Michael Bowker

Adsorbate diffusion on single crystal surfaces: I.The influence of lateral interactions

Abstract Two independent methods have been devised to analyse the process of surface diffusion at a planar surface. The first is a Monte Carlo simulation which provides a flexible routine for simulating the process under varying conditions of adsorbate-adsorbate nearest-neighbour lateral interactions. The second method uses order-disorder theory in the quasi-chemical approximation to provide an analytical expression for the influence of lateral interactions on surface diffusion profiles. Results obtained by the two methods are in good agreement, and indicate that in cases where the model is appropriate experimental diffusion profiles may be used to evaluate lateral interaction energies.

The adsorption and decomposition of formic acid on Cu {110}

Abstract The reactive adsorption of formic acid (HCOOH) on oxygen dosed Cu(110) has been studied using a molecular beam system, TPD, LEED and STM. At low temperature the reaction is strongly oxygen coverage dependent. All coverages result in high reaction probability (0.8 at room temperature) for formic acid and, for less than 0.25 monolayers of oxygen there is complete oxygen clean-off, leaving formate on the surface in a c(2 × 2) structure. At higher coverages the situation is more complex, with some oxygen remaining coadsorbed with the formate. The two adsorbates are then mainly phase separated into islands of c(6 × 2) oxygen and (3 × 1) formate. The two phases mutually compress each other due to pressure at the phase boundaries. The reaction stoichiometry is 2:1 formic acid:oxygen atoms in this temperature range. At higher temperatures (> 450 K) the formate itself is unstable and decomposes during adsorption which results in a change of stoichiometry of the reaction; one molecule of formic acid removes an oxygen atom as water, and hydrogen evolution ceases. There is a range of temperature between 350 and 420 K for which the reaction becomes very difficult, and the reaction probability drops to ∼ 0.1. It is proposed that this is due to rapid compression of much of the oxygen adlayer into the unreactive c(6 × 2) structure by small amounts of formate. The reaction proceeds through a highly mobile, weakly held, “precursor” state on the surface, which is able to seek out the active sites on the surface, which are low in coverage at high levels of oxygen. These active sites are the terminal oxygen atoms in the oxygen islands (in the [001] direction), which are only present at step edges or phase boundaries at 0.5 monolayers coverage of oxygen.

CO and O2 adsorption on Rh(110)

The adsorption of CO and oxygen on Rh(110) has been studied using a thermal molecular beam, thermal desorption and LEED. CO adsorbs with an initial sticking coefficient of 0.68 and shows a coverage dependence which is well described by the Kisliuk formalism for precursor kinetics. The desorption shows a main peak at 485 K, which shifts to lower temperature with increasing coverage up to ~ 0.4 monolayers. Above this coverage a shoulder appears at ~ 425 K and a further shoulder at 390 K above about 0.75 monolayers. These effects are due to repulsive lateral interactions in the adlayer, though the only clear ordered pattern seen in the LEED at 320 K is a (2 × 1)p1g1 pattern above 0.9 monolayers. For oxygen the initial sticking coefficient is 0.62 (±0.01) at 310 K which diminishes relatively slowly with increasing oxygen coverage. We believe that this is due to row pairing in the adlayer which produces stretched” rows of higher reactivity adjacent to them. The row pairing is seen in a wide sequence of LEED patterns with increasing coverage — (1 × 3), (2 × 2)plgl, (1 × 2)(1 × 3), (2 × 3)p1g1, c(2 × 6), c(2 × 8). Much higher exposures are require latter 3 structures since the sticking coefficient has diminished to a low value by then. The apparent “saturation” point in the sticking curves is 0.65 monolayers, but oxygen can continue to adsorb slowly. We propose that the last three structures are due to completely restructured surfaces, possibly with subsurface oxygen, though this requires further work. Desorption from these high coverage states begins at the relatively low temperature of 750 K.

The adsorption and decomposition of NO on Pd(110)

The sticking probability of nitric oxide (NO) on Pd(110) and the relative selectivity of the surface to nitrogen (N2) and nitrous oxide (N2O) production has been measured as a function of coverage and as a function of surface and gas temperatures using a molecular beam. It is found that, at low temperatures (<440 K), molecular adsorption occurs with an initial sticking probability of 0.40 ± 0.02, rising quickly to a maximum of about 0.48 ± 0.02 as coverage increases before falling towards saturation. Following adsorption at 170 K four distinct adsorption sites can be identified by subsequent TPD. Hence, if beaming occurs at a temperature above the TPD peak due to a given site, then that site cannot be populated and the saturation coverage is found to be reduced. At higher temperatures (440–650 K) the sticking probability is seen to decrease continuously as a function of coverage. At a given NO uptake, the sticking probability falls with temperature indicating that the rate of NO desorption is significant in this temperature range. In addition, dissociation occurs leading to the desorption of nitrogen and nitrous oxide leaving only oxygen adatoms on the surface. The oxygen adatoms poison further reaction but can be cleaned off, even at the lowest temperature at which dissociation occurs, by hydrogen or carbon monoxide. At the low temperature end of this range more nitrous oxide is produced than nitrogen but this ratio falls with temperature until, above 600 K, there is 100% selectivity to the production of nitrogen which we propose is due to the low lifetime of molecular NO on the surface. However, at such high temperatures, reaction only occurs on a few sites probably located at the few step edges present on the crystal.

The adsorption and decomposition of nitrous oxide on Rh(110) and Rh(111)

The adsorption and decomposition of nitrous oxide on the surfaces of Rh(110) and Rh(111) planes have been investigated using molecular beam adsorption and reaction, LEED, and XPS. A clear structural effect was observed. On Rh(110) nitrous oxide adsorbs, at room temperature, with an initial sticking coefficient of around 0.5. The molecules decompose into molecular nitrogen and atomic oxygen, the former is desorbed while the latter is left on the surface. On Rh(111), on the other hand, the surface is rather inert to the incoming nitrous oxide, little adsorption and decomposition can be found at all temperatures studied even when the dosed gas is heated up to 900 K before being introduced onto the surface.

Ammoxidation of propane to acrylonitrile on FeSbO4

Abstract The properties of FeSbO4 (Fe:Sb = 1:1 and 1:2) for propane (amm)oxidation have been investigated. Both catalysts showed little selective oxidation activity, however, both were active and selective for propane ammoxidation to acrylonitrile. The 1:2 catalyst showed increased selectivity, relative to the 1:1 catalyst, for acrylonitrile but at the expense of propane conversion. XRD and FT-IR showed the major phases present were rutile FeSbO4 and Sb2O4. XPS studies showed the surface of both catalysts to be antimony rich and TPD studies showed the major pathway to acrylonitrile was a two-stage process via propene and an allylic intermediate, although there is a direct contribution from a one-stage mechanism via an alkyl intermediate. It is postulated that the degree of surface oxidation is critical in determining the dominant surface intermediate. The rate limiting step in the ammoxidation reaction is considered to be dissociative adsorption of propane on the surface, the adsorption rate of propane being orders of magnitude lower than for propene, as manifested by TPD experiments. The role of Sb and Fe are discussed, together with the principal surface mechanisms.

Aspects of formaldehyde synthesis on Cu(110) as studied by STM

Abstract The synthesis of formaldehyde from methanol (both CH3OH and CD3OD) on oxygen-predosed Cu(110) surfaces has been studied using scanning tunneling microscopy (STM) and low-energy electron diffraction (LEED). Sequential STM images show the conversion of methanol to methoxy results in the removal of O from the short sides of the ( 2 × 1 ) Cu-O islands present when the surface is dosed with oxygen to give 1 4 monolayer coverage. Methoxy forms a (5 × 2) reconstruction which incorporates added Cu atoms. Detailed observations have been made on the formation of this (5 × 2) methoxy reconstruction. (5 × 2) LEED patterns have also been observed and show diffraction spot absences characteristic of plgl or p2mg symmetries. Models for the (5 × 2) methoxy-induced structure are proposed that are consistent with both the STM and LEED data. The decomposition of methoxy to formaldehyde occurs with STM images showing as an initial step the diminishing of the (5 × 2) methoxy islands. This occurs from island edges. Upon island-breakup the added Cu atoms incorporated in the (5 × 2) reconstruction are released causing the expansion of nearby step edges.

NO adsorption on Rh(110)

Abstract NO adsorption on Rh(110) has been investigated using a molecular beam system and TPD. NO appears to adsorb initially dissociatively at 300 K, followed by molecular adsorption at higher exposures. The initial sticking coefficient is high (0.67) at all temperatures, declines only slowly with increasing coverage and the saturation coverage at 300 K is 1.1 ( ± 0.1) monolayers. This seems to be composed of a mix of atomic adsorption in the trough sites and molecular adsorption on on top sites. If the adsorption measurements are performed above 380 K, N 2 is seen to evolve during the experiments, with kinetics which depend on the binding strength of the N atoms to the surface. TPD shows that is dramatically weakened above a total atomic coverage of ~ 0.5 monolayers, due to the presence of coadsorbed oxygen atoms. Since the saturation atomic coverage under these circumstances can be as high as 1.5 monolayers it appears that this destabilisation is due to a structural rearrangement of the surface which displaces nitrogen atoms to low coordination sites (perhaps on top) while the oxygen remains in the more favourable sites.

Copper-palladium alloy surfaces : I. Cu/Pd[85:15]{110}, surface structure and reactivity

Abstract Both ordered and disordered phases of Cu:Pd[85 :15]{110} have been studied using low-energy electron diffraction (LEED), X-ray photoelectron spectroscopy (XPS), and ultraviolet photoelectron spectroscopy (UPS). Low-energy ion scattering (LEIS) has also been employed to study the ordered surface. In the case of the ordered phase, these techniques have been used to ascertain the structural and electronic influence of the Pd on this surface in comparison to that observed for Cu{110}. In addition temperature-programmed desorption (TPD) has been employed to determine the chemical reactivity of this surface towards the decomposition of adsorbed formic acid (HCOOH) and C-deuterated formic acid (DCOOH).This is compared to the same reaction on Cu{110} and a kinetic isotope effect is noted in the decomposition of the two acids. Segregation of the Cu component of the alloy is observed upon the thermally induced transition from disordered to ordered surface phases, leaving the surface layer essentially free of Pd, while the second layer is enriched in Pd to approximately 50% composition.

Molecular beam studies of ethanol oxidation on Pd(110)

The adsorption and decomposition of ethanol on Pd(110) has been studied by use of a molecular beam reactor and temperature programmed desorption. It is found that the major pathway for ethanol decomposition occurs via a surface ethoxy to a methyl group, carbon monoxide and hydrogen adatoms. The methyl groups can either produce methane (which they do with a high selectivity for adsorption below 250 K) or can further decompose (which they do with a high selectivity for adsorption above 350 K) resulting in surface carbon. If adsorption occurs above 250 K a high temperature (450 K) hydrogen peak is observed in TPD, resulting from the decomposition of stable hydrocarbon fragments. A competing pathway also exists which involves C-O bond scission of the ethoxy, probably caused by a critical ensemble of palladium atoms at steps, defects or due to a local surface reconstruction. The presence of oxygen does not significantly alter the decomposition pathway above 250 K except that water and, above 380 K, carbon dioxide are produced by reaction of the oxygen adatoms with hydrogen adatoms and adsorbed carbon monoxide respectively. Below 250 K, some ethanol can form acetate which decomposes around 400 K to produce carbon dioxide and hydrogen.