MINERALS IN THE INORGANIC CRYSTAL STRUCTURE DATABASE

Rudolf Allmann,

Institut für Mineralogie, Hans Meerwein Str., D35032 Marburg, Germany

Keywords: Minerals, mineral groups, database, IC SD, PDF, printing errors

The Inorganic Crystal Structure Database ICSD contains about 46,000 inorganic crystal structures (excluding metals), including nearly 9,000 mineral structures. It was started by Prof. Bergerhoff, Bonn, and is now maintained by the Fachinformationszentrum (FIZ), Karlsruhe. Within the last two years I could check the mineral structures and insert mineral-specific additions as mineral and group names, localities, rneasured densities, and a cross-reference to the PDF-database of X-ray powder patterns. For the great group of more than 800 zeolites also the subgroups are specified (as in PDF).Of about 500 minerals the mineral names were missing at all, even that their structures were present in ICSD. About 350 overlooked papers with mineral structures were newly typed in.

As compared with the "Structure Reports" (last volume for 1990) ICSD contains the (thermal) displacement factors too, in many cases even the deposited ones, and additional data supplied from the authors. Until now isotropic and anisotropic displacement factors cannot be cited simultaneously in ICSD and the "Retrieve"-program is still lacking a routine for calculating equivalent isotropic B values from the given anisotropic ones. Also the X-ray density DX is not calculated. At FIZ the complete transfer of the corrected and completed mineral data onto CD-ROM (updated twice per year) will soon be carried through. I want to ask all authors to check occasionally the structures in ICSD originating from their own papers and to submit found errors and additional data (very often the non-published displacement factors).

The extracted atomic coordinates in ICSD fortunately show rather few errors because of the careful checking process for these data. The (anisotropic) displacement factors were checked less precisely and in some cases the mixed B_ik-values are still to be bisected. Especially for the B_ik of special positions in trigonal and hexagonal space groups the errors of authors, printers and abstractors sum up. Nearly every 3^rd entry wás erroneous, because the symmetry constraints for these space groups were not handled correctly. For nearly 700 entries these errors were corrected as far as possible.

But more bothering are the rather frequent printing errors in the papers itself. These are above all:

1. missing signs (in one case 30 signs of co-ordinates were missing, which only could be corrected by recourse to the original thesis),
2. missing leading zeros (e.g. .3 instead of .03),
3. interchanged digits (e.g. .169 instead of .196), and
4. wrong duplication of digits (e.g. .336 instead of .366).

Sometimes for an easier comparison with related structures, especially for minerals, unusual space group settings were chosen (e.g. C 1- or F 12/d 1 ), which not in all cases were handled correctly by the abstractors.

As a by-product of the mentioned checking procedure an equivalence list of cross references was drawn up for about 2,400 minerals, wherein the PDF- and ICSD-numbers (with the lowest R-values) are given.