The atomic rates for nickel are less accurate, and less well debugged, than those for other elements. It is recommended that any model be run first with the nickel abundance set to zero. Substantial differences between models with zero and non-zero nickel, in terms of temperature, opacity, etc., should be treated with great caution.
Similar comments apply to elements introduced in version 2.2: Li, Be, B, odd-Z elements between F and K, plus the iron-peak elements (except for Fe itself). Many of the rates for these elements are scaled hydrogenic, most ions with 3 or more electrons have a level structure which is assumed hydrogenic. Quantitative results affected by these elements should be regarded as reliable only for highly ionized models, which are dominated by H- or He-like ions.