Figure 7. Isomer shifts and absolute quadrupole splitting values for the 20 reference complexes (A) and selected signals from FeNC catalysts (B). In (A), calculated values are shown as open circles (ᴏ) and experimental values as filled squares (■); the error bars are taken as the mean absolute deviation. In (B), the labels differ from the nomenclature used in the original papers (S1 corresponds to D1, S2 and S4 to D2-like species, S3 to D3 in porphyrin-type catalysts, S5 to D3 often obtained after an ammonia treatment step or from MOF-based catalysts; roman numerals indicate doublets that originate from a specific treatment or appear only within distinct synthesis routes). The error bars in (B) correspond to standard deviations for the values of different synthesis routes where applicable; for individual data points, e.g. roman numerals, the error bar is taken as the 95 % confidence interval.
The challenge posed by these systems for the computational chemist is to identify structural and electronic models for the various FeNx sites that will permit the identification of the environment of the catalytically active site. Among the characteristic doublets, S1–S4 fall within a similar range of isomer shifts between 0.28–0.54 mm s−1. The variation within a group of Si signals from different preparations appears approximately as large as the deviation between groups. In contrast, signal S5 is well separated at an isomer shift of 0.98–1.01 mm s−1. With the B3LYP trust region for the isomer shift of 0.13 mm s−1, it is evident that S1–S4 will not be distinguishable by different computational models although it can be expected that one will be able to clearly identify S5. The quadrupole splitting values of the S1–S4 and S5 signals are spread over a wider range of 0.71–2.83 mm s−1. This not only allows a better distinction of the individual signals, but also provides clear evidence that different types of local environment are present for the various FeN4 sites observed experimentally. Taking the B3LYP trust region for the quadrupole splitting of 0.45 mm s−1, it appears conceivable that S1, S2 and S3 will be distinguishable while the exclusive consideration of the quadrupole splitting will not enable to differentiate between S2, S4, S5, IVx and IIIb. If for instance upon poisoning, a chemical connection between one of the Si doublets and a species labelled with a roman numeral in Figure 7B can be made, this could be used as additional information to better characterize the geometric and electronic structures of typical FeN4sites.