search

COMCIFS / Powder_Dictionary

CIF definitions for powder diffraction
4 stars 4 forks source link

Should we record specimen displacements? #57

Open rowlesmr opened 1 year ago

rowlesmr commented 1 year ago

an offshoot of #53

Sort of related to angle calibrations is how to record specimen displacements.

  1. Do we need to?
  2. I think there are three displacements (units of mm): 20230105_133845
    • Consider the above dodgy diagram to be the side view of a conventional Th-Th BB diffractometer; X-rays coming from left and going right
    • Equatorial - Displacement perpendicular to 2Th = 0, in the plane of diffraction; ie specimen displacement in Bragg-Brentano. This definition needs some work if you're using a Th-2Th goniometer.
    • Lateral - Displacement parallel to 2Th = 0; ie increasing/decreasing the effective specimen-detector distance. You can get this with capillaries in Debye-Scherrer
    • Axial - Displacement perpendicular to the other two directions. (if equatorial is up/down, and lateral is left/right, axial is toward/away from you)
  3. These displacements also need to have specimen-detector distances recorded. I think that appropriate data items would be
    • _pd_instr.detector_circle_radius
    • _pd_instr.dist_spec_anal
    • _pd_instr.dist_spec_detc
    • _pd_instr.dist_spec_vdetc
  4. Where to put such displacements? I think PD_SPEC is the best place. I don't think the PD_INSTR is appropriate. PD_PROC_LS doesn't seem right, as although it is often a refined parameter, it does refer to a "real" thing. PD_SPEC is a Set category; do we need to be able to provide displacements per detector_id? if so, then we'd need to make something like PD_SPEC_DISPLACEMENT or PD_CALIB_DISPLACEMENT .
  5. Can we provide something like a _pd_spec.displacement_geometry to allow for the automated calculation of corrected 2theta values ? With an enumerated choice: "bragg-brentano", "debye-scherrer", "seeman-bohlin", ... , "other". Other would require _pd_calibration.conversion_eqn to be defined.

(I promise to stop thinking of things now... :))

rowlesmr commented 1 year ago

ooooor, we can just explicitly put in the angle offset due to any kind offset with PD_XCOORD.

jamesrhester commented 1 year ago
  1. I think if sample displacements are a parameter used in refinement or angle correction they should be recorded.
  2. Agreed
  3. Didn't we define these already??
  4. PD_SPEC is good. I don't think a specimen would be per-detector.
  5. I think we go in terms of the three axial directions you've already indicated. _pd_instr.geometry would be the right place to describe the geometry, but this is currently a free-text entry. We should make a new one with recognised values, e.g. _pd_instr.standard_geometry