Any Chemists out there??

[AMDGuy]

Anyone know of a site that lists the heat transfer properties of various liquids? I've been thinking about it more and more as I near the end of my water cooling project and I'm thinking of other substances besides water. Is there anything that transfers heat faster? I've mentioned amonia before and my friend swears it would be better. I'd like to look at a chart showing the differences in substances to make the best determination.

Thanks in advance.

 

[Clemson Physics Geek]

Go to your local library and see if you can put your hands on a CRC

 

[AMDGuy]

Clemson Physics Geek (Jun 01, 2001 01:39 p.m.):
Go to your local library and see if you can put your hands on a CRC

This may be a dumb question, but what's a CRC??]

 

[Mictlan]

You should get the Perry's Chemical Engineering Handbook. It has a lot of resourdes about Heat Transfer Coefficients. I know it as I use it for work.

 

[Szech]

Is this what you mean?

I think mercury would be the best "liquid" to use, but it would be hard to get that much, and you would need a powerful pump to move something that dense.

 

[Slain]

This is probably the best site on the net. If it ain't here, it don't exist.

http://www.matweb.com/

 

[Fink]

Perry's is probaly the best place; the CRC (Chemical Rubber Company) will also list heat transfer data, but even the new versions are 20 years out of date. Easiest to find this data at you local university library, a regular library won't have the $$ for these reference books...

You are really looking for a chemical engineer - they are much more involved with these processes. You might want to take a trip down or email a professor in Chem. Eng at a local college, you will find most people to be very helpful.

 

[William]

my highschool chem lab, we have the Merck and three CRCs!!!!!

 

[William]

Fink (Jun 01, 2001 06:51 p.m.):
Perry's is probaly the best place; the CRC (Chemical Rubber Company) will also list heat transfer data, but even the new versions are 20 years out of date. Easiest to find this data at you local university library, a regular library won't have the $$ for these reference books...

You are really looking for a chemical engineer - they are much more involved with these processes. You might want to take a trip down or email a professor in Chem. Eng at a local college, you will find most people to be very helpful.

well its not like the heat transfer properties of liquids has changed in the past 20 years......

 

[Fink]

Yea, good point

I was implying that many of the newer compounds will not be in the CRC. And for the most part much of the CRC data has been replaced with more accurate values in the literature.

 

[doublec16]

I'm an astrophysicist, not a chemist, so I may not be an authority on this, but I think the rate at which you can move heat away from your CPU depends on a few factors, such as thermal conductivity (how fast the fluid will absorb the heat), heat capacity (how much heat the fluid will absorb), and density (how fast you can pump it). For instance, air has crappy thermal conductivity and heat capacity for the purposes, but a whole lot of it can be blown past the heatsink, whereas water is more difficult to move and requires a stronger pump, but its thermal conductivity and heat capacity make up for that. Thermal conductivity and heat capacity are roughly increasing functions of density, because the closer together the atoms are, the more easily the electrons of those atoms can transfer momentum (heat) to each other, and also the more electrons there are to carry the heat. I guess the ideal overclocking fluid would be one with lots of heat carrying electrons per atom, and is a liquid or gas at room temperature. If you were to use mercury, you'd only have to pump it at a rate 1/15 times how fast you pump water to get the same cooling, but mercury is extremely toxic so perhaps that is not a good idea. Instead of plastic tubing, you should use a metal such as copper so that it not only transfers heat to the cooling fluid more readily, but the tubing itself conducts heat as well. Unfortunately it also conducts electricity quite well... The obvious choice would be to use antifreeze or some such fluid, because of its high heat capacity. I have heard of people using motor oil.

I myself prefer nice, abundant, nontoxic, safe, non electricity conducting air (and a whole bunch of fans) for my cooling.

 

[WyrmMaster]

Maybe the answer is an extremely dense gas, you would have to rig up a way to exchange your pump for fans to move it, but it would be easy to move, and hold alot of heat. Im not sure what gas you would use though, is argon dense, or am i thinking of something else.

 

[William]

WyrmMaster (Jun 02, 2001 05:32 p.m.):
Maybe the answer is an extremely dense gas, you would have to rig up a way to exchange your pump for fans to move it, but it would be easy to move, and hold alot of heat. Im not sure what gas you would use though, is argon dense, or am i thinking of something else.

you could just pressure the system and use ordinary air.

 

[doublec16]

Yeah you could make a refrigeration system, using a pressure pump and a substance with a high latent heat of vaporization. First you pressurize it until it's a liquid, then as it goes past the cpu and heats up it will vaporize, eating up a whole lot of heat as it does. Then the cycle begins again, when you pressurize the gas until it becomes a liquid. I think noble gases are good for that. I once worked on an experiment which used liquid He3 for cooling. It got the detector down to 0.001K, if I am remembering it right (.001 degrees above absolute zero). You can't get much cooler than that.

 

[Spewn]

You can be pretty much guaranteed that as a straight coolant, water is THE best substance out there. Why? Every industry in existance uses it for cooling. Nuclear power plants are cooled with water, in-fact, and you can imagine how much heat nuclear power plants generate. The reason why water is such a good coolant is as follows: Water can basically absorb the most amount of heat energy of ANY substance before raising in temperature. This is important, because as the temperature difference between the coolant, and the thing being cooled decreases, the thermal efficiency decreases. So, even if say ammonia could absorb the heat energy faster than water to begin with, it's temperature would rise MUCH faster, and your cpu would run hotter. In a watercooling setup, you'd have to have one HOT cpu to raise your water more than maybe 5'C-10'C as it passes through the waterblock. They run supercooled ammonia through pipes underneath skating rinks to cool them basically for the reason that you can supercool ammonia, but you can't supercool water(well, not if you want it to remain a fluid).

 

[MsNath]

Spewn (Jun 03, 2001 03:25 p.m.):
You can be pretty much guaranteed that as a straight coolant, water is THE best substance out there. .

I agree totally. As far as liquids are concerned you are not going to better than water. Plus you are more limited by the temp of the liquid than the liquid itself. IE a 50/50 mix of water and antifreeze gives you a liquid that does not conduct heat as well as pure water. But it does allow you to drop that water down to -50c. At -50c the 3 or 4 degrees difference between heat transfer rates becomes irrelevant.

So if your getting bored of straight water. Try adding antifreeze and dropping your water temps below zero. I am currently using a chest freezer to do this and it works great!

OR as others mentioned buy or try and make your own Vapochill. This system works on the same process as your freezer/fridge. There is a small compressor mounted in your case with some copper tubing that runs to your CPU. Freeon<sp?> runs through the tubing just like in your freezer. When the it hits the point where the copper contacts the cpu it cools the cpu well below zero.

To buy the actual Vapochil case its about $750-$850 US dollars. However, I think that it wouldn’t be that hard to build one using parts from an old freezer. You would have to know a bit about how they work, how to solder copper and be able to find someone willing to charge the system for you. We are a resourceful bunch here. I think between the lot of us we could come up with something.

 

[BboySkid]

doublec16 (Jun 03, 2001 02:03 p.m.):

It got the detector down to 0.001K, if I am remembering it right (.001 degrees above absolute zero). You can't get much cooler than that.

Correct me if I'm wrong, but i recall hearing somewhere that absolute zero has not been reached yet. So for someone doing a routine experiment in university or what not to get that close is pretty impressive. Are you sure you are not mistaken, I'm just having a hard time believing that you got that close to absolute zero, based on my knowledge of whatever we are talking about.

 

[doublec16]

It wasn't really a routine university experiment. The detector was part of the payload of a balloon that was launched into the stratosphere and was used to measure the cosmic microwave background radiation. Since the temperature fluctuations we were measuring are on the order of a millionth of a degree, very low temperatures are necessary for peak detection efficiency. There were actually 2 refrigerators: one containing regular helium 4 that lowered the temperature to 1K or so, which surrounded the helium 3 fridge.

Of course none of that is really necessary to cool a CPU, but the concept is the same if you wanted to refrigerate it, making use of the fact that much heat is lost when a liquid vaporizes, instead of just carrying heat away relying on a large heat capacity. I have heard of people using a refrigeration system in their computer, but I imagine it's quite expensive to assemble.

BTW: absolute zero has not yet been reached nor will it ever be reached. It is a quantum mechanical concept, not a true temperature. All we can do is get really close and try to get closer.

 

[Spewn]

helium 3 and helium 4? I think someone's BS'ing...anyway, recompressing a gas into liquid state isn't pracitcal for a computer. It would require a compressor far more expensive than buying a few computers, and be far louder too

 

[David]

Spewn (Jun 04, 2001 07:20 a.m.):
helium 3 and helium 4? I think someone's BS'ing...anyway, recompressing a gas into liquid state isn't pracitcal for a computer. It would require a compressor far more expensive than buying a few computers, and be far louder too

Helium 3 and Helium 4 relate to the size of the atoms. Helium has an atomic number of 2, which means all Helium atoms have 2 protons in the nucleus. They all have 2 circling electrons.

The difference comes with the number of neutrons. Helium 4 has 2 protons + 2 neutrons so has an atomic mass of 4. helium 3 has 2 protons but only 1 neutron so has an atomic mass of 3. sorry the pic wont attcah

proton - a positively charged particle in the nucleus
neutron - a neutral particle in the nucleus
electron - a negative particle circling the nucleus