Solidification problem (Phase change is not observed)

Science Talk
1

Dear LAMMPS Experts and users

I want to simulate a system that contains a ternary eutectic salt mixture with a composition of 46.5% LiF + 11.5% NaF + 42% KF (FLiNaK).
At first, for my initial configuration, I follow the instructions below article,
https://pubs.acs.org/doi/10.1021/acs.jpcb.2c06915?ref=pdf

As they suggested, to reach a ternary mixture, LiF, NaF, and KF crystals should be placed beside each other in a simulation box and heated under NVT up to a much higher temperature (3000K). After reaching a fine mixed and melt system, it has to be cooled down to the desired melt temperatures via NPT. Later the system must be relaxed and then the properties can be calculated.

I did these steps one by one and fortunately, I could regenerate the closed results to validate/verify my simulations.

But after that, I encountered a big problem. This cooling and equilibrating just worked for melt temperatures and when I cooled it below the melting point I couldn’t capture any phase changes (e.g. Jump in energy, enthalpy, density, etc.) and reach a solid phase in solid temperatures. I observed something like supercooling or a lake of nucleation.

I have read the forums, similar articles, etc, and done lots of chances according to recommendations (increase/decrease cooling rate, cool to a much lower temperature, increase/decrease equilibration time, …) but none of them worked.

Besides I have some other problems which I should explain:
⁃ As I explained I have to heat my system first to reach my mixture so my well-mixed and equilibrated system will occur in the cooling process. I need to capture solidification and phase change in the cooling process and I can’t capture only melting and phase change during the heating process due to completing the mixing process.
⁃ I need to equilibrate and calculate several temperatures in both solid and liquid phases near melting point so I can’t use a single temperature coexistence method which is commonly suggested to overcome this problem.

Maybe this question is not exactly a LAMMPS technical/coding problem but of course, many experts and experienced people are in this forum therefore sharing any experience, advice, and suggestions would be greatly appreciated.

Best regards
Amir Abdi

2

It is off-topic for this forum, and while there are people here that are experts in running and developing LAMMPS, the chances to come across somebody that is familiar with your specific area of research are rather slim. Please keep in mind that while there are quite a large number of subscribers on MatSci.org, rather few of those respond to questions.

The most suitable MatSci category would be Science Talk, but that is not quite as popular as LAMMPS. :wink: Nevertheless, I kindly ask you to change the category because we don’t want to have your post be used as an indication that it is OK to ask off-topic questions. This category is about LAMMPS and not an “ask-me-anything”. There are other forums for scientific questions, too.

So you need to find people that have experience in your area of research. That is usually best achieved by having your adviser/supervisor reach out to potential collaborators with the required expertise or by you reaching out yourself.

That said, your description is far too vague to even speculate about what is the cause of your difficulties.

4

simply, it is an MD simulation via LAMMPS in which solidification did not occur during the cooling process. I requested help to capture this phase change from many experts here who of course may have experienced a similar problem during any phase change simulations via LAMMPS. it is both a science and technical question.

even you had much advice in this forum already about phase change simulation for example: … it is an active process … or … a finite size effect is expected in this type of simulation… and coexistence simulation for that is suggested here many times by yourself before.

now I don’t want a scientific guide about the specific area of research. I need a technical guide/advice/… to better capture phase change during cooling via running LAMMPS.

5

The fact that you are using LAMMPS doesn’t make it a LAMMPS problem. Choice like simulation settings and force field details or simulation workflow choices can be implemented with many other software packages as well and the right choices do require domain specific knowledge. And you won’t likely get any advice while you keep being vague about the details of your simulation.