About the available experimental magnetic moment database suitable for vasp's MAGMOM tag.

[hszhao.cn@gmail.com]

Hi there,

Is there any available experimental magnetic moment database suitable for vasp's MAGMOM tag?

Regards,
Zhao

[christopher_sheldon1]

Dear Zhao,

We suggest taking a look at the Bilbao crystallographic server, where there is a growing collection of experimental magnetic moments available. These should be multiplied by a factor of 1.2-1.5 larger than experiment. We've added this to the MAGMOM wiki page.

Best wishes,

Chris

[hszhao.cn@gmail.com]

Dear Chris,

You said:

These should be multiplied by a factor of 1.2-1.5 larger than experiment.

But the the MAGMOM wiki page gives the following tip:

Tip: To converge to the magnetic ground state, we recommend setting the magnetic moments slightly larger than the expected values, e.g., using the experimental magnetic moment multiplied by 1.2 or 1.5. A growing collection of experimental data is available at the Bilbao crystallographic server.

I still have the following questions:

1. 1.2-1.5 is not equivalent to 1.2 or 1.5.
2. For which elements should we use 1.2, 1.5, or a value between them as the multiplication factor, respectively?
3. Currently, there are only 2273 entries in MAGNDATA. So, for which magnetic materials do not exist in MAGNDATA database, how should we determine the corresponding MAGMOM?
4. What's the corresponding relationship between the magnetic moments given on MAGNDATA and the SAXIS tag of vasp?
5. On the other hand, the Materials Project API can also be used to get the magnetic moment of each atom, as shown below:

Code: Select all

In [6]: from mp_api.client import MPRester as mpr
   ...: structure = mpr().get_structure_by_material_id('mp-999516')
   ...: 
   ...: print(f"Structure: {structure.composition.reduced_formula}")
   ...: print("\nMagnetic Site Properties:")
   ...: print(f"{'#':3} {'Element':6} {'magmom':>8}")
   ...: print("-" * 20)
   ...: 
   ...: for i, site in enumerate(structure):
   ...:     magmom = structure.site_properties['magmom'][i]
   ...:     print(f"{i:3} {site.specie.symbol:6} {magmom:8.3f}")
   ...: 
   ...: print("\nMagnetic moments by element:")
   ...: elements = {}
   ...: for i, site in enumerate(structure):
   ...:     el = site.specie.symbol
   ...:     magmom = structure.site_properties['magmom'][i]
   ...:     if el not in elements:
   ...:         elements[el] = []
   ...:     elements[el].append(magmom)
   ...: 
   ...: for el, moments in elements.items():
   ...:     print(f"{el}: {moments}")
   ...: 
Retrieving MaterialsDoc documents: 100%|█████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 12087.33it/s]
Structure: MnP

Magnetic Site Properties:
#   Element   magmom
--------------------
  0 Mn        2.000
  1 Mn        2.000
  2 P        -0.076
  3 P        -0.076

Magnetic moments by element:
Mn: [2.0, 2.0]
P: [-0.076, -0.076]

But it doesn't give the mx, my, mz components as shown on the MAGMOM wiki page.

Regards,
Zhao

[hszhao.cn@gmail.com]

See here for the related discussion.

Regards,
Zhao

[christopher_sheldon1]

Dear Zhao,

1. 1.2-1.5 is not equivalent to 1.2 or 1.5.
It is a suggested range, I have corrected the wiki page to 1.2-1.5. The important thing is to multiply it by a factor.

2. For which elements should we use 1.2, 1.5, or a value between them as the multiplication factor, respectively?
The range is to give a good initial guess in order to obtain a reasonable magnetic structure. What is best depends on what magnetic structure you want to find for your individual system and you have to determine for yourself.

3. Currently, there are only 2273 entries in MAGNDATA. So, for which magnetic materials do not exist in MAGNDATA database, how should we determine the corresponding MAGMOM?
For this, you can follow the procedure in Huebsch, Nomoto, and Suzuki, and Arita, Phys. Rev. X 11, 011031 (2021). Obtaining magnetic moments is discussed for systems without experiment is discussed within this paper.

4. What's the corresponding relationship between the magnetic moments given on MAGNDATA and the SAXIS tag of vasp?
The magnetic moments in MAGNDATA can be taken from the mcif file and then multiplied by a factor of 1.2-1.5 to write as inputs for MAGMOM. MAGMOM defines the magnitude of the magnetic moments, while SAXIS defines its relative orientation to Cartesian coordinates (there is more detail in the SAXIS wiki page). The procedure in Fig. 2 of the Huebsch et al. 2021 shows how to convert between the experimental magnetic moments in MAGNDATA to MAGMOM.

5. On the other hand, the Materials Project API can also be used to get the magnetic moment of each atom. But it doesn't give the mx, my, mz components as shown on the MAGMOM wiki page.
For that, I do not know. Perhaps some other user has a better idea.

Best wishes,

Chris