Catalytic reactions are essential for generating the chemical products required by the modern society. In particular, reactions related to clean energy storage and generation as well as fertilizer production are facilitated by catalysts. However, the processes are often insufficiently understood at a mechanistic level. One of the main reasons is that a holistic investigation of heterogenous catalyst surfaces during reaction conditions requires experimental techniques that combine element specificity, surface sensitivity and can work under operando conditions. While excellent in terms of the first two criteria, x-ray photoelectron spectroscopy (XPS) has traditionally not been compatible with the high pressures and temperatures required for many catalytic reactions; a “pressure gap” opened between the obtainable conditions in the lab and the relevant conditions in a real catalytic reactor.

We have built a scientific instrument, a synchrotron endstation, that addresses this issue and allows operando probing at 100x higher pressure than elsewhere. The POLARIS instrument is located at PETRA III in Hamburg. This work describes the instrumentation and the theoretical background for the technique. The main focus, however, is on the mechanistic discoveries made when operando XPS with POLARIS was applied to hydrogenation of CO, CO₂ and N₂ over single crystal catalysts. The surfaces examined in this work include Fe, Co, Ni, Cu-Zn, Rh and Ru.

Regarding the CO hydrogenation reaction, this work describes how the Fe surfaces facilitate rapid CO dissociation, but slow adsorbate desorption. This combination results in carbide phases and a drastic accumulation of long-chain hydrocarbons. A similar behavior was noted in Ni catalysts at low temperatures, where a non-stoichiometric carbide was formed, but the hydrogenation rate of the carbide was dependent on the temperature and the partial pressure of the reactants. Co surfaces exhibit a mixture of CO and partly hydrogenated hydrocarbons, indicating a slower termination than observed on Ni, but without the drastic carburization noted for Fe. On Rh catalysts, a subset of the non-dissociated CO molecules may hydrogenate, and alkoxy intermediates co-exist with non-saturated hydrocarbons, allowing for selectivity towards oxygenated products. 

For the CO₂ hydrogenation reaction on Rh, the residence time of CO₂ was observed to be short and the coverage of dissociated intermediates was low in the 150 mbar pressure range. However, when switching the pressure rapidly it can be shown that pressures around 2 bar increase the coverage, and reveals other adsorbates than the static pressure study.

A Cu catalyst with surficial Zn was examined in ternary reaction mixtures of CO₂, CO and H₂. Here we noted that CO kept the Zn reduced. 

In the N₂ hydrogenation reaction, the rate of chemisorption and dissociation of N₂ dictate two different rate limiting scenarios. On Ru the reaction is limited by the N₂ dissociation and on Fe it is also limited by the hydrogenation of chemisorbed N.

The significance of operando conditions is particularly manifested with regard to the hydrogen partial pressure and its interplay with the resulting adsorbate distribution.