B_eff vs H_eff, continuity condition in mumax3.10

[Mumax Lu]

Hi, 

I am using mumax3.10 and have a weird result for the effective field B_eff in my simulation. On the border faces between the magnetic regions and the non-magnetic regions of the simulation the continuity condition "scalarproduct(face normal vector, (B_eff_2 - B_eff_1))=0" is not fulfilled. There is a jump in the values of the B_eff component parallel to the face normal.

One idea is that the true B = mu_0 (H + M) is actually continous and the exported value "B_eff" is not the true B value, instead B_eff = mu_0 H_eff, without the magnetization. So, where M has a step, H also has a step to make B continous. Is this idea the solution? Or what else can be the reason for the simulation to break this continuity condition?

I hope you can help me with this.

[Mumax Lu]

I have to add that this jump in B_eff does not happen on all border faces, but on the left and right side of each cuboid (of which I have an array: long stripes next to each other. There must be stray fields between them). The affected faces are not the smallest and not the largest of the cuboid's outer faces. They are the medium sized ones. It is the side where the face normal is (anti-)parallel to B_ext.

[Josh Lauzier]

Hi,

The definition of B_eff in mumax3 does not include M, it is referring to the effective field term used in the LLG equation. You can find the full definition in the mumax paper "the design and verification of mumax3", as well as a breakdown of how each term is handled. (Other micromagnetic programs may vary, Oommf for instance uses H_eff instead). The difference is mainly in units (B_eff in Tesla, H_eff in A/m)

In regards to your other post, mumax allows you to view/save each individual contribution to B_eff seperately as B_anis, B_therm, B_ext, etc. For the magnetization, there is m (for the normalized magnetization) or m_full.

Best,
Josh L.