Hydrogen sorption sites in holmium silicide on silicon(1 1 1) Christopher Eames a, * , Charles Woffinden a , Matthew I.J. Probert a , Steve P. Tear a , Andrew Pratt b a Department of Physics, University of York, York YO10 5DD, United Kingdom b York Institute for Materials Research, Department of Physics, University of York, York YO10 5DD, United Kingdom article info Article history: Received 30 November 2009 Accepted for publication 20 January 2010 Available online 2 February 2010 Keywords: Density functional calculations Metastable induced electron spectroscopy abstract The hydrogen sorption sites on the surface of holmium silicide grown on Si(1 1 1) have been determined using metastable de-excitation spectroscopy, ultraviolet photoemission spectroscopy and density func- tional theory calculations. Comparison of calculated and measured surface density of states spectra allow us to locate the position of the second subsurface hydrogen atom in each unit cell to an interstitial site in the layer of rare earth atoms. The hydrogenation energies indicate a reaction pathway that involves con- comitant site occupation. Ó 2010 Elsevier B.V. All rights reserved. 1. Introduction The interaction of hydrogen with surfaces is of fundamental interest because it is the most basic adsorbate and acts as a proto- type for understanding more complex chemisorption phenomena. An understanding of hydrogen storage in fuel cells is completely reliant on these processes [1]. Silicon device fabrication is also af- fected by the presence of hydrogen [2]. In this work we are concerned with the interaction of hydrogen with silicides of the rare earth metals grown on the Si(1 1 1) sur- face. Rare-earth silicides are of great interest due to the properties they exhibit as metal/semiconductor interfaces in a number of materials systems including heterostructures [3], nanowires [4], and sensors [5]. Depending on the amount of RE deposited, either a 2D or 3D silicide may form with the atomic structure of each shown schematically in Fig. 1. In both cases the surface of the sil- icide consists of a buckled Si bilayer that sits above a flat RE layer. The 3D silicide then has further, alternating layers of graphite-like Si and RE atoms with a network of vacancies distributed through- out the flat Si layers to relieve strain. This leads to a ffiffiffi 3 p  ffiffiffi 3 p   R30  stoichiometry with 1 in every 6 Si atoms missing. The interaction of hydrogen with the 2D and 3D RE silicides on Si(1 1 1) results in fundamental changes to the surface structure and the electronic properties. H adsorption onto a 2D RE silicide re- sults in a transition from a semimetal to a semiconductor [6]. The resulting passivation of the 2D HoSi 2 surface has also been investi- gated as a base for further overlayer growth [7]. Structurally, hydrogenation causes the surface bilayer to flip from type B in which the buckling direction is opposite that of the bulk to type A where they are equivalent [8]. This adsorption results in the interlayer spacing between the top Si bilayer and the flat RE layer expanding by 0.32 Å. In the 3D RE silicides H is known to bond to the top bilayer and to interstitial sites as in the 2D silicides but it has also been suggested that there is additional H passivation of dangling bonds in the vacancy network [9]. To determine the hydrogenation sites in the 2D RE silicides the electronic density of states (DOS) has been calculated with the ex- tended Hückel method and compared with angle-resolved photo- emission spectroscopy (ARPES) data [10]. This work suggests that two H atoms are chemisorbed into each unit cell in the 2D silicides, one bonding atop the silicon bilayer to the dangling bonds protrud- ing into the vacuum and the other to interstitial sites in the RE hex- agonal layer. It is not known definitively which site is occupied by this second, absorbed, H atom. The authors remark that from their ARPES data they could only ‘tentatively assign’ a band to the Er–H interstitial bonds. The Hückel model calculations indicate which of the interstitial sites has the larger hydrogenation energy and a comparison of calculated and measured band structures adds fur- ther evidence as to which interstitial site is occupied. However, there are problems with this method. The structures were not geometry optimised and the structural parameters were assumed using reasonable values extrapolated from experiment. In our experience non-optimised atomic positions can raise the energy of a model structure by more than 1 eV. Also, a complete range of possible model structures was not covered. The authors them- selves remarked that the work could be extended by looking at more models with more precise ab initio techniques. The aims of the density functional theory (DFT) calculations in this work are to provide a more accurate and complete survey of this system and to directly compare it with an experimental tech- nique that has sufficient surface sensitivity to unambiguously determine the second hydrogenation site. In a recent paper [11] we have demonstrated how metastable de-excitation spectroscopy 0039-6028/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.susc.2010.01.016 * Corresponding author. E-mail addresses: ce124@york.ac.uk (C. Eames), spt1@york.ac.uk (S.P. Tear). Surface Science 604 (2010) 686–691 Contents lists available at ScienceDirect Surface Science journal homepage: www.elsevier.com/locate/susc