Determination of Spherosiloxane Cluster Bonding to Si(100)-2×1by Scanning Tunneling Microscopy

Abstract

Scanning tunneling microscopy is used to determine the bonding geometry of the spherosiloxane cluster, H(8)Si(8)O(12) , on Si(100)-2 x 1. The images obtained are consistent with monovertex bonding to the Si(100)-2 x 1 surface via activation of a single Si-H bond. Filled and empty state images show good agreement with calculations of the electron density distribution of the cluster as well as the Psi(2) highest occupied molecular orbital and lowest unoccupied molecular orbital surface plots of the cluster.

Simulated scanning tunneling microscopy images of three-dimensional clusters:H8Si8O12onChen1, Schneider2, Holl3 et al. 2004Phys. Rev. BSelf Cite

“…The second apparent discrepancy is the distance between individual cluster features ͑ϳ5 Å͒ in the filled-states image. If these features arise from the oxygen lone pair electrons, as has previously been reported, 12 one may expect feature separations of ϳ2.8 Å in accordance with the actual cluster geometry. With these ideas in mind, we turn to simulation to examine the two adsorption models and to understand the origin of the image features and dimensions.…”

“…4(a)]. 12 In the empty-states image of the same chemisorbed cluster, Fig. 3(b), only a single continuous oblong feature is apparent.…”

“…The monovertex attachment model contains the following key propositions: (1) the cluster cage remains intact, (2) the cluster is activated at one of the Si-H vertices, and (3) the cluster bonds to a Si dimer atom via a single Si-Si bond. [9][10][11][12] The cracked cluster model features Si-O bond scission along a cluster cage edge, resulting in a bridged bonding configuration across a Si dimer. 13,14 For these models, the Si 9 H 12 cluster was employed to simulate the silicon surface dimer to which the cluster is bonded.…”

“…13,14 Although structural assignments based upon XPS 9,10 and RAIRS 11 data have been controversial, interpretations of STM data, based largely on a comparison between experimental results and electron density maps derived from nonlocal density-functional theory (DFT) calculations, strongly support the monovertex attachment model. 12 However, the simple model previously utilized for the purposes of this comparison does not address the discrepancy between the measured experimental apparent height of the cluster and its actual geometric size. As an example, the distance between molecular features associated with the electron density of cluster oxygen atoms is larger than expected in filled-states experimental STM images.…”

“…Currently, two bonding models for the chemisorbed product have been proposed: experimental evidence has pointed towards the monovertex attachment model [ Fig. 1(a)], [9][10][11][12] whereas theoretical studies have suggested the cracked cluster model [ Fig. 1(b)].…”