C. Cohen et al., STUDY OF ATOMIC RELAXATIONS ON CLEAN AND OXYGEN-COVERED (100)COPPER, (410)COPPER AND (510)COPPER SURFACES BY CHANNELING, Surface science, 339(1-2), 1995, pp. 41-56
Atomic relaxations on Cu(100) and stepped Cu(410) and (510) surfaces w
ere measured using a few hundred keV He+ ions in channeling and double
alignment geometries. On stepped surfaces, we have also attempted to
determine the specific relaxation of edge atoms (both the components p
arallel and perpendicular to the (100) terraces). For this purpose, we
have analyzed blocking effects on particles backscattered near theta(
lab) = 90 degrees, around directions parallel to various major crystal
lographic axes of the terraces, the incident beam being aligned along
the [100] axis perpendicular to the terraces. The studies were perform
ed on clean and on oxygen covered surfaces. On (100) flat surfaces, we
find a small contraction of the surface plane on clean Cu (about 2% o
f the bulk interplanar spacing), and a marked expansion (about 10%) on
oxygen covered Cu (at saturation coverage, whose value was measured b
y nuclear reaction analysis). On stepped surfaces, it appears difficul
t to discriminate between the specific relaxation of the [001] edge ro
w perpendicular to the terraces and the relaxation of the other [001]
rows inside the terraces. Concerning the component of the [001] edge r
ow relaxation in the terrace plane, our results provide an unambiguous
estimate: we find a small contraction (about 0.04 Angstrom) on bare C
u and a larger expansion (about 0.18 Angstrom) under oxygen coverage.
In all cases, we observed that the presence of oxygen induces a defini
te surface disordering which can be described as random small atomic d
isplacements of Cu surface atoms around their mean position (on the or
der of 0.1 Angstrom). The results obtained on the (100) surface covere
d with oxygen can be interpreted in terms of a missing row model which
has been proposed by some authors. However, our data can also be fitt
ed with no missing row. Similarly, on the oxygen covered (410) stepped
surface the data interpretation does not require the assumption of a
missing row, but the results are not inconsistent with a description o
f the terrace where the fourth or possibly the third [001] row, parall
el to step edge, is missing and where the edge [001] row suffers a con
siderable outward displacement in the terrace plane (approximate to 0.
3 Angstrom). The limits of high energy ion channeling for the measurem
ent of surface atomic displacements, in particular when many sites are
involved (for instance in the case of alternate relaxation) are discu
ssed and illustrated.