What about CO2?

[HAL-9000]

Hey all, new poster here. I really like Pro-Cooling.com, folks around here seem to know the technical side of our hobby along with the enthusiasm, which is rare.

Anyways, I was wondering....what about using carbon dioxide as a phase change coolant? Its chemically inert, non-flammable, a crappy electrical conductor, and its boiling temp is -87C; that's chilly in computer-chip cooling world, but about a hundred degrees C "warmer" than LN2. Anyone ever try it?

From what I understand it will solidify into "snowflakes" after the compression-heat exchanger cycle whilst in the capilliary stage, but if the system could be designed in such a way that high velocity ~-80C "snow" was raining down on the CPU...that would be pretty frickin awesome. It would be almost three times lower in temp than a Vapochill! Just an idea...

[bigben2k]

I haven't heard of anyone trying it. I would think that the temp would be an issue: it's awfully cold. Also, what are the pressure characteristics of CO2?

[HAL-9000]

Looks like you need about 870psi to compress it to a liquid, pretty intense. Here's a pump for that purposer: http://co2.com/html/liquid_pumps.html

However, if you look at the pump's specs it can pump ~35lbs a minute, which I think is considerable overkill for a chip!

[Sandman001]

Originally Posted by bigben2k

I haven't heard of anyone trying it. I would think that the temp would be an issue: it's awfully cold. Also, what are the pressure characteristics of CO2?

temperature is fine, people are running cascade coolers colder than that, and it has real adverse effect on the chip. Macci and Oppainter do it for all thier insane 3dmark runs, Macci has one at like -112C I think.

He took his 2.4A precott to 4.8ghz.

[HAL-9000]

Have been reading up on CO2, and if such a coolant were to be used in a 'puter cooler, some issues would have to be addressed.

For one, the system can never reach STP unless the CO2 is in a gaseous state. At STP the compound directly sublimes from a gas to a solid, never develops a liquid phase. So the system would always have to be around half a megapascal of pressure whenever the engineered spec called for the CO2 to be a liquid in the system.

The other issue is thermal shock. Unlike say, a Vapochill, which ramps around -15C and cools down as it runs, is the coolant that comes out when the thing starts up in a CO2 rig will be incredibly cold to begin with. (Think CO2 BB gun cartridge leaking out). If the chip is really ramped, 200 watts of heat could meet 200 watts of cold basically, and crack the die. But if the chip is started to run, and then the coolant loop starts, it probably won't look like a chip anymore by the time the coolant gets there!

I am intrigued by the cascade coolers I see out there, but the reason CO2 interests me is it seems to offer the heat capacity to actually integrate a supercold computer into a reasonable volume, and not fill up a garage and burn multi-KW/hr's out of the wall.

[Brians256]

edit: looks like you guys have found out most of what I found out already.

I just asked my father about CO2 phase change. He says it is indeed used, but it is very expensive. Take a look at R744 systems on the web.

Operating pressures of 1500psi to 2000psi are common, which means that tubing must be stronger, joints must be stronger/tighter, etc... I believe the pressure differential is also higher, but I'm not sure.

On a side note, CO2 has great potential to replace R-134a in cars, due to the greenhouse effects of R-134a (even if R-134a doesn't damage ozone, it is a great greenhouse gas).

[Meethoss]

Um...I saw a post by a guy on here (I think - could have been Extreme Overclocking) where he built a chamber in which he put his components and CO2 (I think). The CO2 ran below 0 but there was no vapor in the system and the gas was cooled using phase change. Hope that's some help.