ABINIT Discussion Forums

Hello All

I want to know what is the general procedure for relaxing and optimizing structure? I followed two below procedures for a specific structure and got two different answers:
1- Firstly,I relaxed the atoms by optcell=0 and ionmov=3 and then proceed to optimize lattice constant by optcell=1 and ionmov=3
while I insert the atom positions resulted from the first calculation in the second one.
2- I applied optcell=0 and ionmov=3 for several lattice constants (multidataset )and finally fitted the energy versus acells to Birch-Murnaghan and extracted optimized lattice constant.

Regards, Payam

Hard to tell without more indications on your system

1) are they very different? Your precision is limited by the tolerances you set on the forces and stresses etc... as well as the way you fit the BM equation. In particular for the latter, if you do not have enough points near the equilibrium volume (or have an anisotropic system see below) you may have a decent BM curve without much precision on the location of the minimum in energy.

2) optcell 1 allows only volume modifications, with "homogeneous dilation" of the 3 lattice constants. If you have a non-cubic system this will not give you the overall minimum - use optcell 2

In principle the minima structures should agree unless you are constraining the searches in some way.

Matthieu