Source code for ase.md.velocitydistribution

# VelocityDistributions.py -- set up a velocity distribution

"""Module for setting up velocity distributions such as Maxwell–Boltzmann.

Currently, only a few functions are defined, such as
MaxwellBoltzmannDistribution, which sets the momenta of a list of
atoms according to a Maxwell-Boltzmann distribution at a given
temperature.
"""
from typing import Optional

import numpy as np

from ase import Atoms, units
from ase.md.md import process_temperature
from ase.parallel import world

# define a ``zero'' temperature to avoid divisions by zero
eps_temp = 1e-12


class UnitError(Exception):
    """Exception raised when wrong units are specified"""


def force_temperature(atoms: Atoms, temperature: float, unit: str = "K"):
    """Force the (nuclear) temperature to a precise value.

    Parameters:
    atoms: ase.Atoms
        the structure
    temperature: float
        nuclear temperature to set
    unit: str
        'K' or 'eV' as unit for the temperature
    """

    if unit == "K":
        E_temp = temperature * units.kB
    elif unit == "eV":
        E_temp = temperature
    else:
        raise UnitError(f"'{unit}' is not supported, use 'K' or 'eV'.")

    if temperature > eps_temp:
        E_kin0 = atoms.get_kinetic_energy() / len(atoms) / 1.5
        gamma = E_temp / E_kin0
    else:
        gamma = 0.0
    atoms.set_momenta(atoms.get_momenta() * np.sqrt(gamma))


def _maxwellboltzmanndistribution(masses, temp, communicator=None, rng=None):
    """Return a Maxwell-Boltzmann distribution with a given temperature.

    Paremeters:

    masses: float
        The atomic masses.

    temp: float
        The temperature in electron volt.

    communicator: MPI communicator (optional)
        Communicator used to distribute an identical distribution to
        all tasks.  Set to 'serial' to disable communication (setting to None
        gives the default).  Default: ase.parallel.world

    rng: numpy RNG (optional)
        The random number generator.  Default: np.random

    Returns:

    A numpy array with Maxwell-Boltzmann distributed momenta.
    """
    if rng is None:
        rng = np.random
    if communicator is None:
        communicator = world
    xi = rng.standard_normal((len(masses), 3))
    if communicator != 'serial':
        communicator.broadcast(xi, 0)
    momenta = xi * np.sqrt(masses * temp)[:, np.newaxis]
    return momenta


[docs]def MaxwellBoltzmannDistribution( atoms: Atoms, temp: Optional[float] = None, *, temperature_K: Optional[float] = None, communicator=None, force_temp: bool = False, rng=None, ): """Set the atomic momenta to a Maxwell-Boltzmann distribution. Parameters: atoms: Atoms object The atoms. Their momenta will be modified. temp: float (deprecated) The temperature in eV. Deprecated, use temperature_K instead. temperature_K: float The temperature in Kelvin. communicator: MPI communicator (optional) Communicator used to distribute an identical distribution to all tasks. Set to 'serial' to disable communication. Leave as None to get the default: ase.parallel.world force_temp: bool (optional, default: False) If True, the random momenta are rescaled so the kinetic energy is exactly 3/2 N k T. This is a slight deviation from the correct Maxwell-Boltzmann distribution. rng: Numpy RNG (optional) Random number generator. Default: numpy.random If you would like to always get the identical distribution, you can supply a random seed like `rng=numpy.random.RandomState(seed)`, where seed is an integer. """ temp = units.kB * process_temperature(temp, temperature_K, 'eV') masses = atoms.get_masses() momenta = _maxwellboltzmanndistribution(masses, temp, communicator, rng) atoms.set_momenta(momenta) if force_temp: force_temperature(atoms, temperature=temp, unit="eV")
[docs]def Stationary(atoms: Atoms, preserve_temperature: bool = True): "Sets the center-of-mass momentum to zero." # Save initial temperature temp0 = atoms.get_temperature() p = atoms.get_momenta() p0 = np.sum(p, 0) # We should add a constant velocity, not momentum, to the atoms m = atoms.get_masses() mtot = np.sum(m) v0 = p0 / mtot p -= v0 * m[:, np.newaxis] atoms.set_momenta(p) if preserve_temperature: force_temperature(atoms, temp0)
[docs]def ZeroRotation(atoms: Atoms, preserve_temperature: float = True): "Sets the total angular momentum to zero by counteracting rigid rotations." # Save initial temperature temp0 = atoms.get_temperature() # Find the principal moments of inertia and principal axes basis vectors Ip, basis = atoms.get_moments_of_inertia(vectors=True) # Calculate the total angular momentum and transform to principal basis Lp = np.dot(basis, atoms.get_angular_momentum()) # Calculate the rotation velocity vector in the principal basis, avoiding # zero division, and transform it back to the cartesian coordinate system omega = np.dot(np.linalg.inv(basis), np.select([Ip > 0], [Lp / Ip])) # We subtract a rigid rotation corresponding to this rotation vector com = atoms.get_center_of_mass() positions = atoms.get_positions() positions -= com # translate center of mass to origin velocities = atoms.get_velocities() atoms.set_velocities(velocities - np.cross(omega, positions)) if preserve_temperature: force_temperature(atoms, temp0)
def n_BE(temp, omega): """Bose-Einstein distribution function. Args: temp: temperature converted to eV (*units.kB) omega: sequence of frequencies converted to eV Returns: Value of Bose-Einstein distribution function for each energy """ omega = np.asarray(omega) # 0K limit if temp < eps_temp: n = np.zeros_like(omega) else: n = 1 / (np.exp(omega / (temp)) - 1) return n
[docs]def phonon_harmonics( force_constants, masses, temp=None, *, temperature_K=None, rng=np.random.rand, quantum=False, plus_minus=False, return_eigensolution=False, failfast=True, ): r"""Return displacements and velocities that produce a given temperature. Parameters: force_constants: array of size 3N x 3N force constants (Hessian) of the system in eV/Ų masses: array of length N masses of the structure in amu temp: float (deprecated) Temperature converted to eV (T * units.kB). Deprecated, use ``temperature_K``. temperature_K: float Temperature in Kelvin. rng: function Random number generator function, e.g., np.random.rand quantum: bool True for Bose-Einstein distribution, False for Maxwell-Boltzmann (classical limit) plus_minus: bool Displace atoms with +/- the amplitude accoding to PRB 94, 075125 return_eigensolution: bool return eigenvalues and eigenvectors of the dynamical matrix failfast: bool True for sanity checking the phonon spectrum for negative frequencies at Gamma Returns: Displacements, velocities generated from the eigenmodes, (optional: eigenvalues, eigenvectors of dynamical matrix) Purpose: Excite phonon modes to specified temperature. This excites all phonon modes randomly so that each contributes, on average, equally to the given temperature. Both potential energy and kinetic energy will be consistent with the phononic vibrations characteristic of the specified temperature. In other words the system will be equilibrated for an MD run at that temperature. force_constants should be the matrix as force constants, e.g., as computed by the ase.phonons module. Let X_ai be the phonon modes indexed by atom and mode, w_i the phonon frequencies, and let 0 < Q_i <= 1 and 0 <= R_i < 1 be uniformly random numbers. Then .. code-block:: none 1/2 _ / k T \ --- 1 _ 1/2 R += | --- | > --- X (-2 ln Q ) cos (2 pi R ) a \ m / --- w ai i i a i i 1/2 _ / k T \ --- _ 1/2 v = | --- | > X (-2 ln Q ) sin (2 pi R ) a \ m / --- ai i i a i Reference: [West, Estreicher; PRL 96, 22 (2006)] """ # Handle the temperature units temp = units.kB * process_temperature(temp, temperature_K, 'eV') # Build dynamical matrix rminv = (masses ** -0.5).repeat(3) dynamical_matrix = force_constants * rminv[:, None] * rminv[None, :] # Solve eigenvalue problem to compute phonon spectrum and eigenvectors w2_s, X_is = np.linalg.eigh(dynamical_matrix) # Check for soft modes if failfast: zeros = w2_s[:3] worst_zero = np.abs(zeros).max() if worst_zero > 1e-3: msg = "Translational deviate from 0 significantly: " raise ValueError(msg + f"{w2_s[:3]}") w2min = w2_s[3:].min() if w2min < 0: msg = "Dynamical matrix has negative eigenvalues such as " raise ValueError(msg + f"{w2min}") # First three modes are translational so ignore: nw = len(w2_s) - 3 n_atoms = len(masses) w_s = np.sqrt(w2_s[3:]) X_acs = X_is[:, 3:].reshape(n_atoms, 3, nw) # Assign the amplitudes according to Bose-Einstein distribution # or high temperature (== classical) limit if quantum: hbar = units._hbar * units.J * units.s A_s = np.sqrt(hbar * (2 * n_BE(temp, hbar * w_s) + 1) / (2 * w_s)) else: A_s = np.sqrt(temp) / w_s if plus_minus: # create samples by multiplying the amplitude with +/- # according to Eq. 5 in PRB 94, 075125 spread = (-1) ** np.arange(nw) # gauge eigenvectors: largest value always positive for ii in range(X_acs.shape[-1]): vec = X_acs[:, :, ii] max_arg = np.argmax(abs(vec)) X_acs[:, :, ii] *= np.sign(vec.flat[max_arg]) # Create velocities und displacements from the amplitudes and # eigenvectors A_s *= spread phi_s = 2.0 * np.pi * rng(nw) # Assign velocities, sqrt(2) compensates for missing sin(phi) in # amplitude for displacement v_ac = (w_s * A_s * np.sqrt(2) * np.cos(phi_s) * X_acs).sum(axis=2) v_ac /= np.sqrt(masses)[:, None] # Assign displacements d_ac = (A_s * X_acs).sum(axis=2) d_ac /= np.sqrt(masses)[:, None] else: # compute the gaussian distribution for the amplitudes # We need 0 < P <= 1.0 and not 0 0 <= P < 1.0 for the logarithm # to avoid (highly improbable) NaN. # Box Muller [en.wikipedia.org/wiki/Box–Muller_transform]: spread = np.sqrt(-2.0 * np.log(1.0 - rng(nw))) # assign amplitudes and phases A_s *= spread phi_s = 2.0 * np.pi * rng(nw) # Assign velocities and displacements v_ac = (w_s * A_s * np.cos(phi_s) * X_acs).sum(axis=2) v_ac /= np.sqrt(masses)[:, None] d_ac = (A_s * np.sin(phi_s) * X_acs).sum(axis=2) d_ac /= np.sqrt(masses)[:, None] if return_eigensolution: return d_ac, v_ac, w2_s, X_is # else return d_ac, v_ac
[docs]def PhononHarmonics( atoms, force_constants, temp=None, *, temperature_K=None, rng=np.random, quantum=False, plus_minus=False, return_eigensolution=False, failfast=True, ): r"""Excite phonon modes to specified temperature. This will displace atomic positions and set the velocities so as to produce a random, phononically correct state with the requested temperature. Parameters: atoms: ase.atoms.Atoms() object Positions and momenta of this object are perturbed. force_constants: ndarray of size 3N x 3N Force constants for the the structure represented by atoms in eV/Ų temp: float (deprecated). Temperature in eV. Deprecated, use ``temperature_K`` instead. temperature_K: float Temperature in Kelvin. rng: Random number generator RandomState or other random number generator, e.g., np.random.rand quantum: bool True for Bose-Einstein distribution, False for Maxwell-Boltzmann (classical limit) failfast: bool True for sanity checking the phonon spectrum for negative frequencies at Gamma. """ # Receive displacements and velocities from phonon_harmonics() d_ac, v_ac = phonon_harmonics( force_constants=force_constants, masses=atoms.get_masses(), temp=temp, temperature_K=temperature_K, rng=rng.rand, plus_minus=plus_minus, quantum=quantum, failfast=failfast, return_eigensolution=False, ) # Assign new positions (with displacements) and velocities atoms.positions += d_ac atoms.set_velocities(v_ac)