Creating Initial and Final POSCARs for NEB calculation>

[eeluno0694]

I am a new user trying to run transition state calculations using CI-NEB and the Dimer method. My fundamental problem is while trying to create the files for the images, how do I create the initial and final structures. I am working on CH3COOCH3 cleavage to CH3COO +CH3. I understand that the initial should be CH3COOCH3 while the final should be CH3COO+CH3. Do I use the optimized CH3COOCH3 structure as POSCAR1 and then that of CH3COO+CH3 (where the two are placed on their individual activation sites separately on the same slab) as POSCAR2 ? What I have been doing prior to now is making a little displacement of the bond length at the cleavage site (O-C bond) of two different intervals and using both for the initial and final structures.

[graeme]

This is a good question. In principle there are many possible product states from a given initial state. Typically, what you would do to study a reaction like to describe is to start by optimizing the CH3COOCH3 molecule above the surface. That optimized structure can be your initial state. Then you can use some chemical intuition to find favorable binding sites for CH3COO and CH3. Again, you would use the optimized structure, but it this case it would give you a possible final state.

It is also possible to search for a final state in other ways. The dimer is one approach, and another that you mention would be to manually break the C-O bond sufficiently so that the fragments relax into a candidate final state.

[eeluno0694]

If I understand you clearly, for the optimized CH3COOCH3 structure for the initial state, I am not to touch the C-O bond length. Then, for the final structure where I have found the favorable binding sites for CH3COO and CH3, does the distance between them matter?
For the dimer method, I understand I use 2 structures, POSCAR1 and 2 to create the MODECAR, then a new structure (POSCAR) to run the calculation. Again, how would I make these structures in this case?
Thank you.

[graeme]

Just keep it simple. Relax the molecule above the surface for the initial state and then the dissociated fragments on the surface for the final state. Then connect them with a band, check to make sure that it looks reasonable, and then run the NEB.