Mismatch between e_above_hull and composition energies

There are some materials that are stable with e_above_hull = 0 but those calculated based on the composition energies obtained via MPRester are not 0.

For example, in the case of Cr-Cd-O, the following energy can be obtained.
However, when the energy of the competing phase of Cr2Cd2O8 (=CrCdO4) is calculated by CdO + 1 / 2Cr4Cd2O8 + 3 / 8O2,
-39.58368804 / 8 * 3-96.60562273 / 2-7.91327389 = -71.05996827
It is lower than -70.89144956 of Cr2Cd2O8.

Below are the values ​​taken by MPRester today.
{Comp: Cd1: Composition Energy (energy = -0.92288976, source = ‘mp-1096861’),
Comp: Cd1 O1: Composition Energy (energy = -7.91327389, source = ‘mp-1132’),
Comp: Cr2: Composition Energy (energy = -19.30609494, source = ‘mp-90’),
Comp: Cr4 Cd2 O8: Composition Energy (energy = -96.60562273, source = ‘mp-19262’),
Comp: Cr4 O6: Composition Energy (energy = -79.8478336, source = ‘mp-19399’),
Comp: Cr20 O48: Composition Energy (energy = -482.1972026, source = ‘mp-773920’),
Comp: Cr2 Cd2 O8: Composition Energy (energy = -70.89144956, source = ‘mp-18781’),
Comp: Cr2 O4: Composition Energy (energy = -44.63814413, source = ‘mp-19177’),
Comp: O8: Composition Energy (energy = -39.58368804, source = ‘mp-12957’)}

This might be related to the previous post.

So sorry if I have stupid mistakes.
Yu Kumagai

Hi @yktit , thanks for posting your question. I don’t know the exact answer but I would advise you to update pymatgen to the latest version if you have not already. We recently changed the energy corrections that are applied by pymatgen when you pull entries via the MPRester. If you are using an older pymatgen, it’s possible that your system is applying the older corrections even though new corrections are being used to compute e_above_hull shown on the website. See Materials Project Database Release Log - #17 by shyamd for more details.

Hi @rkingsbury, thank you for the quick reply.
As you suggested, the corrected energies with MaterialsProject2020Compatibility seems to be consistent with the e_above_hull in the database.

I will check other compounds as well.

1 Like