EFT-ODF calculator
Orbital dependent friction (ODF) allows evaluating electronic friction tensors (EFTs) from electronic structure code, enabling evaluation of full EFTs. For more details, go to NQCDynamics-ODF.
Below are the instructions on how to initialize the ACEds (currently: ACEfriction) ODF calculator, to run molecular dynamics with electronic friction (MDEF) within NQCDynamics.jl using FrictionProviders.jl.
The following instructions will include Julia-based code.
We start with importing NQCDynamics.jl packages and PyCall which allows importing Python-based packages.
using NQCDynamics
using PyCall
using NQCModels
using Unitful
using UnitfulAtomic
using FrictionProviders
using NQCBase: NQCBase
using ACE
using ACEds: ac_matrixmodel
using ACEds.FrictionModels
using ACEds.FrictionModels: Gamma
using JuLIP
using JuLIP: set_positions!
using ASE
using JLD2
# Importing Python modules with PyCall
io = pyimport("ase.io")
Now, we specify the EFT units and indices, and paths to the model. We read the ASE atoms object and we convert it to JuLIP Atoms object.
eft_unit = u"ps^-1"
friction_ids = [length(ase_atoms)-1,length(ase_atoms)]
eft_model_path = "path/to/aceds/model/eft_ac.model"
atoms_path = "path/to/atoms.xyz"
ase_atoms = io.read(atoms_path)
ase_jl = ASE.ASEAtoms(ase_atoms)
julip_atoms = JuLIP.Atoms(ase_jl)
Next, we load the ACEds model and we create ODFriction object with FrictionProviders.jl.
eft_model_aceds = read_dict(load_dict(eft_model_path))
aceds_model = ACEdsODF(eft_model_aceds, Gamma, julip_atoms; friction_unit=eft_unit)
odf_model = ODFriction(aceds_model; friction_atoms=friction_ids)
Together with PES model, the above ODF object (ODFriction) can be then used for MDEF simulations (as documented in NQCDynamics-MDEF).