Publications

It has been recognized that documentation for new customers of ASCI Red, aka janus or the Intel Teraflops at Sandia National Laboratories, has been sadly lacking. This document has been prepared by a team of subject matter experts to fill that void and to provide a starting point for providing a similar document for ASCI Red Storm in the future. This document is intended for SNL users who need to jumpstart their use of Janus and Janus-s.

Abstract not provided.

Carbon is an important support for heterogeneous catalysts, such as platinum supported on activated carbon (AC). An important property of these catalysts is that they decompose upon heating in air. Consequently, Pt/AC catalysts can be used in applications requiring rapid decomposition of a material, leaving little residue. This report describes the catalytic effects of platinum on carbon decomposition in an attempt to maximize decomposition rates. Catalysts were prepared by impregnating the AC with two different Pt precursors, Pt(NH{sub 3}){sub 4}(NO{sub 3}){sub 2} and H{sub 2}PtCl{sub 6}. Some catalysts were treated in flowing N{sub 2} or H{sub 2} at elevated temperatures to decompose the Pt precursor. The catalysts were analyzed for weight loss in air at temperatures ranging from 375 to 450 C, using thermogravimetric analysis (TGA). The following results were obtained: (1) Pt/AC decomposes much faster than pure carbon; (2) treatment of the as-prepared 1% Pt/AC samples in N{sub 2} or H{sub 2} enhances decomposition; (3) autocatalytic behavior is observed for 1% Pt/AC samples at temperatures {ge} 425 C; (4) oxygen is needed for decomposition to occur. Overall, the Pt/AC catalyst with the highest activity was impregnated with H{sub 2}PtCl{sub 6} dissolved in acetone, and then treated in H{sub 2}. However, further research and development should produce a more active Pt/AC material.

Magnesium vanadates are potentially important catalytic materials for the conversion of alkanes to alkenes via oxidative dehydrogenation. However, little is known about the active sites at which the catalytic reactions take place. It may be possible to obtain a significant increase in the catalytic efficiency if the effects of certain material properties on the surface reactions could be quantified and optimized through the use of appropriate preparation techniques. Given that surface reactivity is often dependent upon surface structure and that the atomic level structure of the active sites in these catalysts is virtually unknown, we desire thin film samples consisting of a single magnesium vanadate phase and a well defined crystallographic orientation in order to reduce complexity and simplify the study of active sites. This report describes the use of reactive RF sputter deposition to fabricate very highly oriented, stoichiometric Mg{sub 3}(VO{sub 4}){sub 2} thin films, and subsequent studies of the reactivity of these films under reaction conditions typically found during oxidative dehydrogenation. We demonstrate that the synthesis methods employed do in fact result in stoichiometric films with the desired crystallographic orientation, and that the chemical behavior of the films closely approximates that of bulk, high surface area Mg{sub 3}(VO{sub 4}){sub 2} powders. We further use these films to demonstrate the effects of oxygen vacancies on chemical behavior, demonstrate that surface composition can vary significantly under reaction conditions, and obtain the first evidence for structure sensitivity in Mg{sub 3}(VO{sub 4}){sub 2} catalysts.

A new method to generate chemical reaction network is proposed. The particularity of the method is that network generation and mechanism reduction are performed simultaneously using sampling techniques. Our method is tested for hydrocarbon thermal cracking. Results and theoretical arguments demonstrate that our method scales in polynomial time while other deterministic network generator scale in exponential time. This finding offers the possibility to investigate complex reacting systems such as those studied in petroleum refining and combustion.

Magnesium vanadates are potentially important catalytic materials for the conversion of alkanes to alkenes via oxidative dehydrogenation. However, little is known about the active sites at which the catalytic reactions take place. It may be possible to obtain a significant increase in the catalytic efficiency if the effects of certain material properties on the surface reactions could be quantified and optimized through the use of appropriate preparation techniques. Given that surface reactivity is often dependent upon surface structure and that the atomic level structure of the active sites in these catalysts is virtually unknown, we desire thin film samples consisting of a single magnesium vanadate phase and a well defined crystallographic orientation in order to reduce complexity and simplify the study of active sites. We report on the use of reactive RF sputter deposition to fabricate very highly oriented, stoichiometric Mg{sub 3}(VO{sub 4}){sub 2} thin films for use in these surface analysis studies. Deposition of samples onto amorphous substrates resulted in very poor crystallinity. However, deposition of Mg{sub 3}(VO{sub 4}){sub 2} onto well-oriented, lattice-matched thin film ''seed'' layers such as Ti(0001), Au(111), or Pt(111) resulted in very strong preferential (042) crystallographic orientation (pseudo-hexagonal oxygen planes parallel to the substrate). This strong preferential growth of the Mg{sub 3}VO{sub 4}{sub 2} suggests epitaxial (single-crystal) growth of this mixed metal oxide on the underlying metal seed layer. The effects of the seed layer material, deposition temperature, and post-deposition reactive treatments on thin film properties such as stoichiometry, crystallographic orientation, and chemical interactions will be discussed.

We report propane dehydrogenation behavior of catalysts prepared using two novel synthesis strategies that combine inverse micelle Pt nanocluster technology with silica and alumina sol-gel processing. Unlike some other sol-gel catalyst preparations. Pt particles in these catalysts are not encapsulated in the support structure and the entire Pt particle surface is accessible for reaction. Turnover frequencies (TOF) for these catalysts are comparable to those obtained over Pt catalysts prepared by traditional techniques such as impregnation, yet the resistance to deactivation by carbon poisoning is much greater in our catalysts. The deactivation behavior is more typical of traditionally prepared PtSn catalysts than of pure Pt catalysts.

The conversion of small alkanes into alkenes represents an important chemical processing area; ethylene and propylene are the two most important organic chemicals manufactured in the U.S. These chemicals are currently manufactured by steam cracking of ethane and propane, an extremely energy intensive, nonselective process. The development of catalytic technologies (e.g., selective dehydrogenation) that can be used to produce ethylene and propylene from ethane and propane with greater selectivity and lower energy consumption than steam cracking will have a major impact on the chemical processing industry. This report details a study of two novel catalytic materials for the selective dehydrogenation of propane: Cr supported on hydrous titanium oxide ion-exchangers, and Pt nanoparticles encapsulated in silica and alumina aerogel and xerogel matrices.

The use of hydrous titanium oxide (HTO) ion-exchange materials as supports for iron and chromium based dehydrogenation catalysts is compared to current commercial catalyst systems in order to determine the potential of HTO technology for impacting this important chemical processing area. The best Fe/HTO catalysts synthesized to date achieve ethylbenzene conversions to styrene approaching those of commercial catalysts, even though the Fe/HTO catalysts contain no promoters while the commercial catalysts contain several different promoters, including K, Cr, and Ce. Addition of promoters to Fe/HTO catalyst is expected to result in further conversion improvements such that the activity of the commercial catalysts may be equaled or exceeded. Fe/HTO and Cr/HTO catalysts achieve only modest conversions of isobutane to isobutene that are far below available commercial catalysts. With the Cr/HTO catalysts, however, activity normalized to Cr loading far exceeds that of the commercial catalyst. Since optimum Cr loading conditions have not yet been identified, there is ample room for increases in both Cr loading and catalyst activity. Even if Cr/HTO and Fe/HTO catalysts do not ultimately exceed the performance obtained with commercial catalysts, the ability to cast HTO materials in the form of thin films may present important advantages for catalytic membrane reactor systems. These potential advantages are discussed and evaluated.

This project focuses on the modification of silica and alumina surfaces by titania and hydrous titanium oxide ion-exchange films, and the use of these modified materials as supports for MoS{sub 2} catalysts. FTIR studies of molybdena interaction with {gamma}-Al{sub 2}O{sub 3} demonstrate that at low loadings Mo interacts with the most basic hydroxyl groups, and that these hydroxyls are associated with tetrahedrally coordinated Al. Furthermore, hydrodesulfurization (HDS) activity as a function of Mo loading shows a maximum in specific activity with loading. The Mo species bound to tetrahedrally coordinated Al sites are therefore believed to be inactive for the HDS reaction. Only after the tetrahedral Al sites have completely consumed does molybdena adsorb on the alumina in a manner that leads to an active catalyst. According to this scheme, the activity of alumina supported MoS{sub 2} catalysts could be greatly improved by either titrating the tetrahedral Al sites with a modifier, or by using {alpha}-Al{sub 2}O{sub 3} which contains no tetrahedrally coordinated Al. HDS tests over MoS{sub 2} supported on both {alpha}-Al{sub 2}O{sub 3} and {gamma}-Al{sub 2}O{sub 3} modified by a titania film confirm this hypothesis. Neither support material gives rise to a maximum in activity with Mo loading, but rather exhibits a smooth decrease in activity with loading. Furthermore, for equivalent Mo loadings the activity of both of these support materials exceeds that of unmodified {gamma}-Al{sub 2}O{sub 3} due to the fact that no Mo is tied up in the inactive form. FTIR, XPS, and TEM are currently being used to determine whether the model can indeed account for the observed activity trends. Although the surface area of {alpha}-Al{sub 2}O{sub 3} is too low for use as a commercial catalyst, the titania coated {gamma}-Al{sub 2}O{sub 3} represents an important, practical improvement in support materials for hydrotreating catalysts.

Promoted iron catalysts are commonly used for Fischer-Tropsch (F-T) synthesis. Copper, potassium and silica are frequently employed as promoter species, either singly or in combination. The number of different iron catalyst formulations which have been investigated for F-T synthesis is enormous and there does not yet appear to be a general consensus as to the optimum catalyst composition. In addition, questions regarding the effects of variations in catalyst activation and reaction conditions are still open. Because of the large number of parameters involved in the development of F-T catalysts, a great deal of work remains to be done before the factors affecting catalyst performance are fully understood. In this paper one of these factors, namely the effects of variations in activation procedure on the surface composition of iron based F-T catalysts, will be investigated. Two different catalysts were studied. The first catalyst, with a composition of 100 Fe/5 Cu/4.2 K/25 SiO{sub 2} (parts by weight) whose little variation in activity procedure (1). The second catalyst (100 Fe/3 Cu/0.2 K) displays wide variations in activity with activation procedure (2). Surface compositions of these two catalysts were measured, after the activation treatments described above, using Auger electron spectroscopy (AES). It will be shown that the variations in catalyst activity observed by Bukur, et al. (1,2), correlate well with variations in surface composition, offering insights into the optimum conditions for catalyst activation. 9 refs., 6 figs.