predicts dispersions and a density of states that are shifted and/or scaled in frequency and show different features compared to the dispersion branches and dos predicted using the finite displacement (FD) method. However, when I revert to a regular PAW-PE pseudopotential as an input to DFPT calculation in VASP, the resulting dispersion and density of states agrees with the original FD calculation and is no longer offset in frequency. Does anyone know of a particular aspect of the PAW-PE-GW POTCAR files that would result in this offset of the dispersion and dos using the DFPT method? I examined the energy of each atom (EATOM) in the POTCAR files but found that this offset occurs even when the energy of each atom is equivalent between the PAW-PE-GW and PAW-PE POTCAR’s. (LA-UR-20-25087)Phonon dispersion discrepancy between FD and DFPT when using PAW-PE-GW pseudopotentials
Moderators: Global Moderator, Moderator
-
ekoskelo
- Newbie
- Posts: 1
- Joined: Thu Jun 18, 2020 6:33 pm
Phonon dispersion discrepancy between FD and DFPT when using PAW-PE-GW pseudopotentials
I have been working on calculating the phonon dispersions and density of states for a diatomic crystalline material. When using the PAW-PE-GW pseudopotentials, DFPT in VASP (IBRION = 7 or predicts dispersions and a density of states that are shifted and/or scaled in frequency and show different features compared to the dispersion branches and dos predicted using the finite displacement (FD) method. However, when I revert to a regular PAW-PE pseudopotential as an input to DFPT calculation in VASP, the resulting dispersion and density of states agrees with the original FD calculation and is no longer offset in frequency. Does anyone know of a particular aspect of the PAW-PE-GW POTCAR files that would result in this offset of the dispersion and dos using the DFPT method? I examined the energy of each atom (EATOM) in the POTCAR files but found that this offset occurs even when the energy of each atom is equivalent between the PAW-PE-GW and PAW-PE POTCAR’s. (LA-UR-20-25087)
predicts dispersions and a density of states that are shifted and/or scaled in frequency and show different features compared to the dispersion branches and dos predicted using the finite displacement (FD) method. However, when I revert to a regular PAW-PE pseudopotential as an input to DFPT calculation in VASP, the resulting dispersion and density of states agrees with the original FD calculation and is no longer offset in frequency. Does anyone know of a particular aspect of the PAW-PE-GW POTCAR files that would result in this offset of the dispersion and dos using the DFPT method? I examined the energy of each atom (EATOM) in the POTCAR files but found that this offset occurs even when the energy of each atom is equivalent between the PAW-PE-GW and PAW-PE POTCAR’s. (LA-UR-20-25087)-
martin.schlipf
- Global Moderator
- Posts: 542
- Joined: Fri Nov 08, 2019 7:18 am
Re: Phonon dispersion discrepancy between FD and DFPT when using PAW-PE-GW pseudopotentials
Can you be more specific which POTCAR you are talking about or is this a general behavior you found for all "GW" POTCARs?
Does this behavior appear only in the material you currently investigate or can you reproduce it for other materials containing this POTCAR?
Does this behavior appear only in the material you currently investigate or can you reproduce it for other materials containing this POTCAR?
Martin Schlipf
VASP developer
-
alex
- Hero Member
- Posts: 586
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
Re: Phonon dispersion discrepancy between FD and DFPT when using PAW-PE-GW pseudopotentials
Hello ekoskelo,
you are right. This cannot be.
My first guess since you are talking about a crystal of a diatomic molecule: Do you have the same groundstate in all frequency calculations, e.g. the same energy?
Hth,
alex
you are right. This cannot be.
My first guess since you are talking about a crystal of a diatomic molecule: Do you have the same groundstate in all frequency calculations, e.g. the same energy?
Hth,
alex