I'm curious to know if you are doing a particle study. In a particle study, you can specify that gravity will have an effect on the particles within a fluid. I can only assume that a rotational frame of reference would have a similar effect to gravity and that the particles would separate. I think you would get more worthwhile results using this method.
Also, you are correct. Flow can not have two fluids in the same domain of disimilar phase (ie. water and air).
All the best,
I think that modeling this as an airflow problem with dust as a particle study is a better fit for your need.
I don't think the mix of air and dust acts like a continuum in the physical sense which is how it would be treated the way you defined the setup. The concentrations will change, but based on mixing rules.