Phase identification by X-ray diffraction means comparing unknown measured data with known reference data. This reference data is usually taken from one or several databases. The comparison process is often called search-match as well.
The quality and the contents of the different reference databases vary considerably, depending on the data quality and on the editorial process. Generally speaking there is a relation between price and quality.
Measured and calculated reference data
Measured reference patterns often score better for phase identification. The data source is a real powder diffraction measurement made on a pure, carefully selected and characterized material. This is possible even when the crystal structure is unknown. Some databases contain the original powder scan (reference scan) next to the stick pattern.
Calculated reference patterns always come with crystal structure data, as required for Rietveld refinements or structure analysis. The data source is a single crystal experiment most of the time. The theoretical stick pattern is calculated from the crystal structure, and, sometimes, optimized for search-match.
Commercial and free reference patterns
The reference patterns in the commercial databases have gone through a profound editorial process with automatic checks on data quality and consistency. They have quality marks and cross references attached and they are reedited from time to time. Typical examples are the PDF2 (measured data) and the PDF4+ (measured and calculated data) products from the ICDD.
A free reference pattern is either just a measurement of a pure phase used for search-match purposes; or it is calculated from a known (single) crystal structure. The quality is usually not documented and often lots of additional (physical and chemical) data are absent. Collections of free measured reference patterns are often small and serve dedicated purposes like for instance kidney stone characterization.
Crystal structure only databases
These databases cannot be used directly for phase identification. Typical examples are the commercial ICSD and the free COD crystal structure databases. They are used for standardless phase quantifications with the Rietveld method.
A free-of-charge alternative is based on the COD crystal structures and can be downloaded ready for use, from the COD webpage.
Collections of free calculated patterns can be big. Their size only depends on the number of crystal structures.
Special and user reference databases
These databases are often small and contain measured pure substances or even phase mixtures. The HighScore software supports any number of user-defined reference databases to be used either alone or in any combination. Another example of a commercial database for a specific purpose is Malvern Panalytical's CanDI-X. This search-match database is based on measured controlled substances, which have been characterized with other methods.
Reference databases for phase identification can be installed on all supported Microsoft Windows operating systems.
The database sizes vary considerably, depending on release year and on database contents (with or without structural data, with or without experimental data, with more or less chemical and physical properties).
PANalytical supports the yearly database releases without requiring software updates. Databases you can not directly read or convert yourself are issued once a year by Malvern Panalytical (COD).
The following search-match reference databases are supported by the HighScore software. They can be used either alone or in combination**:
- PDF-4+, yearly releases by the ICDD: ca. 350,000 reference patterns; 250,000 with atomic coordinates; ca. 8.0 GB disk space (+ 2.0 GB for optimized access*).
- PDF-2, yearly releases by the ICDD: ca 260,000 reference patterns; no atomic coordinates (but links to 150,000 NIST and ICSD patterns); ca 4.0 GB disk space (+ 1.0 GB for optimized access*)
- PDF-4/Organics, yearly releases by the ICDD: ca. 500.000 reference patterns; 50,000 with atomic coordinates; ca 20 GB disk space (+ 3.0 GB for optimized access*).
- PDF-4/Minerals, yearly releases by the ICDD: ca. 45,000 reference patterns; 30,000 with atomic coordinates; ca. 2.0 GB disk space (+ 1.0 GB for optimized access*).
- Free Crystallographic Open Database, yearly conversions from Malvern Panalytical on the COD website: > 300,000 reference patterns; >300,000 with atomic coordinates; available in two zipped downloads, ca. 4.0 GB disk space**.
- Any number of user reference datatabases, either created in HighScore from experimental peak lists or with the Plus option from available structure data in CIF format.
The following crystal structure databases are only supported by the Plus option for HighScore:
- ICSD, inorganic crystal structures database from FIZ in Karlsruhe, Germany, 3 GB disk space
- COD, free crystal structure database as a collection of CIF files, ca. 18 GB.
(Note: These databases contain only crystal structure data and no reference patterns; they are NOT suitable for search-match purposes!)
* Due to the license construction the original ICDD databases must be kept installed in their unconverted format, next to the converted format which enables a much faster access from HighScore. This requires some additional space on the hard disk.
** ICDD does not allow the use of their database products in combination with the COD database.