ABINIT Discussion Forums

Hello,

I would like to ask to help me to understand my results. I run optimization of the unit cell and coordinates (Abinit 7.2.1)
of FCC solid containing 2 Na and 34 Si atoms in reduced FCC cell. Here are my results:

Run 1. optimize acell and coordinates starting with experimental geometry and use Abinit geometry builder. Works great.
Run 2. optimize acell and coordinates using output coordinates (in Angstroms) and acell from Run 1. As expected, the
results are the same as in Run 1. GREAT!
Run 3. scf run, using coordinates from Run 1 and optimized acell from Run 1. (two runs, using xred and xangst coordinates)
Result: there is a small difference in total energy between Runs 3 and 2 (or 1):

Run 1 Run 2 Run 3
Etotal (Ha) -1.34862119866008E+02 -1.34862119865606E+02 -1.34863954659665E+02
Pressure -8.5203E-03 -8.5337E-03 -4.3472E-01

Question: Does Abinit scale cell size by dilatmx for optimization runs? If so, k-points (in atomic units) will be slightly different
that for scf run, what would explain the difference in the total energy and pressure. Am I correct?

Thank you,
Jan Gryko

Actually, abinit scales the set of planewaves (defined by ecut), so that it is sufficient even with the maximum dilatation.
Hence, if you use a bigger value of ecut, the difference between the runs should get smaller.

Hello Gabriel,

Thank you very much for your answer. Indeed, increasing ecut to 30 Ha (in my case) reduced the difference between opt and scf runs to very small
value.

Best,
JG