: specifies the
utilization of QP-corrections for X, G0, or K
The variable QPdb is one of two options to impose QP corrections. The other
option utilizes scissors-shift for conduction band states and a stretch correction
for the band width of the LDA band width of conduction and valence bands (cf.
variable QP_E ).
enables the inclusion of QP corrections obtained in a preceding
calculation as stored in db.QP file and applies them to the corresponding physical
quantity (response function X, green's function G or BSE
The general format of the QPdb
- none: QP-correction will not be applied
- <WHAT>_<_<DB>_<OPTION: The quantity specified by
<WHAT> (see below)
will be read from . is the db.QP file name including the relative
path (e.g. "SAVE/ndb.QP"). Mind the spaces _ to separate the
The possible options for <WHAT
- "E" QP-energy corrections (real)
- "Z" Renormalization factors
- "W" QP-width (i.e. imaginary self-energy)
while the only <OPTION
> supported now is "iu
activates the interpolation procedure for the missing QP states.
"E < example/db.qp" - QP-energy corrections will be read from the database file
"db.qp" stored in the directory "example/".
Interpolation of missing QP-values
For a consistent calculation all states included either X, G, or K must be corrected
on the same footing. However, to provide QP-values for all states via the
QP-database may be to costly in some cases. Therefore, an interpolation of the
missing QP-values on the basis of the QP-database is available via the token
. For each band state at a given k-point the missing QP-value (E,Z, or W)
will be calculated as an average of the values provided for the QP_N
neighbouring k-points (i.e. those with smallest
distance). The QP-values thus needs to be provided only at a number of k-points for
the bands in question.