Introduction

The activation of oxygen by transition metal oxides is of fundamental importance in organic synthesis, catalysis and biochemistry. Efficient catalysts are found in nature for the hydroxylation and oxidation of organic molecules. Examples are cytochrome P450, methane monooxygenase (MMO), and the antitumor drug bleomycin (BLC), all of which have low-valent iron ions as the active centers that can be turned into highly reactive high-valent iron-oxo, iron-peroxo or iron-hydroperoxo complexes. Also Fenton's reagent, a mixture of iron(II) ions and hydrogen peroxide, is widely applied to oxidize organic substrates. Although H.J.H. Fenton himself used the unchelated Fe$^{2+}$/H$_2$O$_2$ mixture in acidic aqueous solution when he first recorded the oxidative potential of this mixture in 1876[128,155], many modifications have since then been applied in order to alter the reactivity, to enhance the solubility in other solvents, to increase the pH range, and to avoid precipitation of the metal catalyst. This has led to a great variety of ligated iron(II) complexes in combination with H$_2$O$_2$ (or HOCl), or even with different transition metals, such as Mn(II) or Fe(III) which are all often referred to as Fenton's reagent. To distinguish between the iron(II) and iron(III) combinations, we will however follow the convention of using Fenton-like reagent for the Fe$^{3+}$/H$_2$O$_2$ mixture and restrict the use of Fenton's reagent to denote the Fe$^{2+}$/H$_2$O$_2$ mixture. The Fenton-like reagent is also capable of oxidizing organic substrates, but it is somewhat less reactive than Fenton's reagent. As iron(III) can be produced in applications of Fenton's reagent, Fenton chemistry and Fenton-like chemistry often occur simultaneously. Interestingly, Fenton and Fenton-like chemistry are generally believed to proceed via similar mechanisms as oxidation reactions with the aforementioned complex and bulky enzymes.

Until recently, Fenton chemistry was still far from being fully understood. Numerous reaction mechanisms have been proposed based on different active intermediates such as OH. and OOH. radicals and the earlier mentioned high-valent iron species. Haber and Weiss' OH. radical mechanism[156] is probably the most popular candidate for the Fenton reaction:

$${\rm Fe}^{2+} + {\rm H}_2{\rm O}_2 \rightarrow {\rm Fe}^{3+} + {\rm OH}^- + {\rm OH}\mbox{\.{}}$$ (66)

followed by the alternative mechanism first suggested by Bray and Gorin[157], in which the ferryl ion, [Fe$^{\rm {IV}}$O]$^{2+}$, is supposed to be the active intermediate:

$${\rm Fe}^{2+} + {\rm H}_2{\rm O}_2 \rightarrow [{\rm Fe^{IV}O}]^{2+} + {\rm H}_2{\rm O}$$ (67)

In both mechanisms, the hydrogen peroxide O-O lysis forms the essential step. We have recently performed static density functional theory (DFT) calculations to study the active species produced by the hydrated Fenton's reagent [Fe$^{\rm {II}}$(H$_2$O)$_5$ (H$_2$O$_2$)]$^{2+}$ in vacuo[145] as well as ab initio (DFT) molecular dynamics (AIMD) simulations of Fe$^{2+}$ and H$_2$O$_2$ in aqueous solution[171,146]. In this work, we showed that the ferryl ion is easily formed when hydrogen peroxide coordinates to an iron(II) ion in water, which confirms the reaction mechanism first proposed by Bray and Gorin (reaction 6.2). Moreover, formation of the ferryl ion from the hydrated Fenton's reagent in vacuo was found to be energetically favored over the formation of free hydroxyl radicals (reaction 6.2).

For the Fenton-like reagents it is believed that initially no O-O bond breaking takes place, but instead an iron(III)hydroperoxo intermediate is formed as the first step via hydrolysis:

$${\rm Fe}^{3+} + {\rm H}_2{\rm O}_2 \rightarrow [{\rm Fe^{III}OOH}]^{2+} + {\rm H}^+$$ (68)

This intermediate might be able to react with organic substrates or break up in smaller active species in a second step. The iron(III)hydroperoxo may e.g. homolyze at the Fe-O bond[181,182,132]

$$[{\rm Fe^{III}OOH}]^{2+} \rightarrow {\rm Fe}^{2+} + {\rm OOH}\mbox{\.{}}$$ (69)

generating iron(II) and producing the reactive OOH. radical, or at the O-O bond producing the ferryl ion and an OH. radical[132,158]:

$$[{\rm Fe^{III}OOH}]^{2+} \rightarrow [{\rm Fe^{IV}O}]^{2+} + {\rm OH}\mbox{\.{}}$$ (70)

Alternatively, O-O bond heterolysis could take place, producing the highly reactive Fe$^{\rm {V}}$ species[183,132]:

$$[{\rm Fe^{III}OOH}]^{2+} \rightarrow [{\rm Fe^{V}O}]^{3+} + {\rm OH}^-$$ (71)

Probably many of the proposed mechanisms compete with each other depending on the reaction conditions, such as the metal ligands, the solvent, the pH and the organic substrate to be oxidized. This is of course one of the reasons why Fenton-like chemistry still holds secrets. Another important reason is the very elusive nature of the active species, which live too short to have been definitively observed yet in experiments. Obviously, that is not a problem for computer simulations.

In the present work, we will investigate the active species for the Fenton-like chemistry, and we will focus on the differences between iron(II) and iron(III) in activating hydrogen peroxide. To simulate the generation, evolution and termination of reactive intermediates such as hydroxyl radicals we need an accurate description of the aqueous environment. The method of choice is ab initio (DFT) Molecular Dynamics method (AIMD), using the Car-Parrinello technique, which has already proven to be a very useful tool for our type of systems.[171,144,52,184] Density functional theory (DFT) has already proven to be very useful for the unraveling of the activation mechanisms of MMO[163,185,164,186,187,188] and P450[161,159,189,190,160] and the oxidation of methane and benzene with the bare ferryl ion (Fe$^{\rm {IV}}$=O)[170,191,192,193,194,195,168].

The paper is organized as follows. First, we briefly summarize the computational details in section 6.2. In section 6.3, we present our results, starting with the static DFT calculations of the energetics of the proposed elementary Fenton and Fenton-like reactions of the hydrated iron hydrogen peroxide complexes in vacuo. Clearly, the different nature of the Fe(III)/H$_2$O$_2$ reagent compared to the Fe(II)/H$_2$O$_2$ reagent is revealed, showing that OH. radical or ferryl ion formation via O-O lysis as the first mechanistic step is energetically very unfavorable and instead hydrolysis (eq 6.3) is indeed predicted to be the most likely initial step. However, realistic modeling of the hydrolysis, i.e. simulating the donation of H$^+$ by the iron(III) complex to the aqueous solution, requires the explicit inclusion of the solvent molecules. This is done in subsection 6.3.2 by first presenting three illustrative AIMD simulations of the reaction between Fe$^{\rm {III}}$ and H$_2$O$_2$ in water, which show the spontaneous formation of the iron(III)hydroperoxo species, followed by the characterization of Fe(III)OOH(aq). We have computed the vibrational spectra for the iron(III)hydroperoxo species in both the high-spin ($S=5/2$) state and the low-spin ($S=1/2$) state, and have compared to experimental Raman spectra. As the second step in the Fenton-like mechanism, the O-O homolysis of the iron(III)hydroperoxo species, producing the ferryl ion and a hydroxyl radical, is predicted by our static DFT calculations in vacuo. The free energy barrier for this step was computed in aqueous solution using the method of constrained molecular dynamics and thermodynamic integration, the results of which are presented in subsection 6.3.3. The paper ends with conclusions in section 6.4.