DOI: 10.1021/acs.chemrev.7b00488
Part of this summary is inspired by this article.
ORR pathway and mechanism
2-electron pathway involves *OOH as a reaction intermediate. For 4- electron pathway, the associative mechanism (equation 4) involves 3 different intermediates: *OOH, *O and *OH while the dissociative mechanism (equation 5) involves only *O and *OH.
In the computational hydrogen electrode (CHE) model, the adsorption free energies of a reaction intermediate (with n proton-electron pairs) is given by:
where ΔEele is the DFT calculated binding energy, ΔEw is the adsorbate solvation, ΔEfield is the electric field effects, ΔZPE is the zero-point enegy, -TΔS is the entropic corrections and -eU is the free energy of a single proton-electron pair relative to H2 in the gas phase at standard conditions, where U is the electrode potential vs RHE.
From Fig. 2, at U = 1.23 V (green lines), the reduction of *OH to H2O is uphill in energy, indicating that the surface will be covered by these species and will be inactive for O2 adsorption. The maximum potential where all steps are downhill in free energy for Pt (111) is ca. 0.8 V, and is known as the thermodynamic limiting potential (UL). The difference between the equilibrium potential and the limiting potential is called the theoretical overpotential, and ηtheo for Pt(111) is therefore 0.4 V. Note that ηtheo and UL is just a measure of the activity of a catalyst and should not be compared directly with a measured overpotential, which depends on current density. For 2 e– processes, the ηtheo for PtHg4 is ~ 0.07V.
ORR trends for transition metals: Linear scaling relationship and the volcano plots
The binding energies for *OOH, *O, and *OH are strongly correlated and change monotonically for different metals. These linear relationships arise because all the adsorbates bind to the surface through an O atom. In particular, the *OOH vs *OH line has a slope close to unity, indicating a similar metal−oxygen single bond for both adsorbates. In contrast, the *O vs *OH scaling line has a slope close to 2, in line with a picture where the adsorbed O has a double bond to the surface while *OH binds through a single bond.
As the theoretical overpotential is a function of the binding energies of *OOH, *O, and *OH, we can use the scaling relations in (a) to define the limiting potentials for the four steps (UL1−L4) as a function of the OH free energy of adsorption as:
UL1 = – ΔGOH + 1.72
UL2 = – ΔGOH + 3,3
UL3 = ΔGOH
UL4 = ΔGOH
The lowest limiting potential for the full catalytic reaction defines the overall limiting potential for the reaction, and it is indicated by the blue and green solid lines in (b). For metals that bind *OH strongly, *OH→ H2O is potential limiting (solid blue line), whereas for the weakly bonding metals the activation of O2, O2 → *OOH, is potential limiting (solid green line). From (b) we can see that, Pt(111) is the closest to the top of the limiting potential volcano compared to the (111) facets of other transition metals. That’s why Pt is so ideal for ORR!
For 2 e– processes, the analysis is similar and the resulting plot is shown in (c). The peak of the volcano crosses the equilibrium potential at 0.70 V. This indicates that it is possible, in principle, to find a catalyst with an ideal activity if it binds that single intermediate with optimal strength, not too weak nor too strong.
Kinetics of ORR
The barriers for proton transfer to surface-bound intermediates have been estimated to be quite small, essentially of the same order of magnitude as proton transfer barriers in water. Subsequently, a microkinetic model for the ORR based on the DFT calculations was shown to reproduce the experimental kinetic observations:
Using DFT calculations and thermodynamic analysis, it has been suggested that the selectivity toward H2O2 is determined by the two competing reactions:
The key to avoid the four- electron pathway is to prevent the O−O bond dissociation in the adsorbed *OOH. This simple analysis excludes all catalysts with strong oxygen binding energies (due to favorable O* formation) and limits the search to materials that bind oxygen weakly.
Improving ORR activity: alloys and core-shell catalysts
Although Pt(111) already lies close to the top of the thermodynamic volcano, weakening the binding energy of *OH (and subsequently *O or *OOH) will move it closer to the top. Specifically, for a catalyst that follows the above scaling relations, an increase in ΔGOH of ≈ 0.1 eV is optimum. Detailed examples are given in the original text and is not included in this summary.
For strongly binding metals (e.g., Pd(111), Pt(111), etc.), the removal of *OH (i.e., ΔGOH) is responsible for UL, while the kinetically relevant step is the rate of O2 binding to free surface sites. As the availability of free sites is related to the coverage of *OH, which is also determined by ΔGOH, the predictions of the two approaches are very similar. For the weakly binding catalysts, the *OOH binding energy gives rise to the limiting potential (green line, Figure 6), while kinetic analysis indicates that the activation energy for the first O2 protonation to form *OOH is rate limiting. It has been found very generally that activation energies for surface processes scale with the reaction energy. Thus, increasing the stability of the *OOH (and thus *OH, through scaling) will give rise to a lowering of the activation energy and increased activity
Perspective and Outlook
We need to break the scaling between *OOH and *OH activation energies. A few promising strategies shown in 7(b).
Suggested Reading: Catalyst Design for Electrochemical Oxygen Reduction toward Hydrogen Peroxide (discuss the state-of-art understanding of material modification toward 2 e– process and summarize on the catalyst engineering strategy for electrochemical hydrogen peroxide production)