[QE-users] Discrepancy in value of 'a' while convergence test

[Pooja Vyas]

Dear users,
I'm trying to perform a convergence test for CaSiO3 tetragonal phase.
Following is my input :
#!/bin/sh

for aa in 5.6131
do
for cc in 7.181 7.182 7.183 7.184 7.185
do
for k in 8
do
for ecut in 30
do
cat > ${aa}_${cc}_${k}_${ecut}.in << EOF
 &control
    calculation = 'scf',
    prefix = '${aa}_${cc}_${k}_${ecut}'
    tprnfor=.true.
    tstress=.true.
    outdir = '/home/user/PNV/CaSiO3_tetragonal/'
    pseudo_dir = '/home/user/PNV/CaSiO3_tetragonal/pseudo/'
 /
 &system
    ibrav = 7,
    A=$aa
    C=$cc
    nat = 10,
    ntyp = 3,
    ecutwfc = $ecut,
/
&electrons
    diagonalization='david'
    mixing_mode = 'plain'
    mixing_beta = 0.7
    conv_thr = 1.0d-12
/
ATOMIC_SPECIES

Ca 40.078 Ca.pz-spn-kjpaw_psl.1.0.0.UPF
Si 28.0855 Si.pz-n-kjpaw_psl.0.1.UPF
 O 15.999 O.pz-n-kjpaw_psl.0.1.UPF

ATOMIC_POSITIONS {crystal}
Ca 0.250 0.750 0.500
Ca 0.750 0.250 0.500
Si 0.000 0.000 0.000
Si 0.500 0.500 0.000
O 0.222 0.722 0.9441
O 0.250 0.250 0.000
O 0.278 0.222 0.500
O 0.722 0.778 0.500
O 0.750 0.750 0.000
O 0.778 0.278 0.0559

K_POINTS (automatic)
  $k $k $k 1 1 1

EOF
mpirun -np 40 -machinefile x2 /apps/codes/qe/6.5/bin/pw.x <
${aa}_${cc}_${k}_${ecut}.in > ${aa}_${cc}_${k}_${ecut}scf.out
done
done
done
done

To perform the test, I fixed c=7.182 angstrom and varied 'a'. The value of
'a' for which I attained zero pressure was 5.6131 angstrom, which is quite
different from other theoretical finding which is a=4.9812 angstrom (
https://sci-hub.se/https://doi.org/10.1107/s0108768106035762). The error in
'a' is about 10-12%, which is quite large.

Can I know why I am getting this much discrepancy in lattice constant? And
how can I get the correct value of 'a'?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210629/9ab313ff/attachment.html>