Hi there, I'm trying to simulate a system containing sodium and carbon atoms. Since there is no Na-C potential on NIST website, I'm thinking using LJ potential. But where can I find the LJ potential parameters like Sigma and Epsilon? Is there any database or can I calculate these parameters through DFT? Thank you!
The interaction of Na-C is definitely not van der Waals-dominated, therefore a simulation with pure LJ will not provide results with any helpful value.
Of course the best option would be to perform DFT calculations (periodic, if your system is a crystalline solid). You can check out what software fits your needs here:
https://en.wikipedia.org/wiki/List_of_quantum_chemistry_and_solid-state_physics_software
If you don't have the computational capacities for DFT, you can use semiempiric methods such as PM7 as implemented in MOPAC. The older version 7 is still available as open-source, newer versions are commercialized:
https://sourceforge.net/projects/mopac7/
Jürgen Weippert Thank you very much sir. What I want is to simulate one sodium atom penetrate through a cubic-like structure(which containing thousands of carbon atoms). I think DFT can't simulate such big system, and I don't want any reaction between sodium and carbon. I'm thinking either using LJ potential or using Li-S potential(which is available on NIST potential website and Li-S is the most similar system compared with Na-C )
By "penetrate", do you mean a Na atom diffusing through solid carbon or a high-energy Na atom from an effusion/ion source going through with high velocity?
In any case, LJ does not describe these interactions correctly. A former colleague of mine used the aforementioned Mopac with PM7 to simulate carbon atoms or molecular fragments crashing into a surface, so if you have the beam the high-energy scenario you could check out Mopac.
Edit: I just had a look in the Mopac manual, http://openmopac.net/manual/
The keywords you can check out are "DRC" and "velocity".
Jürgen Weippert Thank you for your advice, I'm not familiar with MOPAC, I'll try to learn it. It would be better to realize this simulation in LAMMPS.
By "penetrate", for example in LAMMPS, I use displace_atom command to do that. As can be seen in the graph attached, one sodium atom will go through sp3 hybridization carbon atoms.
Well, above you said you don't want reactions to happen - a change of hybridization IS absolutely a reaction. In that case using a Lennard-Jones potential would be even more useless since it does not do anything about that.
Any other force field method won't help, either, because in these you usually fix the bonding situation and bond formation/breakage and hybridization changes cannot be accounted for.
Semiempiric methods are the minimum level of theory for a meaningful statement here and to my extent of knowledge LAMMPS wasn't made for that.
Edit: one more thing from a more chemical perspective: carbon going from sp3 to sp2 is an oxidation process, but Na is a reducing agent, so why do you think that the reaction will go this way? Essentially what you are simulating there is an implantation process, see e.g.
https://www.sciencedirect.com/science/article/abs/pii/092596359390236U
or
https://apps.dtic.mil/dtic/tr/fulltext/u2/a243060.pdf
Jürgen Weippert Thank you very much for your kind help! I'll reconsider the whole simulation.
- Badges
- Science topic