about eam/alloy potential function

Dear all,

I am investigating the phase transformation in Ni-Ti alloy. The simulation procedure is heating the simulated sample to 400 K and then quenching to 50 K. B2 structure (BCC) should transform to B19 structure ( Monoclinic ). Eam/alloy potential function is used. Structures are identified by compute RDF. But the problem is that after quenching no phase transformation and structural change is observed. what can be the problem? Is it from the potential function? I mean, can eam/alloy simulate the phase transformation and change in structure? Or the potential has been chosen correctly and the problem is probably because of other parts?

Your help would be really appreciated,

Phase transitions generally depend crucially upon interaction potential. you have to consider the interaction potential as accurately as possible. there may be other problems also. you can visualize the system and see whats going wrong. i personally believe that the problem is mostly with your interaction potential.


Hi Nazanin,

Although I’m not an expert on phase transformations, here are some checks you can do to test your potential:

  1. Construct and equilibrate (via the box/relax command) zero temperature, zero pressure configurations of both the BCC and B19 structures. Whichever has the lower potential energy is the preferred state at zero-PT equilibrium. According to the information you’ve given below, that should be B19. If it’s BCC, then the potential is to blame.
  2. Take those same configurations and bring them up to temperatures above and below the transition. Then, run NPT dynamics and evaluate the free energy of the two structures (not just potential energy). The preferred state at any temperature will have the lower free state. Some exploration of the temperature space should tell you whether a phase transition is even possible with the potential you’re using. FYI, free energy is not a trivial calculation but there are tools within LAMMPS to estimate it. Take a look.

Jon Zimmerman