#!/usr/bin/env python """ Lattice constant Test Driver for Hexagonal Structure Computes the lattice constant for: 1. Any element, 2. HCP crystal, 3. At 0 K, 0 GPa, by simplex minimization. Structure Overview: 1. main() deals with I/O and get information from getLatticeConstant() 2. getLatticeConstant() uses an initial guess at the lattice constant 'a' and calls searchLatticeConstants() 3. searchLatticeConstants() minimizes getEnergy() with simplex algorithm 4. When lower energy is obtained with separation > init_upper_guess, collapseHandler() is called Date: 2015/08/25 Author: Junhao Li Last Update: 2016/02/26 Junhao Li 2017/08/28 Daniel S. Karls """ # Import External Modules # ASE Modules from ase.build import bulk from ase import Atoms # KIM Modules from ase.calculators.kim.kim import KIM # Other Python Modules import sys import json import math import jinja2 import os import copy import numpy as np from scipy.optimize import fmin from collections import OrderedDict # Constants for result.edn SPACE_GROUP = 'P63/mmc' WCYKOFF_CODE = '2a' KEY_SOURCE_VALUE = 'source-value' KEY_SOURCE_UNIT = 'source-unit' KEY_SOURCE_UNCERT = 'source-std-uncert-value' UNIT_LENGTH = 'angstrom' UNIT_TEMPERATURE = 'K' UNIT_PRESSURE = 'GPa' UNIT_ENERGY = 'eV' STRUCTURE_PROP_ID = 'tag:staff@noreply.openkim.org,2014-04-15:property/structure-hexagonal-crystal-npt' ENERGY_PROP_ID = 'tag:staff@noreply.openkim.org,2014-04-15:property/cohesive-free-energy-hexagonal-crystal' # Constants frequently used in calculation SQRT3 = math.sqrt(3.0) PERFECT_CA = math.sqrt(8.0 / 3.0) HCP_CUBIC_CELL = np.array([1.0, SQRT3, 1.0]) HCP_CUBIC_POSITIONS = np.array([ [0.0, 0.0, 0.0], [0.5, 0.5 * SQRT3, 0.0], [0.5, 0.5 / 3.0 * SQRT3, 0.5], [0.0, (0.5 + 0.5 / 3.0) * SQRT3, 0.5] ]) HCP_CUBIC_NATOMS = 4 # Minimization Convergence Criteria FMIN_FTOL = 1e-10 FMIN_XTOL = 1e-10 def getEnergy(cellVector, meta, cache): # cellVector: 1d numpy array of size 1(2), containing a (and c) # meta: various information... # cache: cache previously created atoms # Determine a and c based on the length of the cellVector a = cellVector[0] if len(cellVector) == 2: c = cellVector[1] else: c = a * PERFECT_CA # Default to perfect c/a ratio cellSize = HCP_CUBIC_CELL * [a, a, c] repeat = (1, 1, 1) nAtoms = HCP_CUBIC_NATOMS # Get atoms from cache or create a new one if haven't been created sizes = cache['sizes'] if repeat in sizes: sizeId = sizes.index(repeat) atoms = cache['atoms'][sizeId] else: # create a new atoms and save to cache print 'Creating new atoms:', repeat sizes.append(repeat) atoms = meta['prototype'].copy() atoms *= repeat atoms.set_calculator(KIM(meta['model'])) cache['atoms'].append(atoms) # Scale atoms with the scaled cellSize atoms.set_cell(cellSize, scale_atoms = True) energy = atoms.get_potential_energy() / nAtoms return energy def searchLatticeConstants(latticeConstantsGuess, meta, cache): # Simplex searching lattice constants that minimize potential energy of atoms # Searching starts from latticeConstantsGuess print 'Simplex search starting from:', latticeConstantsGuess tmpLatticeConstants, tmpEnergy = fmin( getEnergy, latticeConstantsGuess, args = (meta, cache), full_output = True, ftol = FMIN_FTOL, xtol = FMIN_XTOL, )[:2] return tmpLatticeConstants, tmpEnergy def getInterlayerDist(latticeConstants): # Calculate the distance between atom 0 and 1 (refer to HCP_CUBIC_POSITIONS) hrDist = latticeConstants[0] / SQRT3 vDist = latticeConstants[1] / 2 return math.sqrt(hrDist * hrDist + vDist * vDist) def collapseHandler(): # Called when lower energy is obtained by separating atoms beyond cutoff print 'ERR: Lower energy is obtained by separating atoms beyond cutoff.' print 'ERR: System Collapsed. Structure Maybe Unstable.' sys.exit(0) def getLatticeConstant(elem, model): # Initialize meta and cache meta = { 'model': model } cache = { 'sizes': [], 'atoms': [] } # Create prototype atoms atoms = Atoms(elem + '4', positions = HCP_CUBIC_POSITIONS, cell = HCP_CUBIC_CELL, pbc = [1, 1, 1] ) meta['prototype'] = atoms # Relaxation with c/a Ratio Fixed to PERFECT_CA print 'Relaxation with c/a ratio fixed' init_upper_guess = 5.0 minEnergy = 0 for i in [x * 0.05 + 0.5 for x in range(-2, 3)]: print 'Simplex search starting from: %r * %i' % (init_upper_guess, i) tmpLatticeConstants, tmpEnergy = searchLatticeConstants( [init_upper_guess * i], meta, cache ) print 'Tmp Lattice Constants:', tmpLatticeConstants print 'Tmp Energy:', tmpEnergy if tmpEnergy < minEnergy and tmpLatticeConstants[0] < init_upper_guess: # Make sure there are interatomic forces holding the structure minLatticeConstants = tmpLatticeConstants minEnergy = tmpEnergy print '--------' if minEnergy == 0: collapseHandler() # Relaxation With c/a Ratio Relaxed print 'Relaxation with c/a ratio relaxed' tmpA = minLatticeConstants[0] minEnergy = 0 for i in [x * 0.05 + 1 for x in range(-4, 5)]: print 'Simplex search starting from c/a ratio: 1.633 *', i tmpLatticeConstants, tmpEnergy = searchLatticeConstants( [tmpA, tmpA * PERFECT_CA * i], meta, cache ) print 'Tmp Lattice Constants:', tmpLatticeConstants print 'Tmp Energy:', tmpEnergy if tmpEnergy < minEnergy and \ tmpLatticeConstants[0] < init_upper_guess and \ getInterlayerDist(tmpLatticeConstants) < init_upper_guess: # Make sure there is both intralayer and interlayer forces minLatticeConstants = tmpLatticeConstants minEnergy = tmpEnergy print '--------' if minEnergy == 0: collapseHandler() print 'Lattice Constants:', minLatticeConstants print 'Energy:', minEnergy # Return Results return minLatticeConstants[0], minLatticeConstants[1], minEnergy def V(value, unit = '', uncert = ''): # Generate OrderedDict for KIM JSON Output res = OrderedDict([ (KEY_SOURCE_VALUE, value), ]) if unit != '': res.update(OrderedDict([ (KEY_SOURCE_UNIT, unit), ])) if uncert != '': res.update(OrderedDict([ (KEY_SOURCE_UNCERT, uncert) ])) return res def main(): # Input Parameters elem = raw_input('Element = ') lattice = raw_input('Lattice = ') model = raw_input('Model = ') # Parameters for Debugging # ======== # elem = 'Co' # lattice = 'hcp' # model = 'EAM_Dynamo_PurjaPun_Mishin_Co__MO_885079680379_001' # elem = 'Si' # lattice = 'hcp' # model = 'EDIP_BOP_Bazant_Kaxiras_Si__MO_958932894036_001' # elem = 'Al' # lattice = 'hcp' # model = 'EAM_Dynamo_Mishin_Farkas_Al__MO_651801486679_001' # elem = 'C' # lattice = 'hcp' # model = 'EAM_Dynamo_Hepburn_Ackland_FeC__MO_143977152728_001' # model = 'Tersoff_LAMMPS_Erhart_Albe_CSi__MO_903987585848_001' # ======== # Print Inputs print 'Element:', elem print 'Lattice:', lattice # Not used here print 'Model:', model # Obtain Lattice Constants and Cohesive Energy a, c, energy = getLatticeConstant(elem, model) print 'Lattice Constants:', a, c # Output Results structureResults = OrderedDict([ ('property-id', STRUCTURE_PROP_ID), ('instance-id', 1), ('cauchy-stress', V([0.0, 0.0, 0.0, 0.0, 0.0, 0.0], UNIT_PRESSURE)), ]) cohesiveEnergyResults = OrderedDict([ ('property-id', ENERGY_PROP_ID), ('instance-id', 2), ('cohesive-free-energy', V(-energy, UNIT_ENERGY)), ]) commonResults = OrderedDict([ ('short-name', V(['hcp'])), ('species', V([elem, elem])), ('a', V(a, UNIT_LENGTH)), ('c', V(c, UNIT_LENGTH)), ('basis-atom-coordinates', V([[0.0, 0.0, 0.0], [2.0/3, 1.0/3, 0.5]])), ('space-group', V(SPACE_GROUP)), ('temperature', V(0, UNIT_TEMPERATURE)), ]) structureResults.update(commonResults) cohesiveEnergyResults.update(commonResults) results = [ structureResults, cohesiveEnergyResults, ] resultsString = json.dumps(results, separators = (' ', ' '), indent = 4) print resultsString with open(os.path.abspath('output/results.edn'), 'w') as f: f.write(resultsString) if __name__ == '__main__': main()