Work on CIF began in the late 1980’s as a mechanism to communicate small-molecule single-crystal crystallographic results from the International Union for Crystallography (IUCr). Indeed entire short crystallographic articles could be placed in a CIF for submission to IUCr Journals. At that time CIF was an abbreviation for Crystallographic Information File, so a reference to a “CIF file” was a bit redundant.
CIF is intended to be read and even edited by humans. The basic structure of CIF is that labels, known in CIF as data names, are associated with values, as shown in these examples:
_cell_length_c 12.32318(8)
_cell_angle_alpha 90.0
_symmetry_cell_setting orthorhombic
_space_group_name_H-M_alt ’P 41 21 2’
_computing_structure_refinement
; GSAS-II (Toby, Brian H. & Von Dreele, Robert B.;
Journal of Applied Crystallography 46, 544-549, 2013)
;
The above examples show inclusion of numerical values with and without a s.u. value, and several representations for character strings. Note that quotation marks are required when a string contains spaces, and the semicolon notation is used for strings that span multiple lines of text. The use of white space including line breaks is solely for the convenience of human readers. A single space serves the same purpose as multiple spaces and new-line characters are not distinguished from spaces for the purpose of parsing a file.
To specify a series of values to be associated with a data name, the loop_ keyword is used:
loop_ _space_group_symop_operation_xyz
’x, y, z’
’-x, -y, z+1/2’
’-y+1/2, x+1/2, z+1/4’
A table of values for several data names can also be created using the loop_ keyword:
loop_
_atom_site_type_symbol
_atom_site_fract_x
_atom_site_fract_y
_atom_site_fract_z
Er 0.87730(5) 0.35493(5) 0.13550(5)
Ge 0.90666(11) 0.15231(13) 0.61955(12)
There are two special placeholder values used in CIF. A value of “?” (a question mark, which must be used without quotes) indicates that a value is unknown or is unspecified. A value of “.” (a period, which must also be used without quotes) indicates that the data name does not have an appropriate meaning in the specified context:
_diffrn_ambient_temperature ? # not measured
_chemical_melting_point . # decomposes on heating
CIFs are organized in blocks, where each block is an independent collection of data items and their values. Note that the beginning of a block is indicated by an name that begins with the string data_, but the name that follows the underscore is considered arbitrary and may be changed when the CIF is archived. This is not a complete guide to the syntax of CIF, but this should provide a good overview as to what is required.
CIF was unusual when created, if not unique, in that every CIF data name was given a formal and permanent definition in a set of documents called the CIF data dictionaries. Should a definition need to be changed, a new CIF data name is coined.
By the mid-1990’s, the utility of CIF had been demonstrated and work was began to develop CIF data dictionaries for applications beyond small-molecule crystallography. As this was done, CIF was renamed as the Crystallographic Information Framework, because CIF was taking on the role of becoming a formal language for communication of crystallography, not just a file format. This also has the benefit that the term “CIF file” was no longer a redundancy.
The pdCIF dictionary, with definitions for many items needed for powder diffraction, was one of the first to be started and took me the better part of a decade. One of the major problems was that a data name can only be associated with a single value or a single series of values within a CIF block. That rule means that, for example, lattice parameters can only be defined once in a block. Thus, to report a powder refinement with two crystallographic phases, a separate block would be needed for each phase. A mechanism was developed for pdCIF to establish logical connections between CIF blocks. Another reason that this project took so long was that, as it progressed, more stringent rules were created for the CIF data dictionary to enforce database conventions, so completed work sometimes needed revisions.
CIF has been a major success story. For at least a decade, CIF has been the universal standard for communication of crystallographic information. Most journals require deposition of crystallographic results in CIF for electronic archival. Better journals require deposition of a CIF that includes the data used for the fit and documents the analysis process. I frequently hear people use the term CIF as a synonym for a crystal structure. To the best of my knowledge, the only other field that has been as successful as crystallography in sharing data has been astronomy, but astronomers do have the advantage that they are all looking at the same specimen (the sky).