def __init__(self, structure):
"""
Conservative defect charge generator based on the oxidation statess
determined by bond valence. Targetted materials are wideband
semiconductors and insulators. AxBy where A is cation and B is
anion will have charge assignments {A: [0:y], B:[-x:0]}. For these
systems, antisites typically have very high formation energies and
are ignored.
Args:
structure: pymatgen structure object
"""
struct_species = structure.types_of_specie
if len(struct_species) == 1:
oxi_states = {struct_species[0].symbol: 0}
else:
vir = VIRE(structure)
oxi_states = vir.valences
self.oxi_states = {}
for key, val in oxi_states.items():
strip_key = ''.join([s for s in key if s.isalpha()])
self.oxi_states[str2unicode(strip_key)] = val
self.min_max_oxi = {}
for s in struct_species:
if isinstance(s, Specie):
el = s.element
elif isinstance(s, Element):
el = s
else:
continue
max_oxi = max(el.common_oxidation_states)
min_oxi = min(el.common_oxidation_states)
self.min_max_oxi[str2unicode(el.symbol)] = (min_oxi, max_oxi)
def write_structure(structure, filename):
"""
Write a structure to a file based on file extension. For example, anything
ending in a "cif" is assumed to be a Crystallographic Information Format
file. Supported formats include CIF, POSCAR, CSSR and pymatgen's JSON
serialized structures.
Args:
structure (Structure/IStructure): Structure to write
filename (str): A filename to write to.
"""
fname = os.path.basename(filename)
if fnmatch(fname, "*.cif*"):
writer = CifWriter(structure)
elif fnmatch(fname, "POSCAR*") or fnmatch(fname, "CONTCAR*"):
writer = Poscar(structure)
elif fnmatch(fname.lower(), "*.cssr*"):
writer = Cssr(structure)
elif fnmatch(fname, "*.json*") or fnmatch(fname, "*.mson*"):
with zopen(filename, "wt") as f:
f.write(str2unicode(json.dumps(structure, cls=MontyEncoder)))
return
else:
raise ValueError("Unrecognized file extension!")
writer.write_file(filename)
def write_mol(mol, filename):
"""
Write a molecule to a file based on file extension. For example, anything
ending in a "xyz" is assumed to be a XYZ file. Supported formats include
xyz, Gaussian input (gjf|g03|g09|com|inp), and pymatgen's JSON serialized
molecules.
Args:
mol (Molecule/IMolecule): Molecule to write
filename (str): A filename to write to.
"""
fname = os.path.basename(filename)
if fnmatch(fname.lower(), "*.xyz*"):
return XYZ(mol).write_file(filename)
elif any([fnmatch(fname.lower(), "*.{}*".format(r))
for r in ["gjf", "g03", "g09", "com", "inp"]]):
return GaussianInput(mol).write_file(filename)
elif fnmatch(fname, "*.json*") or fnmatch(fname, "*.mson*"):
with zopen(filename, "wt") as f:
return f.write(str2unicode(json.dumps(mol, cls=MontyEncoder)))
else:
m = re.search("\.(pdb|mol|mdl|sdf|sd|ml2|sy2|mol2|cml|mrv)",
filename.lower())
if m:
return BabelMolAdaptor(mol).write_file(filename, m.group(1))
raise ValueError("Unrecognized file extension!")
def write_structure(structure, filename):
"""
Write a structure to a file based on file extension. For example, anything
ending in a "cif" is assumed to be a Crystallographic Information Format
file. Supported formats include CIF, POSCAR, CSSR and pymatgen's JSON
serialized structures.
Args:
structure (Structure/IStructure): Structure to write
filename (str): A filename to write to.
"""
fname = os.path.basename(filename)
if fnmatch(fname, "*.cif*"):
writer = CifWriter(structure)
elif fnmatch(fname, "POSCAR*") or fnmatch(fname, "CONTCAR*"):
writer = Poscar(structure)
elif fnmatch(fname.lower(), "*.cssr*"):
writer = Cssr(structure)
elif fnmatch(fname, "*.json*") or fnmatch(fname, "*.mson*"):
with zopen(filename, "wt") as f:
f.write(str2unicode(json.dumps(structure, cls=MontyEncoder)))
return
else:
raise ValueError("Unrecognized file extension!")
writer.write_file(filename)
def write_mol(mol, filename):
"""
Write a molecule to a file based on file extension. For example, anything
ending in a "xyz" is assumed to be a XYZ file. Supported formats include
xyz, Gaussian input (gjf|g03|g09|com|inp), and pymatgen's JSON serialized
molecules.
Args:
mol (Molecule/IMolecule): Molecule to write
filename (str): A filename to write to.
"""
fname = os.path.basename(filename)
if fnmatch(fname.lower(), "*.xyz*"):
return XYZ(mol).write_file(filename)
elif any([
fnmatch(fname.lower(), "*.{}*".format(r))
for r in ["gjf", "g03", "g09", "com", "inp"]
]):
return GaussianInput(mol).write_file(filename)
elif fnmatch(fname, "*.json*") or fnmatch(fname, "*.mson*"):
with zopen(filename, "wt") as f:
return f.write(str2unicode(json.dumps(mol, cls=MontyEncoder)))
else:
m = re.search("\.(pdb|mol|mdl|sdf|sd|ml2|sy2|mol2|cml|mrv)",
filename.lower())
if m:
return BabelMolAdaptor(mol).write_file(filename, m.group(1))
raise ValueError("Unrecognized file extension!")
def get_charges(self, defect_type, site_specie=None, sub_specie=None):
"""
Based on the type of defect, site and substitution (if any) species
the defect charge states are generated.
Args:
defect_type (str): Options are vacancy, antisite, substitution,
and interstitial
site_specie: Specie on the host lattice site
For interstitials, use this
sub_specie: Specie that is replacing the site specie.
For antisites and substitution defects
"""
if defect_type == 'vacancy':
vac_symbol = get_el_sp(site_specie).symbol
vac_oxi_state = self.oxi_states[str2unicode(vac_symbol)]
if vac_oxi_state == 0:
return [-1, 0, 1]
else:
minval = min(-vac_oxi_state, 0)
maxval = max(-vac_oxi_state, 0)
return [c for c in range(minval - 1, maxval + 2)]
elif defect_type in ['antisite', 'substitution']:
#TODO: may cause some weird states for substitutions. Worth updating in future.
vac_symbol = get_el_sp(site_specie).symbol
vac_oxi_state = self.oxi_states[str2unicode(vac_symbol)]
as_symbol = get_el_sp(sub_specie).symbol
as_oxi_state = self.oxi_states[str2unicode(as_symbol)]
expected_oxi = as_oxi_state - vac_oxi_state
if expected_oxi == 0:
return [-1, 0, 1]
else:
minval = min(expected_oxi, 0)
maxval = max(expected_oxi, 0)
return [c for c in range(minval - 1, maxval + 2)]
elif defect_type == 'interstitial':
return [-1, 0, 1]
def __init__(self, structure):
"""
Args:
structure: pymatgen structure object
"""
struct_species = structure.types_of_specie
if len(struct_species) == 1:
oxi_states = {struct_species[0].symbol: 0}
else:
vir = VIRE(structure)
oxi_states = vir.valences
self.oxi_states = {}
for key, val in oxi_states.items():
strip_key = ''.join([s for s in key if s.isalpha()])
self.oxi_states[str2unicode(strip_key)] = val
def test_str2unicode(self):
if sys.version_info.major < 3:
self.assertEqual(type(str2unicode("a")), unicode)
else:
self.assertEqual(type(str2unicode("a")), str)
def test_str2unicode(self):
if sys.version_info.major < 3:
self.assertEqual(type(str2unicode("a")), unicode)
else:
self.assertEqual(type(str2unicode("a")), str)
def __init__(self,
structure,
max_min_oxi=None,
substitutions=None,
oxi_states=None,
cellmax=128,
antisites_flag=True,
include_interstitials=False,
interstitial_elements=None,
intersites=None,
standardized=False,
struct_type='semiconductor'):
"""
Args:
structure (Structure):
the bulk structure.
max_min_oxi (dict):
The minimal and maximum oxidation state of each element as a
dict. For instance {"O":(-2,0)}. If not given, the oxi-states
of pymatgen are considered.
substitutions (dict):
The allowed substitutions of elements as a dict. If not given,
intrinsic defects are computed. If given, intrinsic (e.g.,
anti-sites) and extrinsic are considered explicitly specified.
Example: {"Co":["Zn","Mn"]} means Co sites can be substituted
by Mn or Zn.
oxi_states (dict):
The oxidation state of the elements in the compound e.g.
{"Fe":2,"O":-2}. If not given, the oxidation state of each
site is computed with bond valence sum. WARNING: Bond-valence
method can fail for mixed-valence compounds.
cellmax (int):
Maximum number of atoms allowed in the supercell.
antisites_flag (bool):
If False, don't generate antisites.
include_interstitials (bool):
If true, do generate interstitial defect configurations
(default: False).
interstitial_elements ([str]):
List of strings containing symbols of the elements that are
to be considered for interstitial sites. The default (None)
triggers self-interstitial generation,
given that include_interstitials is True.
intersites ([PeriodicSite]):
A list of PeriodicSites in the bulk structure on which we put
interstitials. Note that you still have to set flag
include_interstitials to True in order to make use of this
manual way of providing interstitial sites.
If this is used, then no additional interstitials are generated
beyond the list that is provided in intersites.
standardized (bool):
If True, use the primitive standard structure as unit cell
for generating the defect configurations (default is False).
The primitive standard structure is obtained from the
SpacegroupAnalyzer class with a symprec of 0.01.
struct_type (string):
Options are 'semiconductor' and 'insulator'. If semiconductor
is selected, charge states based on database of semiconductors
is used to assign defect charges. For insulators, defect
charges are conservatively assigned.
"""
max_min_oxi = max_min_oxi if max_min_oxi is not None else {}
substitutions = substitutions if substitutions is not None else {}
oxi_states = oxi_states if oxi_states is not None else {}
interstitial_elements = interstitial_elements if interstitial_elements is not None else []
intersites = intersites if intersites is not None else []
self.defects = []
self.cellmax = cellmax
self.substitutions = {}
self.struct_type = struct_type
for key, val in substitutions.items():
self.substitutions[str2unicode(key)] = val
spa = SpacegroupAnalyzer(structure, symprec=1e-2)
prim_struct = spa.get_primitive_standard_structure()
if standardized:
self.struct = prim_struct
else:
self.struct = structure
struct_species = self.struct.types_of_specie
if self.struct_type == 'semiconductor':
self.defect_charger = DefectChargerSemiconductor(
self.struct, min_max_oxi=max_min_oxi)
elif self.struct_type == 'insulator':
self.defect_charger = DefectChargerInsulator(self.struct)
elif self.struct_type == 'manual':
self.defect_charger = DefectChargerUserCustom(
self.struct, oxi_states=oxi_states)
elif self.struct_type == 'ionic':
self.defect_charger = DefectChargerIonic(self.struct)
else:
raise NotImplementedError
if include_interstitials and interstitial_elements:
for elem_str in interstitial_elements:
if not Element.is_valid_symbol(elem_str):
raise ValueError("invalid interstitial element"
" \"{}\"".format(elem_str))
sc_scale = get_optimized_sc_scale(self.struct, cellmax)
self.defects = {}
sc = self.struct.copy()
sc.make_supercell(sc_scale)
self.defects['bulk'] = {
'name': 'bulk',
'supercell': {
'size': sc_scale,
'structure': sc
}
}
# If interstitials are provided as a list of PeriodicSites,
# make sure that the lattice has not changed.
if include_interstitials and intersites:
for intersite in intersites: #list of PeriodicSite objects
if intersite.lattice != self.struct.lattice:
raise RuntimeError(
"Discrepancy between lattices"
" underlying the input interstitials and"
" the bulk structure; possibly because of"
" standardizing the input structure.")
vacancies = []
as_defs = []
sub_defs = []
VG = VacancyGenerator(self.struct)
print("Setting up defects...")
for i, vac in enumerate(VG):
vac_site = vac.site
vac_symbol = vac.site.specie.symbol
vac_sc = vac.generate_defect_structure(sc_scale)
#create a trivial defect structure to find where supercell transformation moves the lattice
struct_for_defect_site = Structure(
vac.bulk_structure.copy().lattice, [vac.site.specie],
[vac.site.frac_coords],
to_unit_cell=True,
coords_are_cartesian=False)
struct_for_defect_site.make_supercell(sc_scale)
vac_sc_site = struct_for_defect_site[0]
charges_vac = self.defect_charger.get_charges(
'vacancy', vac_symbol)
vacancies.append({
'name': "vac_{}_{}".format(i + 1, vac_symbol),
'unique_site': vac_site,
'bulk_supercell_site': vac_sc_site,
'defect_type': 'vacancy',
'site_specie': vac_symbol,
'site_multiplicity': vac.multiplicity,
'supercell': {
'size': sc_scale,
'structure': vac_sc
},
'charges': charges_vac
})
if antisites_flag:
for as_specie in set(struct_species):
SG = SubstitutionGenerator(self.struct, as_specie)
for i, sub in enumerate(SG):
as_symbol = as_specie.symbol
as_sc = sub.generate_defect_structure(sc_scale)
# create a trivial defect structure to find where supercell transformation moves the defect
struct_for_defect_site = Structure(
sub.bulk_structure.copy().lattice, [sub.site.specie],
[sub.site.frac_coords],
to_unit_cell=True,
coords_are_cartesian=False)
struct_for_defect_site.make_supercell(sc_scale)
as_sc_site = struct_for_defect_site[0]
#get bulk_site (non sc)
poss_deflist = sorted(
sub.bulk_structure.get_sites_in_sphere(
sub.site.coords, 0.01, include_index=True),
key=lambda x: x[1])
if not len(poss_deflist):
raise ValueError(
"Could not find substitution site inside bulk structure for {}?"
.format(sub.name))
defindex = poss_deflist[0][2]
as_site = sub.bulk_structure[defindex]
vac_symbol = as_site.specie
charges_as = self.defect_charger.get_charges(
'antisite', vac_symbol, as_symbol)
as_defs.append({
'name':
"as_{}_{}_on_{}".format(i + 1, as_symbol, vac_symbol),
'unique_site':
as_site,
'bulk_supercell_site':
as_sc_site,
'defect_type':
'antisite',
'site_specie':
vac_symbol,
'substituting_specie':
as_symbol,
'site_multiplicity':
sub.multiplicity,
'supercell': {
'size': sc_scale,
'structure': as_sc
},
'charges':
charges_as
})
for vac_symbol, subspecie_list in self.substitutions.items():
for subspecie_symbol in subspecie_list:
SG = SubstitutionGenerator(self.struct, subspecie_symbol)
for i, sub in enumerate(SG):
sub_symbol = sub.site.specie.symbol
#get bulk_site (non sc)
poss_deflist = sorted(
sub.bulk_structure.get_sites_in_sphere(
sub.site.coords, 0.1, include_index=True),
key=lambda x: x[1])
if not len(poss_deflist):
raise ValueError(
"Could not find substitution site inside bulk structure for {}?"
.format(sub.name))
defindex = poss_deflist[0][2]
sub_site = self.struct[defindex]
this_vac_symbol = sub_site.specie.symbol
if (sub_symbol != subspecie_symbol) or (this_vac_symbol !=
vac_symbol):
continue
else:
sub_sc = sub.generate_defect_structure(sc_scale)
# create a trivial defect structure to find where supercell transformation moves the defect
struct_for_defect_site = Structure(
sub.bulk_structure.copy().lattice,
[sub.site.specie], [sub.site.frac_coords],
to_unit_cell=True,
coords_are_cartesian=False)
struct_for_defect_site.make_supercell(sc_scale)
sub_sc_site = struct_for_defect_site[0]
charges_sub = self.defect_charger.get_charges(
'substitution', vac_symbol, subspecie_symbol)
sub_defs.append({
'name':
"sub_{}_{}_on_{}".format(i + 1, subspecie_symbol,
vac_symbol),
'unique_site':
sub_site,
'bulk_supercell_site':
sub_sc_site,
'defect_type':
'substitution',
'site_specie':
vac_symbol,
'substitution_specie':
subspecie_symbol,
'site_multiplicity':
sub.multiplicity,
'supercell': {
'size': sc_scale,
'structure': sub_sc
},
'charges':
charges_sub
})
self.defects['vacancies'] = vacancies
self.defects['substitutions'] = sub_defs
self.defects['substitutions'] += as_defs
if include_interstitials:
interstitials = []
if interstitial_elements:
inter_elems = interstitial_elements
else:
inter_elems = [elem.symbol for elem in \
self.struct.composition.elements]
if len(inter_elems) == 0:
raise RuntimeError("empty element list for interstitials")
if intersites:
#manual specification of interstitials
for i, intersite in enumerate(intersites):
for elt in inter_elems:
name = "inter_{}_{}".format(i + 1, elt)
if intersite.lattice != self.struct.lattice:
err_msg = "Lattice matching error occurs between provided interstitial and the bulk structure."
if standardized:
err_msg += "\nLikely because the standardized flag was used. Turn this flag off or reset " \
"your interstitial PeriodicSite to match the standardized form of the bulk structure."
raise ValueError(err_msg)
else:
intersite_object = Interstitial(
self.struct, intersite)
# create a trivial defect structure to find where supercell transformation moves the defect site
struct_for_defect_site = Structure(
intersite_object.bulk_structure.copy().lattice,
[intersite_object.site.specie],
[intersite_object.site.frac_coords],
to_unit_cell=True,
coords_are_cartesian=False)
struct_for_defect_site.make_supercell(sc_scale)
site_sc = struct_for_defect_site[0]
sc_with_inter = intersite_object.generate_defect_structure(
sc_scale)
charges_inter = self.defect_charger.get_charges(
'interstitial', elt)
interstitials.append({
'name':
name,
'unique_site':
intersite_object.site,
'bulk_supercell_site':
site_sc,
'defect_type':
'interstitial',
'site_specie':
intersite_object.site.specie.symbol,
'site_multiplicity':
intersite_object.multiplicity,
'supercell': {
'size': sc_scale,
'structure': sc_with_inter
},
'charges':
charges_inter
})
else:
print(
"Searching for interstitial sites (this can take awhile)..."
)
for elt in inter_elems:
#TODO: Add ability to use other interstitial finding methods in pymatgen
IG = InterstitialGenerator(self.struct, elt)
for i, intersite_object in enumerate(IG):
name = intersite_object.name
# create a trivial defect structure to find where supercell transformation moves the defect site
struct_for_defect_site = Structure(
intersite_object.bulk_structure.copy().lattice,
[intersite_object.site.specie],
[intersite_object.site.frac_coords],
to_unit_cell=True,
coords_are_cartesian=False)
struct_for_defect_site.make_supercell(sc_scale)
site_sc = struct_for_defect_site[0]
sc_with_inter = intersite_object.generate_defect_structure(
sc_scale)
charges_inter = self.defect_charger.get_charges(
'interstitial', elt)
interstitials.append({
'name':
name,
'unique_site':
intersite_object.site,
'bulk_supercell_site':
site_sc,
'defect_type':
'interstitial',
'site_specie':
intersite_object.site.specie.symbol,
'site_multiplicity':
intersite_object.multiplicity,
'supercell': {
'size': sc_scale,
'structure': sc_with_inter
},
'charges':
charges_inter
})
self.defects['interstitials'] = interstitials
print("\nNumber of jobs created:")
tottmp = 0
for j in self.defects.keys():
if j == 'bulk':
print(" bulk = 1")
tottmp += 1
else:
print(" {}:".format(j))
for lis in self.defects[j]:
print(" {} = {}".format(lis['name'],
len(lis['charges'])))
tottmp += len(lis['charges'])
print("Total (non dielectric) jobs created = {}\n".format(tottmp))
def get_charges(self, defect_type, site_specie=None, sub_specie=None):
"""
Based on the type of defect, site and substitution (if any) species
the defect charge states are generated.
Args:
defect_type (str): Options are vacancy, antisite, substitution,
and interstitial
site_specie: Specie on the host lattice site
For interstitials, use this
sub_specie: Specie that is replacing the site specie.
For antisites and substitution defects
"""
if defect_type == 'vacancy':
vac_symbol = get_el_sp(site_specie).symbol
vac_oxi_state = self.oxi_states[str2unicode(vac_symbol)]
if vac_oxi_state < 0:
min_oxi = max(vac_oxi_state, self.min_max_oxi[vac_symbol][0])
max_oxi = 0
elif vac_oxi_state > 0:
max_oxi = min(vac_oxi_state, self.min_max_oxi[vac_symbol][1])
min_oxi = 0
else: # most probably single element
oxi_states = get_el_sp(site_specie).common_oxidation_states
min_oxi = min(oxi_states)
max_oxi = max(oxi_states)
return [-c for c in range(min_oxi, max_oxi + 1)]
elif defect_type == 'antisite':
vac_symbol = get_el_sp(site_specie).symbol
vac_oxi_state = self.oxi_states[str2unicode(vac_symbol)]
as_symbol = get_el_sp(sub_specie).symbol
if vac_oxi_state > 0:
oxi_max = max(self.min_max_oxi[as_symbol][1], 0)
oxi_min = 0
else:
oxi_max = 0
oxi_min = min(self.min_max_oxi[as_symbol][0], 0)
return [c - vac_oxi_state for c in range(oxi_min, oxi_max + 1)]
elif defect_type == 'substitution':
site_specie = get_el_sp(site_specie)
sub_specie = get_el_sp(sub_specie)
vac_symbol = site_specie.symbol
vac_oxi_state = self.oxi_states[str2unicode(vac_symbol)]
max_oxi_sub = max(sub_specie.common_oxidation_states)
min_oxi_sub = min(sub_specie.common_oxidation_states)
if vac_oxi_state > 0:
if max_oxi_sub < 0:
raise ValueError("Substitution seems not possible")
else:
if max_oxi_sub > vac_oxi_state:
return list(range(max_oxi_sub - vac_oxi_state + 1))
else:
return [max_oxi_sub - vac_oxi_state]
else:
if min_oxi_sub > 0:
raise ValueError("Substitution seems not possible")
else:
if min_oxi_sub < vac_oxi_state:
return list(range(min_oxi_sub - vac_oxi_state, 1))
else:
return [min_oxi_sub - vac_oxi_state]
elif defect_type == 'interstitial':
site_specie = get_el_sp(site_specie)
min_oxi = min(min(site_specie.common_oxidation_states), 0)
max_oxi = max(max(site_specie.common_oxidation_states), 0)
return list(range(min_oxi, max_oxi + 1))
