Figure 3. Energy diagram for nucleophilic attack of water molecule on oxyl oxygen in HO-FeIV-O• moiety to form Fe-OOH at the Fe1 site : (a) initial complex containing adsorbed water molecule, (b) transition-state complex, and (c) resulting hydroxylated complex. Integer numbers in energy diagram show the total energy differences (in kcal/mol).
Hydroxylation
Terminal oxyl oxygen disappears under the electrophilic water attack since the oxygen is exposed to the proton moiety of the water molecule. The only way to arrange such an approach of water molecule is to adsorb it first on the neighboring metal cation (Figure 4a) to avoid undesired interaction with hydroxyl group on Fe1. When oxyl oxygen in the HO-FeIV-O• moiety abstracts proton from adsorbed water molecule, two hydroxyl groups appear on metal sites (Figure 4c), rather than the Fe-OH moiety and free OH radical as it might be expected for the water oxidation route by analogy with the abstraction of hydrogen from methane on oxyl oxygen producing methyl radical.[15] Therefore, the low barrier process (9 kcal/mol) has to be assigned to the dissociative adsorption or hydroxylation.
The same is true for electrophilic water attack on ferryl oxygen (Figure 5). For the latter case, the abstraction of hydrogen by ferryl oxygen to form two hydroxyl anions goes through even a lower barrier of 4 kcal/mol (Figure 5). This might be explained by more nucleophilic nature of ferryl oxygen preferable for proton abstraction. Nucleophilicity of ferryl and oxyl oxo sites can be compared by means of the difference between the energy of their calculated 1s(O) level. The latter is 1.2 eV lower than the former, which means less negatively charged oxyl oxygen (Table S1). This explanation is in line with intuitive picture of the oxyl oxygen appearance via “back” transfer of electron from from ferryl-state oxygen to iron center. Mulliken charges on oxyl and ferryl oxygen agree surprisingly good with the core level energies.
Although the hydroxylation of oxyl oxygen in the HO-FeIV-O• group and neighboring FeIII center has to form two chemically equivalent sites having two hydroxo ligands (Figure 4c), their spin density (and so the oxidation state) remains different at all stages of process (Figure 4). However, a crucial change takes place for the Fe2 site despite the fact that is not directly involved in the process. The spin density at Fe2 drops from 4.21 to 3.40. Taking into account that energies of 1s level for the iron centers with spins 3.34 (Fe1) and 3.40 (Fe2) (Figure 4c) are equal within 0.01 eV (Table S1), one might guess that the oxidation state of Fe2 is 4+ coinciding with that for Fe1 (Figure 4a). What is interesting, initial oxyl containing cubane has oxidation states Fe4(IV,III,III,III) (Figure 4a). Therefore, the hydroxylation changes this configuration to Fe4(IV,IV,III,III) (Figure 4c) implying delocalization of spin over the oxo bridges. This result reveals unusual effect that the dissociative adsorption of water (which normally proceeds without any electron transfer) on cluster affects the oxidation states of connected iron centers. This effect is certainly connected with the partial disruption of cubane structure as seen from the elongation of one of edges by almost 1 Å (Figure 4c). In case of hydroxylation of the ferryl-oxo cubane the oxidation scheme is Fe4(IV,IV,III,III) from the beginning at each steps of the process (Figure 5).
Worthwhile noting that the account of solvation does not change much the structure and relative energies of above given process of water dissociation. For the processes (in both oxyl and ferryl cases) modeled without such account, the barrier is only 1-2 kcal/mol larger than that for the model with solvation (Figure S4-S5).