AVS2015 Session SS+AS+EN-TuM: Mechanistic Insight of Surface Reactions: Catalysis, ALD, etc. - I

Nitrogen-doped carbon materials have been found to demonstrate high electrocatalytic activity for oxygen reduction reaction (ORR) as the non-metal catalysts but the active site is still under debate. This is due to the complexity of the real catalysts, such as mixing of different type of N and inhomogeneity in both structure and conductance. Here we designed the nitrogen doped graphite (HOPG) model catalysts with different type of N dominance and its concentration to directly clarify the ORR active site. ORR measurements showed that active site was created by pyridinic N (N bonded to two carbon atoms). The ORR active site was ascribed to the carbon atom with Lewis base property created by neighbour pyridinic N based on the investigations of intermediate state of ORR, localized electronic states at carbon next to pyridinic N and CO₂ adsorption property by X-ray photoelectron spectroscopy (XPS), scanning tunneling spectroscopy (STS) and temperature programmed desorption (TPD), respectively. The ORR activity of model catalyst per pyridinic N concentration was then found to be in good agreement with that for real nitrogen-doped graphene catalyst.

Cobalt (II) oxide is a promising electrode material for Li-ion conversion batteries, undergoing the following reversible redox reaction upon exposure to lithium:

2Li + CoO ↔ Li₂O + Co⁰_.

In order to characterize the phase progression and morphology of the Li-CoO reaction, epitaxial CoO(100) and (111) films were exposed to lithium in an ultra-high vacuum chamber. The early stages of the reaction were then characterized with scanning tunneling microscopy (STM), while the diffusion of Li into the films and resultant reduction of CoO was quantified using angle-resolved x-ray photoemission spectroscopy (ARXPS). From these measurements, a model of the Li-CoO reaction was constructed for each orientation.

For CoO(111) films, the conversion reaction initiated at step edges and defect sites before proceeding across the surface of the film. STM images of CoO(111) after 0.2 ML of Li exposure suggest that the conversion reaction products initially assumed a periodic structure which was in registry with the CoO(111) surface. For larger Li exposures, ARXPS measurements indicated that the reaction proceeded in a layer-by-layer fashion into the bulk, maintaining a planar interface between reacted and unreacted CoO.

The reaction of the CoO(100) surface with 0.1 ML of Li resulted in the formation of 2-3 nm Co metal nanoparticles which decorated the CoO step edges. Upon further lithiation, the conversion reaction proceeded into the film preferentially at step edges. ARXPS measurements suggested that the reaction penetrated deep into the CoO film from these nucleation points before spreading across the rest of the surface. These combined results show the importance of crystallographic orientation in determining the reaction kinetics in a Li-ion battery.

Understanding water/solid interactions is a current critical scientific challenge with important implications for a variety of fundamental and applied processes. Here we study the interactions of water with RuO₂, which has a wide range of applications in photocatalytic water splitting, heterogeneous catalysis, electrochemistry and many other energy-related areas. We prepared stoichiometric (s-), reduced (r-) and oxidized (o-) RuO₂(110) surfaces and studied water adsorption, dissociation, and diffusion using time-lapsed scanning tunneling microscopy and density functional theory calculations. On s-RuO₂(110) we show that water monomers become mobile above 238 K and form dimers which are immobile below 273 K. More importantly, we find that the mobile water dimers dissociate readily to form Ru-bound H₃O₂ and hydroxyl species (HO_b) on bridging oxygen (O_b) rows. The onset for diffusion of H₃O₂ on s-RuO₂(110) is observed at ~273 K, indicating a significantly higher diffusion barrier than that for water monomers. The experimentally determined diffusion barriers are in agreement with those obtained from the DFT calculations. The observed behavior is compared and contrasted with that observed for water on isostructural rutile TiO₂(110) where both molecularly-bound monomers and dimers are in equilibrium with their deprotonated states. In contrast with TiO₂(110), the larger separation of Ru atoms induces the segmentation of water chains at high water coverages. On slightly oxidized o-RuO₂(110), water molecules react with oxygen adatoms (O_a’s) on Ru rows and form pairs of terminal hydroxyl groups which can reversibly dissociate back to a water molecule and O_a. This process results in the displacement of O_a’s along the Ru rows. Along- and across-row diffusion of isolated water molecules is tracked at room temperature on both slightly, and heavily oxidized o-RuO₂(110) by following the position of hydroxyl pairs. On r-RuO₂(110), we find that water molecules readily dissociate at bridging oxygen vacancies and form bridging hydroxyl groups. The mechanism of along- and across-row diffusion of the bridging hydroxyl protons is also studied at room temperature. The atomically-detailed, quantitative assessment of binding and diffusion of the surface species formed upon water adsorption on RuO₂(110) represent a critical step in achieving fundamental level understanding of the role RuO₂ plays as H₂ and O₂ evolution co-catalysts in photocatalytic water splitting reactions.

The nonlinear optical technique of broadband sum-frequency generation (BB-SFG) has been used to study the surface reactions during atomic layer deposition (ALD). Vibrational BB-SFG spectroscopy is excellently suited for in-situ studies of the surface chemistry governing ALD because of its inherent interface selectivity, submonolayer sensitivity, and short acquisition times. In contrast to BB-SFG, conventional absorption spectroscopy, based on the so called “differential” measurements, monitors only changes on the surface. On the other hand, due to its surface selectivity, BB-SFG reveals information about both persistent and changing surface groups. Therefore, with this technique, open questions can be addressed such as the origin of the decrease in growth per cycle (GPC) at low temperatures of the ubiquitous process of thermal ALD of Al₂O₃ from Al(CH₃)₃ and H₂O. So far, a complete picture of the surface chemistry explaining the reduced GPC is missing and the exact cause of the limited growth at low temperatures remains unclear.

More particularly, the surface chemistry of thermal ALD Al₂O₃ was followed by monitoring the density of the -CH₃ surface groups. In contrast to ALD at high temperatures, below 200^oC it was observed that a significant amount of -CH₃ could not be removed during the water half-cycle. The observed kinetics could not be explained by a thermally-activated first-order reaction with a constant cross section. We investigated the temperature dependence of the reaction kinetics further by measuring the -CH₃ coverage as a function of precursor and co-reactant exposure at different temperatures. It found that the absolute cross section obtained for the TMA half-cycle was independent of temperature, indicating that the chemisorption of TMA is not a thermally activated process. The behavior during the water half-cycle was found to be more complex showing a strong dependence on temperature; it cannot be described as a reaction simply obeying Arrhenius behavior. This is in line with the more complex behavior predicted by recent DFT work carried out by Shirazi and Elliott [Nanoscale 2015] where a so-called “cooperative” effect was observed leading to a coverage dependent reactivity. The observations presented in this work are direct experimental evidence of such a “cooperative” effect and were only possible due to the inherent surface selectivity of BB-SFG.

The Cu/Al₂O₃/ZnO(0001)-Zn ternary model catalysts were prepared and characterized by XPS and LEISS. The Al₂O₃/ZnO was prepared by depositing Al onto the ZnO surface in O₂ atmosphere at 523 K, and Cu/ZnO was prepared by depositing Cu onto ZnO surface at room temperature. It was found that Al₂O₃ grew on the ZnO surface by a layer-by-layer model, while Cu formed two-dimensional islands only at low coverage and three dimensional clusters at high coverage. For Cu/Al₂O₃/ZnO(0001)-Zn, the XPS and LEIS spectra showed that the copper islands were preferred on the interfaces of Al₂O₃/ZnO. Comparing to the Cu/ZnO binary model catalyst, the addition of Al₂O₃ obviously slowed down the reduction of Cu/Al₂O₃/ZnO by H₂. More significantly, the existence of Al₂O₃ in the ternary model catalyst leaded to an increase of Cu⁺ concentration. The enhancement of Al₂O₃ in Cu/Al₂O₃/ZnO(0001)-Zn for methanol synthesis may origin from that the Al₂O₃ helps to stabilize the surface Cu⁺ which has been proposed as one of the active sites.