(Keynote) Interpretation of Oxygen Reduction Activity Volcano Correlations for MN4 Molecular Catalysts Compared to Those for Metallic Electrodes

Abstract

A classical well-known reactivity index in electrocatalysis is the binding energy of key intermediates to the active sites. For example, this is well documented for the catalytic activity of metal electrodes for the oxygen reduction reaction (ORR)[1], alloys and metal oxides and less studied for molecular catalysts [2]. The activity at constant potential (log(i)E) plotted versus the M-O2 binding energy has the shape of a volcano. This is well documented in several papers, especially by the group of Norksov for metallic electrodes [1] and can be applied to many electrochemical reactions. This has been extended to MN4 molecular catalysts. In contrast to metallic electrodes MN4 molecular catalysts have discrete energy levels and the active site is the central metal, surrounded by an organic ligand with an MN4 central moiety. Essentially one of the key intermediates is the binding of the reacting O2 molecule to the active sites at the rate determining step as: [MN4]ad + O2(aq) + e- ↔ [RMN4O2 -]ad where MN4 is a surface confined macrocyclic complex or a pyrolyzed catalysts bearing a MN4 active moiety embedded in a graphene or graphitic structure. As the ORR reaction involves the transfer of several electrons ( 2, 2+2 or 4 electrons), several adsorbed intermediates can be involved. If the controlling step is the first step, it is expected that for the binding step when DGad = 0 a maximum activity should be observed and the partial coverage of adsorbed intermediate is = 0.5. On thermodynamic ground this corresponds to a thermoneutral condition for the best catalyst. In this work we have tested this hypothesis by studying ORR in alkaline media using several iron porphyrins and iron phthalocyanines as catalysts immobilized on graphite and carbon nanotubes. The trends in reactivity versus the Fe(III)/(II) redox potential for ORR describe a typical volcano correlation. However if the currents are divided by the θFe(II), the fraction of active sites calculated at E=-0.24 V vs SCE using the Nernst equation, log(i/ θFe(II),)E vs. E°Fe(III)/(II) gives a straight line of slope -0.120V/decade. The maximum corresponds to θFe(II) = 0.5 and occurs at a potential close to that used for comparing the activities. A similar behaviour is observed for the oxidation of hydrazine but the continuation of the straight line is observed at potentials more positive than that of the maximum. Again the maximum is observed at the potential chosen for comparison. Those catalysts having Eº’ Fe(III)/(II) >> θFe(II) higher than -0.56V (for hydrazine oxidation) are in the Fe(III) state as predicted by the Nernst equation. That oxidation state Fe(III) is inactive for both reactions studied as OH- ions are strongly bound to Fe(III), especially in alkaline media. So the falling of the activities in the strong binding side of the volcano can be attributed preferentially to a gradual decline in the number of Fe(II) active sites and not to gradual decrease of the fraction (1-θ ) of empty or available active sites due to occupancy by intermediates. Schmickler and Santos [3] have shown that some volcano correlations for HER on metals that are in an oxidized form, and then inactive for this reason.We conclude that even though there are similarities between the volcano correlations for metallic and molecular catalysts, there some differences in the interpretation of the activity maximum that might not correspond to a thermoneutral condition. Acknowledgements: This work was supported by ANID, Anillo Project ACT 192175 and Fondecyt, Project 1181037