Size Effects in Electrocatalysis of Fuel Cell Reactions on Supported Metal Nanoparticles - Fuel Cell Catalysis: A Surface Science Approach - page 533

goes down. Hence, the peak potential should have shifted negatively rather than posi-
tively. The fact that particle edges do not provide active sites for CO electro-oxidation
may be explained by their different site geometry, compared with steps, as discussed in
Section 15.3.
Mayrhofer and co-workers suggested that the active sites are located at the defect
sites on nanoparticle facets [Mayrhofer et al., 2005a]. Based on ex situ high resolution
TEM (HRTEM) images, they speculated that large particles bear a larger fraction of
defects, and hence show better activity in CO ads electro-oxidation. Although we
believe that this is a reasonable assumption, it should be noted that making con-
clusions on the basis of ex situ characterization is not unambiguous. The low potential
peak in CO stripping voltammograms, which appears in curve 5 of Fig. 15.8a, was first
documented by Cherstiouk and co-workers at high Pt loadings and attributed to Pt
agglomeration [Cherstiouk et al., 2003a]. Maillard and co-workers confirmed that
this low potential peak correlates with the number of Pt agglomerates detected with
TEM [Maillard et al., 2005]. They examined Pt/GC agglomerates by HRTEM, and
found complex irregular structures composed of individual Pt nanoparticles intercon-
nected via grain boundaries (so-called “multigrained” structures). These nanocrystal-
line materials feature a high density of surface and bulk crystalline defects (steps and
intergrain boundaries), which are believed to be the likely reason for their high elec-
trocatalytic activity [Cherstiouk et al., 2000, 2003a; Maillard et al., 2005]. This peak
assignment is proven by CO stripping from electrodeposited Pt with multigrained
structure (curve 5 in Fig. 15.8), which exhibits exclusively the low potential peak.
It has recently been demonstrated that the catalytic activity of bimetallic Pt-Ru cata-
lysts in CO and methanol electro-oxidation also correlates with the concentration of
intergrain boundaries [Gavrilov et al., 2007; Maillard et al., 2007a]. Metal atoms in
the vicinity of grain boundaries usually have a decreased number of neighbors in
their first coordination shell [Gleiter, 1992], and thus are expected to show modified
adsorption properties compared with metal atoms of densely packed terraces. For
further discussion of the electrocatalytic properties of these promising materials,
the reader is referred to Maillard et al. [2005, 2007b]; Gavrilov et al. [2007];
Cherstiouk et al. [2000, 2003a, 2008].
Peak multiplicity in CO ads stripping voltammograms and chronoamperograms has
given rise to some controversy in the literature. Guerin and co-workers ascribed the
low potential peak to electro-oxidation of CO ads on Pt terraces, whereas they attributed
the high potential peak to CO electro-oxidation on edges and corner sites [Guerin
et al., 2004]. Solla-Gullon and co-workers attributed the peak multiplicity in CO strip-
ping voltammograms to electro-oxidation of ordered and disordered CO ads domains on
facets of different crystallographic orientation [Solla-Gullon et al., 2005, 2006]. Arenz
and co-workers reported double-peaked versus single-peaked chronoamperograms
for 2 and 1 nm Pt particles, respectively [Arenz et al., 2005], and explained this finding
by the increasing fraction of surface defects on 2 versus 1 nm Pt/C particles [Arenz
et al., 2005; Mayrhofer et al., 2005a, b]. In our opinion, a more likely reason for the
peak multiplicity in chronoamperograms and in stripping voltammograms of sup-
ported metal nanoparticles is so-called “interparticle heterogeneity,” i.e., the presence
of particles of different sizes and structures, although other factors may also be