Operando Spectroscopy Meets Real-World Catalysis
Designing Improved SCR Catalysts
Current diesel vehicles require catalytic control of the emissions of nitric oxides that is achieved using dedicated catalytic converters based on various material compositions depending on light and heavy duty. Nitric oxides are then abated on the catalytic converter using ammonia generated onboard upon decomposition of an aqueous solution of urea according to the selective catalytic reduction (SCR) reaction scheme.
The knowledge of the exact molecular mechanism of such reaction is crucial to develop not only the composition of the SCR catalyst but also the operation strategy during driving. The reaction mechanisms that have been proposed in the literature rely predominantly on observations obtained either under experimental conditions that are not replicating the conditions experienced by the catalyst during operation or by numerical simulations. While the conditions under which the mechanisms have been derived are solid enough for this purpose and are widely exploited, the detailed complete molecular mechanism may remain partly hidden. Automotive catalytic converters operate typically under unsteady state conditions in response to the driving characteristics, that is they experience periodic and unexpected variations of gas feed composition and rate, temperature and pressure.
Monitoring SCR Catalysts
Time-resolved spectroscopy has been started to be used systematically to monitor the structure of SCR catalysts under transient conditions better simulating this type of operation to shed light on the transient changes of surface species. The strategy consists in the perturbation of the catalyst environment by suddenly switching on or off a gas component (for example a reactant) while following the structure of the catalyst using a structural probe such as a spectroscopy method (fig.1). Sub-second time-resolution is used to follow tightly the structural changes during the perturbation of the environment.
As an example, it has been demonstrated by using time-resolved vibrational spectroscopy that ammonia coordinated to the surface of vanadium-based catalysts in the form of protonated ammonia may only play the role of an ammonia reservoir .
On the contrary, ammonia molecularly coordinated to vanadium atoms reacts faster to a dose of nitric oxide. In this study, vibrational spectroscopy was used specifically to follow the fate of ammonia molecules during addition of nitric oxide and consumption (reaction) of coordinated ammonia species on the catalyst.
The observation of the ammonia storage on the catalyst surface has practical implications for the operation of the catalytic converter. This was recently shown for a copper exchanged SSZ-13 catalyst (Cu-SSZ-13) . In this example, time-resolved x-ray absorption spectroscopy at the Swiss Light Source (SLS) was used to follow the structure of the copper atoms because copper ions in different coordination environments are considered as active catalytic sites in the reduction of nitric oxide with ammonia. So far, the process has been implemented under a constant supply of ammonia to the catalyst. The role of strong ammonia coordination to copper centers in oxidation state +1 and +2 in the reaction mechanism has been clearly established, albeit under steady-state reaction conditions, hence different conditions than those experienced during operation.
The nature of the active site changes with reaction temperature. Below 250°C the active copper center is always coordinated to ammonia. At higher temperature, the copper ions occupy cationic positions in the zeolite structure since most of the ammonia has been removed from their coordination sphere. While it is known that the two regimes follow distinct mechanisms and differ in reaction intermediates, some aspects of the reaction mechanism have remained elusive.
The Cu-SSZ-13 catalyst was exposed in the powder form to x-rays to follow the structure of the copper active site, where the SCR takes place, and was subject to ammonia cut-off experiments from a mixture of the reactants (ammonia, nitric oxide and oxygen). The evolution of crucial species in the process during ammonia cut-off was followed in a time-resolved manner while monitoring the gas phase composition and hence the evolution of the SCR reaction (operando methodology).
Understanding the Catalytic Mechanism
Through analysis of the x-ray absorption near edge structure (XANES) spectra using linear combination fit of adequate spectra of reference copper species, it was possible to confirm by repeating the experiments at various temperatures that the re-oxidation of copper intermediates in the oxidation state +1 is the rate-limiting step at low temperature. Moreover, an ammonia inhibiting effect upon its cut-off at low temperature has been demonstrated. When ammonia is cut-off from the gas stream, the concentration of nitric oxide passes through a local minimum (increased nitric oxide conversion) while a fourfold coordinated copper species in oxidation state +2 is observed to pass through a local maximum simultaneously. Hence, it was possible to identify this species as a reactive intermediate during the relaxation of the inhibition by ammonia that was proposed only upon numerical simulations. Ammonia reduces copper to the oxidation state +1 and locks it in this state until ammonia is supplied from the gas phase. When this is not the case anymore, i.e. when ammonia dosage in the gas phase is interrupted, re-oxidation to copper in oxidation state +2 can occur more easily because ammonia stored on the catalyst is consumed by nitric oxide and the initial active site is restored to start the catalytic cycle again and again.
These new findings are significant for the understanding of the SCR mechanisms, and are critical for designing improved SCR catalysts. The results obtained by x-ray absorption spectroscopy bear important implications for the optimization of the catalyst performance and achieving maximum conversion of nitric oxide under the varying transient conditions experienced during operation. Repetition of the same experiment described above with the washcoated catalyst in a quartz glass reactor showed an identical behavior. Upon cut-off of ammonia from the flowing mixture of the reactants a local maximum of conversion of nitric oxide (decrease of nitric oxide concentration after catalyst) was observed that mirrors the spectroscopic observation.
When ammonia is removed from the continuous feed to the catalyst as it would be the case in practical operation between the intermittent doses of urea, ammonia moves from inactive adsorption sites adjacent to the active copper sites and reacts there with nitric oxide to produce nitrogen and water products. The experiments and the results suggest that not all ammonia that is dosed reacts immediately; rather the excess ammonia is stored near the active centers. This excess ammonia is consumed when ammonia is not dosed. Hence, ammonia dosing can be controlled such that not the stoichiometric amount of ammonia but the amount of ammonia that is effectively required for reaction at any point in time is provided to the catalyst.
Conventional reactor experiments widely available in the open scientific and technical literature operate the catalyst under a gas stream of constant alpha value (the ammonia-to-NO dosing ratio) while heating. Based on the spectroscopic observation, contrary to this approach, experiments have been performed where the washcoated catalyst experiences variable alpha values while heating (fig.2) that are obtained by changing the concentration of ammonia in the gas stream at constant nitric oxide concentration. This operation mode results in about twofold nitric oxide conversion values at low temperature (for instance, around 36% conversion with optimized dosage and 20% with conventional dosage at 200 °C). From a practical point of view, variable dosing of urea can result in improved catalytic activity and nitric oxide conversion rates depending on temperature compared to constant dosing at alpha= 1.
While SCR catalysts have been tested following this variable alpha approach in the past years  and this strategy might be well-known at industrial level to optimize the performance of the catalytic converter, it is now possible to provide the molecular perspective of the process using modern time-resolved spectroscopy methods and the described experimental approach. The ammonia inhibition and the resulting perturbation of the equilibrium between copper species of different oxidation state observed by time-resolved x-ray absorption spectroscopy are also valid in a realistic catalytic converter.
Adrian Marberger 1 , Andrey Petrov 1 , Patrick Steiger 1 , Martin Elsene r1, Oliver Kröcher1, Maarten Nachtegaa l1, Davide Ferri1
1Paul Scherrer Institut, Villigen Switzerland
Dr. David Ferri
Paul Scherrer Institute
 A. Marberger, D. Ferri, M. Elsener, O. Kröcher, Angewandte Chemie, International Edition, 55 11989 (2016)
 A. Marberger, A. Petrov, P. Steiger, M. Elsener, O. Kröcher, M. Nachtegaal, D. Ferri, Nature Catalysis, 1, 221 (2018)
 M. Kleemann, M. Elsener, M. Koebel, A. Wokaun, Appied Catalysis B, 27, 231 (2000)