This thread is locked.Only browsing is available.
Top Page > Browsing
Periodic boundary conditions
Date: 2007/10/24 23:10
Name: Florian Rovere   <>

Dear Prof. Ozaki,

I use OpenMX for calculations of metal-nitride systems.
I realized that although the code works reasonably well for fcc structures, I always end up in troubles when it comes to hcp structures.
Somehow the potential curve always looks messy when I am trying to relax either a or c , in particular around the expected minimum values.
It came to my attention that the relaxation seems to move atoms out of the unit cell every now and then. If this is the reason for my problem (breaking of symmetry), how can I avoid that?
I am using EF geometric optimization.

Best regards,
Florian Rovere
Page: [1]

Re: Periodic boundary conditions ( No.1 )
Date: 2007/10/31 02:20
Name: Florian Rovere  <>

Hi everybody,

the problem is solved and it turns out that the solution was well hidden in the manual. "A tip for calculating the energy curve for bulks" deals with energy discontinuities due to sudden changes of the number of real space grids. As a matter of fact, I have seen such discontiniuties while changing lattice parameters a and c (hex). As suggested in the manual an explicit specification of the number of real space grids in a, b, c -direction via "scf.Ngrid #a #b #c" solves the problem.

Florian Rovere

Page: [1]