This
work presents investigations of the interaction of hydrogen with a Pd(210) and
Ni(210) surface using LEED (low energy electron diffraction), thermal
desorption spectroscopy (TDS), work-function measurements (DF),vibrational
loss measurements (HREELS), and isotope exchange experiments.
The
interaction of hydrogen with both surfaces is very similar.
For
temperatures above 100K, hydrogen chemisorbs spontaneously forming atoms in
three binding states (b1, b2, b3). The atomic
adsorption leads to a work function increase. If the temperature is reduced to
50K, hydrogen additionally chemisorbs into two molecular states (g1, g2). In contrast to
the b-adsorption, the
population of g-states induces a
work-function decrease.
The
molecular nature of the g-states was confirmed by observation of the H-H vibration mode in
HREELS, by the detection of the s-H2/Pd- and s*-H2/Pd-bonding in UPS, and by H2/D2-exchange
experiments.
The
energy range of the HREELS- and UPS-signals and the relatively high desorption
temperature in TD-spectra verify a real chemisorptive interaction between
molecule and surface. We consider a side-on complex configuration which allows
a s/s*-synergism
equivalent to the Blyholder backbonding mechanism for CO chemisorption or the
classical Dewar-Chatt-Duncanson model for the bonding situation in olefine or
other organometallic complexes.
A
molecular chemisorption of H2 on a transition metal surface is
unusual. In general, hydrogen adsorbs dissociatively. While on noble metals the
dissociation is hindered by a sizable energy barrier, it occurs spontaneously
on transition metal surfaces. If molecular adsorption states exist, they are
usually very weakly bound in shallow physisorption wells; beyond, it requires
surface temperatures below 20K to stabilize those states.
At
Ni(510), molecular chemisorption has been observed at surface temperatures up
to 125K, presumably at the steps, but only after the surface was passivated
with a dense atomic layer.
We
made similar observations, namely the coexistence of chemisorbed molecular and
atomic hydrogen on the relatively open Pd(210) and Ni(210) surfaces although
the surfaces were not fully passivated. While the atomic adsorption takes place
in high coordinated sites, the molecule adsorbs on top.
It
is very difficult to identify experimentally the exact location and nature of
the hydrogen adsorption states. In order to obtain this microscopic
information, A. Groß et al. performed DFT calculations. The results suggest
that the hydrogen molecule is first attracted to the top Pd atoms. At the clean
Pd(210) surface, no stable molecular adsorption state should exist and the
hydrogen molecule dissociates into the higher coordinated sites. Due to the
presence of atomic hydrogen on the surface however, this behavior changed
considerably. The presence of hydrogen atoms leads to a molecular adsorption
state at the top site. That the top site`s reactivity is hardly influenced by
the pre-adsorbed hydrogen atoms can be traced back to the induced change in the
local density of states.
In
the case of Pd, we observed the absorption of hydrogen atoms which reside
?close to the surface?. Evidence for such a ?subsurface state? is a
low-temperature thermal desorption feature which cannot be saturated, combined
with a vanishing work function change and negligible vibrational loss
contributions of the respective H state. We discuss our observations using the
stamping model developed by Okuyama.
The
hydrogen absorption has already been observed on other open palladium surfaces
and is considered as a specific feature of the element palladium. But in
contrast to other open Pd surfaces, the Pd(210) surface need not to reconstruct
for absorption. Futhermore, the absorption velocity depends on the preparation
of the sample. Both observations give rise to the assumption that a certain
ensemble of substrate atoms necessary for chemisorption exists on the
non-reconstructed Pd(210) surface. We can identify this ensemble as the
four-fold coordinated hollow site and the neighbouring three-fold
coordinated site.
In
addition, we examined the H/CO-coadsorption on the Pd(210) surface. If the Pd
crystal was first exposed to hydrogen and then to carbon monoxid, we observed
the formation of H/CO-complexes. This observation was verified by the detection
of a so-called S desorption signal in TDS and by a O-H
vibrational loss in HREELS. In contrast to other coadsorption systems the S-hydrogen
does not desorb in the low temperature area. Rather, hydrogen desorbs
simultaneously with the strongly bound carbon monoxid, indicating an unusually
stable complex.
