4.6. Advanced functionalities¶

This example is about more advanced functionalities, highlighting the possible exploration methods.

import numpy as np

import eminus
from eminus import Atoms, SCF
from eminus.dft import get_psi
from eminus.localizer import wannier_cost
from eminus.tools import center_of_mass, check_orthonorm, get_dipole, get_ip
from eminus.units import ebohr2d, ha2kcalmol

Start with a simple DFT calculation for neon

If one needs cell parameters from the Atoms object one can use the atoms.build function to generate them

eminus also supports GGA functionals like the Chachiyo GGA

A different cg-form can be used if the system is hard to converge

atoms = Atoms("Ne", [0, 0, 0], ecut=10, center=True)
scf = SCF(atoms, xc="chachiyo")
scf.run(cgform=3)

Calculate the dipole moment

Make sure that the cell is big enough, and that the density does not extend over the borders

Centering the system is recommended to achieve this

dip = get_dipole(scf)
print(f"\nDipole moments = {dip} a0")
print(f"Total dipole moment = {ebohr2d(np.linalg.norm(dip))} D")

Calculate ionization potentials

ip = get_ip(scf)
print(f"\nIonization potential = {ha2kcalmol(ip)} kcal/mol\n")

Transform the orbitals to real-space to get the Kohn-Sham orbitals

Make sure to use orthogonal wave functions to generate them

psi = atoms.I(get_psi(scf, scf.W))

Some functions are controlled with a global logging level that can be changed with

eminus.config.verbose = 3

Check orthonormality of Kohn-Sham orbitals

print("Orthonormality of Kohn-Sham orbitals:")
check_orthonorm(atoms, psi)

Calculate the orbital variance and spread of the orbitals

cost = wannier_cost(atoms, psi)
print(f"\nOrbital variances = {cost} a0^2")
print(f"Total spread = {np.sum(np.sqrt(cost))} a0")

Calculate the center of mass of the density

com = center_of_mass(atoms.r, scf.n)
print(f"\nDensity center of mass = {com} a0")
print(f"Neon position = {atoms.pos[0]} a0")

Write all orbitals to CUBE files

print("\nWrite cube files:")
for i in range(atoms.occ.Nstate):
    print(f"{i + 1} of {atoms.occ.Nstate}")
    atoms.write(f"Ne_{i + 1}.cube", psi[0][0, :, i])

Another useful setting is the number of threads

print(f"\nThreads: {eminus.config.threads}\n")

You can also set them and check the configuration afterwards

eminus.config.threads = 2
eminus.config.info()

eminus uses the NumPy and SciPy packages

These packages will listen to the OMP_NUM_THREADS and/or MKL_NUM_THREADS flags

Setting these flags will control how many threads eminus uses.

Download 06_advanced_functionalities.py