Hello, everyone,
I am trying to determine the melting point of pure Al using a technique called the convergence method. A simulation box is created with a distinct solid and liquid region. Then, an NPT fix at a temperature that is predicted to be the melting point is applied to the entire system. An NPH fix is then applied to the entire system, and the converging temperature is found at each step of the NPH fix. The distinct solid and liquid regions are supposed to remain intact throughout the NPH fix, but my solid region is partly melting.
Does anyone have experience with this? I cannot upload the input script as I’m a new user, but I can send it to you another way if you could take a look at it.
Regards,
Tony
Do you have a reference for that?
What you describe sound awfully similar to what I know as the “coexistence method”.
I am confused by you mentioning NPT first but then only referring to NPH.
If your solid region partly melting, that would indicate that your assumed melting temperature is too high. In the coexistence method, the temperature of the ensemble is modified empirically (optimally, one would use some kind of bisection method to be efficient) until a temperature is found at which the solid and liquid regions maintain their state.
If the use of the NPH ensemble is indeed correct, then I would assume that you would have to monitor the total temperature of the system and that by melting some of the solid some kinetic energy is converted to potential energy and thus the system temperature drops and then this (averaged) temperature at which no further melting occurs would be the melting temperature. But this is my conjecture. You need to check the reference describing the method for details and whether my guess is correct.
Thank you for getting back to me so quickly. People in my research group described this method to me, and the guy who worked on this project before me used the Coexistance and Convergence methods. It’s described in this article in the second paragraph of the ‘Simulation Methodology’ section.
https://www.sciencedirect.com/science/article/pii/S0927025622006127
For my LAMMPS scripts, I create the solid and liquid regions, apply an NPT fix to the entire system at a temperature that I think will be the melting point, and then use the NPH.
It would make sense that my assumed melting temperature is too high. The experimental melting point of Al is 933 K, and my initial guess temperature is 1000 K. I will try lowering the temperature of the NPT fix that I apply to the entire system before applying the NPH. We monitor the temperature vs time step of the NPH fix so that I will keep an eye on this for future simulations. Thank you, Dr. Kohlmeyer!
Regards,
Tony