Internal H-bond in Fe $^{\mathrm{II}}$-H$_2$O$_2$ complex

The isolated $\left[(\mathrm{H}_2\mathrm{O})_5 \mathrm{Fe^{II}}(\mathrm{H}_2\mathrm{O}_2)\right]^{2+}$ complex is internally stabilized by a hydrogen bond between the $\beta$-oxygen of the peroxide and an adjacent water ligand, see figure 5.2. The produced OH. radical is in the gas phase immediately quenched by H abstraction from that water ligand.[145] To investigate this shortest possible OH. radical transfer without any intervening solvent molecules we started a second AIMD simulation, from the last configuration of the equilibration run of the previous simulation and induced this hydrogen bond by constraining the distance from the H$_2$O$_2$ $\beta$-oxygen to an H of an adjacent H$_2$O ligand to the value of $R_{\mathrm{OH}}=2.06$ Å from ref franco1. The Fe-O bond constraints were released after 1.5 ps. and the remaining peroxide $R_{\mathrm{OO}}$ bond and the induced hydrogen bond were released after 1.93 ps. Also in this simulation the peroxide dissociates almost immediately. The hydroxyl radical remains hydrogen bonded to the water ligand for 300 femto-seconds and then abstracts the hydrogen (as in the gas phase computation) to form a water molecule and again the iron(IV)dihydroxo complex. The simulation was continued for 6 pico-seconds more in which occasionally a proton was donated by a H$_2$O ligand to the solvent and vice versa (equation 5.4) but this time no iron(IV)-oxo complex was formed. Also exchange of one of the water ligands with a solvent molecule was observed once in this simulation, but no migration of a ligand to the second coordination sphere.[178]

The spontaneous formation of the internal hydrogen bond in the pentaaqua hydrogen peroxide complex is not expected to be as likely in the solvent as it is in the isolated situation, because the hydrogen peroxide ligand can form hydrogen bonds with solvent waters instead of forming the five membered ring (see figure 5.2) via the adjacent water ligand. However, the reaction path via the internal hydrogen bond illustrates a special case of the OH. radical quenching, by transformation of a water ligand into an hydroxo ligand, namely a path without any solvent molecules involved.

We observe that the iron(IV)(di-/tri-)hydroxo complex, which in the first simulation transformed to the iron(IV)-oxo moiety soon after the first step, does not do so in the present simulation. The reason for this spread in the lifetime of the hydroxo complex can be understood if we assume that reaction 5.5 is only likely to occur if the system finds itself on the left-hand-side of reaction 5.4, i.e. from the dihydroxide. The complex should not already have donated a proton from a water ligand to the solvent to form the trihydroxide. This assumption is endorsed by the reaction energy $\Delta E=65$ kcal/mol of gas phase reaction 5.7,

$$\left[\mathrm{(\mathrm{H}_2\mathrm{O})_3Fe^{IV}}(\mathrm{OH... ...athrm{O})_3Fe^{IV}O}(\mathrm{OH})_2\right] + \mathrm{H_3O^+}$$ (65)

Although this number is of course modified in aqueous solution due to the solvent effects, it is clear that a second hydrolysis after one of the water ligands has already been hydrolyzed is energetically very unfavorable. If hydrolysis of a water ligand has already taken place, it can take some time before the proton traversing through the solvent ends up on the complex again, providing a new possibility for reaction 5.5 to occur in which the iron oxo is formed. At high pH the equilibrium of reaction 5.4 will move to the trihydroxo complex side, which provides us with an explanation for the reduced reactivity of the aqueous Fenton reagent at pH $>$ 5.[176] Note however, that the main reason to employ non-chelated Fenton chemistry at pH as low as 2-3 is the very low solubility of iron at higher pH.