Useful tools integrated for VASP/Wannier90 interface:
-
Give suggestions about key inputs for Wannier90 including
projectionanddis_windows. -
Easy-accessible help function for Wannier90 input.
-
Evaluate Wannier90 interpolation result.
-
low requirement for running. Requirement:
pymatgenandother common python packagesnumpy,matplotlibandscipy.
pre_w90_tool.py offers different outputs with arguments input including DOS (density of states) analysis and suggestions for Wannier90 input.
usage: pre_w90_tool.py [-h] [--path PATH] [--no-soc] [--pick PICK]
[-e ERANGE ERANGE] [--plot] [--extra EXTRA]
mode
Pre analysis before Wannier90 Interpolation.
positional arguments:
mode Mode: dos, kpath, band, template
optional arguments:
-h, --help show this help message and exit
--path PATH Default: .
--no-soc with SOC or not. Default: False
--pick PICK Minimum of selected orbital / max single orbital. default:
0.1
-e ERANGE ERANGE Energy range.
--plot plot the dos distribution
--extra EXTRA Extra input.
The dos mode is the essential utility in pre_w90_tool.py. The program will give the suggestion of projection including total DOS (density of states) and projected DOS within a given energy range with input -e arguments.
First, per_w90_tool.py will give the dos distribution of each orbital at each atom.
⚠ Since the default value of EMIN and the EMAX in VASP is lowest and highest KS-eigenvalue, please mind that there might be some problems for integral calculation outside the EMIN and EMAX interval.
In GaAs example, python pre_w90_tool.py dos -e 2 5 --plot --no-soc --pick 0.15 gives
Calculated DOS Energy Range: 2.0, 5.0
species structure_id orb_id orb_name key_string dos
0 Ga 0 0 s Ga_0_s 0.023871
1 Ga 0 1 py Ga_0_py 0.195084
2 Ga 0 2 pz Ga_0_pz 0.282813
...
⚠ Notice that dos column is normalized with maximum of 1 and the energy doesn't subtract the Fermi level. If we add argument --plot, pre_w90_tool.py will generate a histogram with key_string and dos. In order to not make the figure too long, there is a threshold for histogram plotting that the dos distribution is at least 0.01.
Based on the dos distribution and --pick pick_rate input, we choose the DOS distribution larger than pick_rate as the projection input for Wannier90. Related output as following
Number of Selected Orbitals: 6
Number of Selected WFs: 6
Selected Orbitals:
species site orb
0 Ga -1 [p]
1 As -1 [p]
Wannier90 Projection:
Ga:l=1
As:l=1
⚠ Notice that the number of selected orbitals is not always equal to the number of WFs (Wannier function). When the system includes SOC, there are twice of the number orbitals needed for WFs. This is controlled by the argument --no-soc. It becomes important when we input the argument --extra for dis_windows suggestion.
Make full use of the above message, we can decide the projections for Wannier90 input. These projections can also be parsed to pre_w90_tool.py by following
python pre_w90_tool.py dos -e -6 14 --plot --no-soc --extra 'Ga,0,1-3;As,1,1-3'Calculated DOS Energy Range: -6.0, 14.0
There are 10 states in [-6.0, 14.0] with Fermi level at 3.467.
Suggest dis_froz_min & dis_froz_max as following:
nwann: 6 degenercy: 1 Fermi: 3.467000
WANRING: There are states between given `emin`: -6.0 and lowest `dis_froz_min`: 3.470754. Please carefully treat the suggestion of dis_froz_min / dis_froz_max and check energy range of each bands again. This situation usually happens in no-SOC system with many denegeracy point. But we still want to give you some useful energy window information with states less than number of WFs.
dis_froz_min dis_froz_max pdos tdos percent
4 -6.000000 5.220833 1.172082 2.663917 0.439984
0 3.470754 10.658390 2.108112 6.807999 0.309652
3 7.464829 14.206528 3.352538 10.885037 0.307995
2 3.470755 12.387263 3.097869 10.639155 0.291176
1 3.470755 12.060134 3.153824 11.116874 0.283697
5 -6.000000 7.175487 2.794993 9.946576 0.281000
Lowest `dis_win_max` for -6.0: 13.100133999999999
⚠ If you input the energy range for integral calculation outside the EMIN and EMAX interval, you might get unreasonable result of pdos and tdos.
First, the --extra argument is separated by ; and each part consists of species, structure_id, orbital_id. The setting of structure_id and orbital_id is the same as pymatgen starting from 0. These indices of orbitals are basically the order in
which the orbitals are reported in VASP and have no special meaning. You can use - to replace the continuous number. For example, you can use 0-3 to replace 0,1,2,3 both in structure_id and orbital_id input.
The pdos column gives the summation of all the atom and orbital input. The tdos column gives the total DOS. The percent column gives the quotient of both.
⚠ For system without SOC, there are many degeneracy points affecting our dis energy suggest procedure.
⚠ The percentage of pdos versus tdos doesn't directly mean the quality of fitting.
Since dis_windows require an energy window containing states larger than the number of target WFs. We also give the lowest bound for dis_win_max if you make the lower input energy becomedis_win_min.
The selected projection can also be plotted by adding the argument --plot. Remind of your input energy range.
⚠ Notice that if the input energy is too small to containing enough states, the suggestion will become empty.
$ python pre_w90_tool.py dos -e 2 5 --plot --no-soc --extra 'Ga,0,1-3;As,1,1-3'
Calculated DOS Energy Range: 2.0, 5.0
There are 4 states in [2.0, 5.0] with Fermi level at 3.467.
Empty DataFrame
Columns: [dis_froz_min, dis_froz_max, pdos, tdos, percent]
Index: []
Plot with selected orbitals in `brown` and not selected orbitals in `orange`.
6 orbitals selected
Plot 14 orbitals with 6 selected
pre_w90_tool.py also gives some other useful messages for Wannier90 input.
-
In
kpathmode,pre_w90_tool.pyconverts VASP-type{path}/KPOINTSintoWannier90input . -
In
bandmode,pre_w90_tool.pygeneratesbnd.datfile with k-distance and energy eigenvalue in p4vasp-format in{path}folder. For spinful system,pre_w90_tool.pywill export band data tobnd_up.datandbnd_down.datseparately. -
In
templatemode,pre_w90_tool.pyoffers templates forwannier90.win. Withinextrainput (basic,wann,band), we can choose one of the detailed parts to print.
Since we have some dis_windows suggestions from pre_w90_tool.py, you might needdis_win_suggest.py to offer a more detailed eigenvalue analysis. The eigenvalue file can be EIGENVAL and wannier90.eig.
usage: dis_win_suggest.py [-h] [-i EIG] [--path PATH] [--efermi EFERMI]
[-w NWANN] [-n NBNDS_EXCL] [-d NDEG]
[-e ERANGE ERANGE] [--seperate]
mode
CLI Tool for W90 energy windows.
positional arguments:
mode Mode: report, plot, count, suggest
optional arguments:
-h, --help show this help message and exit
-i EIG Select wannier90.eig file or EIGENVAL file. Default:
EIGENVAL
--path PATH Default: .
--efermi EFERMI Fermi level. Default value is generated from
`vasprun.xml`.
-w NWANN Number of Wannier Functions. Default: 0
-n NBNDS_EXCL Number of bands excluded
-d NDEG Number of degeneracy.
-e ERANGE ERANGE Energy range.
--separate Calculate bands not separately.
The report mode prints a table about the distribution of eigenvalues.
python dis_win_suggest.py report
EFERMI: 3.467000
--------------------------------
Band No. EMIN EMAX
--------------------------------
0~ 2 -11.34863 -11.21636
3~ 4 -11.20861 -11.17761
5~ 5 -8.90720 -6.59268
6~ 8 -3.10932 +3.46676
9~ 35 +3.61018 +41.52101
--------------------------------
So we can clear see the Fermi level is located at 3.467 and there is a gap between -6.59268 and -3.10932. It's also the reason why we use -e -6 15 in previous section for dis_windows suggestion.
Detailed distribution of each band can be achieved by argument --separate. We can also use -e to control the band energy to show.
python dis_win_suggest.py report --separate -e -6 15 -d 1
EFERMI: 3.467000
--------------------------------
Band No. EMIN EMAX
--------------------------------
6 -3.10932 +3.46675
7 -0.24141 +3.46676
8 +0.51331 +3.46676
9 +3.61018 +7.46083
10 +5.22483 +9.11682
11 +7.17949 +13.09613
12 +7.17949 +13.09614
13 +10.66239 +13.96102
14 +12.06413 +16.37072
15 +12.39126 +17.32905
16 +14.21053 +18.69125
17 +14.89751 +19.88154
18 +14.89751 +21.08752
--------------------------------
The result can output to eigenval_dis.png by switching report mode to plot mode.
The count mode gives how many states inside the given energy window.
$ python dis_win_suggest.py count -e -6 15
There are 13 states in [-6.0, 15.0].
The suggest mode gives the dis_windows suggestion the same as pre_w90_tool.py requiring you manually input the number of WFs through argument -w but without tdos and pdos. Since this report only requires EIGENVAL or wannier90.eig, so it's much faster than pre_w90_tool.py especially your vasprun.xml is quiet large.
$ python dis_win_suggest.py suggest -e -6 15 -d 1 -w 6
Lowest `dis_win_max` for -6.0: 13.100133999999999
`dis_froz_min` and `dis_froz_max` Table:
There are 13 states in [-6.0, 15.0] with Fermi level at 3.467.
Suggest dis_froz_min & dis_froz_max as following:
nwann: 6 degenercy: 1 Fermi: 3.467000
WANRING: There are states between given `emin`: -6.0 and lowest `dis_froz_min`: 3.470754. Please carefully treat the suggestion of dis_froz_min / dis_froz_max and check energy range of each bands again. This situation usually happens in no-SOC system with many denegeracy point. But we still want to give you some useful energy window information with states less than number of WFs.
dis_froz_min dis_froz_max
0 3.470754 10.658390
1 3.470755 12.060134
2 3.470755 12.387263
3 7.464829 14.206528
4 9.120819 14.893510
5 9.120819 14.893510
6 13.100137 17.779136
7 -6.000000 5.220833
8 -6.000000 7.175487
After Wannier90 have wannierised and interpolated band structure, cmp_vasp_w90.py can compare the first-principle DFT results from VASP with Wannier90 band data.
usage: cmp_vasp_w90.py [-h] [--efermi EFERMI] [--path PATH] [--vasp VASP] [--ylim YLIM YLIM]
[--kernel KERNEL] [--show-fonts] [--fontfamily FONTFAMILY]
[--fontsize FONTSIZE] [--no-spread] [--no-quality] [--quiet]
name
Comparsion between VASP band and Wannier90 band. `bnd.dat` for VASP band data in p4vasp format
and `wannier90_band.dat`, `wannier90_band.labelinfo.dat`, and `wannier90.wout` are required for
plotting and analysis.
positional arguments:
name name of system
optional arguments:
-h, --help show this help message and exit
--efermi EFERMI Fermi level. Default value is generated from `vasprun.xml`.
--path PATH Default: .
--vasp VASP location of VASP band file in p4vasp format
--ylim YLIM YLIM Energy bound for plot. Default: [E_w90.min - 1, E_w90.max + 1]
--kernel KERNEL kernel function for evaluating diff: type, middle, width. There are two
type of kernel function: unit and gaussian. Defalut=unit,2,5
--show-fonts Show all availabel font families can be used in `rcParams`
--fontfamily FONTFAMILY
Set font family manually. Default: Open Sans
--fontsize FONTSIZE Set font size manually. Default: 18
--no-spread Don't plot spreading
--no-quality Don't show quality of fitting
--quiet Equal to --no-spreading --no-quality
Since the VASP and Wannier90 apply different kpoints in band calculation, we need to use B-spline to interpolate the Wannier90 band data and compare with VASP band data. The quality of Wannier90 result is evaulated from
for each band.
Example usage:
$ python cmp_vasp_w90.py GaAs2 --kernel unit,3.5,1 --ylim -1 1
INFO:cmp:Reading Data from ./bnd.dat
INFO:cmp:k label: ['G', 'X', 'W', 'K', 'G', 'L', 'U', 'W', 'L', 'K', 'U', 'X']
INFO:cmp:Energy range: [-1.0, 1.0]
INFO:cmp:Output figure: ./GaAs2_VASP_W90_cmp.png
INFO:cmp:Evaluating Band Quality:
INFO:cmp:nbnds_excl: 6
INFO:cmp:=== MAX DIFF of VASP vs W90 with each bands (meV) ===
25 ++----------------+-----------------+------------------+-----------------+----------------++
+ + + + + +
| |
| * |
| * * |
20 ++ * * ++
| * * |
| * * |
| * * |
| * * |
15 ++ * * ++
| * * |
| * * |
| * * |
| * * |
10 ++ * * ++
| * * |
| * * |
| * * |
| * * |
5 ++ * * ++
************** * * |
| **** * * |
| ******************* * |
+ + + + *+ +
0 ++----------------+-----------------+------------------+-----------------*******************
1 2 3 4 5 6
INFO:cmp:Average dE (meV): 7.419081825327686
INFO:cmp:Show spreading convergence:
+-----*--------+--------------+---------------+--------------+--------------+--------------+
+ * + + + + + +
| * |
| * |
| * |
30 ++ * ++
| * |
| * |
| * |
| * |
25 ++ * ++
| * |
| * |
| * |
| * |
| * |
20 ++ * ++
| * |
| * |
| *********************************************************************** |
| |
| |
15 ++ ++
| |
+ + + + + + +
+--------------+--------------+---------------+--------------+--------------+--------------+
0 5 10 15 20 25 30
INFO:cmp:MIN_NUM_ITER: 26 SPREAD: 17.0317092107
We use GaAs as example. See input file at ./GaAs folder. DOS is plotted below.
After self-consitant calculation of VASP,
-
python pre_w90_tool.py templateto get how to writewannier90.win. -
python pre_w90_tool.py kpathto getKpoint_Pathfor Wannier90. -
python pre_w90_tool.py band --path ../bndto generate p4vasp-typebnd.datfile. For simplicity, I recommend you copy or link this file towannierfolder. -
Do DOS analysis. This is the key feature for
Wannier90_Toolboxand above description has been very detaild. -
Compare your Wannier90 interpolation result with VASP through
python cmp_vasp_w90.py GaAs2 --kernel unit,3.5,1 --ylim -1 1. ⚠ Remind that the parameters for kernel function is middle point and its half width. You can also check maximum difference between VASP and Wannier90 for each bands and the total spreading convergence during the wannierisation.
See grep TODO *.py