Data Reduction

Data reduction with CrysalisPro

1. Play the frames to get a sense of the diffraction quality.

2. Inspect results (3rd button from top)

• • Bottom of page, check:

    • Completeness (100 % to 0.84 angstroms)

    • Redundancy - rule of thumb: 4 for empirical absorption correction.

    • F2/Sig F2 (signal/noise) - higher is better

    • Rint - rule of thumb: maybe 5% for small molecule (?))

    • Rsigma (?)

  • Red graphs

    • Examine parameters looking for frames to discard, which can be done in refinalize step

3. Refinalize step

• • Space Group determination - Can sometimes be ambiguous. Considerations:

    • Laue symmetry. Removing translations from the space group symmetry and adding a center of symmetry results in a Laue point group, which is the symmetry of the diffraction pattern.

    • Centric or Acentric. Centric space groups include a center of symmetry (that is, an inversion operator, (-x,-y,-z)). Chiral compounds cannot crystallized in centric space groups. There are statistical tests that can distinguish between centric and acentric space gruops, which can help in space group determination.

      • The main criterion for distinguishing between centric and acentric space groups are E-statistics. E^2 is a normalized intensity (E^2=F^2/<F^2>). For centric space groups, <|E^2-1|>=0.986, while for acentric space groups, <|E^2-1|>=0.736. This represents, in some sense, the width of the distribution of normalized intensities, and says that for centric space groups, the intensities are distributed over a wider range of values than for acentric space group. There are theoretical expectations for other powers of E, and these all go into a prediction of whether the space group is centric or acentric. These calculations are carried out within CrysalisPRO, and also in Olex2 and WinGX.

      • Example output from WinGX for a crystal that turns out to be space group P21 (acentric) via Model -> E-Statistics. This output can also be found in Olex2 via Info -> Reflection Statistics -> Cumulative Intensity (note that this plot can also be used to learn whether the crystal is twinned).

• • • All of the programs combine the information from systematic absences, Laue symmetry and E-statistics to come up with a best estimation of the space group. Sometimes it's still ambiguous and the correct space group is only determined via structure determination (i.e. try the structure solving with each of the possible space groups).

• • Exclude frames

    • Outlier rejection

      • Try first without. 

      • If centrosymmetric space group -> Friedel mates equiv

    • Limits etc

      • Exclude bad runs determined above.  Run number starts at zero. (I.e. 41 runs go 0-40)

  • Apply absorption

    • Empirical Absorption (ABSPACK) - Adjusts frame scales to minimize Rint, but not true absorption correction

    • Analytical absorption correction - Model shape of crystal and atomic content to estimate orientation dependent absorption by crystal

  • Output structure factors

    • Check "Add hkl data to CIF" so the reflections are included in the CIF file (makes validation process easier)