Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies IMN Groot, Leiden University, Leiden, The Netherlands © 2018 Elsevier Inc. All rights reserved.

Introduction Bridging the Pressure Gap High-Pressure SPM Instrumentation Requirements for Operando SPM Design of the ReactorSTM Design of the ReactorAFM High-Pressure SXRD Instrumentation Design of the ReactorSXRD Combination of High-Pressure SPM With High-Pressure SXRD High-Pressure Optical Microscopy Instrumentation Catalytic Systems Investigated Under High-Pressure Conditions CO Oxidation on Pd(100) CO Oxidation on Pd Nanoparticles CO Oxidation on Rh(111) NO Reduction on Pt(110) Chlorine Production on RuO2(110)/Ru(0001) Conclusions and Outlook Acknowledgments References

336 336 337 337 338 339 340 340 340 342 343 344 347 347 348 350 351 351 352

Introduction The topic of heterogeneous catalysis has been studied extensively over the last 100 years. Much of our current knowledge, however, has been obtained under conditions that differ significantly from those of practical catalysis. The reason for this discrepancy stems from the fact that most techniques, suitable for obtaining accurate results at the atomic scale, cannot perform under the typical working conditions of industrial catalysis (i.e., high pressures and high temperatures). Surface-sensitive techniques, such as Auger electron spectroscopy (AES), X-ray photoelectron spectroscopy (XPS), or low-energy electron diffraction (LEED), are limited to pressures below 10
5 mbar.1 Although cases exist, where the results obtained at low pressures can be extrapolated to industrial conditions,2,3 more and more examples exist where this “pressure gap” is found to fundamentally change the reaction mechanisms.4–9 The pressure gap seems to be the rule, rather than the exception. Therefore, the investigation of catalysis under more realistic (industrial) conditions, without compromising the sensitivity to the details on the atomic and molecular scale, inescapably forms the next step in this research field. But the difference between laboratory studies and chemical industry in terms of pressure and temperature is not the only gap in heterogeneous catalysis research. Also the nature of the catalyst used is very different. In academic studies the catalyst is often simplified to a model system, usually consisting of a single-crystal metal surface, as opposed to industrial catalysts, which generally consist of metal nanoparticles deposited on oxidic, nanoporous supports and are accompanied by additional elements such as promoters, fillers, and binders. This discrepancy is commonly known as the “materials gap.”

Bridging the Pressure Gap It is now well known that differences exist between reaction mechanisms taking place under (ultra)high vacuum conditions (10
5–10
11 mbar) and under high-pressure conditions (mbardbar range).4–9 Therefore, the last decades have seen a tremendous effort in adapting existing surface-science techniques to high-pressure conditions. When combining ultrahigh vacuum (UHV) chambers with high-pressure chambers using differential pumping, XPS,10,11 low-energy ion scattering,12 and transmission electron microscopy (TEM)13,14 can operate in the mbar regime. A different approach is the use of micro- or nanoreactors that separate the UHV part from the high-pressure volume via ultrathin walls of an inert material, for example, TEM15 and X-ray microscopy.16 Other surface-science techniques exist that are able to operate under atmospheric conditions without facing major limitations. Examples are scanning probe microscopy (SPM),17–21 sum frequency generation (SFG) laser spectroscopy,22–24 polarizationmodulation infrared reflection absorption spectroscopy (PM-IRAS),25 ultraviolet Raman spectroscopy,26–28 ellipso-microscopy,29 surface X-ray diffraction (SXRD),30–34 and X-ray absorption spectroscopy.35,36

336

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

337

Scanning tunneling microscopy (STM) and atomic force microscopy (AFM) are atomically sensitive techniques that in principle are able to operate at pressures varying from 10
11 mbar to above 1 bar, and at temperatures varying from millikelvins to  1000 K.37–39 Therefore, these techniques are suitable to determine the influence of the gas environment on the atomic-scale structure of catalytic surfaces, to identify active sites, and to elucidate the role of promoters, all under the relevant conditions of industrial catalysis. However, there are several practical issues that can influence the imaging quality. The weak interaction of X-rays with gases makes X-ray-based techniques suitable to investigate catalysis under realistic conditions. To study the catalyst’s surface structure under high-pressure, high-temperature conditions, SXRD40,41 and grazingincidence small-angle X-ray scattering (GISAXS)42 are particularly suitable. A very simple, yet elegant way to investigate catalysis over model systems under realistic conditions is the use of optical microscopy. Using the change of reflectivity of a single-crystal surface during reaction, we are able to link the macroscopic scale behavior of the active catalytic surface to the microscopic observations obtained with STM and SXRD.43,44 In this article, I discuss the development of high-pressure, high-temperature SPM, SXRD, and optical microscopy instruments dedicated to catalysis research under relevant industrial conditions. I discuss some catalytic systems investigated under industrial conditions using these instruments: CO oxidation over Pd(100), Pd nanoparticles, and Rh(111); NO reduction over Pt(110); and Cl2 production over RuO2 (110)/Ru(0001).

High-Pressure SPM Instrumentation The development of STM instrumentation suitable for operation under industrial conditions started with the pioneering work of McIntyre and coworkers in 1993.45 In their instrument, the microscope is contained in a 2 L stainless-steel vacuum chamber. The sample can be heated up to 1400 K using an infrared spot heater. To analyze the reactant and product gases, the system is equipped with a differentially pumped quadrupole mass spectrometer (QMS). Later, this design was combined with a UHV chamber for model catalyst preparation, and a gas chromatograph (GC) was used for gas analysis.46 After this successful introduction of high-pressure, high-temperature STM, several other systems have been developed.19,47–50 All these designs have in common that they combine a high-pressure cell with a UHV chamber for catalyst preparation and characterization using standard surface-science techniques. In our group, we have introduced a fundamentally different design. In our ReactorSTM, only the tip and tip holder, instead of the entire microscope, are contained inside the high-pressure environment.17 To be able to investigate the structure–activity relationship of (model) catalysts, a simultaneous imaging of the catalytic surface with a measurement of the reaction rate is needed. This can be done by using the cell containing the microscope as a flow reactor connected to a QMS or GC to analyze the gas leaving the reactor. However, when using a cell large enough to contain the entire microscope, the low conversion rates that should be expected on the small active areas of typical model catalysts make that one can obtain measurable product concentrations only at very low flow rates, which would introduce extremely long response times. Furthermore, a large reactor increases the probability of gases adsorbing and reacting on the reactor walls, compromising the reaction rate measurements. In addition, the exposure of vulnerable parts of the STM to corrosive gases is avoided in our design. The latest version of our ReactorSTM is now able to routinely obtain atomic resolution under harsh reaction conditions.20 In addition to high-pressure STMs, some high-pressure AFMs have been developed as well. The simplest design consists of a standard AFM operated in a high-pressure cell,51,52 limiting the operation temperature and prohibiting the use of corrosive gases. In a more advanced design the piezo element of the AFM is separated from the scanner with a membrane. In this way, the AFM can operate up to 423 K and 6 bar in liquid,53 or up to 350 K and 100 bar in supercritical CO2.54 These instruments are limited to contact-mode AFM and, due to long equilibration times, to constant temperature. In our group, we have developed a ReactorAFM21 based on the same principles as the ReactorSTM.20

Requirements for Operando SPM As mentioned in the previous section, we have developed an unconventional design for SPM instruments suitable for catalytic studies under industrial conditions. We have taken into account the following requirements for the design of the ReactorSTM20 and the ReactorAFM.21

• • • •

The microscope must be suitable for catalytic studies under realistic conditions, that is, high pressures (1 bar and beyond) and elevated temperatures (at least 400 K). Under these conditions, the microscope must be capable of imaging the catalytic surface with atomic resolution (STM) or step resolution (AFM) and low drift. The setup must be suitable for operando studies, that is, studies combining the structural information of the catalyst with its activity. For this, the time resolution for measurements of the reaction product(s) needs to be 10 s or better. The design of the instrument should prevent chemistry taking place anywhere else than on the catalyst surface. Reactions taking place on the reactor walls or elsewhere will not only hamper the accuracy of activity measurements, but may also corrode components of the microscope. Measurements should take place on well-defined model catalysts that are initially atomically clean and highly ordered. The catalyst surface must not be exposed to air between preparation under UHV conditions and characterization under reaction conditions.

338

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

In principle, an STM or AFM can work under atmospheric pressures. Due to the small tip-surface distance no significant interference by the surrounding gas molecules is expected. Due to elevated temperatures, an increase in thermal drift is expected. This thermal drift can be minimized using special materials with low expansion coefficients. In our group, we have shown that stable imaging at temperatures as high as 1300 K can be achieved in this way.55 However, when combining elevated temperatures with atmospheric pressures, difficulties may arise. The piezo element used to actuate the tip must be kept under its Curie temperature. This can be easily achieved under vacuum conditions, where large temperature gradients can be kept over short distances. However, heat transport via gas molecules proposes a challenge under atmospheric pressure conditions. Furthermore, a convective flow in the gas atmosphere will result from a large temperature difference between sample and piezo element, resulting in erratic drift in the images. We have circumvented these problems by keeping most of the components of the microscope in vacuum, while the sample is exposed to reaction conditions. In order to investigate the structure–activity relationship of model catalysts, the catalytic surface must be contained in a reactor volume that houses the microscope and that is connected to a QMS or GC. To optimize both the time resolution and the sensitivity to changes in the concentrations of reactants and products, the ratio between the surface area of the model catalyst and the volume of the reactor must be maximized. Since SPM can only be used on flat surfaces or on nanoparticles on flat supports, we are limited to small surface areas. Therefore, to obtain a high ratio between surface area and reactor volume, the reactor volume must be as small as possible. Under reaction conditions, chemical reactions can take place not only on the surface of the catalyst, but also on the sides and back face of the sample, and on other surfaces such as the reactor walls, the sample holder, and parts of the microscope. These undesired reactions not only influence the activity measurements (e.g., higher reactivity can be observed than if the measured reactivity would only be due to the catalyst), but it can also seriously damage the instrument, especially when corrosive gases are used. To prevent these undesired reactions from happening, the reactor walls consist of inert materials. In addition, the sides and back face of the sample, and as much as possible the elements of the microscope, are kept outside the reactor. To prepare well-defined, highly ordered model catalysts, standard cleaning and preparation procedures are employed. These methods, such as ion sputtering, annealing, physical vapor deposition of metals, and gas exposure, require UHV conditions. Techniques to investigate crystal composition, structure, and ordering, such as LEED, AES, and (standard) XPS, can only be employed under UHV conditions. Since the prepared catalyst samples cannot be exposed to air when transferring them to the high-pressure reactor and SPM, a combination of a UHV chamber with the microscope and reactor is necessary.

Design of the ReactorSTM In this section, I give a short overview of the design of the ReactorSTM. For more details, the reader is referred to Refs.17,20,56. Fig. 1 shows a schematic drawing of the concept of the ReactorSTM, in which all of the above-mentioned design criteria are met. The reactor volume is placed inside a UHV system, used for the preparation and characterization of model catalysts. Two gas lines, one for the inlet of gases, one for the outlet, are connected to the reactor. The inlet line is connected to a gas system which controls the flow, mixing ratio, and pressure of the reactant gases (the gas system is described in detail in Ref.20). The outlet line is connected to a small chamber housing the QMS for analysis of the reactants and products that leave the reactor. The only elements of the microscope that are inside the reactor volume are the tip and tip holder. The reactor is sealed off from the surrounding UHV using two flexible o-rings. The lower o-ring (Viton) separates the reactor from the piezo element that is used to actuate the motion of the tip. The upper o-ring, consisting of Kalrez (Dupont), a chemically resistant material, is pressed between the sample and upper part of the reactor body, thus sealing off the high pressures inside the reactor from the surrounding UHV environment. The sample can be radiatively heated from the rear (the upper side in Fig. 1). The UHV setup consists of three chambers separated by gate valves. The first chamber is the XPS chamber. Here, the XPS for use under UHV conditions is located. Furthermore, this chamber houses a sample library and a sample load-lock that ensures the

Fig. 1 Conceptual drawing of the ReactorSTM. The STM tip is contained in a small high-pressure volume, while the STM scanner is not exposed to the gas environment. The sample forms one side of the reactor, while the other walls are chemically inert. Two polymer o-rings seal off the high-pressure volume from the UHV system around it. Reproduced from Herbschleb, C. T.; van der Tuijn, P. C.; Roobol, S. B.; Navarro, V.; Bakker, J. W.; Liu, Q.; Stoltz, D.; Cañas-Ventura, M. E.; Verdoes, G.; van Spronsen, M. A.; Bergman, M.; Crama, L.; Taminiau, I.; Ofitserov, A.; van Baarle, G. J. C.; Frenken, J. W. M.; Rev. Sci. Instrum. 2014, 85, 083703.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

339

loading of samples without breaking the vacuum. The second chamber is the preparation chamber. It contains an ion sputtering gun, a combined LEED/AES system, and an electron-beam evaporator for preparation of metallic or metal oxide nanoparticles. In addition, gas lines with leak valves are present for sample cleaning and catalyst preparation. The third chamber is the STM chamber, which houses the microscope integrated with a small flow reactor. On the top flange of this chamber a seal library has been installed for the Kalrez seals that close off the reactor. This seal consists of a custom-made Kalrez ring (Dupont) that is vulcanized onto a stainless-steel holder, enabling exchange of the seals without breaking the vacuum. Using a wobble stick seals can be placed on and removed from the top plate of the reactor. Due to the specifications of the Kalrez seal, the operation temperature of the STM is limited to 600 K. The STM body is fabricated from Zerodur (Schott), a chemically resistant glass type with low thermal expansion coefficient, minimizing thermal drifting of the STM during temperature changes. Sample transfer between the chambers is performed with a rack-and-pinion transfer rod. The sample holder is made out of Invar, a low-expansion steel, again to minimize thermal drift. It is strongly pressed against the top of the STM body, making hard mechanical contact via three adjustable screws. In this way, a short and stiff mechanical loop is established between the sample and the tip, which is essential for high-quality imaging. When pressing the sample holder against the STM body, the Kalrez seal is compressed to 80% of its original thickness, providing a leak-tight seal. When the reactor is exposed to pressures up to 6 bar, vacuum conditions are maintained in its surroundings. A single piezo tube (EBL2, EBL Products) is used for both the coarse approach and the fine scanning motion of the tip. The STM tip is clamped in a steel holder, which is pulled against two steel tracks using a SmCo magnet (IBS Magnet). The tip holder and tracks are Au-coated to ensure chemical inertness and to optimize the stick-slip behavior of the motor. The electrical connection to the tip, necessary to measure the tunneling current, is established via the tip holder, tracks, and the aluminum tube in which the tracks are clamped. An additional aluminum piece provides electrical shielding. The two aluminum parts are electrically isolated from each other and from the piezo tube using two insulating Macor rings (Corning Inc.). The piezo tube is glued to a titanium base, which has a thermal expansion coefficient that compensates the expansion of the piezo tube during temperature changes. The central STM part, including vibration isolation, eddy current damping, all electrical connections, and the gas capillaries, is mounted on a CF-200 flange. The STM body that holds the microscope, the reactor, and the sample holder is suspended by a set of springs combined with an eddy current damping system, in order to isolate the STM from external, mechanical vibrations. When transferring samples, the STM body can be locked via controlled inflation of a bellow. A second bellow can be inflated via a thin capillary to press the reactor body to the Kalrez seal, thus closing off the reactor volume from the UHV surroundings. The STM is controlled using fast analog/digital SPM control electronics (Leiden Probe Microscopy B.V.) capable of video-rate STM imaging.57,58

Design of the ReactorAFM With the ReactorSTM (see “Design of the ReactorSTM” section), we are able to bridge the pressure gap. Scanning tunneling microscopy, however, can only be performed on conductive samples, limiting the experiments to model catalysts consisting of metal single crystals or thin oxide films deposited on metal surfaces. To study metallic nanoparticles deposited onto porous oxide supports, and thereby also bridge the so-called materials gap, a different scanning probe technique is required: the atomic force microscope (AFM). This technique uses the interaction force between tip and sample to probe the surface, and is therefore independent of its conductivity. Therefore, to be able to investigate more realistic catalysts, we have developed the ReactorAFM.21 Its design is based on the proven concept of the ReactorSTM (see “Design of the ReactorSTM” section), but its capability to image supported nanoparticles adds unique value for operando catalysis research. Full details of the ReactorAFM are given in Refs.21,56. In short, the ReactorAFM must be able to resolve nanoparticles supported on flat surfaces under industrially relevant conditions with sufficient detail. The minimum requirements that we have defined are a lateral resolution of 1 nm, and a vertical resolution of 0.1 nm, with a range of at least 1 mm in each direction. Furthermore, the scanner should be sufficiently stable to enable imaging of a single feature on the surface for at least 1 h, thus placing constraints on the thermal drift of both the scanner and force sensor. The AFM scanner, however, differs significantly from the STM scanner. The AFM scanner is based on the piezoelectric read-out of a quartz tuning fork (QTF). Due to the small volume of the reactor, no optical access to the tip is possible, ruling out laser deflection techniques commonly used in AFM. Since quartz is chemically inert, and exceptionally high resolution has been reported using QTF-based AFM,59,60 this is the best choice for the ReactorAFM. The high stiffness makes the QTF also relatively insensitive to the presence of the gas atmosphere, thus enabling us in principle to extend imaging with good resolution to reaction conditions. The ReactorAFM operates in non-contact or frequency-modulation (FM) mode. In this mode, the cantilever is oscillated at resonance with an amplitude in the range of 10 pm–100 nm. When the tip is close to the surface, the tip–sample interaction force gradient will influence the effective spring coefficient of the mechanical oscillator, resulting in a shift of the resonance frequency. In the case of dissipative forces, there is also a decrease of the amplitude and an additional phase shift. The resonance frequency is measured by a phase-locked loop. The output signal of this loop is used as the input for the height feedback loop of the AFM scanner in order to trace the surface at constant frequency shift. A separate feedback system adjusts the drive amplitude to keep the oscillation amplitude constant, thereby ensuring that the surface of constant frequency shift corresponds to a surface of constant force gradient. The drive signal of this amplitude feedback loop is recorded in a separate channel and can be used to derive the dissipative force.

340

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

High-Pressure SXRD Instrumentation The weak interaction of X-rays with gases makes X-ray-based techniques suitable to investigate catalysts under realistic conditions. To study the surfaces of catalytically active materials under reaction conditions, SXRD40,41 and GISAXS42 are very suitable. The typical geometry of an SXRD-type reactor consists of a vessel with walls transparent to X-rays (e.g., beryllium or aluminum), into which gas mixtures with variable composition can be introduced. Since a UHV environment is required for proper sample preparation, these reactors also contain typical UHV equipment such as an ion gun and leak valves to introduce gases. Therefore, these reactors have a large volume (> 1 L) and can only be operated in batch mode. Operating in batch mode has the major disadvantage that the gas composition in the reactor changes over the course of a measurement. Nevertheless, interesting results have been obtained for batch reactors.8,61–63 Another method often used is the transfer of the sample from a UHV preparation chamber to the high-pressure reactor. This has the drawback that sample realignment with respect to the X-ray beam is required after each transfer. To circumvent these disadvantages, a novel reactor setup was developed in the framework of a collaboration between Leiden University and the ID03 beamline at the European Synchrotron Radiation Facility.34 This setup combines a small-volume flow reactor with sample preparation under UHV conditions. With this instrument, the surface structure and morphology under reaction conditions can be investigated using SXRD and GISAXS. The activity of the sample can be measured simultaneously by mass spectrometry.

Design of the ReactorSXRD In the design of the ReactorSXRD34 we have combined the requirements of UHV sample preparation on the one hand and high-pressure, high-temperature reaction conditions on the other hand. To avoid sample realignment after every preparation cycle, we have chosen not to use a sample transfer mechanism between the two environments. Instead, we keep the sample fixed inside the instrument and move the upper part of the setup around the sample. The design is shown in Fig. 2. The chamber consists of two steel plates connected by a bellow. The lower plate is mounted on a five-axis positioning system that is part of a six-circle diffractometer.64 The lower plate includes the sample holder and the mass spectrometer. The upper plate includes the ion gun and evaporator, and can be translated vertically. If the bellow is completely extended, as in Fig. 2A, the setup is in the sample preparation configuration. When the top part of the chamber is lowered over the sample, as in Fig. 2B, the small volume around the sample is fully separated from the UHV surroundings. The upper part of this reactor volume is defined by a beryllium dome with good transparency for X-rays.65–67 The gases are mixed in a gas system. Under the sample holder support, the gas inlet and outlet lines are mounted. Gas analysis is performed via a leak into the UHV part of the instrument that contains a mass spectrometer. The sample is heated by a graphite heating element embedded in boron nitride.68

Combination of High-Pressure SPM With High-Pressure SXRD In view of the complexity of catalytic processes, techniques have to be combined in order to arrive at a sufficiently complete description of the working mechanisms of a catalyst. However, every technique introduces constraints that often make it difficult to study

Fig. 2 (A) Cut view of the setup in the UHV sample preparation geometry. (B) Cut view of the setup with closed reactor, 90 rotated with respect to the view of Fig. 2A. The beam is located 170 mm above the diffractometer sample stage surface. The labels denote: (1) turbo pump, (2) mass spectrometer, (3) manual UHV valve, (4) guiding rods for vertical movement of top part of chamber, (5) sample holder foot, (6) sample holder, (7) X-ray beam height, (8) evaporator port, (9) water-cooled top flange, (10) beryllium dome, (11) ion gun port, (12) reactor gas exhaust line, (13) UHV leak valve, (14) Huber five-axis positioning system, (15) cold cathode pressure gauge, (16) blind flange, (17) electromotor and drive shaft, (18) threaded drive rods for vertical movement of top part of chamber, (19) chain drive mechanism for vertical movement, (20) gas entry line, (21) UHV vent valve, and (22) steel bellow. Reproduced from van Rijn, R.; Ackermann, M. D.; Balmes, O.; Dufrane, T.; Geluk, A.; Gonzalez, H.; Isern, H.; de Kuyper, E.; Petit, L.; Sole, V. A.; Wermeile, D.; Felici, R.; Frenken, J. W. M. Rev. Sci. Instrum. 2010, 81, 014101.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

341

catalysts in precisely the same environment. Here I present an instrument that has successfully circumvented these constraints for SXRD/GISAXS and STM/AFM.69 We have integrated a scanning probe microscope (SPM) with an X-ray scattering platform in a gas flow reactor combined with a UHV system that can be mounted on a diffractometer. With this experimental setup we can directly combine X-ray scattering data with real-space information obtained with the microscope. As the basis for our design, we have used the ReactorSXRD setup that has been developed previously for high-pressure, high-temperature SXRD experiments on model catalysts (mentioned earlier). This instrument satisfies all requirements for in situ X-ray scattering experiments.34 For X-ray scattering experiments, we use a beryllium dome, in view of its excellent transmission for X-rays. We have developed an alternative, dedicated aluminum dome for the combined X-ray and SPM experiments. A miniature SPM unit has been constructed that can operate in a chemically harsh, high-temperature, high-pressure environment. It is mounted on top of the aluminum dome. The geometry of the combined instrument has been kept completely modular, which enables us to exchange components easily and switch between STM, AFM, or SXRD modes of operation. The basic configuration of the SPM part of the system is the same as that of the ReactorSTM20 and the ReactorAFM.21 Central to the design is a cylindrical piezo element that is used for the X-, Y-, and Z-motion. The same piezo element serves as the actuator of a stick-slip translation stage for the coarse approach of the tip or the QTF with tip to the surface. The complete SPM unit including the aluminum dome has a height of 20 cm (see Fig. 3). This compact SPM design naturally leads to high mechanical resonance frequencies, which is beneficial for making the instrument minimally sensitive to external vibrations. Switching between STM and AFM modes is straightforward, as it merely requires replacing the slider carrying the probedeither the tip or the QTF with tipdand the electronics that connect to the probe. The mechanical loop that determines the sensitivity of the microscope to external vibrations includes the sample, the sample holder, the holder’s supporting structure, the top flange of the UHV chamber, the aluminum dome, the base plate on which the piezo tube is mounted, the piezo element itself, the aluminum tube, the slider and, finally, the tip or combination of tuning fork with tip. In order to eliminate all flexibility in this loop, the top flange of the UHV chamber has to be maximally lowered.34 Unfortunately, this reduces the leak from the reactor volume to the UHV chamber effectively to zero, so that we cannot use the mass spectrometer on that chamber for gas analysis simultaneously with SPM imaging. We have solved this by attaching an additional UHV chamber with residual gas analyzer to the exhaust gas line of the reactor. The microscope is controlled by SPM electronics57,58 from Leiden Probe Microscopy B.V., which features a digitally controlled fast analog scan generator and a high-bandwidth analog feedback system for high-speed imaging.

Fig. 3 Schematic cross section (left) and photograph (right) of the X-ray-transparent aluminum dome that defines the reactor volume. In the cross section, the sample is indicated together with incident and reflected X-rays (red arrows). The SPM part of the system is mounted on top of the dome and partially reaches into the dome. The STM tip or the AFM tuning fork with tip is located on the central axis of the dome and the sample. A flexible seal separates the high-pressure reactor volume from the vacuum that surrounds the piezo element. The cables at the top of the photograph connect the SPM unit to the control electronics, including a preamplifier that is positioned at 30 cm distance. Reproduced from Onderwaater, W. G.; van der Tuijn, P. C.; Mom, R. V.; van Spronsen, M. A.; Roobol, S. B.; Saedi, A.; Drnec, J.; Isern, H.; Carla, F.; Dufrane, T.; Koehler, R.; Crama, B.; Groot, I. M. N.; Felici, R.; Frenken, J. W. M. Rev. Sci. Instrum. 2016, 87, 113705.

342

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

High-Pressure Optical Microscopy Instrumentation Reflectance difference measurements, where the change in reflectivity is measured in real time during some surface-altering process, have been used successfully in the past to monitor very subtle surface processes. The non-invasiveness of the technique and the fact that light is only weakly scattered at ambient gas pressures has led to the use of optical tools such as reflectance difference measurements, reflection anisotropy microscopy, and ellipsometry for crystal growth70,71 and catalysis processes.72–74 Although the diagnostic power of reflectance difference spectroscopy is greatly enhanced by the use of multiple wavelengths, much of the relevant surface behavior can be inferred already by using a single wavelength, that is, monochromatic light. Surface anisotropy, when the surface shows different characteristics along its in-plane axes, can be explored by investigating the response of the two different polarization directions. In previous experiments, we had noticed that after exposure of Pd(100) to CO oxidation conditions, the surface had changed visibly. It had turned from an almost perfect mirror into a slightly dulled mirror. The origin of this transition was unknown since our SPM and SXRD setups do not provide optical access to the sample in its reactor environment. In order to analyze and quantify these optical changes during the catalytic reaction we built a new experimental setup, using simple optics. The aim of this setup is to image the optical reflectance of a complete, 1-cm-diameter model catalyst sample, with high temporal resolution. We are interested in both the specularly and the diffusely scattered light from the sample surface. Resolving the angular intensity dependence of the diffusely scattered light, which is related to the characteristic length scales of the surface, would enable us to study the possible existence of mirror planes on the rough surface, which might be relevant information for the catalytic process. To measure the sample reflectance we use a home-built reflectometer.43 The sample is housed in a small flow reactor, designed by the company Leiden Probe Microscopy B.V. (LPM), shown in Fig. 4. This mini-reactor is a simplified version of the ReactorSXRD chamber,34 with the same sample mounting stage and similar reactor volume but without the UHV sample preparation environment. The chosen geometry enables a direct comparison of the optical data with the SXRD results obtained previously. With an LPM gas supply system we can set the total flow rate, total pressure (120–2000 mbar (the latter is chosen as an upper limit because of safety specifications of the reactor window)) and partial pressure ratios between all constituent gases at ratios ranging from 100:1 up to 1:100. The maximum flow rate per constituent gas is 10mLn/min. With a reactor volume of  16 mL the refresh rates are in the order of 1 min. This has to be taken into account when switching the gas composition in the reactor. The composition of the gas mixture that leaves the reactor is measured with an LPM high-pressure inlet residual gas analyzer, T100 residual gas analyzer.75 The reflectometer itself consists of two stages. The pre-sample stage starts with a light source. A collimator collects the light and a spatial filter creates a parallel beam. The post-sample stage collects the light, chooses a specific imaging mode, and records the result with a camera. As a light source we use a Thorlabs M625L3 LED with a 625 nm central wavelength. The light is first collimated after the source with an aspherical lens with a 40 mm focal length and a numerical aperture of 0.554 to capture a large fraction of the emitted light. The light passes a diffusor to eliminate any spatial pattern arising from the light source; then the light is spatially filtered with a lens with 10 cm focal length that focuses the light on a 150 mm pinhole. After the pinhole, a lens with 6 cm focal length collimates the beam, recreating a parallel beam. A diaphragm after the spatial filter allows us to set the beam diameter. We do not measure the LED intensity and therefore cannot correct the measured data for fluctuations in the light. A schematic design of the light source and spatial filter is shown in Ref.43 A pellicle beam splitter reflects 50% of the light toward the sample. 50% of the light that is reflected by the sample passes through the beam splitter toward the camera. Behind the beam splitter the post-sample stage starts. Fig. 5 shows four different lens configurations that can be applied to disentangle the different reflected components. In Fig. 5A all reflected light is used to create an image

Fig. 4 Cross section of the Leiden Probe Microscopy mini-flow reactor. A transparent window closes the reactor from the top. The sample is placed on top of the boralectric heater that can heat the sample up to 800 C. With a thermocouple attached to the sample we measure the temperature. The feedthroughs for the gases, heater, and the thermocouple are below the sample holder in the bottom plate. Gas flows in from the gas supply system (blue arrow) and out toward the gas analyzer (red arrow). The sample holder and reactor enclosure are similar to those of the ReactorSXRD chamber van Rijn, R.; Ackermann, M. D.; Balmes, O.; Dufrane, T.; Geluk, A.; Gonzalez, H.; Isern, H.; de Kuyper, E.; Petit, L.; Sole, V. A.; Wermeile, D.; Felici, R.; Frenken, J. W. M. Rev. Sci. Instrum. 2010, 81, 014101 and can be exchanged. Reproduced from Onderwaater, W. G.; Taranovskyy, A.; Bremmer, G. M.; van Baarle, G. J. C.; Frenken, J. W. M.; Groot, I. M. N.; Rev. Sci. Instrum. 2017, 88, 023704.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

343

Fig. 5 Schematics of several possible configurations of the optical system. A parallel light beam arrives from the left where it hits the beam splitter. Half of the intensity is reflected downwards and half is transmitted. After reflection on the sample surface (gold) the light is transmitted through the beam splitter. Four analysis methods are indicated. (A) Imaging of the surface. (B) In a bright-field configuration diffusely reflected light (purple) is blocked by the pinhole. (C) In a dark-field configuration specularly reflected light (red) is blocked by the beam stop. (D) In an angular configuration the distance to the center of the camera is determined by the angle of scattering. Reproduced from Onderwaater, W. G.; Taranovskyy, A.; Bremmer, G. M.; van Baarle, G. J. C.; Frenken, J. W. M.; Groot, I. M. N.; Rev. Sci. Instrum. 2017, 88, 023704.

of the sample. Fig. 5B corresponds to a bright-field configuration where we use a pinhole in the focal plane of lens L4, in order to select only the specularly reflected components. In the dark-field configuration (Fig. 5C) a beam stop is placed in the focal plane of lens L4. In this way the specularly reflected beam is blocked and only the light that is diffusely scattered from the sample is allowed to pass. L5 behind the beam stop focuses the remaining light on the camera. In Fig. 5D the camera is placed in the focal plane of L4. This creates a far-field image that registers the angles at which the light is reflected at the surface. In principle, the optical system could have been designed such that it integrated various imaging modes in the same setup. For simplicity, we chose to simply reconfigure the optical setup between different types of experiments. Since the illumination stage is identical, data collected in different configurations can be directly compared.

Catalytic Systems Investigated Under High-Pressure Conditions In this section I discuss some catalytic systems that have been investigated using high-pressure STM, SXRD, and/or optical microscopy. I start out with a short summary of effects observed for CO oxidation on Pd(100). To move forward to more realistic model catalysts we also studied this reaction on Pd nanoparticles using SXRD. This technique was also used to investigate CO oxidation on Rh(111). The second reaction that is relevant for automotive catalysis is the reduction of NO by either CO or H2. Here I discuss NO reduction using H2 on Pt(110) investigated with STM and SXRD. Recently, it has become possible to also study chemical reactions in atmospheric pressures of highly corrosive gases. The example I discuss here concerns the production of chlorine from HCl and O2 (the so-called Deacon reaction), investigated mainly using SXRD. Some preliminary STM results are shown as well.

344

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

CO Oxidation on Pd(100) CO oxidation on Pd(100) was investigated at a total pressure of 1.25 bar and a temperature of 408 K using the ReactorSTM.76–78 Fig. 6 shows the partial pressures of the reactants (CO and O2) and the product (CO2) (upper panel) and characteristic STM images (lower panel). At t ¼ 0 s the CO pressure was slowly decreased. Image A shows the metallic Pd(100) surface, consisting of flat terraces separated by monoatomic steps. This surface has the same structure as a clean Pd(100) surface imaged in UHV. When decreasing the CO pressure, the CO2 pressure decreases accordingly. However, at t ¼ 1612 s, a sudden increase in CO2 production is observed, coinciding with a significant change of the structure of the Pd(100) surface (Image B, top). A large number of clusters is observed, with heights equal to that of a monoatomic step of Pd(100). This formation of clusters is found to accompany the initial stages of Pd(100) oxidation.79 Therefore, the authors conclude that in Image B the surface oxidizes to form PdO, and that this causes the observed increase in reactivity. In time, roughness develops and a polycrystalline, granular structure is formed (Images C to E). In this stage, the CO2 production decreases proportional to the decreasing CO pressure. At t ¼ 4700 s the CO partial pressure was increased again. The Pd smoothens and the flat, metallic surface is restored (Images F and G). With the increase in CO pressure, a modest increase in CO2 production is observed. At t  6700 s the O2 flow was switched off and the CO pressure was increased (Image H). Switching from an oxygen-rich to a CO-rich flow, the CO2 production peaks at t ¼ 7010 s. As the O2 was pumped out, the CO2 production steadily decreased. When comparing Images G and H, the authors observed no essential structural changes when switching from a modest to a high ratio between the partial pressures of CO and O2.

Fig. 6 (Upper panel) Mass spectrometer signals of the reactants (CO and O2) and the product (CO2) during CO oxidation on the Pd(100) surface at a total pressure of 1.25 bar and a temperature of 408 K. The labels (AdH) refer to the STM images shown in the lower panel. The arrows mark the changes in reactivity from low to high and vice versa. (Lower panel) STM images (140 nm  140 nm, Vbias ¼ 112 mV, Itunnel ¼ 200 pA) obtained during the CO oxidation reaction. Reproduced from Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M.; Surf. Sci. 2004, 552, 229.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

345

Fig. 7 Spontaneous oscillations in the CO oxidation rate on Pd(100) measured by (A) SXRD and (B) mass spectrometry. Reproduced from Hendriksen, B. L. M.; Ackermann, M. D.; van Rijn, R.; Stoltz, D.; Popa, I.; Balmes, O.; Resta, A.; Wermeille, D.; Felici, R.; Ferrer, S.; Frenken, J. W. M. Nat. Chem. 2010, 2, 730.

As discussed earlier, the transition from a metallic Pd(100) surface to the PdO phase increases the reactivity for the CO oxidation reaction. Under conditions in which both the oxide and metal phase can exist, oscillations in the reaction rate were observed.76,77 Using operando SXRD, it was shown that the density of steps on the catalyst surface influences the stability of the metal and oxide phases, and therefore that steps regulate the self-sustained oscillations during CO oxidation on Pd(100) at atmospheric pressure.80 Fig. 7A shows the SXRD data for the (h,k,l) ¼ (1,0,2) diffraction peak of Pd(100) during CO oxidation at 447 K and a CO partial pressure of 25 mbar and an O2 partial pressure of 500 mbar. Fig. 7B shows the partial pressures of CO and CO2 simultaneously measured with a mass spectrometer. In the low-reactivity phase, the diffraction peak had a high intensity, indicating that the surface was metallic, while in the high-reactivity phase the diffraction peak had a low intensity, indicative of the oxide surface. The observed increase of the inverse value of the full width at half maximum (FWHM) of the metal diffraction peak indicated that the surface roughness or steppdensity ffiffiffi pffiffiffidecreased, until the transition to the high-reactivity phase occurred. This transition coincided with the formation of the 5  5 ultrathin Pd surface oxide, followed by the formation of the epitaxial bulk-like PdO film. During the Mars-van Krevelen-type redox reaction,81 a small fraction of the Pd atoms in the oxide film becomes sufficiently undercoordinated that they become mobile and diffuse over the surface until they are re-oxidized and immobilized on top of the oxide film.5,6 At the low operation temperature of the experiment, the oxide is not mobile enough to counteract the roughening of the surface. As a result, roughness gradually accumulates, approximately linear with the amount of generated CO2. In the metallic phase, on the other hand, a high surface mobility is observed, and the surface roughness decreases in time. Therefore, the authors postulate that variation in the roughness, that is, step density, induces the transitions from metallic to oxidic phase and vice versa, thereby regulating the oscillations in CO2 production. Using optical microscopy, we investigated CO oxidation on Pd(100) during a linear temperature ramp from room temperature to 575 K and back.44 We started the experiment with a reduced Pd sample in the metallic state. During the temperature ramp, we kept the gas mixture constant at a total pressure of 200 mbar, with an O2 flow rate of 3 mLn/min, and a CO flow rate of 1.5 mLn/min. Panel A of Fig. 8 shows the behavior of the surface reflectivity, averaged over a small part of the surface. In the episode labeled 1 in Fig. 8 we start at low temperature. The increase in temperature is initially accompanied with a modest, slow decrease of DR, which can be ascribed to thermally induced drift of the Pd sample. In episode 2 we see an increase of the CO2 production. In episode 3 the CO2 production shows a sharp increase. Coinciding with this chemical change in episode 2, DR shows a sudden, small decrease of 0.5%. The resulting DR is maintained for only a few seconds, until episode 4 starts, in which the partial pressure of CO2 remains constant, but DR decreases rapidly with time. The start of episode 5 marks the point in time where the reduction of DR slows down significantly. In this episode, the composition of the gas that leaves the reactor is constant. Episode 6 starts at the point in time when we have reached the maximum temperature and start cooling. Initially, the reduction of DR slows

346

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

Fig. 8 CO oxidation on Pd(100) during large temperature variations. (A) Variation of DR with time, averaged over a small sample area. (B) Production of CO2 as measured with the T100 residual gas analyzer. (C) Setpoint temperature and actual sample temperature. The different episodes indicated are discussed in the main text. The insets show enlarged views of two episodes of interest with rapid changes in reactivity and reflectance. Reproduced from Onderwaater, W. G.; Taranovskyy, A.; van Baarle, G. J. C.; Frenken, J. W. M.; Groot, I. M. N. J. Phys. Chem. C 2017, 121, 11407 (submitted).

down and DR levels off. In the second half of episode 6, DR increases somewhat. In episode 7 we see the combination of a temporary decrease in CO2 production and a very small upward step in DR, while in episode 8 the CO2 production rate is at its previous level. This cycle repeats itself at episodes 9 and 10, this time with a more pronounced increase in DR. In episode 11, the CO2 production severely decreases. This coincides with an almost full recovery of the reflectivity. In the final episode 12, the measured CO2 partial pressure falls in the noise level. The slow increase in reflected intensity results from the reverse thermal drift of that experienced in episode 1. Armed with only the DR and gas composition information in Fig. 8, it is impossible to give a full description of all changes in surface structure and reaction mechanisms during the experiment. In our interpretation, we make use of previous experiments on Pd(100), performed under comparable conditions, with other structurally sensitive techniques, in particular STM and SXRD.76,80,82–84 This brings us to the following scenario. In episode 1 the sample is CO-poisoned and therefore in a metallic state. The CO2 production follows the Langmuir–Hinshelwood mechanism. As the temperature increases, the surface is increasingly populated by O atoms. We speculate that in the last part of episode 2, the reaction rate increases as a result of the presence of an alternative and more reactive co-adsorption structure of O and CO on the metal surface. We interpret the sudden decrease of DR, accompanied by the jump in CO2 partial pressure, as the consequence of the sudden formation of a surface oxide, as has been identified in SXRD and STM experiments.76,83 This surface oxide forms the starting point for the rapid growth of a thin film of a few nanometers thick bulk-like PdO. This goes hand-in-hand with the build-up of additional roughness and, hence, a further reduction of DR (episode 4). This growth of the bulk-like oxide is self-terminating and therefore the rapid reduction of DR also comes to an end (episode 5). Under these conditions, the reaction proceeds according to the Mars-van Krevelen mechanism. As we have seen earlier, this reaction mechanism leads to a steady roughening of the surface and the corresponding steady reduction of DR. When the temperature is decreased again in episode 6, the reaction rate, which is mostly limited by the diffusion of CO in the gas phase, slows down. This progressively reduces the rate at which DR drops. We interpret the effects in episodes 7–10 as reduction and oxidation cycles, induced by the interplay of the decreasing temperature and the changes in surface roughness. In episode 11 the surface is fully reduced. With the removal of the rough oxide the diffusivity of the palladium atoms is increased and the surface smoothens. In the gas analysis we can see that upon reduction of the surface the reactivity is decreased. The reaction is back to the initial Langmuir–Hinshelwood mechanism.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

347

CO Oxidation on Pd Nanoparticles To bridge the materials gap, CO oxidation was also investigated on size-selected 15 nm Pd nanoparticles supported on a - Al2O3 using SXRD, focusing on the reaction conditions where spontaneous reaction oscillations occur.85,86 Few attempts have been made to systematically investigate these oscillations on nanoparticle catalyst systems. Detailed mass spectrometry analysis of oscillations during CO oxidation on platinum particles87 suggested that the mechanism described earlier for Pd(100) could also apply to these nanoparticles. The upper panel of Fig. 9 shows the periodic variations in the CO2 and CO partial pressures during spontaneous reaction oscillations on Pd/a-Al2O3. The particles were exposed to a constant gas flow of CO, O2, and Ar at partial pressures of 9, 210, and 281 mbar, respectively. The sample temperature was  588 K. The SXRD data taken simultaneously with the mass spectrometer data (see bottom panel of Fig. 9) exhibited periodic variations in the diffraction intensities at angles corresponding to either the crystal structure of metallic Pd or that of PdO. These variations were out of phase: when the metallic Pd signal showed a low intensity, the PdO signal showed a high intensity, and vice versa. These oscillations were completely synchronized with those in the partial pressures of CO and CO2. Fig. 9 shows that the high-reactivity phase coincided with the presence of PdO. The sample temperature varied in response to the changes in turnover rate, since CO oxidation is an exothermic reaction. The oscillations had a period of a few minutes, depending on the temperature, and could be sustained indefinitely within the time scale of the synchrotron experiments. The presence of the X-ray beam only quantitatively affects the oscillations: when the X-ray beam is off, the oscillation period is a factor 2 longer and the amplitude of the temperature variations increases by 50 %. In addition to SXRD, Fig. 9 also shows the results of GISAXS experiments. Also the GISAXS signal varied synchronously with the reaction oscillations.

CO Oxidation on Rh(111) Lundgren and coworkers investigated the CO oxidation reaction over Rh(111) at 513 K using high-pressure SXRD combined with mass spectrometry.61 They found that, similar to Pd(100), an increased CO2 production coincides with the formation of a surface

Fig. 9 Spontaneous reaction oscillations observed using X-ray scattering. The upper panel shows the CO and CO2 partial pressures as a function of time. The middle panel shows (a low resolution measurement of) the sample temperature. The gas feed was kept constant at 9 mbar CO, 210 mbar O2, and 281 mbar Ar at a total flow of 60 mLn/min, and the sample heater was operated at constant power (its value was adjusted slightly around t ¼ 20,000 s). Note that the absolute value of the partial pressures of CO and CO2 contains an error due to varying background level for these gases in the mass spectrometer. The amplitude of the oscillations in these signals is not affected by this. The bottom panel shows the periodic intensity variations in three different X-ray signals: the metal Pd peak at the 2q ¼ 17:6 diffraction angle, the oxide PdO peak at 2q ¼ 15 , and the integrated intensity from a region in a series of GISAXS patterns. A high metal intensity is correlated with a low turnover rate, whereas a high oxide intensity correlates with a high turnover rate. Also the changes in the GISAXS signal are synchronized with the reaction oscillations. Reproduced from Roobol, S. B. The structure of a working catalyst, PhD thesis, Leiden University, 2014; Onderwaater, W.G.; CO oxidation catalysis at multiple length scales, PhD thesis, Leiden University, 2016.

348

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

Fig. 10 The CO oxidation reaction on Rh(111) at near-atmospheric pressure and 513 K. (A) Partial pressures of CO, O2, and CO2. (B) Consecutive SXRD scans showing the presence and absence of the surface oxide. Reproduced from Gustafson, J.; Westerström, R.; Mikkelsen, A.; Torrelles, X.; Balmes, O.; Bovet, N.; Andersen, J. N.; Baddeley, C. J.; Lundgren, E. Phys. Rev. B 2008, 78, 045423.

oxide. This surface oxide is built up by three close-packed hexagonal layers of O–Rh–O.88 Fig. 10 shows the results of the CO oxidation experiment on Rh(111). At t ¼
1000 s, 300 mbar of O2 was present in the reactor, and a surface oxide is detected (see Fig. 10B). At t ¼ 0 s, 330 mbar CO was added to the reactor (see Fig. 10A). When CO was introduced, it started consuming oxygen from the surface oxide to form CO2. As seen from Fig. 10, the CO2 signal of the mass spectrometer started to slowly rise. The consumption of oxygen from the surface oxide led to a change of the Rh(111) surface to the metallic state, as observed from the SXRD signal (Fig. 10B). Clearly, in this regime of high pressure and elevated temperature, the reaction follows the Mars-van Krevelen mechanism. At t ¼ 3300 s, a significant increase in the CO2 production was observed. This change in reactivity coincided with the return of the surface oxide on Rh(111), as can be seen from Fig. 10B. The increased CO2 production introduced a temperature rise of  80 K, due to the exothermic nature of the reaction. In a subsequent study, Lundgren and coworkers concluded that the reactivity is always very high when the surface oxide is present, but low otherwise, that is, in the metallic and the bulk oxide phase.89 The same behavior was observed for Rh(100), vicinal Rh surfaces, and Rh nanoparticles.

NO Reduction on Pt(110) The catalytic conversion of nitrogen oxides (NOx) is one of the three processes taking place on the three-way car catalyst.90,91 The major exhaust pollutants of cars are hydrocarbons, CO, and NOx. The NOx is formed during the phase of very high temperature (> 1500  C) of the combustion process resulting in thermal fixation of the nitrogen in air.92 The exhaust of NOx into the air, combined with sunlight, results in the formation of ozone, a major component of smog. It was estimated that by the year 2000, over 800 million tons of combined pollutants of hydrocarbons, CO, and NOx have been abated using automotive catalysts.93 Considering the importance of the NO reduction reaction (either by H2 or CO), it is very surprising that almost no in situ or operando studies have been performed for this system. To the best of our knowledge, the only results for NO reduction on platinum under reaction conditions have been obtained in our group.94–96 Here we discuss the reduction of NO by H2 over Pt(110). This reaction is not very selective over Pt,97 and many products, for example NH3, H2O, N2, and N2O, can be formed. For application as a car catalyst these products need to be avoided, and typically rhodium is added to enhance selectivity toward N2 production.98 Fig. 11 shows the results obtained with high-pressure STM when exposing the Pt(110) surface to H2 and NO. This surface easily reconstructs, and we have observed several new surface reconstructions under high-pressure conditions that were not seen before in UHV. In Fig. 11A the gas composition is shown as a function of time. Gray bars show the times at which the STM images were measured. Prolonged exposure to pure H2 at room temperature at a pressure of 1.2 bar resulted in a row structure with a (1  4) periodicity (see Fig. 11B). Also, some deeper missing-row features can be observed. Subsequently, this surface has been exposed to NO at increasing NO: H2 ratios (Fig. 11C–E). The blue dot in the first image (B) indicates the switching point from pure H2 to a mixture with a NO:H2 ratio of 1:7.2. Exposure to NO slowly lifts the missing-row reconstruction, leaving flat terraces on the surface. Using high-pressure SXRD, we have observed many different surface structures on Pt(110), depending on the reaction conditions.96 We classify them into two groups: surface reconstructions, which require modest rearrangement of platinum atoms; and faceting of the surface, corresponding to massive material transport over the surface.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

349

Fig. 11 Gas composition as a function of time and in situ STM images during NO reduction on Pt(110) at a total pressure of 1.2 bar and room temperature. (A) Stepwise change from H2 (blue) to NO (green) at room temperature and a total pressure of 1.2 bar. Gray regions indicate at which point in time under what conditions the STM images were obtained; (B) pure H2 followed by a switch to a partial pressure ratio of NO:H2 ¼ 1:7.2, 80 nm  82 nm, Vbias ¼
0.70 V, Itunnel ¼ 139 pA; (C) Pt(110) at a partial pressure ratio NO:H2 ¼ 1.1:1.0, 73 nm  80 nm, Vbias ¼
0.70 V, Itunnel ¼ 141 pA; (D) pure NO atmosphere, 78 nm  80 nm, Vbias ¼
0.70 V, Itunnel ¼ 141 pA; (E) after prolonged NO exposure, 158 nm  160 nm, Vbias ¼
0.70 V, Itunnel ¼ 139 pA. Reproduced from van Spronsen, M. A.; van Baarle, G. J. C.; Herbschleb, C. T.; Frenken, J. W. M.; Groot, I. M. N. Catal. Today 2015, 244, 85.

When the partial pressures of NO and H2 are varied, several reconstructions of the Pt(110) surface are observed at 1 bar total pressure between 300 and 400 C. Fig. 12 shows an overview of the specific partial pressures and temperature conditions at which the various reconstructions were observed. The (13) reconstruction observed in H2 is probably the same missing-row reconstruction as observed in vacuum on the freshly prepared sample, with periodicities in l similar to those reported for the (13) missingrow reconstruction by Robinson and coworkers.99 However, none of the other reconstructions that we observe in this work have been identified before in UHV studies. The presence of such a variety of reconstructions can be explained taking into account the higher mobility of surface atoms at ambient pressure and the presence of adsorbed species, which under ambient pressure

350

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

Fig. 12 Observed reconstructions (periodicities) on Pt(110) as a function of gas composition (mixture of NO and H2 at 1 bar total pressure). The sample temperature was in the range 300–400 C, except for the “0.22” reconstruction which was observed at approximately 100–150 C. Reproduced from Roobol, S. B.; Onderwaater, W. G.; van Spronsen, M. A.; Carla, F.; Balmes, O.; Navarro, V.; Vendelbo, S.; Kooyman, P. J.; Elkjaer, C. F.; Helveg, S.; Felici, R.; Frenken, J. W. M.; Groot, I. M. N. Phys. Chem. Chem. Phys. 2017, 19, 8485.

conditions may form different adsorbate structures. One of these reconstructions corresponds to an incommensurate (43) structure. However, the structure is not stable and it always transformed into a commensurate (43) structure over time. The “0.22” reconstruction, which is also incommensurate, seems to be completely different. It appears to be stable over time. Since the intensity of the l-scans (see Fig. 5 of Ref. 96) strongly peaks at integer l values, meaning that the periodicity matches that of the bulk Pt lattice, this suggests that it derives from a significant restructuring of the Pt surface rather than an adsorbate overlayer pattern. Near the end of the experiment, under conditions already probed before, the surface started faceting (see Fig. 12 for the condi  tions). The facets are tilted by 8–12 degrees away from the (110) surface in the 11 0 direction. The facet angle seems to be negatively correlated to the NO partial pressure, with higher NO pressures resulting in smaller angles. It took approximately 4 h for the system to reach a stable state, indicating that the facets were growing in average size. The tilt of approximately 10 implies that the facets typically have the (320) structure, a crystal plane that makes an 11.3 angle with the (110) plane. The higher index orientations (430) and (540), having a structure similar to (320) but with four and five atoms per terrace and tilts of 8.1 and 6.3 , respectively, are likely to be present as well, especially at higher NO pressures. We suggest two possible mechanisms for the development of faceting under high-pressure conditions: the strong binding of an adsorbate to the step edges,100,101 and in this case NO, H2O, NH3, or atomic oxygen are likely candidates, or adsorbate-induced stress102–104 due to the strong repulsive interaction between NO molecules at higher coverages.

Chlorine Production on RuO2(110)/Ru(0001) Chlorine is one of the most important compounds produced in the chemical industry. Its worldwide annual production is approximately 50 megaton, and it is responsible for  50% of the total turnover of the chemical industry.105 It is usually produced from HCl or chloride salts. The production method of choice is presently electrochemical reduction of these reactants, but this is highly energy demanding and it generates large quantities of CO2. Due to these high economical costs and undesired effects on the environment, interest has been growing in the heterogeneously catalyzed oxidation of HCl, the so-called Deacon process.106,107 According to Sumitomo Chemicals the Deacon process can reduce the power consumption from 1100 to 165 kwh per ton Cl2 produced.108 Several catalysts have been developed for the industrial Deacon process. Examples are CuCl2,106,107 CuCl2–KCl/SiO2,109,110 and Cr2O3/SiO2.111–113 Recently, Sumitomo Chemicals and Bayer Materials Science (now Covestro) have developed a very active new catalyst, based on RuO2/TiO2 108,114 and RuO2/SnO2.115–117 All these catalysts have been employed with varying degrees of success. Severe problems are loss of catalytic activity due to volatilization of the active phase, sintering of active sites and supports, and corrosion issues due to the combined presence of HCl and H2O. At the moment, the catalysts employed in the industry (i.e., Sumitomo Chemicals and Covestro) are based on RuO2. Due to the relevance of this chemical reaction for industry, academic studies into the delicate interplay between the atomic-scale structure, the reaction mechanisms, and the performance of the catalyst have been performed recently. Over and coworkers investigated this highly corrosive reaction using high-pressure SXRD in a batch reactor over RuO2(110) and RuO2(100) model catalysts.118,119 When exposing the model catalysts to mixtures of HCl and O2 at temperatures above 500 K, part of the bridging O atoms are replaced by chlorine, a selective and self-limiting process. These chlorinated RuO2 surfaces are stable under reaction conditions varying from 1:4 to 4:1 HCl:O2 at mbar pressures and up to 685 K. The chlorination process is mediated by the formation of bridging H2O.120 At 500 K, water desorbs. Above 650 K, on-top Cl atoms recombine and desorb. This produces

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

351

Fig. 13 Structure–activity measurements of (A) RuO2(110) and (B) RuO2(100) exposed to 2 mbar HCl and 0.5 mbar O2 at 650 K. The oxide surface is stable under reaction conditions as indicated by the invariance of h-scans around the oxide peak (for clarity consecutive h-scans are shifted vertically by a constant value) and by the constant integral intensity of the oxide-related diffraction peak as a function of time (bottom inset). The decline of the oxygen partial pressure over time is shown in the top inset. Reproduced from Zweidinger, S.; Hofmann, J. P.; Balmes, O.; Lundgren, E.; Over, H. J. Catal. 2010, 272, 169.

vacant coordinatively unsaturated sites (cus). Bridging Cl then shifts to the vacant cus and recombines with adjacent on-top Cl to produce Cl2, that immediately desorbs.121,122 Now HCl from the gas phase adsorbs on the cus by transferring H to the newly formed bridging O atoms, forming water. Reactivity measurements in the batch reactor (see Fig. 13) found a turnover frequency of 0.6 Cl2 molecules per site per second at 650 K and partial pressures of HCl ¼ 2 mbar and O2 ¼ 0.5 mbar, independent of the model catalyst investigated. First results for the Deacon process on the ReactorSTM have shown that the setup is capable of dealing with these very aggressive reaction conditions. However, some experimental issues were observed and have been dealt with. These include corrosion of the W tip (we now switched to PtIr), corrosion of the Cu seal in the leak valve of the QMS chamber (we now switched to a Teflon-type seal), and decomposition of the Kalrez seal due to the combination of elevated temperatures and corrosive gases.

Conclusions and Outlook Using the ReactorSTM, ReactorAFM, ReactorSXRD, and the optical microscopy setup developed in our group, we are able to cast a direct view on catalyst surfaces under industrially relevant conditions of high pressure and temperature. As is clear from the examples discussed here, we find pressure gaps for most reaction systems, simply by observing surface structures under reaction conditions that have been observed never before in UHV. The effect of the high pressure is to thermodynamically or kinetically stabilize new surface structures and compositions that incorporate one or more of the reactants. These observations show that a catalyst is more than a mere spectator in chemical reactions: It actively participates! With the ongoing development in our group, for example the ReactorAFM and the combination of SXRD/GISAXS with STM/AFM, we will be able in the near future to also bridge the materials gap. Furthermore, the use of highly corrosive gases in our systems, such as HCl, H2S, or CH3SH, has shown no sign of degradation to the instruments. Therefore, soon there will be virtually no limit to the catalytic systems that we will be able to study.

Acknowledgments The work described in this article would not have been possible without the contributions of many people. I would like to thank Joost Frenken for giving me the opportunity to start my own research line in high-pressure catalytic surface science. Furthermore, I would like to thank Marcelo Ackermann, Gertjan van Baarle, Johan Bakker, Olivier Balmes, Mirthe Bergman, Stefania Bobaru, Marien Bremmer, Marta Cañas–Ventura, Francesco Carlà, Bert Crama, Knut Deppert, Jakub Drnec, Thomas Dufrane, Roberto Felici, Salvador Ferrer, Arjen Geluk, Johan Gustafson, Bas Hendriksen, Kees Herbschleb, Helena Isern, H. Y. Kim, Raymond Koehler, Ewie de Kuyper, Qian Liu, Edvin Lundgren, Maria Messing, Martin Moene, Rik Mom, Violeta Navarro, Alexei Ofitserov, Willem Onderwaater, Lucien Petit, Ioana Popa, Carlos Quirós, Peter Rasmussen, Andrea Resta, Richard van Rijn, Odile

352

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

Robach, Sander Roobol, Marcel Rost, Amirmehdi Saedi, Fred Schenkel, Matthijs van Spronsen, Dunja Stolz, Andriy Taranovskyy, Xavier Torrelles, Peter van der Tuijn, Didier Wermeille, and Rasmus Westerström. This work is financially supported by NanoNextNL, a micro- and nanotechnology consortium of the Government of the Netherlands and 130 partners, by a SmartMix grant and the NIMIC partner organizations through NIMIC, a public–private program, and by STW, which is financially supported by the Netherlands Organization for Scientific Research (NWO).

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54.

Wintterlin, J. Adv. Catal. 2000, 45, 131. Stoltze, P.; Nørskov, J. K. Phys. Rev. Lett. 1985, 55, 2502. Rodriguez, J. A.; Goodman, D. W. Surf. Sci. Rep. 1991, 14, 1. Su, X.; Cremer, P. S.; Shen, Y. R.; Somorjai, G. A. J. Am. Chem. Soc. 1997, 119, 3994. Over, H.; Kim, Y. D.; Seitsonen, A. P.; Wendt, S.; Lundgren, E.; Schmid, M.; Varga, P.; Morgante, A.; Ertl, G. Science 2000, 287, 1474. Hendriksen, B. L. M.; Frenken, J. W. M. Phys. Rev. Lett. 2002, 89, 046101. Over, H.; Muhler, M. Prog. Surf. Sci. 2003, 72, 3. Ackermann, M. D.; Pedersen, T. M.; Hendriksen, B. L. M.; Robach, O.; Bobaru, S. C.; Popa, I.; Quirós, C.; Kim, H.; Hammer, B.; Ferrer, S.; Frenken, J. W. M. Phys. Rev. Lett. 2005, 95, 25505. Westerström, R.; Gustafson, J.; Resta, A.; Mikkelsen, A.; Andersen, J. N.; Lundgren, E.; Seriani, N.; Mittendorfer, F.; Schmid, M.; Klikovits, J.; Varga, P.; Ackermann, M. D.; Frenken, J. W. M.; Kasper, N.; Stierle, A. Phys. Rev. B 2007, 76, 155410. Bluhm, H.; Hävecker, M.; Knop-Gericke, A.; Kiskinova, M.; Schlögl, R.; Salmeron, M. MRS Bull. 2007, 32, 1022. Salmeron, M.; Schlögl, R. Surf. Sci. Rep. 2008, 63, 169. Jansen, W. P. A.; van der Gon, A. W. D.; Wijers, G. M.; Rikers, Y. G. M.; Brongersma, H. H.; van der Hoogen, P. W.; de Laat, J. A. M.; Maas, T. M.; Dekkers, E. C. A.; Brinkgreve, P. Rev. Sci. Instrum. 2002, 73, 354. Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.; Clausen, B. S.; Topsøe, H. Science 2002, 295, 2053. Gai, P. L.; Boyes, E. D.; Helveg, S.; Hansen, P. L.; Giorgio, S.; Henry, C. MRS Bull. 2007, 32, 1044. Creemer, J. F.; Helveg, S.; Hoveling, G. H.; Ullmann, S.; Molenbroek, A. M.; Sarro, P. M.; Zandbergen, H. W. Ultramicroscopy 2008, 108, 993. de Smit, E.; Swart, I.; Creemer, J. F.; Hoveling, G. H.; Gilles, M. K.; Tyliszczak, T.; Kooyman, P. J.; Zandbergen, H. W.; Morin, C.; Weckhuysen, B. M.; de Groot, F. M. F. Nature 2008, 456, 222. Rasmussen, P. B.; Hendriksen, B. L. M.; Zeijlemaker, H.; Ficke, H. G.; Frenken, J. W. M. Rev. Sci. Instrum. 1998, 69, 3879. Weeks, B. L.; Durkan, C.; Kuramochi, H.; Welland, M. E.; Rayment, T. Rev. Sci. Instrum. 2000, 71, 3777. Tao, F.; Tang, D.; Salmeron, M.; Somorjai, G. A. Rev. Sci. Instrum. 2008, 79, 084101. Herbschleb, C. T.; van der Tuijn, P. C.; Roobol, S. B.; Navarro, V.; Bakker, J. W.; Liu, Q.; Stoltz, D.; Cañas-Ventura, M. E.; Verdoes, G.; van Spronsen, M. A.; Bergman, M.; Crama, L.; Taminiau, I.; Ofitserov, A.; van Baarle, G. J. C.; Frenken, J. W. M. Rev. Sci. Instrum. 2014, 85, 083703. Roobol, S. B.; Cañas-Ventura, M. E.; Bergman, M.; van Spronsen, M. A.; Onderwaater, W. G.; van der Tuijn, P. C.; Koehler, R.; Ofitserov, A.; van Baarle, G. J. C.; Frenken, J. W. M. Rev. Sci. Instrum. 2015, 86, 033706. Su, X.; Cremer, P. S.; Shen, Y. R.; Somorjai, G. A. Phys. Rev. Lett. 1996, 77, 3858. Rupprechter, G.; Dellwig, T.; Unterhalt, H.; Freund, H.-J. J. Phys. Chem. B 2001, 105, 3797. Rupprechter, G. MRS Bull. 2007, 32, 1031. Beitel, G. A.; Laskov, A.; Oosterbeek, H.; Kuipers, E. W. J. Phys. Chem. 1996, 100, 12494. Li, C.; Stair, P. C. Catal. Lett. 1996, 36, 119. Stair, P. C.; Li, C. J. Vac. Sci. Technol. A 1997, 15, 1679. Li, C.; Stair, P. C. Catal. Today 1997, 33, 353. Rotermund, H. H. Surf. Sci. 1997, 386, 10. Bernard, P.; Peters, K. F.; Alvarez, J.; Ferrer, S. Rev. Sci. Instrum. 1999, 70, 1478. Isern, H.; Peters, K.; Steadman, P.; Robach, O.; Alvarez, J.; Lundgren, E.; Ferrer, S. Surf. Sci. 2001, 482–485, 101. Saint-Lager, M. C.; Bailly, A.; Dolle, P.; Baudoing-Savois, R.; Taunier, P.; Garaudée, S.; Cuccaro, S.; Douillet, S.; Geaymond, O.; Perroux, G.; Tissot, O.; Micha, J. S.; Ulrich, O.; Rieutord, F. Rev. Sci. Instrum. 2007, 78, 083902. Ferrer, S.; Ackermann, M. D.; Lundgren, E. MRS Bull. 2007, 32, 1010. van Rijn, R.; Ackermann, M. D.; Balmes, O.; Dufrane, T.; Geluk, A.; Gonzalez, H.; Isern, H.; de Kuyper, E.; Petit, L.; Sole, V. A.; Wermeile, D.; Felici, R.; Frenken, J. W. M. Rev. Sci. Instrum. 2010, 81, 014101. Knop-Gericke, A.; Hävecker, M.; Schedel-Niedrig, T.; Schlögl, R. Catal. Lett. 2000, 66, 215. Evans, J.; Puig-Molina, A.; Tromp, M. MRS Bull. 2007, 32, 1038. Chen, C. J. Introduction to Scanning Tunneling Microscopy, Oxford University Press: New York, 1993. Kuipers, L.; Loos, R. W. M.; Neerings, H.; ter Horst, J.; Ruwiel, G. J.; de Jongh, A. P.; Frenken, J. W. M. Rev. Sci. Instrum. 1995, 66, 4557. Hoogeman, M. S.; van Loon, D. G.; Loos, R. W. M.; Ficke, H. G.; de Haas, E.; van der Linden, J. J.; Zeijlemaker, H.; Kuipers, L.; Chang, M. F.; Klik, M. A. J.; Frenken, J. W. M. Rev. Sci. Instrum. 1998, 69, 2072. Feidenhans’l, R. Surf. Sci. Rep. 1989, 10, 105. Robinson, I. K.; Tweet, D. J. Rep. Prog. Phys. 1992, 55, 599. Renaud, G.; Lazarri, R.; Leroy, F. Surf. Sci. Rep. 2009, 64, 255. Onderwaater, W. G.; Taranovskyy, A.; Bremmer, G. M.; van Baarle, G. J. C.; Frenken, J. W. M.; Groot, I. M. N. Rev. Sci. Instrum. 2017, 88, 023704. Onderwaater, W. G.; Taranovskyy, A.; van Baarle, G. J. C.; Frenken, J. W. M.; Groot, I. M. N. J. Phys. Chem. C 2017. 2017, 121, 11407. McIntyre, B. J.; Salmeron, M.; Somorjai, G. A. Rev. Sci. Instrum. 1993, 64, 687. Jensen, J. A.; Rider, K. B.; Chen, Y.; Salmeron, M.; Somorjai, G. A. J. Vac. Sci. Technol. B 1999, 17, 1080. Laegsgaard, E.; Österlund, L.; Thostrup, P.; Rasmussen, P. B.; Stensgaard, I.; Besenbacher, F. Rev. Sci. Instrum. 2001, 72, 3537. Kolmakov, A.; Goodman, D. W. Rev. Sci. Instrum. 2003, 74, 2444. Rössler, M.; Geng, P.; Wintterlin, J. Rev. Sci. Instrum. 2005, 76, 023705. Tao, F.; Nguyen, L.; Zhang, S. Rev. Sci. Instrum. 2013, 84, 034101. Lievonen, J.; Ranttila, K.; Ahlskog, M. Rev. Sci. Instrum. 2007, 78, 043703. D’Agostino, D.; Jay, D.; McNally, H. Microsc. Microanal. 2010, 16, 1042. Higgins, S. R.; Eggleston, C. M.; Knauss, K. G.; Boro, C. O. Rev. Sci. Instrum. 1998, 69, 2994. Lea, A. S.; Higgins, S. R.; Knauss, K. G.; Rosso, K. M. Rev. Sci. Instrum. 2011, 82, 043709.

Operando Scanning Probe Microscopy, Surface X-Ray Diffraction, and Optical Microscopy for Catalysis Studies

55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122.

353

Dong, G.; Fourré, E. B.; Tabak, F. C.; Frenken, J. W. M. Phys. Rev. Lett. 2010, 104, 096102. Frenken, J. W. M.; Groot, I. M. N. Operando Research in Heterogeneous Catalysis, Springer Verlag: Berlin, 2017. Rost, M.; Crama, L.; Schakel, P.; van Tol, E.; van Velzen-Williams, G.; Overgauw, C.; ter Horst, H.; Dekker, H.; Okhuijsen, B.; Seynen, M.; Vijftigschild, A.; Han, P.; Katan, A. J.; Schoots, K.; Schumm, R.; van Loo, W.; Oosterkamp, T. H.; Frenken, J. W. M. Rev. Sci. Instrum. 2005, 76, 053710. Rost, M. J.; van Baarle, G. J. C.; Katan, A. J.; van Spengen, W. M.; Schakel, P.; van Loo, W. A.; Oosterkamp, T. H.; Frenken, J. W. M. Asian J. Control 2009, 11, 110. Giessibl, F. J.; Hembacher, S.; Bielefeldt, H. Science 2000, 289, 422. Gross, L.; Mohn, F.; Moll, N.; Liljeroth, P.; Meyer, G. Science 2009, 325, 1110. Gustafson, J.; Westerström, R.; Mikkelsen, A.; Torrelles, X.; Balmes, O.; Bovet, N.; Andersen, J. N.; Baddeley, C. J.; Lundgren, E. Phys. Rev. B 2008, 78, 045423. Kasper, N.; Stierle, A.; Nolte, P.; Jin-Phillipp, Y.; Wagner, T.; de Oteyza, D. G.; Dosch, H. Surf. Sci. 2006, 600, 2860. Over, H.; Balmes, O.; Lundgren, E. Surf. Sci. 2009, 603, 298. HUBER Diffraktionstechnik, www.xhuber.de. Brush Wellman, www.berylliumproducts.com. Nowak, D. E.; Blasini, D. R.; Vodnick, A. M.; Blank, B.; Tate, M. W.; Deyhim, A.; Smilgies, D.-M.; Abruna, H.; Gruner, S. M.; Baker, S. P. Rev. Sci. Instrum. 2006, 77, 113301. Sakata, O.; Tanaka, Y.; Nikolaenko, A. M.; Hashizume, H. J. Synchrotron Radiat. 1998, 5, 1222. Métaux Céramiques Systèmes Engineering, www.mcse.fr. Onderwaater, W. G.; van der Tuijn, P. C.; Mom, R. V.; van Spronsen, M. A.; Roobol, S. B.; Saedi, A.; Drnec, J.; Isern, H.; Carla, F.; Dufrane, T.; Koehler, R.; Crama, B.; Groot, I. M. N.; Felici, R.; Frenken, J. W. M. Rev. Sci. Instrum. 2016, 87, 113705. Aspnes, D. E. Surf. Sci. 1994, 307–309, 1017. Borensztein, Y. Physica A 1994, 207, 293. Rotermund, H. H. Surf. Sci. Rep. 1997, 29, 265. Rotermund, H. H. Surf. Sci. 2009, 603, 1662. Dicke, J.; Erichsen, P.; Wolff, J.; Rotermund, H. H. Surf. Sci. 2000, 462, 90. Leiden Probe Microscopy B.V., http://www.leidenprobemicroscopy.com/. Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M. Surf. Sci. 2004, 552, 229. Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M. Catal. Today 2005, 105, 234. Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M. Top. Catal. 2005, 36, 43. Zheng, G.; Altman, E. I. Surf. Sci. 2002, 504, 253. Hendriksen, B. L. M.; Ackermann, M. D.; van Rijn, R.; Stoltz, D.; Popa, I.; Balmes, O.; Resta, A.; Wermeille, D.; Felici, R.; Ferrer, S.; Frenken, J. W. M. Nat. Chem. 2010, 2, 730. Mars, P.; van Krevelen, D. W. Chem. Eng. Sci. 1954, 3, 41. van Rijn, R.; Balmes, O.; Resta, A.; Wermeille, D.; Westerstrom, R.; Gustafson, J.; Felici, R.; Lundgren, E.; Frenken, J. W. M. Phys. Chem. Chem. Phys. 2011, 13, 13167. Lundgren, E.; Gustafson, J.; Mikkelsen, A.; Andersen, J. N.; Stierle, A.; Dosch, H.; Todorova, M.; Rogal, J.; Reuter, K.; Scheffler, M. Phys. Rev. Lett. 2004, 92, 046101. Gustafson, J.; Shipilin, M.; Zhang, C.; Stierle, A.; Hejral, U.; Ruett, U.; Gutowski, O.; Carlsson, P. A.; Skoglundh, M.; Lundgren, E. Science 2014, 343, 758. Roobol, S. B. The Structure of a Working Catalyst. PhD Thesis, Leiden University, 2014. Onderwaater, W. G. CO Oxidation Catalysis at Multiple Length Scales. PhD Thesis, Leiden University, 2016. Jensen, R.; Andersen, T.; Nierhoff, A.; Pedersen, T.; Hansen, O.; Dahl, S.; Chorkendorff, I. Phys. Chem. Chem. Phys. 2013, 15, 2698. Gustafson, J.; Mikkelsen, A.; Borg, M.; Lundgren, E.; Köhler, L.; Kresse, G.; Schmid, M.; Varga, P.; Yuhara, J.; Torrelles, X.; Quirós, C.; Andersen, J. N. Phys. Rev. Lett. 2004, 92, 126102. Gustafson, J.; Westerström, R.; Resta, A.; Mikkelsen, A.; Andersen, J. N.; Balmes, O.; Torrelles, X.; Schmid, M.; Varga, P.; Hammer, B.; Kresse, G.; Baddeley, C. J.; Lundgren, E. Catal. Today 2009, 145, 227. Heck, R. M.; Farrauto, R. J. Appl. Catal., A 2001, 221, 443. Shelef, M. Catal. Rev. Sci. Eng. 1975, 11, 1. Zeldovich, J. Acta Physicochim. USSR 1946, 21, 577. Heck, R. M.; Farrauto, R. J. Automotive Engineering 1996, 93–96. Herbschleb, C. T.; Bobaru, S. C.; Frenken, J. W. M. Catal. Today 2010, 154, 61. van Spronsen, M. A.; van Baarle, G. J. C.; Herbschleb, C. T.; Frenken, J. W. M.; Groot, I. M. N. Catal. Today 2015, 244, 85. Roobol, S. B.; Onderwaater, W. G.; van Spronsen, M. A.; Carla, F.; Balmes, O.; Navarro, V.; Vendelbo, S.; Kooyman, P. J.; Elkjaer, C. F.; Helveg, S.; Felici, R.; Frenken, J. W. M.; Groot, I. M. N. Phys. Chem. Chem. Phys. 2017, 19, 8485. Nieuwenhuys, B. E. Adv. Catal. 1999, 44, 259. Heck, R. M.; Farrauto, R. J. Catalytic Air Pollution Control, Wiley: New York, 2002. Robinson, I. K.; Eng, P. J.; Romainczyk, C.; Kern, K. Phys. Rev. B 1993, 47, 1070. Thostrup, P.; Christoffersen, E.; Lorensen, H. T.; Jacobsen, K. W.; Besenbacher, F.; Norskov, J. K. Phys. Rev. Lett. 2001, 87, 126102. Hammer, B. J. Catal. 2001, 199, 171. Hanesch, P.; Bertel, E. Phys. Rev. Lett. 1997, 79, 1523. Grossmann, A.; Erley, W.; Ibach, H. Surf. Sci. 1994, 313, 209. Sander, D.; Linke, U.; Ibach, H. Surf. Sci. 1992, 272, 318. Eurochlor, Chlorine Industry Review (2007–2008). Deacon, H. U.S. Patent 85370 (1868). Deacon, H. U.S. Patent 165802 (1875). Iwanaga, K.; Seki, K.; Hibi, T.; Issoh, K.; Suzuta, T.; Nakada, M.; Mori, Y.; Abe, T. Sumitomo Kagaku 2004. 2004-I. Johnson, A. J.; Cherniavsky, A. J. U.S. Patent 2542961 (1951). Company, S. O. U.S. Patent 3210158 (1965). Kiyoura, T.; Fujimoto, N.; Ajioka, M.; Suzuki, T.; Kogure, Y.; Kanaya, K.; Nagayama, T. European Patent EP184413A (1984). Itoh, H.; Kono, Y.; Ajioka, M.; Takenaka, S.; Kataita, M. U.S. Patent 4803065 (1989). Kiyoura, T.; Kogure, Y.; Nagayama, T.; Kanaya, K. U.S. Patent 4822589 (1989). Hibi, T.; Nishida, H.; Abekawa, H. U.S. Patent 5871707 (1999). Wolf, A.; Mleczko, L.; Schlüter, O. F.; Schubert, S. European Patent EP2026905 (2006). Wolf, A.; Kintrup, J.; Schlüter, O. F.; Mleczko, L. European Patent EP2027062 (2006). Wolf, A.; Mleczko, L.; Schubert, S.; Schlüter, O. F. European Patent EP2027063 (2006). Zweidinger, S.; Hofmann, J. P.; Balmes, O.; Lundgren, E.; Over, H. J. Catal. 2010, 272, 169. Hofmann, J. P.; Zweidinger, S.; Seitsonen, A. P.; Farkas, A.; Knapp, M.; Balmes, O.; Lundgren, E.; Andersen, J. N.; Over, H. Phys. Chem. Chem. Phys. 2010, 12, 15358. Hofmann, J. P. Structural Dynamics of Chlorinated Ruthenium Dioxide Model Catalysts Under Reaction Conditions. PhD Thesis, JLU Giessen, 2009. Zweidinger, S.; Crihan, D.; Knapp, M.; Hofmann, J. P.; Seitsonen, A. P.; Weststrate, C. J.; Lundgren, E.; Andersen, J. N.; Over, H. J. Phys. Chem. C 2008, 112, 9966. López, N.; Gómez-Segura, J.; Marín, R. P.; Pérez-Ramírez, J. J. Catal. 2008, 255, 29.