The elastic constants listed in the new web site for rutile TiO2 mp-2657: TiO2 (Tetragonal, P4_2/mnm, 136) show c11 != c22 == c33. This is weird, because normally I’d expect the zz components of a tetragonal structure to be the different one, and indeed the lattice constants have a == b != c.
Am I correct that this is weird, perhaps wrong? I wasn’t sure where to open a PR, since I didn’t see a repo associated with the actual web site data.
Yeah - I think this is wrong, my guess is there’s a bug in the IEEE conversion routine. Here’s a MWE, I think it should be flagged on pymatgen.
from pymatgen.ext.matproj import MPRester
from pymatgen.symmetry.analyzer import SpacegroupAnalyzer
from pymatgen.analysis.elasticity import ElasticTensor
import numpy as np
with MPRester() as mpr:
data = mpr.query({"material_id": "mp-2657"}, ['elasticity', 'structure'])
et_orig = ElasticTensor.from_voigt(data[0]['elasticity']['elastic_tensor_original'])
conv = SpacegroupAnalyzer(data[0]['structure']).get_conventional_standard_structure()
print(et_orig.convert_to_ieee(conv).voigt.round(2))
Another reason might be: the elastic tensor reported on the website is transformed according to rotations defined by the IEEE standard. However, the structure and the lattice parameters shown on the website may not correspond to the IEEE standard. We are looking for better ways to show tensorial properties (elastic, dielectric, …).
But this may not be the problem in this case. We’ll look into it.
However, the structure and the lattice parameters shown on the website may not correspond to the IEEE standard.
On this point, we could certainly report the Structure in the IEEE standard in the Elasticity section on the website, and perhaps this could remove one possible avenue of confusion.