Dave,

The only thing that I can think of is that the boiling points of oxygen and argon are very close together.

The composition profile in the low-pressure oxygen column is interesting.  In this column, waste nitrogen comes off the top and your product oxygen comes off the bottom.  Argon tends to concentration about 2/3 of the way down the column.  This is commonly referred to as the "argon cloud" or "argon bubble".  At this point, the argon-rich gas is taken into the crude argon column to separate the argon from the oxygen.

I hope this helps.

NF3/CF4 also form an azeotrope or pseudo-azeotropic mixture at cryogenic temperatures, one that is esential impossible to split without the use of a co-solvent/stripping agent like HCl.

DrFluor,

I have two questions.  

One, what is NF3 used for?

Two, have you ever considered breaking the azeotrope with a membrane?  At first glance, this would seem to be a workable membrane separation.

Dave

1. NF3 was originally used by the US government as an energy intensive oxidizer for rocket fuels. I believe it then migrated to being used as an F laser source gas for "Star Wars" applications, where it is still used today. Currently the major application is as a plasma cleaning gas for cleaning the semi-conductor manufacturing tools and now the big TFT-LCD manufacturing tools. Think 100's of metric tons/year.
2. Is the membrane question in reference to Ar/O2 or NF3/CF4? In both cases I believe you face the same problem: The boiling points and molecular sizes are almost identical. In the case of NF3/CF4 they are -129/-128C and they differ by about 0.1 Angstrom is size. The NF3 has a dipole (0.235 debye), but is of little help in trying to remove CF4 from NF3, (data from Matheson 6th ed.). [The problem with electronics is everybody wants higher purity; 99.995+% purity or less than 50ppm total contaminants, and preferably less than 10 ppm CF4.]

Thanks for the info.

My interest in the separation is professional curiosity, I work in gas separation membranes (usually at 35-50° C)  but know little about cryogenics.

Gas separations are done by two mechanisms, solubility and diffusion.  Clearly a diffusion controlled membrane would not work well and that is where I do most of my work.  Solubility driven membranes are usually tracked with critical temperature.

I am also not sure how much separation is necessary, as the goal is not to do the separation but to replace a stripping column, but I looked up the critical temperatures, and they are pretty close (6 °C) so this is probably a dead horse.

Dave