Electrocatalysis of Oxygen Reduction in Polymer Electrolyte Fuel Cells: A Brief History and a Critical Examination of Present Theory and Diagnostics - Fuel Cell Catalysis: A Surface Science Approach - page 10

Mechanistic conclusions from DFT calculations: The recent DFT calculations
aided in the clarification of the relative probabilities of various proposed multistep
ORR routes from reactant O 2 to product H 2 O. This has been enabled by the accurate
calculation of formation energies for relevant intermediates in any mechanism con-
sidered, where the interactions of an intermediate with both the Pt metal surface and
the adjacent molecular layers of water are both considered. For example, the
probability that the first step in the ORR process is direct, dissociative chemisorption
of O 2 to form two O ads species on the Pt metal surface appears to be low, not only in
light of the low rates of this process at the Pt/oxygen gas interface at such low temp-
eratures, but also in light of the deep energy sink associated with chemisorption of
an oxygen atom on Pt [Nørskov et al., 2004]. Accordingly, there seems to be an
advantage for the ORR route starting with a single electron and proton transfer to
O 2 to generate the HOO ads intermediate and thereby facilitate subsequent breaking
of the O22O bond, which in HOO ads has an order significantly lower than 2
[Panchenko et al., 2004]. OH ads , rather than O ads , seems the likely next intermediate
on the way to the final H 2 O product, as can be concluded, again, from the deep
energy sink associated with O ads on Pt. The recent DFT calculations further support
the O 2 -to-OOH ads -to-OH ads route by showing that OOH is indeed an intermediate
of significant bonding energy to Pt, at least on some Pt crystal surfaces [Panchenko
et al., 2004]. Another important contribution of the theoretical calculations has been
in providing an explanation for the higher measured ORR activity of Pt-alloy catalysts
versus unalloyed Pt. In accordance with the premise that the M22Ox bond strength is
an effective yardstick for predicting relative ORR rates at metal/electrolyte interfaces,
the key to the higher ORR activity observed following proper alloying of Pt seems to
be the calculated increase in d-band vacancy which causes reduction in the affinity of
Pt surface sites to chemisorbed OH or O species.
Paradoxically, all these significant recent contributions to the theory of the ORR,
together with most recent experimental efforts to characterize the ORR at a fuel cell
cathode catalyst, have not led at all to a consensus on either the mechanism of the
ORR at Pt catalysts in acid electrolytes or even on how to properly determine this
mechanism with available experimental tools. To elucidate the present mismatch of
central pieces in the ORR puzzle, one can start from the identification of the slow
step in the ORR sequence. With the O 2 -to-HOO ads -to-HO ads route appearing from
recent DFT calculations to be the likely mechanism for the ORR at a Pt metal catalyst
surface in acid electrolyte, the first electron and proton transfer to dioxygen, according
to the reaction

O 2 þ H þ þ e ! OOH ads (1 : 2)

becomes the likely slow step that determines the overall interfacial ORR rate at a Pt
catalyst. One reason for this tentative assignment is the demanding requirement of
forming a significant bond to the Pt catalyst surface before O22O dissociation can
take place—a requirement that can apparently be fulfilled on Pt surface sites of specific
atomic geometries, available on some Pt crystal surfaces [Panchenko et al., 2004]. The
second reason for assigning (1.2) as the slow step is the results of past experimental