80 81 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 Looking exhaust catalysis up and down Related publication: Dann E. K., Gibson E. K., CatlowC. R. A., CelorrioV., Collier P., EralpT., AmboageM., Hardacre C., Stere C., Kroner A., Raj A., Rogers S., Goguet A. &Wells P. P. Combined spatially resolved operando spectroscopy: New insights into kinetic oscillations of CO oxidation on Pd/γ- Al2O3. J. Catal. 373 , 201 (2019). DOI: 10.1016/j.jcat.2019.03.037 Publication keywords: XAFS/DRIFTS; CO oxidation; Pd/Al 2 O 3 ; Operando spectroscopy T he catalysts that remove harmful pollutants fromvehicle exhaust are exceedingly dynamic. To truly understand how they operate and improve their efficiency, we need to knowhowandwhy their structure changes. This requires the use of combinedmethods to provide complementary data that can track changes in both space and time. One of the processes that occurs in automotive exhausts is CO oxidation. This reaction has a quirky behaviour where the performance of the catalyst oscillates between high and low limits. Despite this being an extensively studied reaction, this phenomenon is not well understood. Researchers performed a combined study using X-ray Absorption Fine Structure (XAFS) to see how the catalyst structure is changing, and Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) which gives information about molecules adsorbed on the catalyst surface. Using the I20-EDE beamline, they were able - for the first time - to observe the structural changes that drive the oscillatory performance of palladiumnanoparticles.They found that the structural changeswere spatially localised to aminor component of the overall catalyst profile. Their experiment shows that developing X-ray methods for probing catalysts in space and time is essential to the understanding of mechanisms at play in catalysis. These methods will allow us to design better catalytic processes for environmental protection, emerging energy technologies, and the production of sustainable chemical feedstocks. One of the most widely studied catalytic reactions is the oxidation of CO to CO 2 , which is an important process in cleaning the exhaust fumes from automotive vehicles. The traditional class of catalyst used for this process is supported metal nanoparticles, normally Pd, Pt, or Rh. Formulating the metal as a nanoparticle is important, both to thrift the expensive metals used and to control the catalytic properties. One of the fascinating catalytic properties of Pd nanoparticles supported on Al2O3, or for that matter Pt or Rh, is the process known as the ‘CO oxidation kinetic oscillations’ 1 . In this process the catalytic performance spontaneously see-saws between high and low limits, even though the temperature and gas feed applied remain constant. Spectroscopy Group Beamline I20-EDE The proposed rationale for this puzzling phenomenon is that the catalyst structure oscillates between reduced, more metallic, and oxidised forms at these low and high points of conversion, respectively. At a temperature lower than the catalytic ‘light off’, the catalyst surface is saturated with CO in a reduced state. As the temperature is increased there is enough energy in the system to release some of the adsorbed CO from the surface, which allows the excess O 2 in the surrounding atmosphere to fill these vacant sites. For this catalytic reaction to happen it requires both CO and O 2 to be adsorbed on the surface simultaneously. Once the reaction starts, it propagates further as the exothermic reaction releases energy during the process. This explanation is relatively straight-forward, however, these structural changes have proved elusive to conclusively track down 2 . For such an extensively studied process why is this? It is well known that the supported metal nanoparticles are dynamic and readily adapt their structural properties dependent upon the precise conditions. When catalysis is performed in a fixed bed reactor, similar to that found in automotive catalytic converters, the conditions across the length of the bed are not uniform. At the inlet of the reactor the atmosphere is made up of reactants, in this case CO and O 2 , and at the outlet of the reactor the exhaust is dominated by products, for this reaction, CO 2 . Moreover, we also need to consider the temperature fluctuation imparted by the exothermic reaction. If this phenomenon was to be understood, we needed a method that could probe the spatial profile of the reactor assessing both the Pd nanoparticle structure and the adsorbed gases at each point. To achieve this, we developed a reactor that can measure X-ray Absorption Fine Structure (XAFS) spectroscopy and Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) data simultaneously across the length of the reactor, from inlet to outlet of the catalyst bed. The XAFS method provides information on the nanoparticle structures and the DRIFTS inform us about the surface adsorbates. To measure XAFS data of these systems, an intense source of photons (X-rays) provided by a synchrotron is required. Further to this, to carefully map the rise and fall of the kinetic CO oscillations we needed a form of XAFS analysis called Energy Dispersive EXAFS (EDE) that can measure data much faster (for our needs <1 second) than traditional approaches 3 . The infrared spectrometer, for the DRIFTS analysis, is an external machine which can be positioned on the beamline for the combined analysis. In this work we designed a customised square cross-section aluminium reactor (Fig. 1). The aluminium walls were thinned along the length of the catalyst bed to allow the X-rays to pass through the reactor without too many losses. A top window of CaF 2 was affixed to the reactor to allow for our DRIFTS studies; CaF 2 is an infrared transparent material within our region of interest. During our experiments we assessed multiple points along the reactor profile during a temperature ramp from low to high temperature that induced catalysis. Although these oscillations are known to occur at set temperatures (isothermal conditions), the temperature ramp experiment allowed us to ‘sweep’the range of temperature conditions that could be experienced. When the catalyst was operating, e.g. CO 2 was being produced, we observed the change from reduced to oxidised Pd nanoparticles that arises as the catalyst transitions from the CO poisoned catalyst to the highly active, oxidised Pd surface. This transition occurs initially at the end of the catalyst bed, nearest the outlet, and propagates upstream with increasing temperature of the reactor. Most importantly, we were able to observe the oscillatory behaviour of the catalyst structure and adsorbed CO that changed with the same frequency as the CO oxidation kinetic oscillations (Fig. 2). These structural oscillations can be easily visualised by looking at the intensity of specific features in the X-ray Absorption Near Edge Structure region of the XAFS spectrum during the temperature ramp. What was most interesting was that these changes only occurred at the very inlet (first 1 mm) of the catalyst bed. We proposed that these kinetic oscillations can only occur at the front of the catalyst bed where there is sufficient concentration of CO in the gas phase to compete with O 2 for adsorption sites at the catalyst surface. In this work we further demonstrated the complex nature of the evolving catalyst structures and surface reactivity during catalytic operation. Furthermore,this isaclear indicationoftheneedforspatiallyresolvedmethods for understanding and optimising highly active catalysts essential for future sustainable technologies. References: 1. Ertl G. Oscillatory Kinetics and Spatio-Temporal Self-Organization in Reactions at Solid Surfaces. Science 254 , 1750 (1991). DOI: 10.1126/science.254.5039.1750 2. Figueroa S. J. A. et al. What drives spontaneous oscillations during CO oxidation using O2 over supported Rh/Al2O3 catalysts? J. Catal. 312 , 69 (2014). DOI: 10.1016/j.jcat.2014.01.006 3. Newton M. A. et al. Bringing time resolution to EXAFS: recent developments and application to chemical systems. Chem. Soc. Rev. 31 , 83 (2002). DOI: 10.1039/B100115A Funding acknowledgement: UK Catalysis Hub Consortium and EPSRC (portfolio grants EP/K014706/1, EP/ K014668/1, EP/K014854/1, EP/K014714/1 and EP/I019693/1. Corresponding authors: Dr Peter Wells, University of Southampton,
[email protected] and Prof Alexandre Goguet, Queen’s University Belfast,
[email protected] Figure 1: Annotated image of our square cross-section Al reactor for spatially resolved XAFS/DRIFTS. Figure 2: Bottom panel shows EDE Pd K-edge white line intensity (green) and DRIFTS CO adsorption intensity (purple) of catalyst Pd/Al 2 O 3 at spatial position 1 (nearest to the reactor inlet) of the fixed catalyst bed in reactant (1% CO/3% O 2 /Ar) gas feed during temperature ramp experiment (100 - 135°C). Top panel shows the simultaneous end-pipe mass spectrometry signals for O 2 (red), CO (black) and CO 2 (blue) concentrations of the reactor exhaust.