How to loop unit cells & temperature values

[briantoby]

Back to working on CIF exports from GSAS-II.

Looking at a bug related to reporting the data measurement temperature, I am realizing that have a fundamental conflict in how to report that and unit cell constants from multiple dataset fits. Obviously one wants them to be listed in the same block with fractional coordinates -- without them the latter are useless, but when one has a multiple dataset fit, one likely has only one set of coordinates, but likely more than set of one unit cell constants and quite possibly more than one temperature value.

IMHO the right way to record this is with a loop_ structure with _cell_measurement_temperature, _cell_length_a,... and that loop can be included in the same block as the crystal structure, but I don't think these CIF items are allowed in loops. Probably makes sense to put a block pointer to the block with the raw data in that loop as well.

Temperature can certainly be included in the block with raw data, but the unit cell dimensions (possibly for more than one phase) do not belong there.

[jamesrhester]

Last year we put quite a few data names in place in both the core dictionary and powder dictionary to properly handle the various scenarios that come up in powder diffraction (see pdCIF issues and pull requests). We are, however, in dire need of a complex example, so I plan over the next month or so to create some examples, perhaps based on the GSAS-II tutorial data.

We have built flexibility in to the design, so that something like the loop you speak of (temperature and cell lengths in a single table) is allowed if the value of _audit.schema is not Default. But note that _cell_measurement.temperature is now deprecated, and only used if the cell is measured at a different temperature to the main experiment - you should use _diffrn.ambient_temperature for powder and any modern single crystal experiment.

[briantoby]

_diffrn*ambient_temperature removed from GSAS-II CIF exporter (see AdvancedPhotonSource/GSAS-II#78) (FWIW _cell_measurement.temperature was already in use).

[jamesrhester]

Looking at that code I think it is the other way around, i.e. cell_measurement. temperature has been removed (as it should have been.)

[briantoby]

Yes, indeed. Juggling too many open windows.

[jamesrhester]

A draft proposal covering some of these issues is available for comment and improvement.

[briantoby]

I did not look closely at this, but the CIF contents do look pretty familiar. What, if anything, has been changed?

[jamesrhester]

The CIF examples are those I used in the discussion around COMCIFS/Powder_Dictionary#171, adjusted to take into account your suggestion to put diffraction conditions together with the diffraction pattern (regardless of repetition) and likewise cell/space group information together in a single data block. So I believe the result will essentially be what you already do in GSAS-II. And yes, the CIF examples are based on GSAS-II tutorial examples.