Environmental Engineering Reference
In-Depth Information
charge, decreasing to 0.30 eV at a surface charge corresponding to a potential about
1.5 V more positive than the potential at which the surface is uncharged. They also
found that the reaction of CO with O in the presence of water has a much higher barrier
than the reaction of CO with OH.
The above experiments show convincingly that the active sites for CO oxidation are
the step sites and that the mobility of adsorbed CO on Pt(111) terraces should be con-
sidered high. The idea of a highly mobile CO chemisorbate is always inconsistent with
a N&G model, because in the N&G model the assumption of immobile reaction
partners is essential. This does not prove that the mean field model is quantitatively
correct, although clearly the fit is very good, but merely that on the electrodes that
we have studied, there is no physical basis for the key assumption in the N&G
model, namely the supposed surface immobility of adsorbed CO. Interestingly, in
later work on stepped rhodium single-crystal electrodes [Housmans et al., 2004;
Housmans and Koper, 2005a, b], experiments do suggest a low CO mobility. The
chronoamperometric transients for rhodium are very different from those obtained
for platinum, and a consistent kinetic modeling is quite a bit more involved, as will
be discussed in Section 6.2.2.
Close inspection of the experimental transients shows that the initial part of the
transient cannot be modeled by a mean field or N&G model (see, e.g., Fig. 6.2b).
After the double-layer charging has died out, the current shows an almost constant
value for a certain period of time, before it begins to rise as predicted by the mean
field and N&G models. We have ascribed this constant current (the “plateau”) at
the beginning of the transient to an incipient oxidation of the CO adlayer during
which the CO adlayer relaxes such that no new free sites are created for OH adsorption
[Bergelin et al., 1999; Lebedeva et al., 2002a, c]. This would explain the constant
current and the implied zeroth-order reaction kinetics in u
CO
. Once the CO coverage
is below a certain critical coverage, which we expect to be around 2 - 3% below the
saturation coverage, the second-order LH mechanism sets in (6.2). The fact that the
potential dependence of this “plateau” process, as discerned from a Tafel slope of
70 - 80 mV/dec for the various stepped electrodes [Lebedeva et al., 2002c], is similar
to that of the potential dependence of the process associated with the chronoam-
perometric peak suggests that the mechanism is similar but that the reaction order is
different. The current in the plateau was also found to depend on step density. A
large increase in plateau current is observed when switching from Pt(111) to
Pt(15, 15, 14), while further increases in the step density lead to much less significant
current enhancements. Various authors have suggested that this CO oxidation initiation,
which is also seen in voltammetry as a prewave (see Section 6.2.1.2), may be ascribed
to a kind of Eley - Rideal mechanism in which the adsorbed CO is attacked directly by
a water molecule from the double layer, and not by adsorbed OH. There are at least
two reasons why we believe this interpretation is less likely. First, for such a reaction,
a different potential dependence would be expected, and one would expect it to be less
sensitive to the step density. Second, DFT calculations have shown that the activation
energy for such a reaction would be unrealistically high [Dunietz et al., 2002].
6.2.1.2 CO Stripping Voltammetry
As discussed in the previous section, the
chronoamperometric transients can be modeled using the LH mechanism. Using the
Search WWH ::
Custom Search