[QE-users] QE convergence problem for molecular crystal

[Si, Pradip]

Hello QE users,
I am trying to do relax followed by vc-relx calculation for a molecular crystal with QE. There are 13 polymorphic crystals. For 8 out of 13 structures, relax calculations converged with this input
&CONTROL
    !
    calculation = 'relax'
    prefix = 'polymorph'
    pseudo_dir = '/home/ps0579/SSSP'
    outdir = './out_ROY_R'
    nstep = 150
    !
/
&SYSTEM
    !
    ibrav = 0
    nat = 216
    ntyp = 5
    vdw_corr = 'grimme-d3'
    occupations = 'fixed'
    ecutrho = 480.0
    ecutwfc = 60.0
    !
/
&ELECTRONS
    !
    conv_thr = 1e-06
    electron_maxstep = 300
    diagonalization = 'cg'
    mixing_beta = 0.4D0
    !
/
&IONS
    ion_dynamics = 'bfgs'
    !
/
&CELL
/
K_POINTS automatic
1 1 1 0 0 0


But the convergence was not achieved for the other 5 structures and failed after a few BFGS steps, and SCF doesn't converge within 300 steps. What parameters should I change to converge? I tried increasing ecut and lowering mixing_beta and mixing_mode to local-TF, but still, it doesn't converge.

Best,
Pradip Si
University of North Texas

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20250522/bf5ffdb8/attachment.html>