I'm generating amorphous silica using LAMMPS and I would like to decrease the computation time. I'm currently using a melt/quench process with stages heat, heat sustain, cool, and cool sustain. I've considered decreasing the run time of these stages as well as decreasing the cutoff range and using the radial distribution function to check the accuracy. I'm also unsure the minimum acceptable size of the silica sample where these checks would be effective. There are many different parameters I could change, so doing these checks (especially if I control for only one each time) would also take a very long time. I haven't found much previous work on this topic. Is there literature on the relation between these parameters (e.g. run time for the stages, cutoff range) and the accuracy (i.e. how well the result models amorphous silica)? What recommendations would you have for checking this efficiently? Thanks!
https://docs.quantumatk.com/tutorials/amorphous_structures/amorphous_structures.html
Hi Ephraim,
There are a few aspects here. First of all you have to consider the type of force field you are using. Then of course the size of the system. For example, in this work https://pubs.acs.org/doi/abs/10.1021/acs.langmuir.9b03755 we show that an amorphous structure with few hundreds of atoms can be comparable to a larger one. Lastly, since most of the running time during quenching, you can adjust the cooling rate, but this will likely affect how "amorphous" are the structures. In any case, I guess it depends on the model you would like to produce and the properties you need to measure. Good luck with your simulations!
Best regards, Julio.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders