VASP piezoelectric calculations
Posted: Mon Apr 13, 2020 5:55 pm
Dear VASP users,
I am trying to understand how piezoelectric constants, among others are computed in VASP. If I understand correctly, LCALCEPS and IBRION=6, the piezoelectric tensor is computed into two parts: the clamped-ion part through the PEAD scheme and ion relaxation contribution from the product of the internal strain tensor matrix (L), the pseudo-inverse of the Hessian matrix (K-1) and the Born effective charge tensor (Z). My main question concerns the clamped-ion piezoelectric tensor. I am a bit confused by what is done in the PEAD_ELMIN routine. It seems to me that the initial stress tensor is obtained from minimizing the wavefunction in absence of the electric field, however the forces derived from this minimized wavefunction are obtained in presence of the electric field. Am I correct? The other stress tensor for the SCF response is obtained from minimized wafunction and derived forces in presence of the electric field. Is this also correct. Finally, I am wonder why the Ewald stress tensor in presence of the electric field is added to the piezoelectric tensor but to none of the stress tensors aforementioned? I would be grateful if someone could help me understand why this is done the way it is.
Best regards,
Pierre
I am trying to understand how piezoelectric constants, among others are computed in VASP. If I understand correctly, LCALCEPS and IBRION=6, the piezoelectric tensor is computed into two parts: the clamped-ion part through the PEAD scheme and ion relaxation contribution from the product of the internal strain tensor matrix (L), the pseudo-inverse of the Hessian matrix (K-1) and the Born effective charge tensor (Z). My main question concerns the clamped-ion piezoelectric tensor. I am a bit confused by what is done in the PEAD_ELMIN routine. It seems to me that the initial stress tensor is obtained from minimizing the wavefunction in absence of the electric field, however the forces derived from this minimized wavefunction are obtained in presence of the electric field. Am I correct? The other stress tensor for the SCF response is obtained from minimized wafunction and derived forces in presence of the electric field. Is this also correct. Finally, I am wonder why the Ewald stress tensor in presence of the electric field is added to the piezoelectric tensor but to none of the stress tensors aforementioned? I would be grateful if someone could help me understand why this is done the way it is.
Best regards,
Pierre