cmgtam
commented
11 years ago - Relax perfect cell e.g. 0 0 Fe 0 0.3 Fe 0 0.6 Cr 0.3 0 Fe 0.3 0.3 Ni 0.3 0.6 Fe 0.6 0 Fe 0.6 0.3 Fe 0.6 0.6 Cr
- In input file, have vacancy groups, e.g. D1 and D2 D1 V 0 0 Fe A 0.3 0 Cr A 0.6 0 Ni D2 V 0.6 0 Fe A 0 0 Cr A 0.3 0 Ni
So these defect groups consist of two antisites and one vacancy. The two antisites both slide over at the same time into the vacancy position.
The neb section of the input file has a subsection for each neb neb_1: 0.3 0 Cr 0 0 Cr 0.6 0 Ni 0.3 0 Ni to show how these atoms are going to move.
- Relax perfect bulk
- Apply defect groupings D1 and D2. Regroup the POSCAR by elements, with vacancy lines removed and interstitial lines appended to the bottom of their respective element groups.
- Relax the defects.
- For each defect grouping in the NEB (one grouping for initial state, one grouping for final state), match the relaxed coordinates (by tolerance matching) to the NEB lines specified for that NEB hop. Remove these lines.
- Reorder the rest of the lines, grouping by elements. Reinsert the removed atoms at the bottom of their element groups, one pair at a time.
- Regular NEB between the two POSCAR files.
Is POSCAR being interpolated correctly? Are atoms keeping their positions? How can kickout or crowdion movements be handled? Need to be able to specify which atoms are moving to where? Simple reordering/nearest-neighbor is not good enough. Also dumbbell movement.