Anatoly I. Frenkel, Syed Khalid, Jonathan C. Hanson, and Maarten Nachtegaal
1.1 Introduction
Mechanisms of catalytic reactions are often very complex and elusive, due, in part, to the multiple length and timescales that characterize chemical transformations [1]. Characteristic length scales relevant for catalysis range from several millimeters (typical sample sizes) to micrometers (the size of the support) to nanometers (the typical size of catalytic nanoparticles) to picometers (the root mean square bond length disorder). The timescales range from minutes (e.g., reduction–oxidation [redox] reactions) to milliseconds (typical turnover rates) to pico- and femtoseconds (photoexcitation processes). Furthermore, the only possibility to have a glimpse at the activity of a catalyst is to follow the chemical reaction in real time, as opposed to prenatal and postmortem investigations [2]. In that sense, catalytic investigations in chemical and energy sciences are akin to biological catalysis studies that have similar challenges due to the large range of typical length and timescales, and similar solutions, for example, the use of in-situ spectroscopic and scattering methods. Mechanistic investigations of enzymatic catalysis have a unique advantage over nanoparticle catalysis due to the well-defined positions, and a small number, of active sites (metal ions) in the enzymes as opposed to a much larger number of active surface sites in the nanoparticles. Therefore, the modeling of many enzymatic processes, including the mapping of their energy landscapes, characterizing transition states, and studying reaction kinetics, can be successfully handled by first-principle calculations [3].
Among the indirect methods capable of resolving structural environment and electronic properties of active sites in catalytic materials, synchrotron-based X-ray absorption spectroscopy (XAS) has become one of the methods of choice, due to its excellent spatial, temporal, and energy resolutions [4]. The extended X-ray absorption fine structure (EXAFS) region refers to the oscillations observed in the X-ray absorption coefficient measured within 1000–1500 eV of the X-ray absorption edge energy, that is, the excitation energy of the core-level electron. In the EXAFS region, information about the local structural environment of the X-ray absorbing atom is extracted from the fine structure oscillations of the absorption coefficient. This fine structure signal is adequately described by the photoelectron scatterings from the neighboring atoms. The EXAFS measurement is, therefore, capable of probing atomic structure within the distance range of approximately 6–8 Å from the resonant (X-ray absorbing) atom [5]. Being a local structural probe, XAS does not require a long-range order. In addition, it is element specific (every element has its own core-level excitation energies) and can be used to study elements at low concentrations (such as promoters or noble metals) and low metal loading. The X-ray absorption near-edge structure (XANES) portion of the XAS signal is assigned to the energies between the first symmetry-allowed unoccupied state and the continuum states, that is, approximately, from 30 eV below to 40 eV past the absorption edge (e.g., Fig. 1.2a). XANES contains information about the electronic structure, density of unoccupied states, and bonding geometry around the absorbing atom.
XAS experiments have extremely short (10−16–10−15 s) characteristic time (which is the duration of the interference between the outgoing and the incoming photoelectron waves, limited by the core hole lifetime and other losses) and thus can be used to probe catalytic processes in the course of the reaction. Due to the relatively large penetration depth of hard X-rays, this technique can also be applied in situ, under reaction conditions (e.g., controlled gas atmosphere or pressure, or sample temperature) [6–8].
Crystallographic techniques are superior in quantitative and direct determination of the sample structure in comparison with XAS. However, the applications of these techniques are limited to ordered phases. These techniques would therefore be useless when reaction intermediates are dilute, strongly disordered, or dispersed [9]. However, these are the most common conditions that occur in chemical or enzymatic catalysis. XAS applications are particularly powerful in these rapidly emerging areas of science, since they are capable to determine the transient states, and their evolution, in the process of catalytic reaction.
Timescales in homo- and heterogeneous catalysis range from milliseconds (nucleation and growth, sintering, particle morphology changes) to seconds (reaction turnover times) to minutes (reduction and oxidation reactions) to hours to days (catalyst degradation and aging). Time-resolved X-ray absorption spectroscopy (TR-XAS) is particularly useful in these studies since it contains relevant information about real-time catalyst structure in the course of reaction. The in-situ TR-XAS technique is among the most commonly used structural methods to date for probing intermediate states during real-time transformations in a large variety of systems of interest in structural biology [9, 10] and materials science [11–15]. In structural biology studies, TR-XAS is often carried out by the freeze–quench method [16], which enables access to similar reaction rates (from milliseconds to seconds) but better data quality compared to the capabilities of the alternative, energy-scanning, or dispersive XAS modes. In the case of inorganic catalysts used in the industry and in model studies, the main effort in the TR-XAS methodology has been the push for fast data collection methods due to the need for in-situ or operando investigations [1].
To study the structural changes in the reactions that take place in the subsecond regime, TR-XAS was developed in energy-dispersive EXAFS (EDE) and energy-scanning, or quick EXAFS (QEXAFS) modes. In EDE, described in greater detail in Chapter 3 of this book, the sample is illuminated by the polychromatic beam and thus one can take repeated snapshots of the entire EXAFS spectrum at a timescale limited by the detector readout time and the photon flux, a few milliseconds [17–22]. EDE can be efficiently used in transmission mode only, and is thus limited to samples that have a relatively large concentration of absorbing atoms (in the percentage range) and are uniform in thickness. Detector normalization problems, temporal changes in beam flux, and spatial beam stability can decrease data quality. QEXAFS was developed by Frahm and coworkers [23] and applied to solving structures of materials over the past couple of decades by many groups [1, 23–31]. The newest version of eccentric cam-driven monochromators allow for much larger spectral ranges than previously used piezo-driven ones [32] and can collect EXAFS data with a time resolution as fast as a few tens of milliseconds per spectrum [33]. Ultrafast spectroscopic methods (in the femto- and picosecond ranges) by means of pump–probe schemes have also been recently developed [34–36], but are not yet applied to in-situ studies of catalytic systems and thus fall outside the scope of this chapter.
In Section 1.2, we will give an overview of the existing implementations of the QEXAFS technology. Section 1.3 will highlight the most commonly used methods in processing and analyzing TR-XAS data. In Section 1.4, we will survey the different applications of QEXAFS to the problems of heterogeneous catalysis that will demonstrate the advantage of in-situ and operando investigations by combined techniques versus static (e.g., only prenatal and postmortem) measurements. Finally, Section 1.5 will present the summary and future directions.
1.2 Implementation
In its simplest form, conventional QEXAFS or, alternatively, “on-the-fly XAS,” can be performed on all XAS beamlines when moving the double-crystal monochromator or a channel-cut monochromator (semi-)continuously through the energy range (Bragg angle) of interest, while the encoder readout of the Bragg angle and the detectors are sampled simultaneously. This mode of QEXAFS has been introduced by Frahm [23] at Hasylab in Germany and is now implemented at many beamlines, for example, at the DUBBLE beamline of the European Synchrotron Radiation Facility (ESRF) [37] in France, at BL18 [38] at Diamond in the United Kingdom, or at BL01 at the SPring-8 [39] in Japan, to name a few. With conventional QEXAFS, one typically obtains a time resolution in the range of a few seconds for the XANES region and up to a few minutes for the EXAFS region. In this chapter we will discuss only the implementation and technical details of dedicated QEXAFS systems that reach a subsecond resolution for a full EXAFS scan.
A beamline for subsecond QEXAFS is characterized by an intense and continuous photon source, a channel-cut crystal monochromator that moves in an oscillatory motion, and a data acquisition system that simultaneously samples several detectors and encoders.
The ideal X-ray source for QEXAFS is a tapered or helical undulator, a wiggler, or a supercooled bending magnet. What is important is that the source delivers a continuous spectrum, a flux at the sample of a minimum of 1011 photons/second and, ideally, deposits not too much heat on the monochromator crystal. After the source, the beam is collimated with a collimating mirror in order to obtain the best possible energy resolution. The best place to install a QEXAFS monochromator is downstream of a collimating mirror and upstream of a focusing mirror unit to minimize the movement of the beam caused by moving the channel-cut crystal. Alternatively, the channel-cut crystal of the QEXAFS monochromator could be made with a very small gap and be placed right before the sample.
The heart of a QEXAFS setup is the channel-cut monochromator [40]. In all dedicated QEXAFS monochromators that will be described later, the crystal moves in an oscillatory motion around a preselected Bragg angle, driven by an actuator such as an eccentric cam or galvano scanner. The angular range is chosen to cover the spectral range of the element of interest. The first developments of QEXAFS monochromators come from the Frahm group at the University of Wuppertal, Germany. Their latest “Frahm-type” monochromator [33] consists of a channel-cut crystal, mounted on a tilt table or cradle that is connected to an eccentric cam mechanism to oscillate the crystal with an angular amplitude that can be tuned dynamically between 0 and ±1.5° and with a speed up to 40 Hz.
The main Bragg angle is selected on a goniometer, the tilt table is rocked around this angle, and the angular offsets of the crystal table are determined with an angular encoder [41]. The energy calibration is determined from a QEXAFS spectrum covering two metal absorption edges. The apparent angular distance between both edges can be used to determine the step width of the angular encoder. Finally, from the signal of the angular encoder and a reference foil, the absolute energy can be back-calculated for each spectrum [41]. The channel-cut crystal is cooled indirectly using water or liquid nitrogen.
A first commercial version of this monochromator has been installed at the SuperXAS beamline of the Swiss Light Source (SLS) [42]. A QEXAFS monochromator based on the “Frahm design” has been constructed at the SAMBA beamline at the Soleil synchrotron facility in France. Both at SuperXAS and SAMBA, the QEXAFS monochromators are installed at (supercooled) bending magnet beamlines, where a final monochromatric flux of ∼1011–1012 photons/second is achieved. At both beamlines, the QEXAFS monochromators are placed parallel to a conventional double-crystal monochromator, so that switching between QEXAFS and conventional XAS or QEXAFS in the second-to-minute range takes less then 5 min.
At two bending magnet beamlines, X18A and X18B, of the National Synchrotron Light Sourc...