mat-md-probability-density

MD Probability Density Visualization

Goal

To visualize the spatial probability density of mobile ions (e.g., Li, Na) from an MD simulation trajectory. This helps in understanding conduction pathways and identifying preferred occupation sites within the crystal structure. The output is a volumetric data object in CHGCAR format, which can be easily visualized using VESTA.

Instructions

1. MD Simulation: Run an MD simulation at an appropriate temperature to observe sufficient diffusion events.

   Note: Short MD trajectories (e.g., ≤10 ps) often have too few discrete ion hops to naturally form continuous probability density tubes. The resulting density will look like isolated blobs exactly at the crystal lattice sites. To visualize continuous macroscopic diffusion pathways for short trajectories, use the --log compression flag to mathematically connect the sparse pathways.

   • The trajectory is typically saved to trajectory.traj.
   • Ensure supercell_min_length is reasonably large (>10 Å) to avoid finite-size artifacts in the density mapping.
   • Allow the simulation to run long enough so that the ions sample the entire available volume (e.g., 50-100 ps or more).

2. Calculate Probability Density: Use the provided script to extract the fractional coordinates of the targeted species over time and convert them into a spatial density grid.

   # Env: base-agent
   python .agents/skills/mat-md-probability-density/scripts/calculate_probability_density.py \
       results/md_600K/trajectory.traj \
       --species Li \
       --interval 0.2 \
       --ignore_ps 5.0 \
       --output_chgcar results/md_600K/CHGCAR_proba
   

   • --species: The specific diffusing ion to visualize.
   • --interval: Grid spacing in Angstroms (0.1 to 0.5 is recommended). Smaller values give smoother isosurfaces but take longer to process and generate larger files. Defaults to 0.2 Å.
   • --ignore_ps: The equilibration time to discard from the beginning of the trajectory.
   • The script will automatically detect the frame time step if a .log file is available alongside the .traj file.