Refinement Options

These are the options that can be passed from to the refinement_dict

f0j_source

Source of the atomic form factors. The computation_dict will be passed on to this method. See the individual files in f0j_sources for more information, by default ‘gpaw’ Options: ‘gpaw’, ‘iam’, ‘gpaw_mpi’ Some limitations: ‘gpaw_spherical’

reload_step

Starting with this step the computation will try to reuse the density, if this is implemented in the source, by default 1

core

If this is implemented in a f0j_source, it will integrate the frozen core density on a spherical grid and only use the valence density for the updated atomic form factor options are ‘combine’, which will not treat the core density separately, ‘constant’ which will integrate and add the core density without scaling parameter and ‘scale’ which will refine a scaling parameter for the core density which might for systematic deviations due to a coarse valence density grid (untested!) By default ‘constant’

extinction

Use an extinction correction. Options: - ‘none’ means no extinction correction - ‘shelxl’ use the (empirical) formula used by SHELXL to correct to correct for extinction - ‘secondary’ see Giacovazzo et al. ‘Fundamentals of Crystallography’ (1992) p.97

By default ‘none’

max_dist_recalc

If the max difference in atomic positions is under this value in Angstroems, no new structure factors will be calculated, by default 1e-6

max_iter

Maximum of refinement cycles if convergence not reached, by default: 100

min_iter

Minimum refinement cycles. The refinement will stop if the wR2 increases if the current cycle is higher than min_iter, by default 10

core_io

Expects a tuple where the first entry can be ‘save’, ‘load’, ‘none’ which is the action that is taken with the core density. The second argument in the tuple is the filename, to which the core density is saved to or loaded from

cutoff

Expects a tuple of three values where the first two are strings and the last one is a float value. First string is a cutoff mode. Currently available are ‘none’ where all reflections are used, ‘sin(theta)/lambda’ where the cutoff is set according to a user given resolution, ‘fraction(f0jval)’ where the resolution cutoff is set to include a certain fraction of the mean absolute valence atomic form factors. ‘I/esd(I)’ can be used for excluding values based on the value over estimated standard deviation The second string can be either be ‘above’ or ‘below’ and denominates in which direction values will be excluded. The final value is the actual cutoff value.