Using a high-stability quartz-crystal microbalance to measure and model the chemical kinetics for gases in and on metals: Oxygen in gold

Abstract

This paper describes the use of a high-stability quartz-crystal microbalance (QCM) to measure the mass of a gas absorbed on and in the metal electrode on the quartz oscillator, when the gas pressure is low and the gas can be considered as rigidly attached to the metal, so the viscosity effects are negligible. This provides an absolute measure of the total mass of gas uptake as a function of time, which can be used to model the kinetic processes involved. The technique can measure diffusion parameters of gases in metals close to room temperature at gas pressures much below one atmosphere, as relevant to surface processes such as atomic layer deposition and model studies of heterogeneous catalysis, whereas traditional diffusion measurements require temperatures over 400 °C at gas pressures of at least a few Torr. A strong aspect of the method is the ability to combine the “bulk” measurement of absorbed mass by a QCM with a surface-sensitive technique such as Auger electron spectroscopy in the same vacuum chamber. The method is illustrated using atomic oxygen, formed under O2 gas at 6 × 10−5 Torr in the presence of a hot tungsten filament, interacting with the gold electrode on a QCM crystal held at 52 to 120 °C. Some of the incident oxygen forms a surface oxide which eventually blocks more uptake, and the rest (about 80%) indiffuses. Surprisingly, the rate of oxygen uptake initially increases with the amount of oxygen previously absorbed; therefore, the measured oxygen uptake with time is reproducible only if preadsorption of oxygen conditions the sample. Temperatures above 130 °C are necessary for measurable thermal desorption, but all the oxygen can be removed by CO scavenging at all temperatures of these experiments. Simple kinetic models are developed for fitting the experimental QCM data to extract parameters including those controlling adsorption of oxygen, the CO scavenging probability, and limits on the diffusion jump frequency of dissolved oxygen. The reproducibility of the data and the good model fits to it provide proof-of-principle for the technique.