54 Pt-La and Pt-Ce alloys: the active chemical phase
form a thick oxide layer 144, 145.
Our experimental evidences are in clear discrepancy with the results and interpretation
of Kim and co-workers for Pt3La 99. On the basis of their DFT calculations
on a closely packed structure of Pt3La they claim that La atoms would be
present on the surface without forming any Pt overlayer. The alloying energy would
stabilize this structure not only in vacuum but also under reaction conditions: no
oxidation or dissolution of La would occur. They support this interpretation with
AR-XPS measurements, showing that La atoms are present on the surface and that
their chemical phase does not vary before and after electrochemical measurements.
However, the peak shape and position of their La 3d XPS peaks does not appear as
metallic: it is instead similar to those we observed for La oxide in Figure 3.4 c,d.
In general, these spectra are typical of compounds where La is in an oxidized state,
with very intense fN+1 components, almost as intense as the main La 3d peaks
130, 141, 146. This observation together with the fact that no explicit details
about the purity of their fabricated alloys or any surface cleaning procedures are
described in their work 99 suggest the formation of a rather thick La oxide layer.
It should be noted that the XRD measurements reported by Kim and co-workers
exhibit the same crystal structure of pure Pt with the presence of some lattice strain.
This indicates that their Pt3La phase should consist of a solid solution 147, 148,
meaning that the overall bulk composition might still be constant and close to the
stochiometry of Pt3La, but with a random substitution of Pt and La atoms at any
given atomic coordinate. Solid solutions are typical of metal compounds with weak
interactions between its constituents, as in the case of Pt3Co, in particular when
the annealing temperatures are not sufficiently high 147. In the case of Pt and La
these interactions should be strong, providing a significant driving force towards
the formation of an ordered intermetallic compound. In this case, the energetically
unfavorable La-La bonds would be minimized resulting in a compound with the
highest possible number of Pt-La nearest neighbours. Interestingly, the phase diagrams
found in literature for the Pt-La intermetallic compounds do not report the
existence of any stable Pt3La phase 128, 129. The XRD measurements on our
polycrystalline sample with nominal composition of Pt3La are in agreement with
these diagrams; they allowed to identify two coexisting phases: Pt2La and Pt5La
that are both reported as stable intermetallic compounds. In our view, Kim and coworkers
obtained a meta-stable solid solution due to their particular preparation
conditions: they fabricated thin films of Pt3La by magnetron sputtering at room
temperature, without any heat treatment that could facilitate the formation of an
ordered intermetallic compound. For such a meta-stable phase the low stability of
their Pt3La sample to oxidation in air is unsurprising.
Focusing now on Pt5La and Pt5Ce, and in particular on the AR-XPS results after
ORR testing (see Figures 3.4 e,f and 3.5 c,d), it is clear that both the La and
Ce oxides, as well as their metallic phases to some extent, are dissolved from the
near-surface region in acid. This tendency can be understood in terms of simple
thermodynamic considerations. First considering the Pt5La case, La2O3 is not ther