Hello everyone,
I have been doing DFT calculation using Quantum Espresso for DOS and Bands calculation of Mg2Si bulk. The exchange correlation energy of pseudopotential that I use for both Mg and Si are using PBE which belongs to GGA functional.
For the occupations input for $SYSTEM of PWscf file input that I use is 'fixed' for vc-relax and scf -- and also 'tetrahedra' for scf(dos) and nscf(dos). For the scf(bands) and nscf(bands) I didn't write the occupations.
From the literature, Mg2Si is a semiconductor with indirect band gap.. but, my calculation result from the band plot tell that it is direct band gap.
What factor/parameter which causing it happened? Can anyone briefly explain the reason why it can be happened? Did I do something wrong and what input parameter should I change?
Thank you
Finally, I got the solution for this problem.
If you face the same problem, it might be help:
So, at first before.. I set the ibrav=0 and add some CELL_PARAMETERS {angstrom} and write the 12 atoms' positions in ATOMIC_POSITIONS {crystal} in the file input.
Then I tried using ibrav=2 since the Mg2Si is fcc, and do not forget to add the lattice parameter [celldm (in Bohr) or A (in Angstrom)]. I also changed the ATOMIC_POSITIONS {alat} with only 3 atom and their position.
And it's working!
Monalisa Nur Ramadhan
Good that you written how did you approach to solve the problem for others. Appreciate it, which many people doesn't do.
I request you may join the new google groups below to share the knowledge and ask questions.
https://groups.google.com/forum/#!forum/vasp-users
https://groups.google.com/forum/#!forum/quantumespresso
Regards
Thank you sir/madam SK M . I will be happy if my problem-solving can help others.. and thank you for the grup recommendation.
Why DFT doesn't give accurate band gap?, so is the band gap property of the ground state? no, it's an excitation, basically, it's adding or removing an electron or for example an optical excitation. So that, not the ground state energy. the DFT is not designed to do that, it gets close, you can do something as an approximation, like adding Hubbard correction (lda_plus_u= .TRUE.). DFT is designed to study the lowest energy of the system.
SK M Materials modelling using density functional theory properties and predictions. By FELICIANO GIUSTINO.
Page 173; 9.5 The band gap problem
- Badges
- Science topic
- Similar topics
- Filing