Kinetic Energy Effects on the Oxidation of Ni(111)

Kinetic Energy Effects on the Oxidation of Ni(111)
Using Molecular Oxygen Beams

ABSTRACT - The oxidation kinetics of the Ni(111) surface have been quantitatively examined utilizing kinetic energy selected supersonic beams of molecular oxygen. Using in situ high-resolution electron-energy-loss spectroscopy, we have observed notable differences in the oxidation mechanism for this interface as a function of incident beam energy. Exposure of a 300 K surface to a relatively low energy 60 meV O₂ beam leads to oxidation kinetics which follow an island growth model, qualitatively similar to what is seen with simple ambient gas dosing. In contrast to this, exposure to a relatively high energy 600 meV O₂ beam yields fundamentally different oxidation behavior: the kinetics of oxidation no longer follow an island growth model but rather behave with a Langmuir-like sticking model, implying key differences in the nucleation stage for interface oxidation. Cryogenically cooled Ni(111) could not be oxidized using either of these incident beam conditions, indicating that the energetic constraints needed to move from oxygen chemisorption to actual metallic oxidation cannot be simply overcome using incident O₂ kinetic energy.

The kinetic energy dependent oxidation data were obtained by alternately dosing the crystal with a supersonic beam and monitoring the dipole-allowed oxygen vibrations (~65 meV energy loss) in situ with HREELS. Because the spectra were acquired with a low electron beam

energy of 3.5 eV at a small current density of 2 nA/cm², the probe electrons did not cause any extraneous effects on the oxidation kinetics. Two different beams were used to oxidize the crystal: one beam, neat O₂ expanded from a room temperature nozzle, had a mean kinetic energy of 60 meV; the other beam, a 5% mixture of O₂ seeded in He, had a mean kinetic energy of 600 meV. Curve fits to time-of-flight data taken with the QMA are shown in figure 1. The beam flux for each condition was determined from time-of-flight data and from the pressure rise in the third differential region of the beamline. Fluxes for the low and high energy beams were adjusted to be equal simplifying analysis of the comparative kinetics. Each beam was used to dose the nickel surface at two different substrate temperatures, 300 K and 120 K. Figure 2(a) shows typical HREEL spectra during dosing. The growth of the Ni-O stretch signal at ~65 meV due to increased oxygen adsorption can be clearly seen. The inset in this figure shows the shifts in the Ni-O peak position as a function of exposure. There shifts are indicative of the different stages of oxygen overlayer formation. Figure 2(b) was constructed by plotting the ratio of the integrated intensity of the Ni-O peak to the integrated

specular peak, versus oxygen exposure in units of Langmuirs (1 L = 10^-6 torr sec). The HREELS intensity ratio is calibrated to monolayers of NiO using the known saturation limit of 3 ML at room temperature. This calibration gives comparable results to the frequently used Auger method of determining concentrations of surface oxygen. Different phases of oxygen uptake are indicated in figure 2(b). Figure 3 shows the oxygen uptake on Ni(111) versus exposure to both low and high kinetic energy beams with the substrate at either 120 K or 300 K. For all four of these dosing conditions, oxygen uptake has the same behavior below 25 L exposure: rapid chemisorption to ca. 0.25 ML. Beyond 25 L, the behavior of the room and low temperature surfaces deviates significantly. The low temperature substrate saturates at the chemisorption limit for both beam conditions, a behavior seen previously under exposure to O₂ from background dosing. In contrast, the oxygen uptake on the room temperature surfaces continues beyond the

chemisorption limit for both beam energies. For both beam conditions, oxidation begins quickly and then slows near the saturation limit of oxygen uptake. The shapes of the oxygen uptake curves differ significantly for each dosing energy. We first attempted to fit the oxidation data of figure 3 for both the high and low energy beams with the island growth model [P.H. Holloway and J.B. Hudson, Surf. Sci. 43 (1974) 123 and 141], a model originally developed to explain oxidation in the thermal regime using ambient gases. While the low energy beam data is well reproduced by this model, it provides a poor description of the high energy beam data. The inability to fit both data sets with the same kinetic model suggests that qualitatively different mechanisms of oxide growth occur when using low and high energy incident fluxes of O₂. This has led us to develop a new model based on Langmuir kinetics. For a detailed description of the island growth and Langmuir models please refer to "Kinetic Energy Effects on the Oxidation of Ni(111) Using O₂ Molecular Beams" B.D. Zion, A.T. Hanbicki and S.J. Sibener, Surface Science Letters 417 L1154-L1159 (1998). The functional difference for the rate of oxygen uptake between these two models is that the island growth model is second order with respect to exposure whereas the Langmuir model is first order. The simplicity in the difference between models belies the fundamental difference in their derivations. Each model has been used to fit the 60 and 600 meV beam data as shown in figure 4. From a least squares analysis, and indeed as is evident upon inspection, it is determined that the island growth model provides a clear fit to the data obtained with the lower energy beam. To describe the high energy beam data it is necessary to invoke the Langmuir growth model which more accurately follows the rapid rise at oxide onset and subsequent tail off in slope near 3 ML coverage.

One conclusion of this work in the comparison of the low energy beam data and previous work done with simple background dosing of the sample. The data for oxygen uptake on the room temperature substrate under exposure to the low energy beam is best fit with the same model originally derived to fit the background dosing data, implying no qualitative changes in the kinetics of nickel oxidation. However, a significant difference in the value of the rate constants for background and beam dosing suggests a quantitative difference in oxide growth under beam exposure. A likely explanation for this difference is that the collimated beam limits the angle of incidence whereas background dosing exposes the surface to molecules from all possible angles, with commensurate differences in sticking.

The primary finding from this study is that the kinetics of oxidation actually change functional form under exposure to high kinetic energy molecular beams. For low energy dosing, oxidation is well described by the island growth model; however, the high kinetic energy data cannot be fit with this model. In this regime the oxidation mechanism follows a kinetic scheme better described by the proposed Langmuir model. In this picture, oxidation may occur anywhere on the exposed surface, not just at island perimeters. This model may also be thought of as oxide nucleation site formation occurring anywhere on the surface. For this model to be true, we require that the rate at which oxygen is incorporated into the new nucleation sites be much faster than the rate at which oxygen is incorporated into oxide islands. This interpretation implies that there is an energy barrier to oxide nucleation that is overcome by the use of the high energy incident molecules.

Return to HREELS and Molecular Beam Laboratory