Unit cell parameters were found from geometric optimization in VASP using PAW data sets with PW-91 xc functional. Silicon unit cell contains 8 atoms, thus 24 frequencies (phonons), three acoustic close to zero.
For this example vasp_raman.py
script version [0.6.0] (https://raw.github.com/raman-sc/VASP/51627f0c9ce24526ef757c1d8212cfa2dbd84648/vasp_raman.py) was used.
This example calculates the Raman activity tensor of bulk Si.
The phonon modes are calculated by a finite difference method (http://theory.cm.utexas.edu/vtsttools/dynmat/) which is implemented through VASP.
The Raman activity tensor is calculated following the completion of the dynamical matrix calculation using the script vasp_raman.py. as explained in the CPD example [Cyclopentadiene example](https://github.com/ raman-sc/VASP/tree/master/Cyclopentadiene).
Several proceedural changes are necessary to complete setup the calculation from DMFDM output as outlined below:
##Generating normal mode frequency files and eigenvector files Necessary scripts from vtstscripts package: dymmatrix.py
Run dymmatrix.py which creates the hessian matrix and gives output files: freq.dat, modes_sqrt_amu. dat which will be used here, others output files are not relevant for this calcualtion.
Include freq.dat and modes_sqrt_amu.dat in your working directory
With the post-processing of the DM VASP calculation complete,
The order of these modes is from smallest to largest wavenumbers in the DMFDM freq.dat file
Create a directory which contains the following:
POSCAR.phon -used in DM VASP run
OUTCAR -from DM VASP run
INCAR.raman -should contain LEPSILON=.TRUE. or LCALCEPS=.TRUE. because we want 'MACROSCOPIC STATIC DIELECTRIC TENSOR' in the OUTCAR
KPOINTS
POTCAR
vasp_raman.py -version 0.6.0
DMFDM.py -support script to vasp_raman.py which parses DMFDM output
raman_sub.sh -changed to link DMFDM output files
raman.sub
collect.sh
freq.dat
modes_sqrt_amu.dat
Run raman_sub.sh to setup and exectute vasp_raman.py through local queueing system
Run collect.sh after all VASP calculations are complete. Contents of 'collect.sh':
#!/bin/bash
backup_dir="Raman_Backup"
mkdir -p ${backup_dir}
rm -f vasp_raman.dat
for dir in ./Modes*
do
cat "${dir}"/vasp_raman.dat >> vasp_raman.tmp
cp "${dir}"/OUTCAR.* "${backup_dir}"
done
head -n 1 vasp_raman.tmp > vasp_raman.dat
sed '/^#/d' vasp_raman.tmp > vasp_raman.unsorted
sort -k 2 -n vasp_raman.unsorted >> vasp_raman.dat
rm -f vasp_raman.tmp
rm -f vasp_raman.unsorted
The remaining steps are the same as in the CPD example Cyclopentadiene example.
Resulting freq.dat output from DMFDM calculation with dr = 0.005 Angstoms, PW-91 xc, 4x4x4 kpt sampling:
#freq ... 1=imaginary, 0 = real
0.013841 cm^{-1} ... 1
0.004294 cm^{-1} ... 1
0.000438 cm^{-1} ... 1
141.905626 cm^{-1} ... 0
141.908828 cm^{-1} ... 0
141.911775 cm^{-1} ... 0
141.932114 cm^{-1} ... 0
141.933816 cm^{-1} ... 0
141.939454 cm^{-1} ... 0
412.092481 cm^{-1} ... 0
412.102361 cm^{-1} ... 0
412.104556 cm^{-1} ... 0
412.105947 cm^{-1} ... 0
412.108592 cm^{-1} ... 0
412.125218 cm^{-1} ... 0
463.614422 cm^{-1} ... 0
463.630034 cm^{-1} ... 0
463.639223 cm^{-1} ... 0
463.645767 cm^{-1} ... 0
463.647886 cm^{-1} ... 0
463.660010 cm^{-1} ... 0
515.536841 cm^{-1} ... 0
515.549200 cm^{-1} ... 0
515.550550 cm^{-1} ... 0
output from 'vasp_raman.dat':
# mode freq(cm-1) alpha beta2 activity
4 141.90563 0.0003618 5.6200775 39.3405484
5 141.90883 0.0052292 54.6599140 382.6206283
6 141.91178 -0.0037038 160.0769321 1120.5391422
7 141.93211 0.0000371 78.9567767 552.6974371
8 141.93382 0.0085383 83.6444174 585.5142023
9 141.93945 0.0270537 57.5587791 402.9443895
10 412.09248 0.0000720 171.8463528 1202.9244699
11 412.10236 -0.0020808 66.9794704 468.8564879
12 412.10456 0.0066282 25.5925725 179.1499847
13 412.10595 -0.0114676 4.1176669 28.8295864
14 412.10859 -0.0527437 9.8348916 68.9694266
15 412.12522 -0.0054311 18.4996014 129.4985374
16 463.61442 0.0069102 89.3339891 625.3400725
17 463.63003 -0.0030733 109.7296067 768.1076720
18 463.63922 0.0007015 135.9888766 951.9221585
19 463.64577 -0.0027378 62.7335161 439.1349496
20 463.64789 -0.0022364 38.0224668 266.1574928
21 463.66001 -0.0006996 79.5765437 557.0358282
22 515.53684 -0.0005787 21.8868936 153.2082706
23 515.54920 0.0390079 30.0167765 210.1859082
24 515.55055 -0.0580559 19.9332233 139.6842350
An intense peak is expected at 520 cm-1 from experiment and calculation from VASP-phonon calculation wiht PAW@PBE, dr = 0.01, 4x4x4 kpts shows triple degenerate mode at 504 cm-1 has the largest intensity.
From experiment, Raman spectrum of the bulk silicon contains only one intense peak at around 520 cm-1 (J.H. Parker, et al., Phys Rev, 155, 712 (1967))
Shannon Stauffer (UT Austin, PI: G. Henkelman): Email.
Alexandr Fonari (Georgia Tech, PIs: J.-L. Bredas/V. Coropceanu): Email