Making Liquid Air

Most recent answer: 01/22/2014

Q:
My son is trying to liquify air for his science project. That would require a temperature of 77K. We would like to know if it is possible to achieve with a compression of 7 atmospheres because that's commonly available with air compressors. In fact, we bought on that goes up to 150psi, or 10atm. While I haven't taught him the calculations yet, I came up with: PV=nRT V2 = V1*(P1/P2)^(1/gamma) for adiabatic compression. This means that air compressed to 7atm reaches T2 = 476K. If we shed that heat and decompress, it theoretically reaches 156K. That neglects losses. Right now, we are trying as a first step compressing to 80psi, sending air through a copper coil in an ice bath (~3C) Then into a thermos where the air leaks out and decompresses. We are not even achieving temperatures colder than the ice bath. We are about to try a bigger leak on the theory that a slow leak isn't doing enough work. We also have a Peltier cooler that seems capable of 30 degree F difference. If we put the warm end in an ice bath and reached -20C, that could be a second stage before the expansion in the thermos. The Linde process carries the gas in the thermos back along the incoming line, cooling it down. Is there anything better we could do? Are we operating with way too little pressure?
- Dov Kruger (age 47)
Teaneck, NJ, USA
A:

This is a really interesting project, and it sounds like you guys have put some real thought into it. Unfortunately, I don't think that a single-stage compressor cooler of the type you're using will be able to get cold enough. 

Let's follow through on that adiabatic calculation you were doing. It has a lot of approximations and idealizations, but it's still a good starting point. You want to reduce the absolute temperature about a factor of 4 to get liquification at atmospheric pressure. If we treat the gas as ideal with a fixed gamma of γ=7/5, we end up with p proportional to T(γ/(1-γ)). That would require p of 128 atmospheres, way beyond what your compressor can do. 

In principle you could do this with multiple stages of compression, with each stage taking a portion of the material from the previous stage and cooling it further. I suspect that's much too elaborate for a home project. 

Still, the project sounds worth pursuing to see how cold you can get it. I'm not sure why your current set-up isn't getting the air colder than the ice bath. Your idea of playing around with different types of leaks sounds good.

Mike W.


(published on 01/22/2014)

Follow-Up #1: limits on cooling air

Q:
We have one compressor up to 150psi, 7 Peltier coolers including 1 with heat sinks that we have tested and an air conditioner. The compressor is hooked to a 20 foot long copper pipe through an ice bath The concept was to cool the high pressure gas, and then release it in the thermos. That doesn't seem to work at all. The best temperature we have achieved is just under 0C, so the decompression barely has any effect. We're still trying to figure out why. It's possible we are releasing too much gas. We have seen some flecks of ice coming out so humidity is being frozen out. We also have a small 12V compressor that doesn't have much flow but can reach 250psi. With a multi-stage peltier cooler, we should be able to reach 100C cooler than the high temperature. The air conditioner should give us approximately 0C if we contain the cold air. Any guidance you can give as to the multiple stages, or why we're not getting ANY cooling as the gas expands, would be greatly appreciated. From videos I have seen, it appears that multiple stages aren't necessary, but the videos are frustratingly vague on the specifics.
- Dov
Teaneck, NJ, USA
A:

Here's a first thought. You mentioned flecks of ice. The latent heat for converting water vapor to ice is very high. You could be removing heat from the gas but instead of lowering the temperature much it's just causing bits of ice to freeze out.  If it turns out that's what's happening, it's a pretty interesting result even if it's not what you were aiming for. Maybe this is a long shot, but if you could start with dehumidified air (from the airconditioner?)  maybe you could at least see more noticeable cooling.

Mike W.


(published on 02/08/2014)