Environmental Engineering Reference
In-Depth Information
electrochemical kinetics can be performed by applying appropriate mathematical
models, allowing separation of the various contributions. Since mathematical
models include many unknown parameters (agglomerate structure, diffusivities in
ionomer films, etc.), quantification of intrinsic catalyst activity may be associated
with large uncertainties. Hence, unambiguous analyses require that kinetic measure-
ments be performed in the low overpotential interval. As we will see from our further
analysis of the literature, this is not always the case, and conclusions concerning PSEs
are sometimes made in the potential interval of hindered mass transport.
The fourth limitation is related to the fact that the catalyst utilization factor u may
fall well below 1. This factor is commonly defined as the ratio of the metal surface
atoms participating in an electrochemical reaction to the overall number of metal sur-
face atoms. There is currently no agreement in the literature regarding the utilization
factor achieved in state-of-the-art PEMFCs. Indeed, Gasteiger and co-workers suggest
that u may be close to 1 [Gasteiger et al., 2005], while, according to Eikerling and
co-workers, the maximum achievable u in a random CL cannot exceed about 0.4
[Eikerling et al., 2005]. This apparent discrepancy is likely to be due to somewhat
different meanings assigned to this parameter by different authors. It was formerly
widely accepted that only surface sites located at the three-phase boundary, i.e., at
the interface between catalytic particles, proton-conducting electrolyte, and gas-
filled pores, are electrochemically active. However, a number of recent fuel cell as
well as model studies suggest that proton transfer may rather efficiently occur in a
thin aqueous film in the absence of a proton-conducting ionomer [Paulus et al.,
2003; Eikerling et al., 2005]. Paulus and co-workers have shown that metal utilization
is high for electrodes where Pt particles form a contiguous network (e.g., Pt black) and
much lower if metal particles are distributed on a carbon surface [Paulus et al., 2003].
Thus, proton conduction may occur in a network of hydrogen-bonded water molecules
adsorbed on the Pt surface. Rao and co-workers have demonstrated that u may
significantly vary depending on the porous structure of carbon materials [Rao et al.,
2005]. This adds to the uncertainty in studying size and structural effects with the
single-cell approach.
The fifth limitation is related to the heterogeneity of the catalyst. The point is
that metal nanoparticles are usually deposited on high surface area carbon supports,
and, in order to provide high currents at low ohmic losses, often contain the active
component (e.g., Pt) in a high percentage. This gives rise to polydispersed catalysts
containing particle agglomerates. The presence and structure of the latter may have
dramatic influence on catalytic activity, as discussed in Cherstiouk et al. [2000,
2003a]; Maillard et al. [2004b, 2005, 2007a]; Gavrilov et al. [2007]. In the investi-
gation of PSEs, both commercial and “home-made” catalysts are used. The latter
are prepared through various synthetic approaches, including colloidal, microemul-
sion, incipient wetness impregnation, adsorption, electrodeposition, and chemical
and physical vapor deposition. The reader is referred to review articles for more infor-
mation about various preparative approaches to supported nanoparticle catalysts
[Aiken and Finke, 1999; Lewis, 1993; Simonov and Likholobov, 2003; Liu et al.,
2006; Ralph and Hogarth, 2002a, b; Hogarth and Ralph, 2002]. A detailed discussion
is beyond the scope of this chapter.
Search WWH ::
Custom Search