Compensation charge over fine fft grid in nscf run with PAW
Posted: Thu Mar 16, 2017 5:27 am
Dear All
I've tried to generate the input files for the Boltzrap code in the calculation using PAW, with Abinit 8.0.8b.
I used the the input files from "tpaw1" as my study. I added to the input file a nscf calculation (dataset2)
following a usual ground state scf (dataset1), at higher k-grid density. The input file is listed as below.
The log file are attached.
The calculation was completed and all the needed files were generated. I checked the output file for the "PAW TEST" section
and found:
-for dataset1 (scf)
for dataset2 (nscf)
I am not sure if the "Compensation charge over fine fft grid = 0.000000000000000" is what we normally
have when doing nscf in calculations using PAW? , or if I did something wrong?
Kind Regards,
Thanusit
I've tried to generate the input files for the Boltzrap code in the calculation using PAW, with Abinit 8.0.8b.
I used the the input files from "tpaw1" as my study. I added to the input file a nscf calculation (dataset2)
following a usual ground state scf (dataset1), at higher k-grid density. The input file is listed as below.
The log file are attached.
Code: Select all
# =======================================
# Input for PAW1 tutorial
# Diamond at experimental volume
# =======================================
ndtset 2
#dataset1: ground state scf
ngkpt1 6 6 6
#dataset2: nscf at a denser k-grid & output the files for Boltztrap code
iscf2 -3
getden2 1
ngkpt2 12 12 12
prtbltztrp2 1
#Cell and atoms definition
acell 3*3.567 angstrom
rprim 0.0 1/2 1/2
1/2 0.0 1/2
1/2 1/2 0.0
ntypat 1 natom 2 typat 2*1
xred 0.0 0.0 0.0
1/4 1/4 1/4
znucl 6 nband 6
#Grid definitions
ecut 15.
ecutsm 0.5
#PAW-Energy CUToff for the Double Grid
pawecutdg 50
#SCF cycle parameters
tolvrs 1.0d-10
nstep 20
#K-points and sym
nsym 0
occopt 1
kptopt 1
nshiftk 4
shiftk 0.5 0.5 0.5
0.5 0.0 0.0
0.0 0.5 0.0
0.0 0.0 0.5
#I/O parameters
prtden 1 prtdos 2
The calculation was completed and all the needed files were generated. I checked the output file for the "PAW TEST" section
and found:
-for dataset1 (scf)
Code: Select all
PAW TEST:
==== Compensation charge inside spheres ============
The following values must be close to each other ...
Compensation charge over spherical meshes = 0.413178580175431
Compensation charge over fine fft grid = 0.413177280327792
for dataset2 (nscf)
Code: Select all
PAW TEST:
==== Compensation charge inside spheres ============
The following values must be close to each other ...
Compensation charge over spherical meshes = 0.413178429102933
Compensation charge over fine fft grid = 0.000000000000000
I am not sure if the "Compensation charge over fine fft grid = 0.000000000000000" is what we normally
have when doing nscf in calculations using PAW? , or if I did something wrong?
Kind Regards,
Thanusit