Data curation

Data for ICSD are usually all published in peer reviewed journals. The crystal structure data is either taken from stored research data in the joint depot from CCDC and FIZ Karlsruhe or it is taken from the publications themselves. In some cases, the authors are asked to provide us with the electronic structure data.
If the ICSD data are generated from crystal structure data deposited by scientists in the joint depot of CCDC (Cambridge Crystallographic Data Centre) and FIZ Karlsruhe, the original data set remains available there as well. These data are linked to each other.

The following diagram shows the steps that a crystal structure has to go through from publication in a scientific journal to inclusion in the ICSD.

In ICSD, all data are available as CIF, the standard format for crystal structure data defined by the IUCr.

An ICSD record consists of the following data at least:

• Space group/symmetry
• Lattice parameters
• Atom positions containing the following information for each position: element, label, multiplicity, Wyckoff position, oxidation number, coordinates (including standard deviations if available), and the site occupation factor
• Experimental information (temperature and pressure, single crystal or powder measurements, etc.)
• Bibliographic information (authors, journal, article title, publication year, volume, page numbers)
• Sum formula, structured formula
• Compound name (but do not expect full-fledged IUPAC names)

A record may also contain additional information (if available):

• Mineral name and locality
• DOI and Abstract
• Keywords
• Isotropic and/or anisotropic displacement parameters with standard deviations
• R-value
• Comments/Remarks
• Structure type assignment by the authors 

The content goes through an intensive review process. First automatic checks are meant to reveal inconsistencies in the data by testing the plausibility of several properties like the elementary cell, space group and symmetry, oxidation states, multiplicities, electroneutrality, sum formula, measured and calculated densities, etc. Failed tests may produce warnings or comments to provide more information on potential problems with the structure.

The review process includes among other things the correction of syntax errors, checks for duplicate records, assignment of structure types and chemical names, allocation of properties or possible application of the materials based on the publication.
The following data are often missing in publications and are generated automatically based on the crystallographic data: Wyckoff sequence, Pearson symbol, ANX formula, AN formula, sum formula, standardized data, density, atomic distances, etc.

However, the published crystal structure data is not changed, the structures are recorded in the ICSD database “as given in article”. Only comments on possible inconsistencies are added. Through our continuous quality assurance, existing content is modified and supplemented, and duplicates are removed. As a result, and by filling gaps from previous years, even the older content is not static.