretty poor at conducting heat so it'll be hard for the other parts of a CPU core to pass on their waste heat to the only area that has a 'direct link' to the heatsink -unless they are in direct contact with the HS themselves. Think of this as a bottleneck problem.
I disagree. if you have just a tad touching that conducts much better, then you are ok. It would easily transfer to the gold foil. I would suspect this would touch a lot of it though.
Bad Maniac
06-22-01, 04:15 AM
The thing is, my co philosopher. Gold leaf IS very adhesive, and so thin, and gold in it self, as I said before a very soft metal. So I would think that you don't need that huge ammount of preassure to get it to stick to the different surfaces. You usually apply it with a paintbrush, and more or less paint it on.
If someone would (please) test my idea, you could pint the back of a lapped heatsink or waterblock with goldleaf, and another gold leaf onto the chip surface (lapp away the HORRIBLE printed text and stuff so the surface is clean) and then stick them toghether with a clamp or the original HSF mounting thingy. that might be enough, and still way thinner than any arctic silver would ever become, we are talking 1000ths of a millimeter here, and due to gold leafs very adhesive nature and softness, it would hopefully mould itself to fit the metal. and it has WAY higher conductivity than any grease ever. and we all know how much difference it makes to add arctic silver instead of some radioshack gunk. and that is just raising the transfer with say 3... gold leaf would raise it a few HUNDRED (given good contact would be accomplished)
Input please. this might be a VERY imortant discovery, made by me, the humble Swedish guy.
Blue Jester_2112
06-22-01, 10:14 AM
Spewn (Jun 20, 2001 07:20 a.m.):
and it's malleability(it's very smushy, so they can make pins out of it/mold it onto pins very easily).
Smushy! Great word!
cjtune (Jun 22, 2001 02:42 a.m.):
Your CPU heats up all over and if you cover only 2.5% of it than you only have a small cold spot while the other parts get burnt on the core!
Yes.. But what did people do before the paste came around? to cover 100% of it? I don't know what (I'm a newbie...). It seems though that at one point in time, the heatsink would be directly touching. They didn't burn up then! (Of course, now our chips run hotter...) Remember, I'm a newbie so I don't know...
JigPu
William
06-22-01, 04:58 PM
JigPu (Jun 22, 2001 04:48 p.m.):
cjtune (Jun 22, 2001 02:42 a.m.):
Your CPU heats up all over and if you cover only 2.5% of it than you only have a small cold spot while the other parts get burnt on the core!
Yes.. But what did people do before the paste came around? to cover 100% of it? I don't know what (I'm a newbie...). It seems though that at one point in time, the heatsink would be directly touching. They didn't burn up then! (Of course, now our chips run hotter...) Remember, I'm a newbie so I don't know...
JigPu
sure you know! You answered your own question, they produced so little heat that they did not need the paste.
I've been following this with great enjoyment, but something just occurred to me. Gold or Silver Leaf will fleck off and the fans will no doubt land one of those flecks on a couple of IC pins, Power Mosfets, Down in an empty PCI connector, etc. I don't mean to be a wet blanket, just thinking it through. The idea is tantalizing all the same.
Hoot
William
06-23-01, 01:10 AM
you could silicone the area around the cpu. Would only a problem if you where using a thermoengine. Would also want to cover those bridges. Good thought Hoot.
coldmop
06-23-01, 02:56 AM
Gold Leaf for Cooling? what about Gold Leaf for Modding? ...do the hole case that would be shweet and not very difficult, monitor too
I read about a guy on the oc.au forums who did this, now either it was a load of bs or true but from what he said he got very good results. Gold leaf is very, very thin and soft, he didn't just put one leaf on he put a bit on. Now silver is better than gold at conducting heat, hell even copper is, but Artic silver or copper pastes are not, so even if the gold was acting as a transition between the cpu and heatsink instead of just filling in the gaps like a paste does then it may still be better. It would be interesting to see a comparrison and if I can get my hands on some gold leaf (from my reliable sources, say no more :) ) then I'll let you know if it works.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.