Wrong structure returned by API for mp-5986


#1

This is supposed to be a tetragonal material, and the website confirms the relaxed structure has all the angles at 90 degrees, but API returns a lower-symmetry structure!

MPRester().get_structure_by_material_id(“mp-5986”,final=True).lattice
Lattice
abc : 4.004456562976344 4.004456562976344 4.2006364192253862
angles : 89.864508091374447 90.135491908625568 90.047460208463463
volume : 67.359629854788295
A : 4.0044532400000001 -0.00165554 -0.00488596
B : -0.00165554 4.0044532400000001 0.00488596
C : -0.0048062699999999996 0.0048062699999999996 4.20063092

Notably, using final=False does return a tetragonal structure (though still different from the one on the website).

Note also that the task labeled on the website in bold (presumably the preferred one?) as “GGA Structure Optimization” does not open:
https://materialsproject.org/tasks/mp-5986#mp-1070716
returns a page with only a header and empty content.
However, task mp-5986 and a few others that I checked (there are too many to check all!) all have a tetragonal structure.

(If this is related to the recent database upgrade, say something went wrong during the upgrade, is there maybe a way to use a previous version?)


Error querying for a specific material ID
Error querying for a specific material ID
#2

This entry has no DOS data.

>>> data = m.get_dos_by_material_id('mp-5986')
>>> print(data)
None

I guess the final structure might be made tetragonal when calculating DOS.


#3

I beg your pardon?

BaTiO3 is tetragonal at room temperature. This is the phase corresponding to mp-5986 as is evident already from the number of ICSD entries for this material. There is no need to make it tetragonal for DOS calculation: it already is, if you start from an ICSD positions and relax using VASP! VASP wouldn’t break the symmetry during relaxation - it can only relax to a higher-symmetry structure. Somebody had to either break the symmetry manually or to assign a completely different phase to “mp-5986” that is clearly associated with a tetragonal phase - judging by ICSD IDs, by the figure, by the labels, CIFs, everything.

(Looks like the web page started to also produce angles !=90 degrees now, for no good reason - CIFs all still show 90 though).

Tetragonal BTO is an extremely important phase, please fix this!!

Thanks a lot.


#4

Yes, this will definitely require further investigation. I’m not surprised a non-tetragonal (but almost tetragonal) BTO is being grouped together with the other BTO tasks. I am, however, surprised this is apparently lower in energy and thus being preferred for the “final structure” being reported on the website.

For the first point (why the structure mentioned above is being grouped with the tetragonal structures) I attached the following two screenshots which show our symmetry analysis.

With the website grouping symmetry tolerance:
36%20AM

With a stricter symmetry tolerance:
18%20AM