PAW control tags

In principle, the PAW method can be used in the same manner as the US-PP method. Only special PAW POTCAR files are required. In principle, also no additional user interference is required. However there are a few flags that control the behavior of the PAW implementation. The first one is LMAXPAW:

 LMAXPAW = l

This flag controls the maximum $l$ quantum number for the evaluation of the on-site terms on the radial support grids in the PAW method. The default for LMAXPAW is $2*l_{max}$, where $l_{max}$ is the maximum angular quantum number of the partial waves. Useful settings for LMAXPAW are for instance:

 LMAXPAW = 0

In this case, only spherical terms are evaluated on the radial grid. This does not mean that a-spherical terms are totally neglected, because the compensation charges are always expanded up to $2*l_{max}$ on the plane wave grid.

Finally, LMAXPAW=-1 has a special meaning. For LMAXPAW=-1, no on-site correction terms are evaluated on the radial support grid, which effectively means that the behavior of US-PP's is recovered with PAW input datasets. Usually this allows very efficient and fast calculations, and this switch might be of interest for relaxations and molecular dynamics runs. Energies should be evaluated with the default setting for LMAXPAW.

An additional flag controls up to which l quantum number the onsite PAW charge densities are passed through the charge density mixer and written to the CHGCAR file:

 LMAXMIX = l

The default is LMAXMIX=2. Higher l-quantum numbers are usually not handled by the mixer, i.e. a straight mixing is applied for them (the PAW on-site charge density for higher l quantum numbers is reset precisely to the value corresponding to the present wavefunctions). Usually, it is not required to increase LMAXMIX, but the following two cases are exceptions:

• L(S)DA+U calculations require in many cases an increase of LMAXMIX to 4 (or 6 for f-elements) in order to obtain fast convergence to the groundstate.
• The CHGCAR file also contains only information up to LMAXMIX for the on-site PAW occupancy matrices. When the CHGCAR file is read and kept fixed in the course of the calculations ( ICHARG=11), the results will be necessarily not identical to a selfconsistent run. The deviations can be (or actually are) large for L(S)DA+U calculations. For the calculation of band structures within the L(S)DA+U approach it is strictly required to increase LMAXMIX to 4 (d elements) and 6 (f elements).

The second switch, that is useful in the context of the PAW method (and US-PP) is ADDGRID. The default for ADDGRID is .FALSE. If

 ADDGRID = .TRUE.

is written in the INCAR file, an additional (third) support grid is used for the evaluation of the augmentation charges. This third grid contains 8 times more points than the fine grid NGXF, NGYF, NGZF. Whenever terms involving augmentation charges are evaluated, this third grid is used. For instance: The augmentation charge is evaluated first in real space on this fine grid, FFT-transformed to reciprocal space and then added to the total charge density on the grid NGXF, NGYF, NGZF. The additional grid helps to reduce the noise in the forces significantly. In many cases, it even allows to perform calculations in which NGXF=NGX etc. This can be achieved by setting

 ENAUG = 1  ; ADDGRID = .TRUE.

in the INCAR file. The flag can also be used for US-PPs, in particular, to reduce the noise in the forces.