Support Interaction and Thermal Stability of Size Selected Cluster Deposited onto Single Crystal Surfaces

Top down nanofabrication in its quest for smaller structures eventually runs into the problem of granularity. Chemists call these grains clusters. Harnessing such clusters and their properties is one of the greatest challenges for atomically precise nanotechnology. Towards this end it is necessary to understand structure and quantum size effects for clusters - both in their isolated state as well as under realistic, surface trapped conditions. We pursue these two goals in this subproject.

We have studied a number of gas-phase, elemental cluster systems, with the aim of determining the evolution of cluster structure with number of constituent atoms. The main experimental techniques we use are gas phase ion mobility spectrometry (IMS) and trapped ion electron diffraction (TIED). In addition, ion chemistry of metal cluster ions is probed in an ion cyclotron resonance mass spectrometer (FT-ICR). Below we provide a few recent examples of our work.

Gold cluster anions Au_n^-(n≤20) show a wide diversity of structural phases ranging from planar (n=4-12) through three-dimensional “flat-stacked” (n = 12-15), hollow (m=16-18), and fcc-like (n=19, 20) structures. Interestingly, most clusters in this size range appear to comprise a single dominant isomer. In a few cases, the experimental data could only be fit by assuming a superposition of multiple coexisting isomeric forms. Typically the experimentally best fitting structure is either the DFT global minimum or an energetic close-lying isomer. However, comparison to experiment has uncovered systematic problems of DFT descriptions for certain cluster sizes, e.g. in the region of the 2D to 3D transition. At larger sizes (n=34), compact surface reconstructed structures were found.[1]

Small mixed gold-silver clusters Au_nAg_m⁺ (n+m = 4-6) are prototypic clusters of a fully miscible alloy. Detailed adsorption kinetics were studied and the resulting absolute rate constants were analyzed to yield binding energies of carbon monoxide. A CO molecule always binds in an atop position on a single gold atom with a binding energy that generally decreases with increasing number of neighboring silver atoms. This can be understood by the increased charge density on the gold atoms due to their higher electronegativity (relative to silver) and the associated weakening the σ-donor bond of CO. In principle, "tunable" binding energies by way of varying the composition may be relevant to catalytic processes where specific reaction steps needs to be properly activated.[2]

We have determined the structures of boron cluster cations using ion mobility spectrometry and quantum chemical calculations: they are planar up to B₁₅⁺, starting at B₁₆⁺ we find cyclindrical structures (double rings comprised of connected B₃-triangles). Icosahedral structures, prominent in bulk phase boron, are completely absent in boron clusters with less than 25 atoms. Instead, we see a tendency towards cylinders having two or more boron ring segments, i.e. towards single-walled boron nanotubes [3].

Tin clusters are promising candidates for cluster materials since their binding energies as function of cluster size quickly converge to the bulk value. We have determined the structures of positively and negatively charged tin clusters by IMS and TIED. Small clusters have (quasi) compact structures. Sn₁₂^-is an icosahedral-cage-structure. For larger clusters Sn₁₃^- - Sn₂₅^- prolate structures were found. Interestingly, these clusters are built from multiply-repeating subunits, which are either face connected or which form dimers/clusters of subunits (mainly tricapped trigonal prism and/or bicapped quadratic antiprism). This rather unusual growth mode can be rationalized by the high stability of the subunit clusters.[4]

B: Size Selected Clusters on Single Crystal Surfaces

Concerning clusters on surfaces, wehave focused on the fabrication of new monodisperse materials consisting of mass-selected non-IPR fullerene cages. The corresponding growth procedure is based on soft landing of the mass-selected cluster ions onto well-defined substrates (i.e. HOPG, Au(111), a-C, and others). The non-IPR carbon cluster building blocks, C_n, were created via electron impact ionization/fragmentation of larger IPR fullerenes (C₆₀ and C₇₀). The thin film deposits were prepared under ultra-high vacuum conditions via selecting the desired ions, C_n⁺ (46 < n < 70) and collecting them on inert surfaces at low incident kinetic energies (< 6 eV). Non-IPR fullerene cages may be thought of as  functionalized (ligand free) carbon clusters, i.e. the non-IPR sites located on the cage “surfaces” exhibit enhanced reactivity allowing e.g., for covalent cage-cage interconnection. Consequently such systems are of potential interest for all-carbon nanotechnology: the sub-nm, building blocks are quite stable and can be easily positioned, manipulated and used as a patterning medium on surfaces. The patterns created by soft-landing can be readily tuned by controlling deposition conditions or by depositing the ions onto intentionally (pre-)patterned surfaces [5].

The thermal stability of the films is governed by the quasi-covalent intercage bonds formed between non-IPR sites on adjacent cages, i.e. in the fully coordinated solid, non-IPR fullerenes can interlink to form  -C_n-C_n-C_n- oligomers in 2- and 3-dimensions. Thermal desorption studies, TDS, reveal that the mean binding energy of the building blocks depends on the size of the individual cages. All non-IPR materials exhibit considerably higher thermal stability than do the corresponding parent IPR solids, C₆₀ and C₇₀[6].

The valence bands of the C_n solids reveal a clear dependence of the DOS on the size of the constituting cages. All non-IPR C_n solids exhibit lower surface ionisation potentials, I_p, than do the IPR films. I_p varies with n in the range from 6.4 eV up to 6.9 eV. All C_n materials appear to be wide band semiconductors with the HOMO-LUMO gap ranging from 0.7 to 1.9 eV.

DFT calculations have predicted that it should be possible to form “Stone-Walls transformed” non-IPR C₆₀ cages (at elevated temperatures ~ 4400 K). These new stable C₆₀ isomers should exhibit two non-IPR sites. We have been able to generate such cage isomers by electron-impact induced fragmentation of C₇₀. By soft-landing these (and selectively subliming the more volatile C₆₀(I_h) component of the deposited isomer mixture), we have fabricated a new monodisperse material consisting of the C₆₀(C_2v) isomers exclusively. This material shows unique electronic properties [7].

The hydrogenation/deuteration of non-IPR materials by reaction with atomic hydrogen has been investigated as a simple indicator of reactivity (and of reactivity differences to IPR materials). Progressing formation of C_nH_x hydrides reduces the mean binding energy of -C_n-C_n- oligomers. Finally, oligomers decompose entirely into hydrogenated C_n cages. The transformation of dendritic islands into compact smooth rimmed islands is an indicator of the associated cleavage of covalent intercage bonds in the oligomers – to yield hydrogenated cages with much weaker van der Waals interactions.

References

List of Publications 2006-2011 as PDF

Subproject Report 2006-2010 as PDF