Publications

Abstract not provided.

Abstract not provided.

Abstract not provided.

Thermionic energy conversion in a miniature format shows potential as a viable, high efficiency, micro to macro-scale power source. A microminiature thermionic converter (MTC) with inter-electrode spacings on the order of microns has been prototyped and evaluated at Sandia. The remaining enabling technology is the development of low work function materials and processes that can be integrated into these converters to increase power production at modest temperatures (800 - 1300 K). The electrode materials are not well understood and the electrode thermionic properties are highly sensitive to manufacturing processes. Advanced theoretical, modeling, and fabrication capabilities are required to achieve optimum performance for MTC diodes. This report describes the modeling and fabrication efforts performed to develop micro dispenser cathodes for use in the MTC.

Density functional theory is used to predict workfunctions, {psi}. For relaxed clean W(1 0 0), the local density approximation (LDA) agrees with experiment better than the newer generalized gradient approximation, probably due to the surface electron self-energy. The large Ba metallic radius indicates it covers W(1 0 0) at about 0.5 monolayer (ML). However, Ba{sup 2+}, O{sup 2-}, and metallic W all have similar radii. Thus 1 ML of BaO (one BaO unit for each two W atoms) produces minimum strain, indicating commensurate interfaces. BaO (1 ML) and Ba (1/2 ML) have the same {psi} to within 0.02 V, so at these coverages reduction or oxidation is not important. Due to greater chemical activity of ScO vs. highly ionic BaO, when mixing the latter with this suboxide of scandia, the overlayer always has BaO as the top layer and ScO as the second layer. The BaO/ScO bilayer has a rocksalt structure, suggesting high stability. In the series BaO/ScO/, BaO/YO/, and BaO/LaO/W(1 0 0), the latter has a remarkably low {psi} of 1.3 V (LDA), but 2 ML of rocksalt BaO also has {psi} at 1.3 V. We suggest BaO (1 ML) does not exist and that it is worthwhile to attempt the direct synthesis and study of BaO (2 ML) and BaO/LaO.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Under ultrahigh vacuum conditions at 300 K, the applied electric field and/or resulting current from an STM tip creates nanoscale voids at the interface between an epitaxial, 7.0 angstroms thick Al2O3 film and a Ni3Al(1 1 1) substrate. This phenomenon is independent of tip polarity. Constant current (1 nA) images obtained at +0.1 V bias and +2.0 V bias voltage (sample positive) reveal that voids are within the metal at the interface and, when small, are capped by the oxide film. Void size increases with time of exposure. The rate of void growth increases with applied bias/field and tunneling current, and increases significantly for field strengths >5 MV/cm, well below the dielectric breakdown threshold of 12±1 MV/cm. Slower rates of void growth are, however, observed at lower applied field strengths. Continued growth of voids, to approximately 30 angstroms deep and approximately 500 angstroms wide, leads to the eventual failure of the oxide overlayer. Density functional theory calculations suggest a reduction-oxidation mechanism: interfacial metal atoms are oxidized via transport into the oxide, while oxide surface Al cations are reduced to admetal species which rapidly diffuse away. This is found to be exothermic in model calculations, regardless of the details of the oxide film structure; thus, the barriers to void formation are kinetic rather than thermodynamic. We discuss our results in terms of mechanisms for the localized pitting corrosion of aluminum, as our results suggest nanovoid formation requires just electric field and current, which are ubiquitous in environmental conditions.

First principles density-functional slab calculations are used to study 5 {angstrom} (two O-layer) Al{sub 2}O{sub 3} films on Ru(0001) and Al(111). Using larger unit cells than in a recent study, it is found that the lowest energy stable film has an even mix of tetrahedral (t) and octahedral (o) site Al ions, and thus most closely resembles the {kappa}-phase of bulk alumina. Here, alternating zig-zag rows of t and o occur within the surface plane, resulting in a greater average lateral separation of the Al-ions than with pure t or o. A second structure with an even mix of t and o has also been found, consisting of alternating stripes. These patterns mix easily, can exist in three equivalent directions on basal substrates, and can also be displaced laterally, suggesting a mechanism for a loss of long-range order in the Al-sublattice. While the latter would cause the film to appear amorphous in diffraction experiments, local coordination and film density are little affected. On a film supported by rigid Ru(0001), overlayers of Cu, Pd, and Pt bind similarly as on bulk truncated {alpha}-Al{sub 2}O{sub 3}(0001). However, when the film is supported by soft Al(111), the adhesion of Cu, Pd, and Pt metal overlayers is significantly increased: Oxide-surface Al atoms rise so only they contact the overlayer, while substrate Al metal atoms migrate into the oxide film. Thus the binding energy of metal overlayers is strongly substrate dependent, and these numbers for the above Pd-overlayer systems bracket a recent experimentally derived value for a film on NiAl(110).

Atoms and molecules adsorbed on metals affect each other even over considerable distances. In a tour-de-force of density-functional methods, the authors establish the nature and strength of such indirect interactions, and explain for what adsorbate systems they can critically affect important materials properties. These perceptions are verified in kinetic Monte Carlo simulations of epitaxial growth, and help rationalize a cascade of recent experimental reports on anomalously low diffusion prefactors. The authors focus their study on two metal systems: Al/Al(111) and Cu/Cu(111).

XPS studies have been carried out on sputter deposited copper on a substantially hydroxylated {alpha}-Al{sub 2}O{sub 3}(0001) (sapphire) surface under ultra-high vacuum (UHV) conditions. XPS-derived Cu uptake curves show a sharp change in slope at a coverage of 0.35 monolayer (on a Cu/O atomic basis), indicative of initial layer-by-layer growth. CU(LMM) lineshape data indicate that, prior to the first break in the curve, Cu is oxidized to Cu(I). At higher coverages, metallic CU(0) is. observed. These data agree with first principles theoretical calculations, indicating that the presence of ad-hydroxyl groups greatly enhances the binding of Cu to bulk sapphire surfaces, stabilizing Cu(I) adatoms over two-dimensional metallic islands. In the absence of hydroxylation, calculations indicate significantly weaker Cu binding to the bulk sapphire substrate and non-wetting. Calculations also predict that at Cu coverages above 1/3 monolayer (ML), Cu-Cu interactions predominate, leading to Cu(0) formation. These results are in excellent agreement with experiment. The ability of surface hydroxyl groups to enhance binding to alumina substrates suggests a reason for contradictory experimental results reported in the literature for Cu wetting of alumina.

In this paper, we include for discussion three topics of current interest in metal oxide surface science. Using first principles density functional theory (DFT) [1] calculations, we have investigated: (1) the atomic-scale structure of experimentally-relevant ultrathin alumina films, (2) the role of common point defects in metal island nucleation on oxide terraces, and (3) the growth and morphology of metals on oxide surfaces which have high concentrations of a common impurity.

The rutile TiO2(110) (1 × 1) surface is considered the prototypical 'well-defined' system in the surface science of metal oxides. Its popularity results partly from two experimental advantages: (i) bulk-reduced single crystals do not exhibit charging, and (ii) stoichiometric surfaces, as judged by electron spectroscopies, can be prepared reproducibly by sputtering and annealing in oxygen. We present results that show that this commonly applied preparation procedure may result in a surface structure that is by far more complex than generally anticipated. Flat, (1 × 1)-terminated surfaces are obtained by sputtering and annealing in ultrahigh vacuum. When re-annealed in oxygen at moderate temperatures (470-660 K), irregular networks of partially connected, pseudohexagonal rosettes (6.5 × 6 Å wide), one-unit cell wide strands, and small (≈tens of Å) (1 × 1) islands appear. This new surface phase is formed through reaction of oxygen gas with interstitial Ti from the reduced bulk. Because it consists of an incomplete, kinetically limited (1 × 1) layer, this phenomenon has been termed 'restructuring'. We report a combined experimental and theoretical study that systematically explores this restructuring process. The influence of several parameters (annealing time, temperature, pressure, sample history, gas) on the surface morphology is investigated using STM. The surface coverage of the added phase as well as the kinetics of the restructuring process are quantified by LEIS and SSIMS measurements in combination with annealing in 18O-enriched gas. Atomic models of the essential structural elements are presented and are shown to be stable with first-principles density functional calculations. The effect of oxygen-induced restructuring on surface chemistry and its importance for TiO2 and other bulk-reduced oxide materials is briefly discussed.

A massively-parallel ab initio computer code, which uses Gaussian bases, pseudopotentials, and the local density approximation, permits the study of transition-metal systems with literally hundreds of atoms. We present total energies and relaxed geometries for Ru, Pd, and Ag clusters with N = 55, 135, and 140 atoms; we also used the DMOL code to study 13-atom Pd and Cu clusters, with and without hydrogen. The N = 55 and 135 clusters were chosen because of simultaneous cubo-octahedral (fcc) and icosahedral (icos) sub-shell closings, and we find icos geometries are preferred. Remarkably large compressions of the central atoms are observed for the icos structures (up to 6% compared with bulk interatomic spacings), while small core compressions ({approx} 1 %) are found for the fcc geometry. In contrast, large surface compressive relaxations are found for the fcc clusters ({approx} 2-3% in average nearest neighbor spacing), while the icos surface displays small compressions ({approx} 1%). Energy differences between icos and fcc are smallest for Pd, and for all systems the single-particle densities of states closely resembles bulk results. Calculations with N = 134 suggest slow changes in relative energy with N. Noting that the 135-atom fcc has a much more open surface than the icos, we also compare N = 140 icos and fcc, the latter forming an octahedron with close packed facets. These icos and fcc clusters have identical average coordinations and the octahedron is found to be preferred for Ru and Pd but not for Ag. Finally, we compare Harris functional and LDA energy differences on the N = 140 clusters, and find fair agreement only for Ag.

Electronic excitations in adsorbate layers stimulate desorption and dissociation of adsorbed molecules as well as chemical reactions between adsorbates. The highest-probability stimulated processes produce neutral desorbates and determine how surface composition is altered by electron or photon radiation. A basic understanding has emerged, due largely to laser resonance-enhanced multi-photon ionization (REMPI) experiments, which provide quantum-state resolution of the gas-phase products. Auger phenomena enter this understanding in several ways. For example, CVV Auger spectroscopy determines the screened hole-hole interaction, U, in adsorbates, which in turn provides insight into the degree of charge-transfer screening from the substrate. In those systems where screening charge is used in excitation Auger decay, screening directly determines the lifetime, which in turn can exponentially affect the yield. Reductions in screening, e.g. induced by coadsorption of electro-negative species, thus can result in giant yield enhancements. As separate issues, a finite U may prevent the fast resonant decay and thus increase the yield from two-hole excitations, as has been suggested for NO2 dissociation on Pt(lll), or may assist in the localization (self-trapping) of two-hole excitations in dense adsorbate layers, as apparently is the case for NO desorption from the same surface. The latter causes the yields from one- and two-hole excitations to differ in their coverage dependence. Finally, CVV Auger spectroscopy, of course, measures the energies of two-hole excitations, which can be correlated with observed stimulated thresholds. © 1992 IOP Publishing Ltd.

We present calculations which show the radial dependence of the KVV and L12VV Auger matrix elements of silicon. We find greatly differing dependences, converging within ~ 1 a.u. of the nucleus in the case of the KVV, but not until — 4 a.u. in the case of the L23VV, well beyond the bond midpoint of — 2.2 a.u. We also find quite different dependences for the various elements within a particular CVV transition. Because the local density of states (LDOS) is dependent on the radius of the sphere of integration, our results suggest that different CVV Auger processes on the same atom in fact probe different LDOSs, as do even different contributions within the same transition. (This effect is separate from the well-known matrix element property which weights angular-momentum components differently.) These results call into question both the single-site LDOS approximation when used in the interpretation of low-energy ( < 100 eV) Auger spectra, and the application to high-energy spectra of local densities of states obtained by integration over muffin-tin or Wigner-Seitz spheres which have a large radius compared to the region probed by the Auger process. © 1992, American Vacuum Society. All rights reserved.