Database Generation

Generate a database generation template file using atl_gen_configuration_file -t database_generation.

Attention

All keys are mandatory unless stated otherwise.

Database - [database]

General settings and file paths for the database.

• database_name:

  • Description: Name of the database to be used for internal reference and as the filename.

  • Type: (str)

  • Example: 'my_material_db'.

• min_num_atoms:

  • Description: Minimum number of atoms allowed in the generated structures.

  • Type: (int)

  • Default: 64.

• max_num_atoms:

  • Description: Maximum number of atoms allowed in the generated structures.

  • Type: (int)

  • Default: 128.

• min_cell_size:

  • Description: Minimum cell size in Angstrom.

  • Type: (float)

  • Default: 5.0.

• relax_struct_path:

  • Description: Path to a folder containing DFT optimized structures.

  • Type: (optional, str)

  • Default: ''.

• database_path:

  • Description: Path where the final database will be saved.

  • Type: (str, PosixPath)

  • Default: ''.

• rng_seed:

  • Description: Numerical value used to fix the RNG seed. If not specified, it will be chosen randomly each run.

  • Type: (optional, int)

  • Example: 42.

• overwrite_db:

  • Description: Allow database overwrite. If false, and the database exists, the new database name will include a timestamp.

  • Type: (optional, bool)

  • Default: False.

Composition - [database.composition]

Settings for the composition of the database.

• size:

  • Description: Maximum number of structures to generate for the database.

  • Type: (int)

  • Default: 7500.

Ratios - [database.composition.ratios]

Fraction of different structure types. The sum of the fractions must be equal to 1.0.

• bulk:

  • Description: Fraction of structures that will be bulk.

  • Type: (float)

  • Default: 0.4.

• surface:

  • Description: Fraction of structures that will be surfaces.

  • Type: (float)

  • Default: 0.6.

• cluster:

  • Description: Fraction of structures that will be clusters.

  • Type: (optional, float)

  • Default: 0.0.

Plot_Db - [database.plot_db]

Display and Export Options for the phase diagram plot.

• show:

  • Description: Whether to display the database with a phase diagram after creation.

  • Type: (optional, bool)

  • Default: True.

• format:

  • Description: Format for the figure.

  • Type: (optional, str)

  • Default: 'png'.

  • Possible values are: png, svg.

Rc_Params - [database.plot_db.rc_params]

Matplotlib rcParams for the plot.

• font.family:

  • Description: Font family for the phase diagram plot.

  • Type: (str)

  • Default: 'monospace'.

• font.size:

  • Description: Font size for the phase diagram plot.

  • Type: (int)

  • Default: 14.

Show_Db_Ase - [database.show_db_ase]

ASE GUI display options.

Attention

This section is optional.

• show:

  • Description: Whether to display the database using ASE GUI after creation.

  • Type: (optional, bool)

  • Default: False.

Export - [database.export]

Export options for the database.

• export:

  • Description: Whether to export the database.

  • Type: (bool)

  • Default: True.

• format:

  • Description: Export format supported by ASE (e.g., ‘extxyz’).

  • Type: (str)

  • Default: 'extxyz'.

• file_path:

  • Description: Path where the exported file will be saved.

  • Type: (str, PosixPath)

  • Default: ''.

• file_name:

  • Description: Name of the exported file.

  • Type: (str)

  • Default: 'export_db_filename'.

Phase_Diagram - [phase_diagram]

Description of the phase diagram.

• material_name:

  • Description: Internal name for the material in the phase diagram.

  • Type: (str)

  • Default: 'default_material_name'.

• element_list:

  • Description: List of elements to include in the phase diagram.

  • Type: (list[str])

  • Example:

['Cu', 'O']

• base_element:

  • Description: Symbol of the most abundant element in the phase.

  • Type: (str)

  • Example: 'Cu'.

Phase Definition - [phase_diagram.phase]

Defines a specific phase within the phase diagram. Multiple phases can be added.

Phase Definition - [phase_diagram.phase.XXXXX]

This key describes settings for dynamic entries. Several entries can be added by using different key names.

The key name (XXXXX) is used as the reference name. Replace XXXXX with a name of your choice.

Accepted parameters for each entry:

• name:

  • Description: Name to be used as reference for the phase.

  • Type: (str)

  • Example: 'alpha'.

• cluster_element:

  • Description: Symbol of the element defining the cluster.

  • Type: (optional, str)

• prototype:

  • Description: Materials Project ID of the prototypical structure.

  • Type: (str)

  • Example: 'mp-30'.

• offset:

  • Description: Fraction of composition allowed over and under the phase limits.

  • Type: (float)

  • Default: 0.1.

• limit_max_num_structures:

  • Description: Maximum number of structures to generate for this phase.

  • Type: (optional, int)

  • Default: 100.

• allow_modifications:

  • Description: Allow modifications (supercells, replacements, etc.) to the base structure.

  • Type: (optional, bool)

  • Default: True.

• use_cache:

  • Description: Store structures in cache to speed up generation. Can consume a lot of disk space.

  • Type: (optional, bool)

  • Default: False.

Parameters using composition. prefix:

Parameters using replacements. prefix:

• replacements.replace:

  • Description: Whether to replace specific elements. Elements in element_list will be considered for replacement and replaced by a single element species.

  • Type: (optional, bool)

  • Default: False.

• replacements.element_list:

  • Description: List of elements to be replaced.

  • Type: (optional, list[str])

  • Example:

['Ti']

• replacements.replace_with:

  • Description: Element to replace with.

  • Type: (optional, str)

  • Example: 'Ir'.

Generation - [generation]

Structure generation settings.

• generate_type:

  • Description: Types of structures to generate.

  • Type: (list[str])

  • Default: ['bulk', 'surface', 'cluster'].

  • Possible values are: bulk, surface, cluster.

Bulk - [generation.bulk]

Bulk structure generation settings.

• num_struct:

  • Description: Number of structures to generate.

  • Type: (int)

  • Default: 25.

• num_repeat:

  • Description: Number of repeats for each structure.

  • Type: (int)

  • Default: 5.

• supercell_max_idx:

  • Description: Maximum Miller index for the bulk supercells.

  • Type: (int)

  • Default: 2.

Surface - [generation.surface]

Surface structure generation settings.

• min_miller_index:

  • Description: Minimum Miller index used to generate surface structures.

  • Type: (int)

  • Default: 1.

• max_miller_index:

  • Description: Maximum Miller index used to generate surface structures.

  • Type: (int)

  • Default: 3.

• min_slab_size_ang:

  • Description: Minimum slab thickness in Angstrom.

  • Type: (optional, float)

  • Default: 7.0.

• min_vacuum_size_ang:

  • Description: Minimum size of the vacuum layer in Angstroms.

  • Type: (float)

  • Default: 12.0.

• get_supercells:

  • Description: Whether to generate supercells for surface structures.

  • Type: (bool)

  • Default: True.

• fixed_layers:

  • Description: Number of fixed layers in the surface slab.

  • Type: (int)

  • Default: 3.

• max_number_supercells:

  • Description: Maximum number of surface supercells to generate.

  • Type: (int)

  • Default: 200.

• save_in_db:

  • Description: Whether to save generated surfaces in the database.

  • Type: (bool)

  • Default: True.

• num_replacements:

  • Description: Number of replacement percentages to generate for each structure.

  • Type: (int)

  • Default: 20.

• num_repeat_replace:

  • Description: Number of repeats for each replacement.

  • Type: (int)

  • Default: 2.

• frac_slabs_save:

  • Description: Fraction of slabs to save after generation.

  • Type: (optional, float)

  • Default: 0.1.

• frac_supercells_save:

  • Description: Fraction of unreplaced supercells to save after generation.

  • Type: (optional, float)

  • Default: 0.1.

• max_slab_num:

  • Description: Maximum number of slabs to gather from the slab generation.

  • Type: (int)

  • Default: 15.

• n_workers:

  • Description: Maximum number of workers for parallel processing.

  • Type: (optional, int)

Deformation - [deformation]

Lattice deformation settings.

Attention

This section is optional.

• lattice_frac_deform_max:

  • Description: Maximum deformation value as a percentage of the lattice side length.

  • Type: (float)

  • Default: 0.05.

• lattice_frac_deform_min:

  • Description: Minimum deformation value as a percentage of the lattice side length.

  • Type: (float)

  • Default: 0.01.

• num_repeats:

  • Description: Number of repeats for each structure with random deformations.

  • Type: (int)

  • Default: 5.

• limit_max_num_deformations:

  • Description: Maximum number of lattice deformations to generate.

  • Type: (int)

  • Default: 100.

Perturbation - [perturbation]

Perturbation settings.

Attention

This section is optional.

• filter_struct_types:

  • Description: Types of structures to which the perturbation will be applied.

  • Type: (list[str])

  • Default: ['bulk', 'surface'].

  • Possible values are: bulk, surface, cluster.

• limit_max_num_perturbs:

  • Description: Maximum number of perturbations to generate.

  • Type: (int)

  • Default: 100.

• num_repeats:

  • Description: Number of repeats for each structure with random perturbations.

  • Type: (int)

  • Default: 1.

• perturbation_ang:

  • Description: Perturbation magnitude in Angstrom.

  • Type: (optional, float)

  • Default: 0.04.

Adsorbates - [adsorbates]

Adsorbate placement settings.

Attention

This section is optional.

• filter_struct_types:

  • Description: Types of structures to which adsorbates will be added.

  • Type: (optional, list[str])

  • Default: ['surface'].

  • Possible values are: bulk, surface, cluster.

• limit_max_num_perturbs:

  • Description: Maximum number of structures with adsorbates to generate.

  • Type: (optional, int)

  • Default: 100.

• num_repeats:

  • Description: Number of repeats for each structure.

  • Type: (int)

  • Default: 1.

• adsorbate_species:

  • Description: List of adsorbate species to consider.

  • Type: (list[str])

  • Example:

['H', 'H2O']

Struct_Filters - [struct_filters]

Settings for filtering out incorrect structures.

Attention

This section is optional.

No_Neighbors - [struct_filters.no_neighbors]

Filter for structures with atoms that have no neighbors.

• cov_rad_multiplier:

  • Description: Multiplier applied to the covalent radii to be used as cutoff radius for the neighbor check.

  • Type: (optional, float)

  • Default: 1.2.

Layer_Distance - [struct_filters.layer_distance]

Filter for layer distances in surface slabs.

Attention

This section is optional.

• max_layer_distance_ang:

  • Description: Maximum accepted distance between layers in Angstrom.

  • Type: (optional, float)

  • Default: 4.0.

Duplicate_Slabs - [struct_filters.duplicate_slabs]

Filter for duplicate slabs.

Attention

This section is optional.

• tolerance:

  • Description: Tolerance for the duplicate slabs filter.

  • Type: (optional, float)

  • Default: 0.2.

Vacancies - [vacancies]

Settings for vacancy generation.

• filter_struct_types:

  • Description: Types of structures to which vacancies will be applied.

  • Type: (list[str])

  • Default: ['bulk', 'surface'].

  • Possible values are: bulk, surface, cluster.

• limit_max_num_vacancies:

  • Description: Maximum number of structures with vacancies to generate.

  • Type: (optional, int)

  • Default: 400.

• num_repeats:

  • Description: Number of repeats for each structure with different random vacancies.

  • Type: (int)

  • Default: 3.

• max_vacancy_percentage:

  • Description: Maximum vacancies to generate as a percentage of the total number of atoms.

  • Type: (float)

  • Default: 0.75.

• min_vacancy_percentage:

  • Description: Minimum vacancies to generate as a percentage of the total number of atoms.

  • Type: (float)

  • Default: 0.025.

• element_list:

  • Description: List of elements to consider for the vacancies.

  • Type: (list[str])

  • Example:

['O']

Targeted_Modification - [targeted_modification]

Settings for targeted structural modifications.

Attention

This section is optional.

Central_Atom_Octahedral - [targeted_modification.central_atom_octahedral]

Apply perturbations to the central atom in octahedral sites.

Attention

This section is optional.

• filter_phases:

  • Description: Only apply the modification to the following phases.

  • Type: (optional, list[str])

  • Example:

['rutile', 'original_IrO2']

• filter_struct_types:

  • Description: Types of structures to which the modification will be applied.

  • Type: (optional, list[str])

  • Default: ['bulk', 'surface'].

  • Possible values are: bulk, surface, cluster.

• central_element:

  • Description: Symbol of the central element of the octahedral site.

  • Type: (optional, str)

  • Example: 'Ir'.

• num_repeats:

  • Description: Number of repeats for each structure with different perturbations.

  • Type: (optional, int)

  • Default: 3.

• limit_max_num_modifications:

  • Description: Maximum number of modified structures to generate.

  • Type: (optional, int)

  • Default: 200.

• max_perturbation_ang:

  • Description: Maximum perturbation movement of the central atom in Angstrom.

  • Type: (optional, float)

  • Default: 0.2.

Concave_Hull - [concave_hull]

Settings for descriptors and concave hull generation.

Attention

This section is optional.

• gen_concave_hull:

  • Description: Whether to generate the concave hull of the descriptors for all structures in the database.

  • Type: (optional, bool)

  • Default: False.

• descriptor:

  • Description: Descriptor to use for the concave hull generation.

  • Type: (optional, str)

  • Default: 'SOAP'.

  • Possible values are: SOAP, MACE.

• dim_reduction:

  • Description: Dimensionality reduction method for the concave hull generation.

  • Type: (optional, str)

  • Default: 'autoencoder'.

  • Possible values are: PCA, autoencoder.

• plot_filename:

  • Description: Filename for the figure displaying the concave hull.

  • Type: (optional, str)

  • Default: 'descriptors_concave_hull.png'.