THE STRUCTURE OF THE P(ROOT-3X-ROOT-3) R30-DEGREES BILAYER OF D2O ON RU(001)

Authors
Citation
G. Held et D. Menzel, THE STRUCTURE OF THE P(ROOT-3X-ROOT-3) R30-DEGREES BILAYER OF D2O ON RU(001), Surface science, 316(1-2), 1994, pp. 92-102
Citations number
43
Categorie Soggetti
Chemistry Physical
Journal title
ISSN journal
00396028
Volume
316
Issue
1-2
Year of publication
1994
Pages
92 - 102
Database
ISI
SICI code
0039-6028(1994)316:1-2<92:TSOTPR>2.0.ZU;2-#
Abstract
We present the first complete LEED-IV analysis of an ordered water lay er adsorbed on a metal surface, the structure of p(square-root 3 x squ are-root 3)R30-degrees-2D2O/Ru(001), commonly assumed to be a bilayer. The best fit geometry is a buckled hexagonal arrangement of water mol ecules adsorbed on top of first-layer ruthenium atoms. The main, at fi rst sight surprising feature of the surface geometry is that the water ''bilayer'' is almost flat: the vertical distance between the O atoms of the two types of molecules in the p(square-root 3 x square-root 3) R30-degrees unit mesh (DELTAz(O-O) is 0.10 +/- 0.02 angstrom. Buckling of the first Ru layer is found in anticorrelation to the water molecu les (0.01 +/- 0.04 angstrom and 0.07 +/- 0.02 angstrom with respect to the uncovered Ru atoms), so that the Ru-O distances for the two types of molecules are still significantly different from each other (2.08 +/- 0.02 angstrom and 2.23 +/- 0.02 angstrom). We conclude from this t hat the construction principle of the bilayer (chemical bond between t he Ru atoms and the O atoms of the closer molecules, and hydrogen bond ing from the latter to the other molecules) is still correct. Dipole-m irror dipole attraction and charge-transfer-induced changes of the D2O bond angle are discussed as possible explanations for the compression of the ideal bilayer structure, which cannot be explained by the infl uence of the lattice mismatch between Ru(001) and ice. The vertical di stance between the first- and second-layer Ru atoms ranges from 2.04 /- 0.04 angstrom to 2.11 +/- 0.03 angstrom due to the first-layer buck ling, i.e. is always - and in part considerably - smaller than in the bulk.