The Deacon Process from the Perspective of Surface Science: HCl Oxidation on CeO2−x(111)/Ru(0001) Thin Films

Abstract

Chlorine is a basic chemical essential for a broad range of products, with a current annual demand of 90 million tons. However, hydrogen chloride is often an inevitable byproduct, accumulating to around ten million tons per year. A sustainable way to solve this waste issue is to recover molecular chlorine by thermal catalysis in the Deacon process (4 HCl + O2 → 2 Cl2 + 2 H2O). Compared to conventional hydrogen chloride electrolysis, only 15% of the energy is required. Cerium oxide is a viable catalyst for this reaction, which is active yet stable, but it is still unclear why. A reaction mechanism has been proposed based on density functional theory (DFT) calculations, but it has yet to be confirmed experimentally. The approach for corresponding model studies in this work is to divide the overall reaction into two half-reactions, based on so-called chemical looping, where the catalyst is alternately cycled between two reactants. Similarly, the proposed mechanism for the Deacon process over CeO2−x(111) can also be divided: first, the catalyst is exposed to HCl, corresponding to an exothermic surface chlorination process, and the byproduct water is formed. Secondly, oxygen is activated upon exposure and adsorbed, and chlorine is formed in an endothermic de-chlorination process. The model catalyst, i.e., single-crystalline reduced CeO2−x(111) thin films, is prepared via physical vapor deposition PVD. The films are characterized using dedicated surface-sensitive methods: X-ray photoelectron spectroscopy XPS, near edge X-ray absorption fine structure NEXAFS, low-energy electron diffraction LEED, X-ray reflectivity XRR, and temperature programmed desorption TPD. Complementary density functional theory calculations DFT+U are realized in collaboration. First, chlorination of the reduced cerium oxide is modeled on reduced single-crystalline CeO2−x(111) model surfaces, which stabilizes various ordered surface structures, e.g., (√7 × √7)R19.1°, (3 × 3), or (4 × 4), depending on the concentration of oxygen vacancies (VO). Saturating these phases with HCl at room temperature, followed by annealing to the Deacon process temperature of 700 K, results for all cases in a uniformly covering (√3 × √3)R30°-Clvac overlayer structure with identical coverage and adsorption geometry of Cl in oxygen vacancies (Clvac). Water or hydrogen formation can be observed depending on the reduction degree x. In order to rationalize why the formation of the (√3 × √3)R30°-Clvac structure on CeO2−x(111) is independent of the original reduction degree x of CeO2−x(111) efficient diffusion of surface and bulk oxygen vacancies is required. Second, as a key experiment, the stepwise re-oxidation of the chlorinated surface is investigated. Here, the displacement of tightly bound chlorine in an oxygen vacancy is predicted to be the most critical step. Synchrotron-based XPS and NEXAFS disentangle the surface and bulk properties of the surface-chlorinated model catalyst during re-oxidation. The re-oxidation process is found to start from the bulk of the catalyst and propagate towards the surface. Chlorine recombines and forms the desired product Cl2 solely during surface oxidation. DFT+U calculation evidence Cl is displaced by an adsorbed peroxo species in a concerted reaction, likely not being the rate determining step. The formation and dissociation of the peroxide species drive the re-oxidation and de-chlorination processes, hence being considered essential for cerium oxide-based catalysis in general.