Papers

AVS1996 Session SS-ThM: Metal Surface Reactions II

Thursday, October 17, 1996 8:20 AM in Room 203B

Thursday Morning

Start	Invited?	Item
8:20 AM		SS-ThM-1 Density Functional Theory of Chemical-Reaction Dynamics M. Scheffler, A. Gross, S. Wilke, C. Wei (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany) Advances in theoretical methods now afford density-functional-theory calculations of complete reaction pathways of small molecules at surfaces. The calculated (high-dimensional) potential-energy surfaces provide insight into, e.g., the surface chemical reactivity, mechanisms of poisoning catalytic reactions, and the site dependence of surface reactions at surfaces of alloys.For chemical reactions involving hydrogen our calculations (treating electrons as well as H nuclei quantum-mechanically) establish the importance of a HIGH-DIMENSIONAL treatment of the molecule-surface interaction DYNAMICS for a proper understanding of sticking, dissociative adsorption, and associative desorption. Quantum effects are found to be significantly more important that hitherto assumed. For example, we identify sharp quantum oscillations in the reaction probability and show how the building up of zero-point effects, generated by the H\sub 2\-surface interaction, affects the chemical processes and rates.The emphasis in this talk will be on the high-dimensional dynamics of the H\sub 2\ interaction with transition metals and on the mechanism of poisoning the catalytic activity of Pd by adsorbed sulfur, but several other reactions will be briefly discussed as well.
9:00 AM		SS-ThM-3 A Monte-Carlo Approach to the Oxidation of CO on Platinum F. Behrendt (Universit\um a\t Heidelberg, Germany) The numerical simulation of transient processes like catalytic ignition requires a detailed knowledge of elementary reaction steps both at the gas-surface interface and in the gas phase as well as of the transport processes between both phases. Using these detailed models, structural effects at the surface (e.g., island formation) are normally not accounted for, i.e., the mean-field approximation for surface coverages is used.In the present paper, flow field and gas-phase reactions of a stagnation-point flow pointing towards a catalytic surface are calculated using an one-dimensional reactive flow code. Here, surface coverages are not calculated using the mean-field approximation, but are taken from averaged results of a Monte-Carlo simulation for each time step of the flow code.In this MC simulation randomly one atom of a two-dimensional array of 50x50 atoms representing a Pt(110) surface is choosen. Proportional to the gas-phase composition at the surface, a second random number decides whether CO or oxygen adsorbs given the atom is not already covered. CO only needs one site to adsorb while O\sub 2\ needs two free adjacent sites. Is this platinum atom already covered by an adsorbent either reaction or desorption is selected by the random number. The probability for a process is proportional to the rate coefficient of the respective reaction. This process is continued for each time step of the flow field calculation until a reasonable margin of error is achieved.Using this procedure, phenomena like island formation are observed. The influence of these effects on the igntion process and their dependence on kinetic parameter are discussed.
9:20 AM		SS-ThM-4 Chemisorption of Co on the Ir(111) Surface: Adsorption and Desorption Kinetics Measured with In Situ Vibrational Spectroscopy M. Sushchikh, J. Lauterbach, W. Weinberg (University of California, Santa Barbara) The kinetics of adsorption and desorption of carbon monoxide on Ir(111) have been investigated using the well-defined and carefully calibrated IR absorption peak position of the CO intra-molecular stretching vibration as a non-intrusive coverage probe. Experiments were performed with time-resolved Fourier transform infrared reflection-absorption spectros-copy (TR-IRAS), which allows high-resolution measurements over a large pressure range between 1 x 10 \super-8\ and 1 x 10 \super-5\ mbar and surface temperature range between 390 and 500 K. The rate of desorption into vacuum is compared with the rate of desorption in the presence of a finite pressure of the desorbing gas. In this case a flux-dependent enhancement of the desorption rate has been observed previously \super1,2\, which, if ex-istent, should have an important impact on heterogeneously catalyzed reactions per-formed under high-pressure conditions. Most of these experiments, however, were performed using thermal desorption spectroscopy, a non-equilibrium method \super1\. We have been able to measure all rate parameters that are essential to describe the desorption in situ with TR-IRAS and have thereby clarified the disagreements which exist in the literature.\super1\ T. Yamada, T. Onishi, and K. Tamaru, Surf. Sci. 133, 533 (1983). \super2\ N. Takagi, J. Yoshinobu, and M. Kawai, Phys. Rev. Lett. 73, 292 (1994).Supported by the National Science Foundation (Grant CHE-9626338)
9:40 AM		SS-ThM-5 Theoretical Study of the Catalytic Oxidation of CO at Ru(0001) C. Stampfl, M. Scheffler (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany) We investigated the carbon monoxide oxidation reaction at Ru(0001) using density functional theory and the generalized gradient approximation. These studies were stimulated by recent high gas partial pressure experiments, the results of which exhibited an anomalous behavior of the kinetics. On the basis of the experimental results, for CO/O\sub 2\ ratios < 2 (oxidizing conditions), it was speculated that the reaction proceeds via the Eley-Rideal (E-R) mechanism, i.e., reaction between gas-phase and chemisorbed particles. Our calculations reveal that a (1x1) monolayer oxygen structure can form on Ru(0001). The interaction of of CO with this surface is initially repulsive, we identify however the reaction pathway of the E-R mechanism which has an energy barrier of about 1.1 eV. The estimated rate is however rather low. It is pointed out that CO molecules adsorbed at O vacancies, will turn efficiently into CO\sub 2\ by the Langmuir-Hinshelwood (L-H) mechanism. Thus, the exceptionally high rate of CO oxidation is actuated by the (1x1) oxygen layer, which enables both E-R as well as L-H mechanism to take place.
10:00 AM		SS-ThM-6 Mechanism For CO Oxidation on Ru(001) M. Wheeler, D. Seets, C. Mullins (University of Texas, Austin) The catalytic oxidation of CO is an important reaction in the reduction of automotive emissions and has been studied on many transition metal surfaces. Previous studies include high pressure and UHV studies on both supported and single-crystalline samples. Additionally, the UHV studies have shed significant insight into the mechanisms by which this reaction occurs under actual catalytic conditions on many surfaces. However, there is a disagreement between results of steady-state CO oxidation on Ru(001) at high pressures (1-5 Torr) and under UHV conditions. This is probably due to a saturation coverage of oxygen that is lower under UHV conditions (~0.5 ML) than under high pressure conditions (~1.0 ML). High pressure studies indicate that the mechanism for CO\sub 2\ formation on Ru(001) is different from that on other metals studied. Evidence for this includes the fact that Ru(001) is most active for CO oxidation when the surface coverage of atomic oxygen approaches one monolayer and CO coverage is negligible. In contrast, other metal surfaces are most active under conditions where there is a high coverage of CO and oxygen adsorption is limited. To lend insight into the mechanism of CO oxidation on Ru(001), we have planned a series of supersonic molecular beam experiments in which CO will be impinged on an O precovered surface. We have created high coverages (~1.0 ML) of atomic oxygen under UHV conditions by using an RF plasma source to supply gas-phase oxygen atoms to the surface. We will compare the reaction probability of CO as a function of oxygen surface coverage and surface temperature to steady-state results previously reported.
10:20 AM		SS-ThM-7 Oxygen and Carbon Monoxide on Rh(110): Adsorption, Desorption and Reaction Kinetics by Real Time X-ray Photoemission Spectroscopy A. Baraldi, S. Lizzit, D. Cocco, G. Comelli, G. Paolucci, M. Kiskinova (Sincrotrone Trieste SCpA, Italy) We present a real time XPS study of the oxygen and carbon monoxide interaction with Rh(110). The adsorption of oxygen, the thermal desorption of carbon monoxide and the removal of the oxygen layer by carbon monoxide have been studied. Real time XPS is allowed by the use of third generation synchrotron radiation sources. Following the surface concentration of each species in its adsorption states as a function of time it is possible to highlight the interplay between the adsorption rate and changes occuring at the surface during adsorption and desorption. Dissociative adsorption of oxygen above 400 K induces substrate reconstruction. In the case of unreconstructive adsorption at 270 K long-bridge sites are occupied at coverages below 0.35 ML and fcc 3-fold sites at higher coverages. In the case of reconstructive adsorption we suggest a mechanism involving buried oxygen in the initial stages, occupation of fcc 3-fold sites in the reconstructed phases and a structure-dependent adsorption rate. For the carbon monoxide thermal desorption, the variation of concentration of the top and bridge bonded CO allows to develop a model for the structural evolution of the adsorbed layer which takes into account the CO bonding configurations and the role of CO-CO repulsive interactions. The reaction of CO with the O saturated layer on both reconstructed and unreconstructed Rh(110) surfaces was studied at temperatures ranging from 200 to 350 K. The effect of the reaction temperature correlates with the deactivation effect of oxygen on CO adsorption. Correlation with the titration rate has shed light on the reactivity of differently bonded O and CO species.
10:40 AM		SS-ThM-8 Cyanide Intermediates in Catalytic Reduction of NO by C\sub 2\H\sub 4\ on Rhodium (111) J. Niemantsverdriet, A. Schmidt, R. Van Hardeveld (Eindhoven University of Technology, The Netherlands) Although the reaction between NO and CO to CO\sub 2\ and N\sub 2\ is commonly held responsible for the removal of NO from automotive exhausts in the threeway catalyst, hydrogen and hydrocarbons make a substantial contribution to NO reduction as well. Temperature programmed reactions of NO and C\sub 2\H\sub 4\ coadsorbed on Rh(111) give rise to the desorption of a number of gases, where H\sub 2\, H\sub 2\O, CO\sub 2\ and N\sub 2\ are the main products at low ethylene coverages. Significant amounts of HCN, however, evolve at higher C\sub 2\H\sub 4\ coverages, while N\sub 2\ desorption becomes suppressed and shifts to unusually high temperatures of up to 900 K. Static secondary ion mass spectrometry indicates the formation of a large inventory of adsorbed CN species, part of which desorbs as HCN around 600 K, while the remainder decomposes and is responsible for the delayed formation of N\sub 2\. Studies in which adsorbed N-atoms are reacted with coadsorbed ethylene indicate that, provided the ethylene coverage is sufficiently high, 100% of the nitrogen may end up in the form of HCN. This suggests that new catalytic reactions based on low temperature C-N coupling and noble metal catalysts should be feasible.
11:00 AM		SS-ThM-9 The Dissociative Chemisorption of Nitrogen on Iron(111) at Elevated Pressures I. Alstrup (Haldor Topsoe Research Laboratories, Denmark); I. Chorkendorff (Technical University of Denmark); S. Ullmann (Haldor Topsoe Research Laboratories, Denmark) It has recently been suggested that the results obtained by Ertl et al. for the dissociative nitrogen chemisorption on Fe(111) may not be applicable to the modeling of the catalytic ammonia synthesis reaction because of low gas temperatures during exposures and a possible gas temperature dependence. We have revisited this chemisorption system using much higher pressures than in previous studies. Adsorption results have been obtained using partial and total pressures from 0.0001 to 500 torr and temperatures in the range 300 - 578 K. No dependence on total gas pressure is seen. This means that the initial sticking probability does not depend on the gas temperature and suggests that the chemisorption of nitrogen proceeds via a precursor mediated process rather than a direct process. A subsurface saturation procedure was used before the final surface cleaning prior to chemisorption in order to minimize migration of nitrogen atoms from the surface into the bulk, as in the work of Ertl et al.. But in the present work much higher nitrogen presaturation pressures were used. After exposures at a nitrogen pressure of 50 torr or higher and at temperatures higher than 570 K a completely new chemisorption state is seen in the TPD spectra and by LEED. The new LEED pattern corresponds to a 5x5 unit cell. The TPD peak temperatures of the two peaks of the new state increases appreciably with coverage. The desorption of nitrogen from the surface with nitrogen in the new state is strongly accelerated when the coverage becomes smaller than a critical value.
11:20 AM		SS-ThM-10 Surface Studies of Model Supported Catalysts: Rh/CeO\sub 2\(100) S. Overbury, D. Huntley, D. Mullins, K. Ailey (Oak Ridge National Laboratory); P. Radulovic (University of Tennessee) It is now well established that interactions between noble metal and reducible oxide supports can synergistically alter the activity for oxidation and reduction reactions, important for emission control catalysis. We have looked for such interaction in Rh supported on ceria. CeO\sub 2\ films were grown epitaxially by laser ablation onto SrTiO\sub 3\(100) substrates. X-ray diffraction and TEM indicate the films are single crystalline with (100) orientation. The surfaces of these films, before and after deposition of Rh, and following exposure of NO were studied by low energy alkali ion scattering, soft x-ray photoemission and temperature programmed desorption. Comparison of experimental and computer simulated ion scattering distributions indicate that annealed CeO\sub 2\ is composed equally of oxygen terminated and Ce terminated domains. Exposures to CO and H\sub 2\ at various temperatures had little effect upon surface structure. Photoemission from the Ce 4d levels exhibited complex features which differed between a partially reduced surface, induced by sputtering and a fully oxidized surface. Deposition of Rh on the annealed substrate resulted in islands of Rh with no preferential azimuthal orientation, and which grow upon O and Ce terminated domains with equal probability. N 1s photoemission clearly distinguished between adsorbed atomic N and molecular NO. NO does not adsorb on the Rh free, fully oxidized surface, although small amounts adsorb upon a sputter reduced surface. NO adsorption on the Rh loaded surface correlated with Rh 3d intensity. The thermally induced conversion from molecular to atomic N was measured as a function of Rh coverage to determine substrate affects on Rh surface chemistry, and the results will be discussed. Research sponsored by Div. of Chemical Sciences, BES/USDOE.
11:40 AM		SS-ThM-11 IRAS Studies of Monometallic and Bimetallic Pd, Cu, and Au Catalysts Supported on Al\sub 2\O\sub 3\ Thin Films D. Rainer, C. Xu, D. Goodman (Texas A&M University) The chemisorptive behavior of CO on several planar model monometallic and bimetallic Pd, Cu, and Au catalysts supported on thin film Al\sub 2\O\sub 3\ has been investigated using infrared reflection absorption spectroscopy (IRAS). Studies of CO chemisorbed on Pd/Al\sub 2\O\sub 3\/Ta(110) model supported catalysts reveal striking similarities between the larger Pd particles (~70 \Ao\ average particle diameter) and the Pd(111) and (100) single crystal surfaces, including a very facet-specific CO ordered overlayer phase transition associated with the <111> surface. These similarities indicate that the particle morphology in this size regime is dominated by <111> and <100> facets. The smaller particles ( 25 \Ao\ average diameter) exhibit a higher ratio of step/edge/defect sites. CO chemisorption on thin film Al\sub 2\O\sub 3\ supported Cu and Au particles was also studied by IRAS. The spectra for each is dominated by features attributable to atop CO bound on defect sites; for Cu, this reflects a high degree of intensity transfer from CO bound on terrace sites to CO bound on the low-coordinated edge sites. For supported Pd-Au and Pd-Cu catalysts, no conclusive indications of ligand effects were apparent in the CO stretching frequencies. The spectra resemble combinations of both analogous monometallic catalysts, even though a significant degree of metal mixing is apparent from thermally induced changes in the IRAS, and from ion scattering spectroscopy (ISS) studies. A peak attributed to Pd diluted in the surface by the presence of a dissimilar metal is apparent for each of these bimetallic supported catalysts.