Relaxation of large system with defect, using new Eigenvalue-Following(EF) method |
- Date: 2008/09/24 17:00
- Name: Heungsik Kim
<heungsikim@gmail.com>
- Dear everyone,
Currently I'm trying to relax hexagonal yttrium manganate supercell with oxygen defect.
The cell size is √3X√3X4, with c-lattice constant 23 angstrom, and 59 atoms within unit cell.
The problem is, when the MD step exceed 6, (the point that force starts to drop down)
the SCF calculation shows severe charge sloshing along c-direction.
The mixing scheme that I have used was Rmm-Diis.
There was no problem in the vacancy-free unit cell.
After the problem, I changed to Rmm-Diisk with Kerker factor 4 to reduce charge sloshing,
but it made worse the problem - charge sloshing problem occurs at the first MD step.
I used LDA+U, with U=6eV for the Mn d-orbitals, and 6*6*1 k-points.
If some of you already had such kind of problem, please comment your experience here.
Thanks in advance.
Heungsik.
| |
This thread is locked.Only browsing is available.
Re: Relaxation of large system with defect, using new Eigenvalue-Following(EF) method