ABINIT Discussion Forums

Has anyone had success with Abinit for molecular systems? I am interested in calculating properties of crystals of small molecules such as 2-methyl-4-nitroaniline (19 atoms, 38 atoms/primitive unit cell). So far, however, I have not been able to get reasonable geometry optimization. The energy and stresses just oscillate and the forces remain above 50 Ha/bohr after 25 Broyd steps, and when I allow the lattice parameter to change, it tries to expand by an order of magnitude. I have tried the default LDA functional with the GTH pseudopotentials and also GGA-PBE with Van der Walls DFT-D2 correction and FHI pseudopotentials.

I will continue trying different parameters and functionals, but I was wondering whether anyone had done anything like this before. A similar problem was reported in this thread: viewtopic.php?f=8&t=2790 so I am starting wonder whether it is even possible with Abinit.


Edit: It was just a mistake. My initial atom coordinates were too close for it to optimize the geometry.