I'm currently doing BSE calculation with bs_coulom_term = 0 (RPA), so I've set up the gamma_WFK and shifted_WFK file as recommended by the tutorial and changed just that input variable.
My calculation stopped by an error
Code: Select all
--- !ERROR
src_file: bethe_salpeter.F90
src_line: 386
mpi_rank: 3
message: |
BS equation with LDA+U not completely coded
...I am using LDA+U, but if I turn off the LDA+U in the BSE equation, will that affect my result or will it be (as expected) showing the dielectric function of my material with LDA+U?
Here's my full input file (the commented ones are previous runs):
Code: Select all
# ----------------------------------------
# INPUTS COMMON TO ALL DATASETS
# ----------------------------------------
#ndtset 3
autoparal 0
paral_kgb 0
# Structural Parameters
ntypat 3
znucl 57 29 8
natom 14
typat 1 1 1 1 2 2 3 3 3 3 3 3 3 3
acell 3*3.808 Angstrom
chkprim 0
rprim
1.0 0.000 0.000
0.000 1.0 0.000
0.000 0.000 3.45851
xred 0.500000 0.500000 0.138998
0.000000 0.000000 0.361002
0.000000 0.000000 0.638998
0.500000 0.500000 0.861002
0.000000 0.000000 0.000000
0.500000 0.500000 0.500000
0.500000 0.500000 0.313883
0.000000 0.000000 0.186117
0.000000 0.500000 0.000000
0.500000 0.000000 0.000000
0.000000 0.000000 0.813883
0.500000 0.500000 0.686117
0.500000 0.000000 0.500000
0.000000 0.500000 0.500000
# Spins on atoms
spinat 0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 -1.0
0.0 0.0 1.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
nspinor 1
nsppol 1
nspden 2
# Convergence parameters
nstep 125
ecut 40
ecutsm 0.5
pawecutdg 90
pawovlp 10
# GGA-PBE + U (1=FLL, 2=AMF)
usepawu 1
lpawu -1 2 -1
upawu 0.0 8.8 0.0 eV
jpawu 0.0 0.8 0.0 eV
# K-points
ngkpt 8 8 8
nshiftk 1
#shiftk 0.0 0.0 0.0
kptopt 1
istwfk *1
ixc -101130
occopt 4
diemac 2
tsmear 0.005
# ----------------------------------------
# DATASETS
# ----------------------------------------
## Dataset 1: Ground State (SCF calculation)
#
# iscf1 17
#tolvrs1 1.d-8
# nband1 96
#prtden1 1
#
## Dataset 2: Creation of WFK file on the symmetric k-mesh
# iscf2 -2
#getden2 1
# nband2 197
#nbdbuf2 5
#tolwfr2 1.0d-8
#
## Dataset 3: Creation of WFK file on shifted k-mesh to break symmetry
# iscf3 -2
# getden3 1
# nband3 101 # reduced nband because this WFK is for transition
# tolwfr3 1.0d-8
# nbdbuf3 5
#chksymbreak3 0 # skip check here
# shiftk3 0.11 0.21 0.31 # breaks symmetry for k-mesh
## Dataset 4: Creation of the screening (eps^{-1}) matrix (WFK not shifted file)
#optdriver 3
# gwpara 2
# inclvkb 2
# awtr 1
# irdwfk 1
# ecuteps 8
# ecutwfn 20
# nband 250 # check gw tutorial 1
# nfreqre 1 # only static limit needed for BSE
# nfreqim 0 # idem,
## zcut4 0.005
##freqremax4 0.45
##gwcalctyp4 2
## gwmem4 0 # for memory shortage set 0
# Dataset 5: BSE run with Haydock
optdriver 99
irdwfk 1
irdscr 1
bs_calctype 1 # KS orbitals
mbpt_sciss 0.8 eV # Scissors operator
bs_exchange_term 1 # With Local Fields
bs_coulomb_term 0 # RPA
bs_coupling 0 # Faster, but not good with EELS
bs_loband 100
nband 250
bs_freq_mesh 0 40 0.1 # Freq mesh
bs_algorithm 2
bs_haydock_niter 200
bs_haydock_tol 0.05 0
zcut 0.15 eV
shiftk 0.11 0.21 0.31
chksymbreak 0
ecutwfn 20.0
ecuteps 8.0
inclvkb 2
Any comments would be appreciated. Thanks a lot
Best,
M. Avicenna Naradipa
National University of Singapore