[QE-users] Phonon calculations, is it possible to calculater lraman (Raman) without calculating the the repentation Self-consistent Calculation again, it is a lot of cost.

ludwigboltzmann.sc10 at nycu.edu.tw ludwigboltzmann.sc10 at nycu.edu.tw
Fri Oct 27 15:42:59 CEST 2023


 Dear users:

I am going to calculate do phonon calculation and Raman spectra calculation of my perovskite(only at Gamma point), 
My procedure is like this: doint: vc-relax & scf→phonon calculation→dynmat
#scf &control    calculation = 'scf'    restart_mode='from_scratch'    prefix = 'perovskite'                               !test100tri    pseudo_dir = "pseudo"    outdir="output"/&system    ibrav=14,                   !triclinic    celldm(1)=16.13770350000000,celldm(2)=0.99156550541343,celldm(3)=1.87963300392941,celldm(4)=-1.94983007225937e-03,celldm(5)=0.13312393987385,celldm(6)=5.68819706737211e-04    nat=94, ntyp=5,    ecutwfc=150    ecutrho=600    vdw_corr=grimme-d2/&electrons    conv_thr=1e-14/ATOMIC_SPECIESSn    118.71000  Sn_ONCV_PBE-1.2.upfH       1.00784  H_ONCV_PBE-1.2.upfC      12.011    C_ONCV_PBE-1.2.upfI     126.90447  I_ONCV_PBE-1.2.upfN      14.0067   N_ONCV_PBE-1.2.upfATOMIC_POSITIONS (crystal)I             0.8103731347        0.6896489973        0.4992596099I             0.1896268653        0.3103510028        0.5007403901C             0.4906900199        1.0018135561        0.8416813990...
K_POINTS (automatic)  5 5 1 1 1 1
---
#phononphonon calculation for diamondg(Only gamma pt).  !use QEsub_ph (ph.x) (rather than pw.x)&inputph  outdir="output"  prefix = ' perovskite',  tr2_ph = 1.0d-14               !Threshold for self-consistency  ldisp = .false  asr = .true  zue = .true  epsil = .true  trans = .true         fildyn = ' perovskite.dyn'  start_q = 1  last_q = 1  "lraman = .true."  recover = .true !start from old/0 0 0                            !gamma pt



#dynmat&input  fildyn = 'perovskite.dyn' !input need to be read  filout = 'PEA2SnI4tri551150d2.out' !output phonon frequencies and norallized phonon displacements(normalized but not orthogonal)  filmol = 'PEA2SnI4tri551150d2.mold'!output file for molden  filxsf = 'PEA2SnI4tri551150d2.axsf'!output file for xcrysden   lperm = .true.  asr = 'crystal'


# mode   [cm-1]    [THz]      IR    1    -39.24   -1.1765    0.0432    2    -27.86   -0.8351    0.0000    3    -23.86   -0.7154    0.0000    4    -15.82   -0.4743    0.0000    5    -12.12   -0.3632    0.0511    6     -0.00   -0.0000    0.0000    7     -0.00   -0.0000    0.0000    8      0.00    0.0000    0.0000    9     13.64    0.4090    0.0000   10     16.10    0.4828    0.0209   11     22.89    0.6863    0.0398   12     29.50    0.8844    0.0000   13     33.06    0.9911    0.0928   14     33.58    1.0068    0.0000

/
I made a mistake without adding  "lraman = .true.", so I got dynmat.out with only IR but without Raman activities. 
Is it possible to calculater lraman (Raman) without calculating the repentation Self-consistent Calculation  again, it is a lot of cost.

Representation #   1 mode #   1
     Self-consistent Calculation
      iter #   1 total cpu time : 21284.5 secs   av.it.:   5.0      thresh= 1.000E-02 alpha_mix =  0.700 |ddv_scf|^2 =  6.481E-09
      iter #   2 total cpu time : 21666.4 secs   av.it.:  17.0      thresh= 8.051E-06 alpha_mix =  0.700 |ddv_scf|^2 =  1.068E-08
      iter #   3 total cpu time : 22023.4 secs   av.it.:  16.0      thresh= 1.033E-05 alpha_mix =  0.700 |ddv_scf|^2 =  3.855E-10
      iter #   4 total cpu time : 22379.9 secs   av.it.:  16.0      thresh= 1.963E-06 alpha_mix =  0.700 |ddv_scf|^2 =  1.560E-11
      iter #   5 total cpu time : 22749.5 secs   av.it.:  16.9      thresh= 3.950E-07 alpha_mix =  0.700 |ddv_scf|^2 =  1.517E-12
      iter #   6 total cpu time : 23124.2 secs   av.it.:  17.0      thresh= 1.232E-07 alpha_mix =  0.700 |ddv_scf|^2 =  1.771E-13
      iter #   7 total cpu time : 23504.1 secs   av.it.:  17.0      thresh= 4.208E-08 alpha_mix =  0.700 |ddv_scf|^2 =  2.903E-14
      iter #   8 total cpu time : 23890.4 secs   av.it.:  17.9      thresh= 1.704E-08 alpha_mix =  0.700 |ddv_scf|^2 =  1.027E-14
      iter #   9 total cpu time : 24267.0 secs   av.it.:  17.0      thresh= 1.013E-08 alpha_mix =  0.700 |ddv_scf|^2 =  5.001E-15
     End of self-consistent calculation
     Convergence has been achieved 

Sincerely,
National Yang Ming Chiao Tung University
HY Lu

  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20231027/a1183d41/attachment.html>


More information about the users mailing list