Hi everyone,

I am trying to simulate the hydrogen desorption process in MgH₂ using LAMMPS (replicating experimental T-P conditions). Here's what I currently do:

  • Initial Configuration: I use a supercell structure of MgH₂ with periodic boundary conditions applied along all three axes (x, y, and z).
  • Potential: I use an ADP potential, which I have verified works well for Mg-Mg, Mg-H, and H-H interactions.
  • Simulation Steps:
    • Energy minimization.
    • NVT heating (up to 700K) in several phases to apply temperature gradually, ensuring system stability.
    • NPT equilibration (pressure up to 0.1MPa), divided into multiple phases.
    • Production run at the same constant temperature and pressure for 1 ns.

    PROBLEMS:

    - The structure distorts significantly during the simulation. The volume expands so much that Mg and H atoms no longer interact properly with their neighbors.

    - Due to the periodic boundary conditions, I cannot track desorbed atoms properly; atoms leaving the simulation cell reappear in the equivalent periodic image, keeping the total number of atoms constant.

    QUESTIONS:

    - Are my steps reasonable for simulating hydrogen desorption in MgH₂?

    - How can I address the structural distortions and excessive volume expansion?

    - What is the best way to handle desorption when using non-periodic boundaries to track desorbed atoms?

    Any advice or general guidelines for setting up a reliable desorption simulation would be greatly appreciated!

    More Eloy Rodríguez's questions See All
    Similar questions and discussions