Cause of the Soret Effect/Thermophoresis?

[Declan]

Hi everyone! I'm simulating a gas mixture of He and Ar, and I'm trying to get them to separate via Thermophoresis. For clarification, http://en.wikipedia.org/wiki/Thermophoresis

When I naively try it with what I have so far in my program, it exhibits Thermophoresis...but with the Ar going to the hot side and He going to the cold side! The wiki says it should generally be the opposite (lighter molecule, He, going to the hot side), and I was told by my professor to expect this as well.

Generally, when something doesn't work, I try to understand the physics behind what should be happening, and go from there. But I've gotta admit...I have no idea why this effect happens! I tried a bit of googlin', but came up with nothing. I can't really think of why the gas would separate.

Right now, I'm putting a 50/50 He/Ar mix between one plate at 500K and another plate at 5K. The distance between the plates is about a micron and a half. This does seem like a really high temperature gradient, so maybe that's messing something up. But then again, I don't really know why this works or why what I'm doing could be bad.

Does anyone know? Can anyone help me?

Thanks!

[Andy Resnick]

Interesting...

The Soret coefficient comes out of the Onsager reciprocal relations, and couples diffusion and thermal conduction. The inverse effect (concentration gradient leading to the temperature difference) is called the Dufour effect.

I don't have much material on it, but there's a section in DeGroot's "Thermodynamics of irreversible processes"

I also found this:

http://www.springerlink.com/content/r144275286w56k64/

and this (measurements at lower temps, but maybe applicable):

http://pof.aip.org/resource/1/pfldas/v4/i10/p1216_s1

[Declan]

Interesting...

The Soret coefficient comes out of the Onsager reciprocal relations, and couples diffusion and thermal conduction. The inverse effect (concentration gradient leading to the temperature difference) is called the Dufour effect.

I don't have much material on it, but there's a section in DeGroot's "Thermodynamics of irreversible processes"

I also found this:

http://www.springerlink.com/content/r144275286w56k64/

and this (measurements at lower temps, but maybe applicable):

http://pof.aip.org/resource/1/pfldas/v4/i10/p1216_s1

Cool, I'll take a look at those. Thanks!

Looking at the working examples my professor showed me, he the gas in between a plate at 300K and one at 2000K...with the plates being 1m apart. So, relatively, that's a waaay smaller temperature gradient. It's possible it breaks or something when you're doing such extreme conditions as mine.

Well, lemme go read those.

[Declan]

Hmmm. Read a bit, can't say I'm clearer on it.

A question about thermophoresis, if you know the answer: once the gas has separated, it should roughly stay in place, right? As in, there isn't much convection going on, right?

[Andy Resnick]

No clue, sorry....