With the help of GULP software, I have fitted the Buckingham parameters for Tl atom splitting it into core-shell for Tl2TeO3, Tl2Te3O7 & Tl2Te2O5 crystalline system. I already have working core-shell potential for Te & O atom. With the obtained potential for Tl I am able to run MD for all the above three structure at 300K while it crashes at 600K and could not even run MD at 300K for amorphous system.
I tried turning off core-shell just for Tl, in that case, MD works perfectly for amorphous system at 300K to 2000K.
What can be possible reason for crash after adding core-shell parameters of Tl?
Hi Raghvender Pandey
I have done some work using a Buckingham potential coupled with a core-shell model and occasionally had similar problems. In our case, we identified that the core and shell would essentially 'separate' if the fitted spring potential was 'soft'. This would cause the code to crash (md was performed using DLPOLY in our case). If this turns out to be the case I have some recommendations:
In the first instance you could manually constrain your spring potential such that it 'harder' (has a larger gradient) and then refit the other potential coefficients. If you are lucky this will be enough.
Alternatively, what we found to work well was to add a strong penalty term in the model Lagrangian if the core and shell become largely separated. Eg
U_{penalty}(r)=a*(r-r_c)^2 ; if r>=r_c. a>0
U_{penalty)(r)=0 ; if r
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics
- Atoms