Question on simulating chiral bubble

[A2025]

Hello everyone,

I am trying to perform micromagnetic simulations of chiral bubble domains / skyrmions, and I am encountering some difficulties in reproducing the experimental observations.

In my experiment, we observe a sign reversal of the THE occurring around the coercive field Hc​. In the same field range, relatively large magnetic bubbles are clearly detected.

However, in my field-sweep micromagnetic simulations, I am unable to reproduce this behavior consistently:

When I tune the parameters to obtain skyrmions/bubbles with a size comparable to the experiment, they only appear at very small magnetic fields, far below the experimentally relevant field range.

If I adjust the parameters so that skyrmions appear around the correct magnetic field, their size becomes extremely small (only a few nanometers), which is clearly inconsistent with the experimental observations.

Therefore, I would like to ask:

Are there recommended strategies or parameter-scaling approaches to simultaneously obtain
(i) skyrmions/chiral bubbles with realistic sizesa nd
(ii) stability in a realistic magnetic field window?

In particular, how should one balance Ku1 and Dind in this situation? Are there common pitfalls that could lead to artificially small skyrmions in simulations?

I also have a question regarding the field-sweep protocol for identifying skyrmion states:

For simulations aiming to mimic experimental field sweeps, is it generally better to first fully saturate the system at a large positive field and then sweep toward negative fields to search for skyrmions,or first prepare a stable skyrmion (or bubble) state at a certain field, and then sweep the field up and down from that point to examine its stability?

Thank you very much for your time and help.

Best regards,
A2025

[Vlad Kuchkin]

Dear A2025,

Could you be more precise about what system parameters you have in the Mumax script and what system size was in the experiment?

From my experience, in experiments, bubble size is determined by the demagnetizing fields, and to simulate bubbles in mumax, one has to be close to the realistic system parameters.

You can have a look at my recent paper 10.1002/adma.202501250

Cheers,

Vlad

 

сб, 13 дек. 2025 г. в 15:07, A2025 <xiaoj...@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "mumax2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mumax2+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mumax2/da97436d-b0f3-4c7c-86fe-a17e7c23d9ccn%40googlegroups.com.

[A2025]

Dear   Vlad,

Thank you very much for your reply.

In my MuMax3 simulations, I use Msat = 4e5 and Aex =5e-12 as constant. I keep these two parameters fixed and vary Ku1 from 5e4 to 2e5 and Dind1 from 0.1e-3 to 2e-3. The simulation geometry is defined as  SetGridsize(256, 256, 4)、SetCellsize(4.000e-09, 4.000e-09, 8.200e-10)  and  periodic boundary conditions are applied in the film plane . 

 My goal is to obtain a chiral bubble or skyrmion with a characteristic diameter of ~100 nm that is stable around an external field of approximately 0.2 T. However, in all cases I have tested, the stabilized objects at 0.1 T remain smaller than ~10 nm, while larger bubbles only appear at much lower magnetic fields.  

Cheers,

A2025

[Anne Bernand-Mantel]

Dear A2025,

1. your exchange length is 7 nm so your cell size of 4 nm is most likely too big to obtain an accurate estimation of the skyrmion energy, you should take 1 nm and in any case you need to vary this parameter and check there are no change in the energy of a domain wall for example to start with

2. by varying the parameters as much as  "Ku1 from 5e4 to 2e5 and Dind1 from 0.1e-3 to 2e-3" you should be able to get pretty much any size you want except you'll never get it at random as you experienced

3. In principle its useful to have a measure of the effective anisotropy of your sample as well and to keep in mind that Keff=Ku1 -1/2µ0Ms²

3. I agree with Vlad, your spin structure is probably a skyrmionic bubble stabilized by demagnetizing field (in addition to DMI) in this case the starting point would be to know the width of the magnetic stripe observed experiementally at zero applied field. You can then fit your missing experimental parameters to reproduce this width using numerical simulations of stripes or /and a analytical model (cf Lemesh, PRB 95, 174423 (2017) or

Bernand-Mantel PRB 111 184423 (2025). If your Ku1 is not too far off, you should get the desired skyrmion size with the right field range at this point. 

Best,

Anne

[A2025]

Dear All,

Thank you for all the helpful replies.

Experimentally, we observe topological Hall signals with opposite signs appearing just before and after the coercive field (H_c). I am trying to reproduce the corresponding domain evolution using MuMax3.

However, for this hard magnetic system, the reversal during field sweeping is either too abrupt (domains switch almost instantaneously) or the domain size becomes much smaller than in FC/ZFC simulations, while experimentally the domain evolution occurs over a wide field range. Any suggestions on how to improve this?

In addition, I would like to introduce some disorder to make the simulations closer to experiment. Are there recommended ways in MuMax3 to implement disorder?

Best regards,
A2025