Starting from hydrogen peroxide coordinated to Fe $^{\mathrm{II}}$

The system of H$_2$O$_2$ coordinated to Fe$^{2+}$ in water was created from a previous CP-PAW MD run of Fe$^{2+}$ (high spin) in a periodic cubic unit cell with 32 water molecules, by replacing one of the six water ligands by hydrogen peroxide. A short MD run of 1.16 ps. averaged out the short memory of any unphysical forces or velocities arising from the construction of the system. The pentaaqua iron(II) hydrogen peroxide complex was found to be a local minumum on the potential energy surface in the gas phase[145] and has also been proposed as a stable intermediate in aqueous solution[167,176]. As we however found that the barrier for oxygen-oxygen lysis in water is very small, we fixed the hydrogen peroxide oxygen-oxygen distance at $R(\mathrm{OO})=1.50$ Å and the Fe-H$_2$O$_2$ distance at $R(\mathrm{FeO})=2.124$ Å, during this period of system equilibration, to prevent the premature breakup of the complex by the unrelaxed environment. Also the five iron-water ligand distances $R$(FeO) were constrained the first 0.65 ps to their value at the first time step ( $R(\mathrm{FeO})\approx2.2$ Å).

After the equilibration, the MD simulation was continued (without any bond constraints) and the evolution of the coordinated Fenton reagent in water was followed for 10.2 ps. Figure 5.3 shows five snapshots of the [(H$_2$O)$_5$Fe $^{\mathrm{II}}$(H$_2$O$_2$)]$^{2+}$ complex in the cubic unit cell at subsequent times during the first part of the simulation. Two of the solvent molecules have also been drawn, since they take part in the reaction to be described below; all other solvent molecules have been left out for clarity. The cubic unit cell is surrounded at all sides by its periodic repetitions, which provides a representation of the infinite solvent environment. In order to aid the discussion, we have drawn in figure 5.3 one periodic repetition of the complex and one water molecule in the cell above the central unit cell. Note that this water molecule is the periodic repetition of the water molecule drawn close to the bottom of the central unit cell. It is interesting to observe that in the first panel the orientations of the H$_2$O$_2$ ligand at the top iron complex, the water at the bottom in the top cell (indicated with ``S1''), and the water in the upper part of the central cell (``S2'') together with a water ligand at the Fe in the central cell (``L1'') are close to creating an H-bond wire. This H-bond wire is fully established in the second panel by slight reorientations of the water molecules.

Figure: Five snapshots of the first reaction step, starting 2 fs after the moment that the $R(\mathrm{OO})$ and $R(\mathrm{FeO})$ constraints of H$_2$O$_2$ coordinated to [Fe(H$_2$O)$_5$]$^{2+}$ were released ($t=0$). The hydrogen peroxide coordinated to the pentaaqua iron complex is drawn in the center of the unit cell with one of its periodic images drawn above. The two solvent waters that are involved in the reaction are also shown, but for simplicity, the other solvent water molecules are left out. See also text.

Almost immediately after the bond constraints were released at t=0 fs, the peroxide dissociates into a hydroxo group coordinated to iron and an OH. radical which attacks a solvent water molecule that was hydrogen bonded to the hydrogen peroxide $\beta$-oxygen (see snapshots 1 and 2 in figure 5.3). Snapshots 2 and 3 show this solvent water molecule (the water molecule in the lower part of the central unit cell, as well as its periodic image drawn in the top cell) moving to the cell boundary, and in panel 4 actually crossing the boundary (at least an OH moiety). This OH moiety leaves the central unit cell across the bottom plane, but of course the periodic image OH in the top cell leaves the top cell and enters the central unit cell across the top plane of the central cell. The H of this H$_2$O is abstracted by the OH. radical coming from the coordinated H$_2$O$_2$. The intertesting point (see snapshots 4 and 5) is that the OH. radical through a chain of H abstractions, running along the preestablished H-bond wire, actually abstracts a H from a water ligand of the neigbouring Fe complex. Closer inspection shows that the radical passage is not a typical chain reaction but rather a concerted process: snapshot 3 reveals that the "chain" starts simultaneously at both ends and is terminated in the middle (snapshot 4 and 5).

$$\left[(\mathrm{H}_2\mathrm{O})_5\mathrm{Fe^{II}}\mathrm{H}_2\math... ...rm{H}_2\mathrm{O})_4\mathrm{Fe^{IV}}(\mathrm{OH})_2\right]^{2+} + \mathrm{H_2O}$$ (61)

The formation of the tetraaqua iron(IV) dihydroxo complex as the first step in the H$_2$O$_2$ dissociation (eq. 5.3) agrees with the energetic requirements derived from the gas phase calculations, see section 5.3.1. Of course, the concerted radical passage could not be observed without explicit inclusion of the aqueous solvent environment. The use of periodic boundary conditions is a means to simulate the "infinite" solvent environment with finite computational resources. It is, however, from this simulation easy to see what the possibilities are in a very large cell, or in a system without periodicity. We have established that O-O dissociation in a coordinated H$_2$O$_2$, leading to one Fe-OH bond and an OH. radical, is only energetically possible if an additional Fe-OH bond is formed. This implies that the concerted radical path has to terminate at a water ligand. This leaves the following possibilities: a) if the pentaaqua iron(II) hydrogen peroxide complex would be in the neighborhood of a hexaaqua iron(II) complex, the OH. passage might be, via two or three solvent waters, to a water ligand of the hexaaqua iron(II), with two mono-hydroxo pentaaqua iron(III) complexes as the result; b) with only one pentaaqua iron(II) hydrogen peroxide complex in such a system, the radical passage can go via a few solvent waters and end up with hydrogen abstraction from one of the water ligands of this same iron complex, resulting in the dihydroxo complex. This is the likely reaction pathway in a typical experimental situation where the iron catalyst is present in a low concentration. Indeed, in preliminary simulations we are performing to study the possibility of H abstraction from organic substrates, we have encountered reaction pathways in which the H abstraction took place from H$_2$O coordinated to the same Fe center (see figure 5.4).

Figure 5.4: Snapshot of an AIMD simulation of the oxidation reaction of pentaaquairon(II) hydrogen peroxide and methane in water (for simplicity, the solvent water molecules are left out). After O$^\alpha$-O$^\beta$ lysis, the O$^\beta$H. radical immediately ``grabs'' the H of an adjacent water ligand, to form a water molecule and tetraaqua dihydroxo iron(IV).

We observed that the formed iron(IV)dihydroxo complex is in equilibrium with its conjugate base by proton donation to the solvent (eq 5.4).

$$\left[(\mathrm{H}_2\mathrm{O})_4\mathrm{Fe^{IV}}(\mathrm{OH})_2\r... ...m{H}_2\mathrm{O})_3\mathrm{Fe^{IV}}(\mathrm{OH})_3\right]^{+} + \mathrm{H_3O^+}$$ (62)

The apparent acidity of the Fe $^{\mathrm{IV}}$ complex obviously can only be seen in a simulation including the solvent. The donated proton is passed on in the solvent to end up again on a hydroxo ligand, which of course does not have to be the same ligand that donated the proton to the solvent initially. Then, hydrolysis of this newly formed or another water ligand may take place, and so forth. As a result of this dynamic equilibrium, the hydroxo ligands are transformed into water ligands and vice versa. Eventually at a time when the system happens to find itself on the left-hand-side of equation 5.4 (in our simulation, 1.7 ps after the formation of the dihydroxo iron(IV) complex, eq. 5.3), a proton is donated by a hydroxo ligand instead of a water ligand and the iron-oxo complex is formed in our simulation.

$$\left[\mathrm{(\mathrm{H}_2\mathrm{O})_4Fe^{IV}}(\mathrm{OH})_2\r... ...m{(\mathrm{H}_2\mathrm{O})_4Fe^{IV}O}(\mathrm{OH})\right]^{+} + \mathrm{H_3O^+}$$ (63)

Such proton donation by a hydroxo ligand has been reported before for e.g. the mechanism for oxo-hydroxo tautomerism observed by high-valent iron porphyrins with an oxo and hydroxo group as axial ligands in aqueous solution.[177] After another two pico-seconds, one of the water ligands leaves the first solvation shell[178], and the formed $\left[(\mathrm{H}_2\mathrm{O})_3\mathrm{Fe^{IV}O}(\mathrm{OH})\right]^{+}$ complex undergoes no more spontaneous chemical changes in the next 6 pico-seconds.

The reported reaction pathway took place spontaneously, once we had constructed the initial reactants configuration. The observed reactions occur apparently without a significant reaction barrier, which puts doubt on the assumption in certain kinetic models[167,176] that there exists a steady state of the concentration of complexed [Fe$^{\rm {II}}$-H$_2$O$_2$]$^{2+}$. In agreement with the energetics obtained in gas phase calculations, we confirm the formation of the iron(IV)-oxo complex from the Fenton reagent, but in view of the ``radical passage'' mechanism of the first step (figure 5.3), the question may be raised whether OH. radicals can be excluded as possible reactive intermediates. In the gas phase, we find that reaction 5.6

$$\left[(\mathrm{H}_2\mathrm{O})_5\mathrm{Fe^{II}}\mathrm{H}_2\math... ...H}_2\mathrm{O})_5\mathrm{Fe^{III}}(\mathrm{OH})\right]^{2+} + \mathrm{HO}\dot{}$$ (64)

is endothermic by 21 kcal/mol (table 5.2). The OH. formation followed by transfer away from the complex into the solvent, where it might react further with an organic substrate, is therefore not likely. Instead, the radical is quenched, according to the pathway in figure 5.3, through a short, energetically favorable, transfer along a hydrogen-bond wire through the solvent to a complex (either the initial one or a neighboring complex) where the necessary energy is gained by formation of a second OH ligand. Only in case of a high organic substrate concentration, i.e. a high chance of finding an organic molecule in the neighborhood of the iron complex where the hydrogen peroxide is coordinated, can we expect the OH. radical to have a significant contribution in the oxidation reactions, because the radical transfer could find a path along an H-bond wire to an organic substrate molecule if the latter is close enough. In all other cases, the ferryl ion is expected to be the active intermediate formed by the Fe $^{\mathrm{II}}$ and hydrogen peroxide in water.

The quenching of the OH. radical by hydrogen abstraction from a H$_2$O ligand explains a notorious problem in the application of the Fenton reagent, namely the degradation of the chelating agents (see e.g. refs. RaRi88,RuKo87,TaRoBo99). Chelating agents are bulky ligand complexes, applied to increase the solubility of the iron catalyst. However, in practice the ligands degrade during the reaction, reducing the number of catalytic cycles per metal complex. The degradation can now be understood as the scavenging process by the very short-lived OH. radical, which is initially formed in the H$_2$O$_2$ oxygen-oxygen cleavage.