TABLE OF CONTENTS
INTRODUCTION
&PROJWFC
prefix | outdir | ngauss | degauss | Emin | Emax | DeltaE | lsym | pawproj | filpdos | filproj | lwrite_overlaps | lbinary_data | kresolveddos | tdosinboxes | n_proj_boxes | irmin | irmax | plotboxes
Notes Format of output files Orbital Order Defining boxes for the Local DOS(E) Important notices
INTRODUCTION
Purpose of projwfc.x:
projects wavefunctions onto orthogonalized atomic wavefunctions,
calculates Lowdin charges, spilling parameter, projected DOS
(separated into up and down components for lSDA)
alternatively, computes the local DOS(E), integrated in volumes
given in input
Structure of the input data:
============================
&PROJWFC
...
/
Namelist: PROJWFC |
|---|
| prefix |
CHARACTER |
| Default: |
'pwscf'
|
Prefix of input file produced by pw.x
(wavefunctions are needed).
|
| outdir |
CHARACTER |
| Default: |
value of the ESPRESSO_TMPDIR environment variable if set;
current directory ('./') otherwise
|
directory containing the input data,
i.e. the same as in pw.x
|
| ngauss |
INTEGER |
| Default: |
0
|
Type of gaussian broadening:
0 ... Simple Gaussian (default)
1 ... Methfessel-Paxton of order 1
-1 ... Marzari-Vanderbilt "cold smearing"
-99 ... Fermi-Dirac function
|
| degauss |
REAL |
| Default: |
0.0
|
gaussian broadening, Ry (not eV!)
|
|
Emin, Emax
|
REAL |
| Default: |
(band extrema)
|
min & max energy (eV) for DOS plot
|
| DeltaE |
REAL |
energy grid step (eV)
|
| lsym |
LOGICAL |
| Default: |
.true.
|
if true the projections are symmetrized
if false partial density of state cannot be computed
|
| pawproj |
LOGICAL |
| Default: |
.false.
|
if true, use PAW projectors and all-electron PAW basis
functions to calculate weight factors for the partial
densities of states. Following Bloechl, PRB 50, 17953 (1994),
Eq. (4 & 6), the weight factors thus approximate the real
charge within the augmentation sphere of each atom.
Only for PAW, not implemented in the noncolinear case
|
| filpdos |
CHARACTER |
| Default: |
(value of prefix variable)
|
prefix for output files containing PDOS(E)
|
| filproj |
CHARACTER |
| Default: |
(standard output)
|
file containing the projections
|
| lwrite_overlaps |
LOGICAL |
| Default: |
.false.
|
if .true., the overlap matrix of the atomic orbitals
prior to orthogonalization is written to the atomic_proj datafile.
|
| lbinary_data |
LOGICAL |
| Default: |
.false.
|
if .true., the atomic_proj datafile is written in binary fmt.
|
| kresolveddos |
LOGICAL |
| Default: |
.false.
|
if .true. the k-resolved DOS is computed: not summed over
all k-points but written as a function of the k-point index.
In this case all k-point weights are set to unity
|
| tdosinboxes |
LOGICAL |
| Default: |
.false.
|
if .true. compute the local DOS integrated in volumes
volumes are defined as boxes with edges parallel
to the unit cell, containing the points of the
(charge density) FFT grid included within
irmin and irmax, in the three dimensions:
from irmin(j,n) to irmax(j,n) for j=1,2,3
(n=1,n_proj_boxes)
|
| n_proj_boxes |
INTEGER |
| Default: |
1
|
number of boxes where the local DOS is computed
|
| irmin |
INTEGER |
| Default: |
1 for each box
|
first point of the given box
BEWARE: irmin is a 2D array of the form: irmin(3,n_proj_boxes)
|
| irmax |
INTEGER |
| Default: |
0 for each box
|
last point of the given box;
( 0 stands for the last point in the FFT grid )
BEWARE: irmax is a 2D array of the form: irmax(3,n_proj_boxes)
|
| plotboxes |
LOGICAL |
| Default: |
.false.
|
if .true., the boxes are written in output as
as xsf files with 3D datagrids, valued 1.0
inside the box volume and 0 outside
(visualize them as isosurfaces with isovalue 0.5)
|
|
|
Notes
Format of output files
Projections are written to standard output,
and also to file filproj if given as input.
The total DOS and the sum of projected DOS are written to file
"filpdos".pdos_tot.
* The format for the collinear, spin-unpolarized case and the
non-collinear, spin-orbit case is:
E DOS(E) PDOS(E)
...
* The format for the collinear, spin-polarized case is:
E DOSup(E) DOSdw(E) PDOSup(E) PDOSdw(E)
...
* The format for the non-collinear, non spin-orbit case is:
E DOS(E) PDOSup(E) PDOSdw(E)
...
In the collinear case and the non-collinear, non spin-orbit case
projected DOS are written to file "filpdos".pdos_atm#N(X)_wfc#M(l),
where N = atom number , X = atom symbol, M = wfc number, l=s,p,d,f
(one file per atomic wavefunction found in the pseudopotential file)
* The format for the collinear, spin-unpolarized case is:
E LDOS(E) PDOS_1(E) ... PDOS_2l+1(E)
...
where LDOS = \sum m=1,2l+1 PDOS_m(E)
and PDOS_m(E) = projected DOS on atomic wfc with component m
* The format for the collinear, spin-polarized case and the
non-collinear, non spin-orbit case is as above with
two components for both LDOS(E) and PDOS_m(E)
In the non-collinear, spin-orbit case (i.e. if there is at least one
fully relativistic pseudopotential) wavefunctions are projected
onto eigenstates of the total angular-momentum.
Projected DOS are written to file "filpdos".pdos_atm#N(X)_wfc#M(l_j),
where N = atom number , X = atom symbol, M = wfc number, l=s,p,d,f
and j is the value of the total angular momentum.
In this case the format is:
E LDOS(E) PDOS_1(E) ... PDOS_2j+1(E)
...
If kresolveddos=.true., the k-point index is prepended
to the formats above, e.g. (collinear, spin-unpolarized case)
ik E DOS(E) PDOS(E)
All DOS(E) are in states/eV plotted vs E in eV
Orbital Order
Order of m-components for each l in the output:
1, cos(phi), sin(phi), cos(2*phi), sin(2*phi), .., cos(l*phi), sin(l*phi)
where phi is the polar angle:x=r cos(theta)cos(phi), y=r cos(theta)sin(phi)
This is determined in file flib/ylmr2.f90 that calculates spherical harmonics.
for l=1:
1 pz (m=0)
2 px (real combination of m=+/-1 with cosine)
3 py (real combination of m=+/-1 with sine)
for l=2:
1 dz2 (m=0)
2 dzx (real combination of m=+/-1 with cosine)
3 dzy (real combination of m=+/-1 with sine)
4 dx2-y2 (real combination of m=+/-2 with cosine)
5 dxy (real combination of m=+/-2 with sine)
Defining boxes for the Local DOS(E)
Boxes are specified using the variables irmin and irmax:
FFT grid points are included from irmin(j,n) to irmax(j,n)
for j=1,2,3 and n=1,...,n_proj_boxes
irmin and irmax range from 1 to nr1 or nr2 or nr3
Values larger than nr1/2/3 or smaller than 1 are folded
to the unit cell.
If irmax<irmin FFT grid points are included from 1 to irmax
and from irmin to nr1/2/3.
Important notices
* The tetrahedron method is presently not implemented.
* Gaussian broadening is used in all cases:
- if degauss is set to some value in namelist &projwfc, that value
(and the optional value for ngauss) is used
- if degauss is NOT set to any value in namelist &projwfc, the
value of degauss and of ngauss are read from the input data
file (they will be the same used in the pw.x calculations)
- if degauss is NOT set to any value in namelist &projwfc, AND
there is no value of degauss and of ngauss in the input data
file, degauss=DeltaE (in Ry) and ngauss=0 will be used
Obsolete variables, ignored:
io_choice
smoothing
|