[QE-users] Structure Relaxation on Charged System Smearing Convergence Question

[Tang, Weilun [M S E]]

Dear all Quantum Espresso users,

I have some questions regarding to structure relaxation on a charged system using smearing.

Objective:
The system I am interested in is a molecule in a cubic cell with large vacuum space. I would like to perform structure relaxation, then use the optimized atomic coordinate to perform self-consistent calculation and phonon calculation for Raman Spectrum using the flag Iraman = .true.. The system has 3- charge, so I used tot_charge = -3.

Problem:
In the structure relaxation calculation, I found that when I used the flag occupations = 'fixed', the energy didn't converged after 200 cycles. By contrast, when I used occupations = 'smearing', the energy converged without issues. I also found that occupations = 'fixed' and occupations = 'smearing' both work in the self-consistent calculation.

1. My first question is whether I need to use the same option of occupations (occupations = 'fixed' or occupations = 'smearing') for both structure relaxation and the self-consistent calculation.

2. My second question is that what kind of output in the output file would provide a hint on why the energy didn't converged in the structure relaxation calculation when using occupations = 'fixed'.

Below is my input file for relaxation:
------------------------------------------------------------------------------------------------------
&CONTROL
  calculation  = "relax",
  prefix       = "ttc",
  pseudo_dir   = "/home/weilunt/pseudopotential/norm_conserving/pz",
  outdir       = "./tmp",
/
&SYSTEM
  ibrav     = 1,
  celldm(1) =25.0,
  nat       = 9,
  ntyp      = 3,
  ecutwfc   = 120.D0,
  tot_charge = -3,
  occupations = 'fixed',
!  occupations = 'smearing',
!  degauss = 0.02,
/
&ELECTRONS
  conv_thr    = 1.D-8,
  mixing_beta = 0.7,
  electron_maxstep = 200,
/
&IONS
/
ATOMIC_SPECIES
  S 32.070  S.pz-hgh.UPF
  N 14.007  N.pz-hgh.UPF
  C 12.010  C.pz-hgh.UPF
ATOMIC_POSITIONS (angstrom)
S   0.0000   -1.2375    0.0000
S  -1.4289    1.2375    0.0000
S   1.4289    1.2375    0.0000
N   0.7145    0.0000    0.0000
N  -0.7145    0.0000    0.0000
N   0.0000    1.2375    0.0000
C   0.0000   -0.4125    0.0000
C  -0.7145    0.8250    0.0000
C   0.7145    0.8250    0.0000

K_POINTS (automatic)
1 1 1 0 0 0
----------------------------------------------------------------------------

3. I performed the self-consistent calcualtion using nbnd = 40 at the Γ point for both occupations = 'fixed' and occupations = 'smearing' and compared the band energy. I found that the band energy are quite different in these two cases. Are they supposed to have huge difference? Below is the output for these two cases:
----------------------------------------------------------------------------  occupations = 'fixed'
k = 0.0000 0.0000 0.0000 (189047 PWs)   bands (ev):

-21.3134  -18.2350  -18.2296  -14.3475  -14.3259  -13.4357  -10.5406  -10.5270
 -9.6984   -7.4329   -6.8305   -5.2388   -5.2188   -4.7928   -4.7839   -4.4779
 -2.2839   -2.2721   -1.7952   -1.7650   -1.7571   -1.4256   -1.3879   -1.3618
 -1.0115   -0.1342   -0.0938   -0.0671   -0.0246   -0.0065    0.3019    0.7790
  0.8220    0.8520    0.8662    0.8663    0.9333    0.9548    1.0616    1.2307

highest occupied, lowest unoccupied level (ev):   -1.3618   -1.0115
----------------------------------------------------------------------------
----------------------------------------------------------------------------   occupations = 'smearing'
k = 0.0000 0.0000 0.0000 (189047 PWs)   bands (ev):

-43.8773  -33.1465  -33.1425  -30.7223  -29.4208  -29.4170  -28.8619  -28.0248
-21.0563  -21.0516  -17.6879  -17.6845  -10.4974  -10.0052  -10.0041   -7.0735
 -7.0715   -5.6177   -5.6166   -5.1731   -4.8122   -3.1190   -1.6271   -1.2365
 -1.1549   -1.1531   -1.0476   -1.0205   -0.7347   -0.6855   -0.6690   -0.6576
 -0.6294   -0.3364   -0.0832    0.1631    0.1694    0.1843    0.2041    0.2461

the Fermi energy is   -1.2926 ev
----------------------------------------------------------------------------

Thank you very much for your help and time in advance,

Weilun Tang

Department of Materials Science and Engineering
Iowa State University

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20250929/5e7476f8/attachment.html>