defective DOS alignment with respect to bulk DOS

[nfpczhangwei]

Dear all,
I read the paper (PHYSICAL REVIEW B 88, 054111 (2013)
Strain-controlled oxygen vacancy formation and ordering in CaMnO3)
which mainly discuss the DOS change with and whithout oxygen vacancies under starin.
The paper said that "The energy scales of the stoichiometric DOS have been aligned with respect to the defective DOS at a Mn p coreb state about 50 eV below the Fermi energy. The origin in energy is set at the Fermi energy of the defective cell". I do not know how they aligned the defective DOS with respect to the bulk DOS under a certain strain. Pls. give me some hints. Thanks a lot.

[graeme]

Typically DOS alignments are done with respect to a common reference. For a slab calculation, the local potential in the vacuum is a good choice. For bulk calculations, the energy levels of low-lying (core) electronic states works pretty well (note that you may have to include s or p electrons as valance to do this well).

[nfpczhangwei]

[quote="graeme"]For bulk calculations, the energy levels of low-lying (core) electronic states works pretty well (note that you may have to include s or p electrons as valance to do this well).[/quote]
Thanks. Dear Prof. Graeme, I want to knwon how to detaily do this alignment in VASP. Is it ture that in this alignment , we assume that the e.g. the energy level of Mn p state or O s atate do not change in the stoichiometric CMO and in the defective (with oxygen vacancy) CMO?

[graeme]

In my experience it is pretty clear when the core states are aligned when comparing the DOS of two structures. While there may be small shifts in the core levels, it is unlikely for there to be a systematic shift in all of the core levels that you can calculate. Anyway, for practical purposes, this alignment is straight forward.

[nfpczhangwei]

[quote="graeme"]In my experience it is pretty clear when the core states are aligned when comparing the DOS of two structures. While there may be small shifts in the core levels, it is unlikely for there to be a systematic shift in all of the core levels that you can calculate. Anyway, for practical purposes, this alignment is straight forward.[/quote]

Dear Prof.Graeme, I still do not know the detailed step. Is true that Ef1-Ep1=Ef2-Ep2? Ef、Ep denote the fermi energy value (directly get by VASP) and the core p state value (directly get by VASP). Thanks

[graeme]

Assuming that the core levels are the same in the two systems then (Ef1-Ep1)-(Ef2-Ep2) = delta Ef will be the difference in the Fermi level.

[nfpczhangwei]

[quote="graeme"]Assuming that the core levels are the same in the two systems then (Ef1-Ep1)-(Ef2-Ep2) = delta Ef will be the difference in the Fermi level.[/quote]
Thanks, Prof. Graeme. Futhermore, I compute the core level under different pressure (not too high) with different crystal structrues (the same symmetry) of the same material, and found the core level does not change too much(~0.5 eV). Can I compare the DOSs of the two structures using the core level as references?

[graeme]

If the core levels are not changing (which is an approximation, but generally good for low-lying energy levels).