Need a method for describing a given neutral species with different equations in different parts of the computational domain.

[One (rejected) way of doing this is to partition the domain geometrically and solve the neutrals by the equation appropriate to each sub-domain, with boundary conditions linking the sub-domains. This is too complicated.]

Partition the neutrals into two co-existing populations:

- kinetic neutrals, to be solved by the Monte Carlo method, described by ,
- fluid neutrals, to be solved by fluid equations, described by , , and .

The required accuracy is achieved by arranging that the fluid density is small where a kinetic treatment is required and vice versa.

The two populations interact via traditional collision terms:

- Self collisions. In the case of fluid neutrals this leads to the usual transport coefficients and for kinetic neutrals it gives a nonlinear collision term.
- Cross collisions, where kinetic neutrals collide with fluid neutrals. This may be handled in exactly the same way as collisions between kinetic neutrals and fluid plasma.

Transfers between the two populations of neutrals are accomplished using the ``reactions'':

- Condensation

which proceeds at which is high for . This is analogous to recombination. - Evaporation

which proceeds at which is high for . This is analogous to ionization.

The reactions between the neutral populations and other species require no special treatment.

Wed Nov 26 10:29:07 EST 1997