[QE-users] Imaginary frequencies away from Gamma point SiO

[Chiara Cignarella]

Dear Cristian,


are you sure about your structure? An imaginary phonon along your BZ means an unstable structure, not at the minimum of the potential energy surface.

If it is not a "real" physical instability, it may be due to not enough relaxation of the forces. Are the 1e-7 threshold you are setting for energy or forces? You should make sure your forces are well relaxed as well.

I see that you are doing restart on your phonons. Sometimes this can also create problems, in case you are not restarting correctly the calculation (for example, the -200 cm^-1 is at the q-point exactly after your restart? Does the restart output give you some warning?)



Hope this helps


Best regards

Chiara Cignarella

PhD EPFL - THEOS

Switzerland


________________________________
Da: users <users-bounces at lists.quantum-espresso.org> per conto di Cristian Mauricio Rebolledo Espinoza via users <users at lists.quantum-espresso.org>
Inviato: domenica 28 gennaio 2024 23:35:55
A: users at lists.quantum-espresso.org
Oggetto: [QE-users] Imaginary frequencies away from Gamma point SiO


Dear QE users,


I've been trying to obtain the phonon dispersion for O passivated Silicon sheets, and while the relaxation appears correct, and the acoustic modes also look correct (~ -1 cm-1) at the gamma point , there is always a single mode with negative frequency away of the gamma point (0.5 0.5 0) with value reaching ~ -200 cm-1.

If I replace the oxygen with 2 H atoms the phonons behave well. Using 2 -OH instead of 1O also results in the same issue.
I have tried:
Relaxations from 1e-4 to 1e-7
More k-mesh density (up to 12x12x1)
Higher cutoffs: up to 150
DIfferent pseudos (currently using Dojo)

Here is one of the relaxation inputs:

 &control
    calculation='relax'
    prefix='SiO_b_4',
    wf_collect = .true.,
    pseudo_dir = './pseudos/dojo/',
    outdir='./'
    etot_conv_thr = 1e-6
    forc_conv_thr = 1e-6
    restart_mode='restart'
 /
&SYSTEM
  ibrav = 0
  ecutwfc = 120.0
  nat = 7
  ntyp = 3
  assume_isolated = '2D'
  nosym = .true.
/
 &electrons
  conv_thr = 1.0d-10
  mixing_mode = 'plain'
  mixing_beta = 0.5
  diagonalization = 'david'
  diago_full_acc = .true.
/
&ions
ion_dynamics = 'bfgs'
/
&CELL
cell_dofree = '2Dxy'
/
CELL_PARAMETERS {angstrom}
  3.8 0.0 0.0
  0.0 3.8 0.0
0.0 0.0 24.0
ATOMIC_SPECIES
  Si   28.08500  Si.upf
   H    1.00750  H.upf
   O 15.999 O.upf
ATOMIC_POSITIONS {angstrom}
Si            0.0016742746       -0.0713893810        8.1568435063
Si            0.3113041995        1.8307404314        9.4771562615
Si            2.6539898849        3.7328987353       12.0865299573
Si            2.1985890646        1.8307732260       10.8584361966
H            -1.4869821232       -0.0709963685        7.9641453095
H             0.5567137638       -0.0718090357        6.7674928212
O             4.1040946046        3.7328817780       12.7505655928
K_POINTS automatic
12 12 1 0 0 0


And the phonon input:

&inputph
  tr2_ph=1.0d-18
  alpha_mix(1)=0.45
  prefix='SiO_b_4'
  ldisp=.true.
  epsil=.true.
  lqdir=.true.
  recover = .true.
  outdir='./'
  fildyn  = 'SiO_b_4.dyn.xml'
  fildvscf = 'dvscf'
  asr=.true.
  search_sym=.false.
  nq1=6, nq2=6, nq3=1,
/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20240129/5b18b4c9/attachment.html>