Input

The fortran namelist format is used for the input of EDI. The default input file name is calcmdefect.dat Required data fils should be specified in the input file. Different calculation options are set in input file as well. Below is the detailed discussion of the input variables and their meanings.

Input file

A brief table showing the meaning of all variables are listed below in the table:

variable	meaning
prefix	qe prefix
outdir	qe outdir
lvacalign	vacuum aligment
vac_idx	vacuum alignment location
lcorealign	core alignment
core_v_d	core alignment value for defect
core_v_p	core alignment value for pristine
wt_filename	weight file for mobility calculation
degauss	gaussian smearing in delta function
noncolin	non-colinear calculation
lspinorb	spin-orbit calculation
calcmlocal	calculate local part M
calcmnonlocal	calculate non-local part M
V_d_filename	defect system local potential
Bxc_1_d_filename	defect system magnetic field along x
Bxc_2_d_filename	defect system magnetic field along y
Bxc_3_d_filename	defect system magnetic field along z
V_p_filename	pristine system local potential
Bxc_1_p_filename	pristine system magnetic field along x
Bxc_2_p_filename	pristine system magnetic field along y
Bxc_3_p_filename	pristine system magnetic field along z
calcmcharge	calculate charged defect
mcharge_dolfa	use LFA approximation in charged calculation
qeh_eps_filename	dielectric function file from QEH
doqeh	use QEH dielectric function
dogw	use BGW dielectric function
k0screen_read	Lindhard model carrier screening
gw_epsmat_filename	BGW dielectric function file for grid q
gw_eps0mat_filename	BGW dielectric function file for small q

An example input file is shown below:

&calcmcontrol
prefix='mos2',
outdir='dout/'
lvacalign=.true.
vac_idx=0
lcorealign=.false.
core_v_d=0.0
core_v_p=0.0
wt_filename='wt.dat'
klist_filename='scfklist.dat'
ev_filename='v.dat'
noncolin =.true.
lspinorb =.true.
calcmlocal = .true.
calcmnonlocal = .true.
V_d_filename='./V_d.dat'
Bxc_1_d_filename='./Bxc_1_d.dat'
Bxc_2_d_filename='./Bxc_2_d.dat'
Bxc_3_d_filename='./Bxc_3_d.dat'
V_p_filename='./V_p.dat'
Bxc_1_p_filename='./Bxc_1_p.dat'
Bxc_2_p_filename='./Bxc_2_p.dat'
Bxc_3_p_filename='./Bxc_3_p.dat'
calcmcharge=.true.
mcharge_dolfa=.true.
qeh_eps_filename='./eps.dat'
eps_type='gw'
dogw=.true.
chidat='./chi.dat'
!eps_type='qeh'
!eps_type='tf'
!doqeh=.true.
!k0screen_read=0.27
gw_epsmat_filename='./epsmat.h5'
gw_eps0mat_filename='./eps0mat.h5'
/

Input parameters

A detailed description of the input parameters is as follows:

QE parameters

These parameters should be the same as used in QE:

prefix, outdir, noncolin, lspinorb

Energy alignment

The energies calculated from different systems may not be able to directly compare. In order to obtain the correct perturbation potential, we need to choose proper energy alignment methods. EDI provides 2 types of energy alignment algorithms:

• vacuum alignment

• core alignment

Vacuum alignment is suitable for 2D materials, a 2D plane chosen from input file will be used to calculate an averaged energy to represent the vacuum energy. Currently, only plane perpendicular to z direction is supported. To use vacuum alignment, set lvacalign to .true.. vac_idx also needs to be set. This parameter sets the location of the vacuum plane, in the form of the FFT grid number index from the DFT calculation.

Core alignment is suitable for 3D bulk materials, the value should be the core level energies of proper element. To use core alignment, set lcorealign to .true.. core_v_d and core_v_p needs to be set for this option. The represent the core level energy of defect and pristine structures respectively.

K point sampling

The initial and final wavefunctions for the scattering process are needed for the calcualtion of matrix elment. The band number and k points are the index for the wavefunctions. Two methods are provided for the sampling of k points: The index of the wavefunction pairs are given in the weight file, which is set by parameter wt_filename. The weight file can be obtained with the provided scripts.

• If uniform grid is used:

  A complete list of \(C_N^2\) kpoint pairs with the gaussian smearing is needed in the weight file.

• If triangular integral method for 2D system is used:

  The wavefunctions pairs are determined using triangular algorithm from the energy conservation term in the Fermi’s golden rule.

Neutral defect perturbtation potential

The neutral defect perturbation potential is separated into local and non-local parts. To calculate matrix element from it, set calcmlocal and calcmnonlocal to .true.. Additionally, the following parameters should be set to determine the files for the potentials.

V_d_filename
Bxc_1_d_filename
Bxc_2_d_filename
Bxc_3_d_filename
V_p_filename
Bxc_1_p_filename
Bxc_2_p_filename
Bxc_3_p_filename

Note

The Bxc file is needed only for SOC calculations.

Charged defect perturbtation potential

If defect is charged, the perturbation potential is represented with a different model from neutral ones. Currently, supported model is Coulomb potential of a point charge, screened by the material. Various screening model is supported by EDI.

To perform this calculation, set the parameter calcmcharge to .true..

Local Fielad Approximation (LFA) is supported for the charged defect systems. To turn on, set the parameter mcharge_dolfa to .true..

Currently, the supported screening models include:

• Thomas-Fermi model with dielectric constant

  Set k0screen_read to use dielectric constant

• Quantum Electrostatic Hetereostructure model (scalar dielectric function)

  Set doqeh to use QEH dielectric function.

  Set qeh_eps_filename for the dielectric function obtainedfrom QEH model

• Lindhard model (matrix dielectric function)

  Set dogw to use BGW dielectric function

  Set gw_epsmat_filename for the full dielectric matrix for q grid obtained from BGW

  Set gw_eps0mat_filename for the full dielectric matrix for small q obtained from BGW