elastic constants (transition metals) with PAW (LDA+U)
Posted: Fri Aug 08, 2014 12:42 pm
hello everyone!
I would like to calculate elastic constants for some oxides of transition metals of d block. I am using 7.8.2 version of abinit.
The geometry is preliminary relaxed with highly converged results.
Using LDA only gives me not realistic results and is problematic no metter how big is plane-waves basis set or how fine is k-point grid, and whether or not I take spin polarization into account (e.g. twice larger values for C11, C12 elements in comparence to the experimental values, negative values for C44, unability to calculate relaxed-ion elastic tensor, anaddb complains about violation of acoustic sum rule, stability of a system ..)
At the same time if I use PAW method (with or without LDA+U) a code crashes with a bug:
Inconsistency between variables choice and ngrhoij !
src_file: m_pawrhoij.F90
I would very appreciate any suggestions, guys.
Here is an example of an input file:
ndtset 2
# Set 1: gs
iscf1 7
kptopt1 1
tolvrs1 1.0d-18
# Set 2 : rf
getwfk2 -1
iscf2 7
ptopt2 2
nqpt2 1
qpt2 0. 0. 0.
rfphon2 1
rfatpol2 1 2
rfstrs2 3
rfdir2 1 1 1
tolvrs2 1.0d-10
#Atomic system data
#so and so
#PW basis
ecut 12.0
ecutsm 0.5
pawecutdg 30.
#k-point grid
kptopt 1
ngkpt 4 4 4
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
#scf
diemac 4.0 #the last thing I was trying to calculate belongs to Mott insulators
iscf 7
nstep 200
optforces 1
pawxcdev 0 #Exchange-corr. part in the spherical part of energy
#is computed on the angular mesh
#Spin
nsppol 2 #FM case
nspden 2
nspinor 1
spinat 0 0 2.0
0 0 0
I would like to calculate elastic constants for some oxides of transition metals of d block. I am using 7.8.2 version of abinit.
The geometry is preliminary relaxed with highly converged results.
Using LDA only gives me not realistic results and is problematic no metter how big is plane-waves basis set or how fine is k-point grid, and whether or not I take spin polarization into account (e.g. twice larger values for C11, C12 elements in comparence to the experimental values, negative values for C44, unability to calculate relaxed-ion elastic tensor, anaddb complains about violation of acoustic sum rule, stability of a system ..)
At the same time if I use PAW method (with or without LDA+U) a code crashes with a bug:
Inconsistency between variables choice and ngrhoij !
src_file: m_pawrhoij.F90
I would very appreciate any suggestions, guys.
Here is an example of an input file:
ndtset 2
# Set 1: gs
iscf1 7
kptopt1 1
tolvrs1 1.0d-18
# Set 2 : rf
getwfk2 -1
iscf2 7
ptopt2 2
nqpt2 1
qpt2 0. 0. 0.
rfphon2 1
rfatpol2 1 2
rfstrs2 3
rfdir2 1 1 1
tolvrs2 1.0d-10
#Atomic system data
#so and so
#PW basis
ecut 12.0
ecutsm 0.5
pawecutdg 30.
#k-point grid
kptopt 1
ngkpt 4 4 4
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
#scf
diemac 4.0 #the last thing I was trying to calculate belongs to Mott insulators
iscf 7
nstep 200
optforces 1
pawxcdev 0 #Exchange-corr. part in the spherical part of energy
#is computed on the angular mesh
#Spin
nsppol 2 #FM case
nspden 2
nspinor 1
spinat 0 0 2.0
0 0 0