Smearing technique
How to set ISMEAR
For the calculation of the total energy in bulk materials, we recommend the tetrahedron method with Blöchl corrections (ISMEAR=-5). This method also gives a good account of the electronic density of states (DOS). The only drawback is that the method is not variational with respect to the partial occupancies. Therefore the calculated forces and the stress tensor can be wrong by up to 5 to 10% for metals. Only for semiconductors and insulators, the forces are correct because the partial occupancies do not vary and are either zero or one. For the calculation of forces and phonon frequencies in metals, we recommend the method of Methfessel-Paxton (ISMEAR>0). The method of Methfessel-Paxton (ISMEAR>0) also results in a very accurate description of the total energy, nevertheless, the width of the smearing (SIGMA) must be chosen carefully. Too large smearing parameters might result in an incorrect total energy, small smearing parameters require a dense mesh of k points. SIGMA should be as large as possible, while keeping the difference between the free energy and the total energy (i.e. the term entropy T*S) in the OUTCAR file negligible (1 meV/atom).
| Warning: Avoid using ISMEAR>0 for semiconductors and insulators, since this often leads to incorrect results (the occupancies of some states might be smaller than 0, or larger than 1). |
Errors for, e.g., phonons frequencies can be substantial, i.e., exceeding 20 %. These errors are very hard to spot if you do not look carefully. For insulators, use ISMEAR=0 or ISMEAR=-5.
The Gaussian-smearing method leads to very reasonable results in most cases. Within this method it is necessary to extrapolate from finite SIGMA results to SIGMA=0 results. You can find an extra line in the OUTCAR file: energy( SIGMA→0 ), giving the extrapolated results. Large SIGMA values lead to a similar error as the Methfessel-Paxton scheme, but in contrast to the Methfessel-Paxton scheme one can not determine how large the error due to the smearing is without systematically reducing SIGMA. In this respect, the method of Methfessel-Paxton is more convenient than the Gaussian smearing method. In addition, in the Gaussian smearing method forces and the stress tensor are consistent with the free energy and not the energy for SIGMA→0. Overall the Methfessel-Paxton method is somewhat easier to use for metallic systems.
Summary
- If you have no a priori knowledge of your system, for instance, if you do not know whether your system is an insulator, semiconductor or metal then always use Gaussian smearing ISMEAR=0 in combination with a small SIGMA=0.03-0.05.
- This is not the default in VASP yet, so to be on the safe side, you might want to include this setting in all your INCAR files.
- For semiconductors or insulators, use the tetrahedron method (ISMEAR=-5), if the cell is too large (or if you use only a single or two k points) use ISMEAR=0 in combination with a small SIGMA=0.03-0.05.
- For relaxations in metals, use ISMEAR=1 or ISMEAR=2 and an appropriate SIGMA value (the entropy term should be less than 1 meV per atom). For metals a reasonable value is often SIGMA= 0.2 (which is the default).
- Mind again: Avoid using ISMEAR>0 for semiconductors and insulators since it might cause severe problems.
- For the calculations of the DOS and very accurate total-energy calculations (no relaxation in metals), use the tetrahedron method (ISMEAR=-5).