Structured Representation of Materials¶
We present an example of our approach towards defining and storing structured data for materials. The aspects presented herein complement the general introduction.
Example Representation¶
In the expandable section below, the user can find an example JSON representation of a face-centered cubic Silicon:
Expand to view
{
"$id": "material",
"$schema": "http://json-schema.org/draft-07/schema#",
"title": "material schema",
"type": "object",
"allOf": [
{
"description": "in-memory entity",
"$ref": "in_memory_entity/named_defaultable_has_metadata.json"
}
],
"properties": {
"formula": {
"description": "reduced chemical formula",
"type": "string"
},
"unitCellFormula": {
"description": "chemical formula based on the number of atoms of each element in the supercell",
"type": "string"
},
"basis": {
"description": "object containing an array of elements and coordinates in the supercell",
"$ref": "properties_directory/structural/basis.json"
},
"lattice": {
"description": "object defining the lattice cell size and shape",
"$ref": "properties_directory/structural/lattice.json"
},
"derivedProperties": {
"description": "object defining the lattice cell size and shape",
"$ref": "properties_directory/derived_properties.json"
},
"external": {
"description": "information about external sources of this material",
"$ref": "system/database_source.json"
},
"src": {
"description": "File information if material was exported from POSCAR, XYZ or UPF file.",
"$ref": "system/file_source.json"
},
"scaledHash": {
"description": "Hash string for a scaled structure with lattice vector a set to 1 (eg. for materials under pressure).",
"type": "string"
},
"icsdId": {
"description": "Corresponding ICSD id of the material",
"type": "integer"
},
"isNonPeriodic": {
"description": "Whether to work in the finite molecular picture (usually with atomic orbital basis)",
"type": "boolean"
}
},
"required": [
"basis",
"lattice"
]
}
{
"_id": "LCthJ6E2QabYCZqf3",
"basis": {
"coordinates": [
{
"id": 1,
"value": [
0.0,
0.0,
0.0
]
},
{
"id": 2,
"value": [
0.25,
0.25,
0.25
]
}
],
"elements": [
{
"id": 1,
"value": "Si"
},
{
"id": 2,
"value": "Si"
}
],
"name": "basis",
"units": "crystal"
},
"createdAt": "2016-04-03T05:25:37.430Z",
"creator": {
"...": "include(system/creator.json)"
},
"derivedProperties": [
{
"name": "volume",
"units": "angstrom^3",
"value": 131.1
},
{
"name": "density",
"units": "g/cm^3",
"value": 2.33
},
{
"name": "symmetry",
"spaceGroupSymbol": "Fd-3m",
"tolerance": {
"units": "angstrom",
"value": 0.3
}
},
{
"degree": 10,
"name": "p-norm",
"value": 0.71
},
{
"element": "Si",
"name": "elemental_ratio",
"value": 0.71
},
{
"name": "inchi",
"value": ""
},
{
"name": "inchi_key",
"value": ""
}
],
"exabyteId": "RKiswuxHKkF3uvLyA",
"formula": "Si",
"hash": "4ac3fa057e3530628eb1295c77110eaa",
"isDefault": false,
"isNonPeriodic": false,
"lattice": {
"...": "include(properties_directory/structural/lattice.json)"
},
"name": "Silicon FCC",
"owner": {
"...": "include(system/owner.json)"
},
"scaledHash": "5bd4eb168e3530628eb1295d88221cbb",
"schemaVersion": "0.2.0",
"slug": "silicon-fcc",
"unitCellFormula": "Si2",
"updatedAt": "2016-04-04T17:58:42.867Z"
}
Explanation of Keywords¶
| Keyword | Short Description | Details |
|---|---|---|
| basis | Crystal basis with explicit identification per atom | The information about the atomic type and coordinates |
| lattice | Crystal lattice in both Bravais and vector notations | Crystal lattice parameters - lattice constants and angles. Components of the corresponding lattice vectors are also included. |
| derivedProperties | descriptive properties derived from lattice/basis (only one example shown above) | Additional properties of the crystal structure under investigation as explained in the section ensuing the present table. |
| hash | Hash string calculated by the Bank Mapping Function | Structure-based hash string for the primitive standard representation of this material, calculated when checking this material against existing entries within the Materials Bank |
| scaledHash | As above, but for the lattice axis scaled to 1.0 (i.e. to identify same structures under different uniform pressure) | This hash string is calculated by scaling all the dimensions of the primitive unit cell representation of the material by the \(a\) lattice constant |
| isNonPeriodic | Boolean value used to describing whether a structure exists in a single unit or repeating units. | The isNonPeriodic Boolean value is assumed to be false by default, indicating that a material is periodic. The value of isNonPeriodic, determines which derivedProperties are calculated for a structure. For example: unit cell volume is calculated for periodic, but not non-periodic structures. Conversely, International Chemical Identifier (InChI)1 are calculated for non-periodic, but not periodic structures. |
Derived Properties¶
As seen above, we use the crystal lattice and basis JSON objects as the main identifying properties. Based upon them, we calculate the derivedProperties, that may include such information as:
- the unit cell volume,
- density,
- chemical formula,
- and a large number of other possibilities.
For every material imported/uploaded to our platform, we pre-calculate a set of such descriptors, and store them inside this "derivedProperties" section. This information can be further used during data analysis or the construction of statistical predictive models.