repository·master·Indexed 40 days ago
https://github.com/abelcarreras/dynaphopyA tool for calculating crystal microscopic anharmonic properties, such as phonon frequency shifts and linewidths, from molecular dynamics data. It utilizes normal-mode-decomposition and quasiparticle theory, providing interfaces for MD outputs from VASP and LAMMPS and integrating with PHONOPY for harmonic phonon modes. It includes capabilities for calculating renormalized force constants, thermal properties (free energy, entropy, heat capacity), and power spectra.
DynaPhoPy is software designed to calculate crystal microscopic anharmonic properties from molecular dynamics (MD) using the normal-mode-decomposition technique.
Key Capabilities:
To install from source, you must have a C compiler available. First, install the necessary dependencies from requirements.txt, then run the setup.py script.
Requirements:
pip install -r requirements.txt --user
python setup.py install --userYou can install DynaPhoPy directly from the PyPI repository using pip. Ensure you have Python 3.5+ and Phonopy 2.0+ installed beforehand.
pip install dynaphopy --userTo calculate anharmonic properties of a material (like GaN) using the LAMMPS interface, you must have the LAMMPS Python interface installed on your system. The workflow involves three main steps:
phonoLAMMPS to obtain these constants.--run_lammps flag: Execute the calculation by providing the input script and MD parameters.Important Notes:
dynaphopy is executed. To avoid redundant calculations, use option flags or interactive mode to calculate multiple properties in a single run.--save_data option to store it in HDF5 format.dynaphopy input_dynaphopy --run_lammps in.lammps 50 0.001 20 --dim 6 6 6 -iThe Quasiparticle class provides a high-level interface for analyzing molecular dynamics (MD) trajectories to extract phonon-related properties, including power spectra, renormalized force constants, and thermal properties.
Key analysis workflows include:
get_power_spectrum_full() to get the full power spectrum, or get_power_spectrum_phonon() for projections onto specific phonon modes.get_renormalized_force_constants(), which relies on get_commensurate_points_data() to identify relevant Q-points.get_thermal_properties() or display_thermal_properties().plot_* methods.The Parameters class is a Singleton used to manage all default parameters for DynaPhoPy. Because it implements the Singleton pattern, any instance of Parameters created will return the same object, ensuring a single source of truth for configuration across your application.
You can initialize it with custom values or update existing values using its properties or by passing a dictionary via get_data_from_dict.
from dynaphopy.parameters import Parameters
# Initialize with specific settings
params = Parameters(silent=True, power_spectra_algorithm=2)
# Since it is a Singleton, subsequent calls return the same instance
params_again = Parameters()
print(params is params_again) # True
# Update via properties
params.spectrum_resolution = 0.01
# Update multiple values from a dictionary
config_dict = {
"spectrum_resolution": 0.02,
"use_NAC": True
}
params.get_data_from_dict(config_dict)When performing long MD simulations, you can configure LAMMPS to store atomic velocities instead of positions (e.g., using an input file like in.md_vel).
Benefits:
This is the preferred approach for large-scale or long-duration MD trajectories.
To prepare a VASP MD simulation, use the dynaphopy input_si command with the -c_poscar flag to specify your unit cell POSCAR file and the --dim flag to define the supercell dimensions. This generates the necessary initial structure for the MD run.
dynaphopy input_si -c_poscar POSCAR --dim 2 2 2Once you have completed an MD simulation (e.g., using LAMMPS), you can compute anharmonic properties such as phonon frequency shifts, linewidths, renormalized force constants, and thermal properties using the dynaphopy input_gan command.
Key flags:
-ts <value>: Time step.-sdata: Output data.-thm: Compute thermal properties.--normalize_dos: Normalize the Density of States (DOS).-sfc <name>: Specify the output filename for renormalized force constants.Example command:
dynaphopy input_gan gan_2000_test.lammpstrj -ts 0.001 -sdata -thm --normalize_dos -sfc renormalized_fcThe Dynamics class is the central object for managing molecular dynamics data. It can be initialized with a structure, trajectory, velocities, energy, time, and supercell information.
If memmap=True is passed, the class will use memory-mapped files for large datasets, utilizing the directory specified by the DYNAPHOPY_TEMPDIR environment variable. If this environment variable is not set, it defaults to an empty directory.
Note: If you initialize without a structure, a warning will be issued.
from dynaphopy.dynamics import Dynamics
dynamics = Dynamics(
structure=my_structure,
trajectory=my_trajectory,
velocity=my_velocity,
time=my_time,
supercell=my_supercell,
memmap=False
)DynaPhoPy provides an interactive menu interface to view renormalized phonon dispersion and Density of States (DOS).
Use the -i flag to enter the interactive mode:
dynaphopy input_gan gan_2000_test.lammpstrj -ts 0.001 --normalize_dos -iTo prepare a LAMMPS simulation, you can generate the initial crystal structure from a POSCAR type file using the dynaphopy input_gan command. Use the -c_lammps flag to specify the output data file and --dim to define the supercell dimensions.
Example for a 3x3x3 supercell:
dynaphopy input_gan -c_lammps data.gan --dim 3 3 3