Methanol, Formaldehyde, and Formic Acid Adsorption/Oxidation on a Carbon-Supported Pt Nanoparticle Fuel Cell Catalyst: A Comparative Quantitative DEMS Study - Fuel Cell Catalysis: A Surface Science Approach - page 416

peak fully overlaps with the initial current spike and can hardly be separated.
Furthermore, the methanol adsorbate oxidation peak is much narrower compared
with the other adsorbate oxidation peaks.
The identification of the actual adlayer oxidation process is much simpler in the
mass spectrometric signals (Fig. 13.2b) than from the faradaic current transients. In
all cases, the sharp initial spike is absent, confirming that this is due to pseudocapaci-
tive contributions (see above). For constant potential oxidation of a saturated CO
adlayer (CO pre-adsorption), the transient shows a small initial peak, which, however,
is much smaller and also wider than the sharp spike in the faradaic current peak.
Therefore, the onset of CO ad oxidation upon the potential step cannot be followed
via the faradaic current, because of the above-mentioned capacitive contributions,
which are often overlooked [Maillard et al., 2004a, 2005; Arenz et al., 2005; Santos
et al., 1991; Petukhov et al., 1998; Koper et al., 1998; Lebedeva et al., 2002;
Andreaus et al., 2006; Andreaus and Eikerling, 2007]. The subsequent induction
period and the bell-shaped CO ad oxidation peak resemble the features observed in
the faradaic current. The observation of small amounts of CO ad oxidation right
upon the potential step (instantaneous CO ad oxidation) agrees well with previous
findings for CO ad oxidation on nanostructured Pt/C model catalysts [Gustavsson
et al., 2004] and on Pt film electrodes [Heinen et al., 2007]. Also, the other transients
(C 1 adsorbate oxidation transients) largely resemble the faradaic current transients,
with the exception of the initial current spike, which is absent in the mass spectro-
metric transients. Because of the increasing overlap of the initial current spike with
the actual adsorbate oxidation peak in the faradaic current signal, the mass spectro-
metric current transients allow a much better identification and quantification of the
adsorbate oxidation current. Furthermore, exact evaluation of CO ad coverage from
the following transient is hardly possible, owing to the capacitive contribution to
the measured faradaic current from anion re-adsorption upon removal of CO ad
[Heinen et al., 2007]. By integrating the mass spectrometric CO 2 signal, we deter-
mined the initial CO ad coverages and their decay during the transient measurements
[Gustavsson et al., 2004; Heinen et al., 2007]. The CO ad coverage at a given
moment is calculated as the (integrated) saturation coverage minus the ratio of inte-
grated ion current to the total mass spectrometric charge. (This evaluation is based
on the assumption that CO ad is the only stable adsorbed species resulting from C 1 mol-
ecule adsorption; see Section 13.3.1.1). The resulting coverage transients are plotted in
Fig. 13.2c. In agreement with the potentiodynamic experiments, the CO ad coverage
decreases in the order CO sat . formaldehyde . formic acid . methanol. The exact
values are collected in Table 13.1. The difference with respect to the values deter-
mined in the potentiodynamic stripping experiments is explained by the different
adsorption potential in the potentiostatic measurements (0.16 V vs. 0.11 V). The
increase in adsorbate coverage is most pronounced for methanol adsorption. This
observation agrees closely with previous findings for potentiodynamic and potentio-
static methanol adsorbate stripping experiments performed after adsorption at different
adsorption potentials [Jusys and Behm, 2001; Lanova et al., 2006]. Recent in situ
ATR-IR experiments equally showed that the potential dependence is most pro-
nounced for methanol adsorption, less for formic acid adsorption, and essentially
absent for formaldehyde adsorption [Chen et al., unpublished results].