[nmag-users] Equation setting for maximum damping and no precession

[Alessandro Schaer]

Dear all,

I was wondering what actually happens to the LLG when setting the damping coefficient to the maximum value of 1 ("llg_damping = SI(1)") and switching the precession off specifying "do_precession=False" but NOT setting the gyromagnetic coefficient gamma to 0 explicitly.

In other words, the material contains the following specifications

llg_damping = SI(1)

do_precession=False

but no llg_gamma_G is specified, i.e. the default value is "used". In addition, no exchange coupling, nor anisotropy are present: both are set to zero. I am simulating relatively "big" structures (largest characteristic length reaches up to 100um) and the idea is to see if there are differences in the simulation results of such structures when using Nmag compared to COMSOL.

The question is: does "do_precession=False" overrides the default value of "llg_gamma_G" and sets it to 0, or is the precession switched off in some other way? If "llg_gamma_G" is actually set to zero, is the LLG still used to solve the problem ore are other equations (Maxwell) to compute the effective field/the magnetization?

Or have misunderstood the precession "concept"?

Thank you in advance

Regards

Alessandro

[Fangohr H.]

Dear Alessandro,

On 3 Jun 2014, at 13:47, Alessandro Schaer <alessand...@gmail.com> wrote:

> Dear all,
>
> I was wondering what actually happens to the LLG when setting the damping coefficient to the maximum value of 1 ("llg_damping = SI(1)") and switching the precession off specifying "do_precession=False" but NOT setting the gyromagnetic coefficient gamma to 0 explicitly.
>
> In other words, the material contains the following specifications
>
> llg_damping = SI(1)
> do_precession=False
>
> but no llg_gamma_G is specified, i.e. the default value is "used”.

This is a typical case, as for example in this example: http://nmag.soton.ac.uk/nmag/current/manual/html/example2/doc.html#relaxing-the-system-faster (although llg_damping=0.5 in that example but you could set it to 1.0).

The nmag switch ‘do_precession’ does not fiddle with alpha or gamma, it just ignores the precession term in the LLG equation of motion.

> In addition, no exchange coupling, nor anisotropy are present: both are set to zero. I am simulating relatively "big" structures (largest characteristic length reaches up to 100um) and the idea is to see if there are differences in the simulation results of such structures when using Nmag compared to COMSOL.

That is outside the remit of micromagnetics and likely to give unphysical results: I expect the demagnetisation field will want to misalign the magnetisation (i.e. connecting north to south poles), but it is not clear what this means for a discretised version of a continuous magnetisation field. Once you switch the exchange interaction off, I am not sure how meaningful the results will be.

>
> The question is: does "do_precession=False" overrides the default value of "llg_gamma_G" and sets it to 0, or is the precession switched off in some other way?

In some other way.

> If "llg_gamma_G" is actually set to zero, is the LLG still used to solve the problem ore are other equations (Maxwell) to compute the effective field/the magnetisation?

> Or have misunderstood the precession "concept”?

Does that answer your query?

Best wishes,

Hans