Hello, I was planning on using B3LYP functional to do some band gap calculations. I have now discovered this isn't an option for B3LYP functionals. So, wondering if anyone has any advice on what functional to use for a non-SCF calculation that maintains quite high accuracies.
Thanks.
In order to use the hybrid functionals in QE, you need to recompile the program using -DEXX option in the DFLAGS of your make.sys.
Thanks Mohammad Saeed Bahramy but I am able to use hybrid functionals in scf calculations. However, QE does not support the use of these functionals in non-scf calculations so I am trying to find out which functionals I can use in non-scf calculations.
OK, if you can run a SCF calculation using hybrid functionals, you can try one of these methods:
1. Manually generate a k-mesh with appropriate weights, and then add additional fake k-points corresponding to the path along which you want to calculate the band structure. For the latter, the weight must be Zero. Put these k-points all together in the input file and run a standard SCF calculation. Once the calculation is converged, you can extract the eigenvalues corresponding to the fake k-points from the output and plot the resulting band structure.
2. Do a Wannier interpolation using wannier90. For this, you don't need fake k-points anymore, but the k-mesh that you use for the SCF and Wannier calculations must be commensurate. To enforce this, the safest option is to turn off symmetry and use a uniform k-mesh grid for both calculations.
3. Force the sci-calculation to generate a coarse q-mesh that is commensurate with your k-mesh and then import the resulting output data using BoltztraP2 to interpolate the band structure.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Quantum Physics
- Quantum Mechanics