Dear all;
I have been trying to relax Tenorite (CuO) with monoclinic structure. During relaxation, the cell parameters change significantly and the final result is not consistent with the literature. (Please see TABLE II in this paper: https://journals.aps.org/prb/abstract/10.1103/PhysRevB.99.035154) I am using the same parameters for my calculations as in the paper; here is my input:
&CONTROL
calculation='vc-relax',
outdir='.',
prefix='pw-paperlike',
pseudo_dir='/home/ma36813/all_pbe_UPF_v1.5',
verbosity='high',
etot_conv_thr=1d-05
/
&SYSTEM
ibrav=0,
celldm(1)=5.4811209354d0,
nat=8,
ntyp=3,
ecutwfc=50,
ecutrho=500,
occupations='smearing',
smearing='mv',
degauss=0.005d0,
nspin=2,
starting_magnetization(1)=0.5,
starting_magnetization(2)=-0.5,
vdw_corr='grimme-d3',
/
&ELECTRONS
conv_thr=1d-08,
mixing_beta=0.7d0,
/
&IONS
/
&CELL
/
ATOMIC_SPECIES
Cu1 63.546000d0 cu_pbe_v1.2.uspp.F.UPF
Cu2 63.546000d0 cu_pbe_v1.2.uspp.F.UPF
O 15.999400d0 o_pbe_v1.2.uspp.F.UPF
ATOMIC_POSITIONS {crystal}
Cu1 0.0000000000d0 0.2500000000d0 0.0000000000d0
Cu2 0.0000000000d0 0.7500000000d0 0.0000000000d0
Cu2 0.5000000000d0 0.0000000000d0 0.5000000000d0
Cu1 0.5000000000d0 0.5000000000d0 0.5000000000d0
O -0.4184000000d0 0.2092000000d0 0.2500000000d0
O -0.4184000000d0 0.7092000000d0 0.2500000000d0
O 0.4184000000d0 -0.2092000000d0 0.7500000000d0
O 0.4184000000d0 0.2908000000d0 0.7500000000d0
K_POINTS {automatic}
5 5 5 1 1 1
CELL_PARAMETERS {alat}
0.807399649995d0 -0.590004919630d0 0.000000000000d0
1.614799299990d0 1.180009839261d0 0.000000000000d0
-0.293063818044d0 0.000000000000d0 1.743801894963d0
*Please note that CuO is antiferromagnetic so it requires to have a supercell twice as large as the unit cell.
My cell parameters are:
•a=4.2949 Angstrom • b=4.0316 Angstrom •c=5.1650 Angstrom •β= 89.744 degrees And the other two angles, instead of being 90, are 89.947, 88.396. I appreciate any comments or suggestions.
1) One thing is to use 'ibrav' and corresponding 'celldm' for monoclinic structures specifically.
2) Also, you may want to keep one of the axis fixed. I had worked on CuO experimentally and feel that it's stability in that particular valence state is difficult to achieve (if I recall right).
In such cases, try to keep an axis or two fixed using cell_do_free.
3) Try to choose a pseudo-potential that is more relevant.
4) In ion_dynamics, try to set different parameters.
Dear Eesha,
Thank you for your comment.
I figured that the results reported in that paper are incorrect and they actually did not do a "full relaxation" whereas they merely moved the ions within a fixed unit cell.
The takeaway message is that PBE cannot predict CuO structural and electronic properties.
- Badges
- Science topic
- Similar topics
- Mathematics