Choice of oscillation range

The choice of oscillation range (=delta-phi for each frame) reflects a compromise between conflicting targets - in order of importance:

1. preventing reflection overlap at high resolution (see table at end of IDXREF.LP) - for a given total rotation range, lower oscillation range is better
2. improving the signal-to-noise ratio - lower oscillation range is better (but see below)
3. minimizing the readout noise (should be 2-10 counts per pixel for CCD, and 0 for pixel array detectors) - larger oscillation range is better (only for CCD)
4. minimizing the disk space and the readout time - larger oscillation range is better

The following aspects need to be considered when choosing the oscillation range:

• crystal mosaicity: choosing the oscillation range less than (crystal mosaicity / 2) will not further increase the signal-to-noise ratio - see Müller et al (2012) Acta Cryst. D68, 42 (open access)
• the maximum oscillation range to prevent angular overlap can be calculated from cell parameters and crystal orientation, and the mosaicity. Specifically, the mosaicity must be subtracted from the value given in the table at the end of IDXREF.LP which takes all of this into account. If the mosaicity is bigger than the table value, of course a positive value for the oscillation range must be chosen, but there will then be overlaps.

In the absence of mosaicity information (or assuming a mosaicity of 0.3° or less), a reasonable default value for crystals that diffract to 2Å resolution or less, and have longest (primitive!) unit cell axis of 150Å or less is, in my opinion and as suggested in publications by Phil Evans and Jim Pflugrath (which should be referenced here ...):

• around 0.25-0.5° for CCD detectors
• around 0.1° for the Pilatus detector