Dear Javier,

The optimization depends on what you want to optimize. Lets assume that you want to optimize the extracted current ( or plasma density)  from the plasma device, then a parabolic profile (rising in the middle of the chamber) is the optimum. Whereas if you want optimize the confinement or looking for high charge states from the plasma device then a MIn-B configration is the best . But to acheive u need both radial and axial magnetic fields.

That is a good idea to use radial magnetic fields at the corners with cyclotron radius less than the radius of the plasma

OK, I will read the paper.

I recommend you using RF power mosfets like that in attached link instead of magnetrons that emit power in a lot of bands

http://www.digikey.com/product-search/en?FV=fff40015%2Cfff802eb&mnonly=0&newproducts=0&ColumnSort=-139&page=1&stock=0&pbfree=0&rohs=0&quantity=&ptm=0&fid=0&pageSize=25

Dear Javier,

Yes may be RF power MOSFETs are a good alternative to magnetron, but if you compare the economy 1.2 KW magnetron is far cheaper than same rating FET ( which may not be available). For pulsed microwave driven sources FETs can be used, but for CW sources, i think its still a challenge. The magnetron frequency jitter does not effect the performance of microwave driven sources because the modes are well seperated in frequency. Hence no question of power going into undesired modes. We also used cober muegge magnetron in our microwave driven ion source

I think the easy way to look at the problem is the vector potential form of inductively coupled plasma (ICP).

I see.

I found that the reflectance of the magnetic mirros R=Bmax/Bmin, so confined particles must fulfill Vt/Vp>1/sqrt(R) accordingly magnetic mirror theory, Vt and Vp are speed with respect magnetic field, so very few particles will be confined if Bmax/Bmin is low. If Bmax=29 and Bmin=17, it would be confined particles with Vt/Vp>0.765 that gives about 40'% of the particles.To maintain 70% of the particles it would be needed at least 10 times less pressure in the central region.

Magnetic bottles seems not to be a good confinement method for fusion due low compression field is allowed in the central part.

Toroids must be better, but it is needed more power that would waste a lot energy for a given magnetic field.

I just added a Technical article with calculus of the confinement time of the two parallel coils.

Article Technical article: Magnetic mirrors confinement time