Atomic-Scale Observations of Catalyst Structures under Reaction

Mar 9, 2016 - School of Engineering of Matter, Transport and Energy, Arizona State University, Tempe, Arizona 85287, United States ... P. Wigner Fello...
1 downloads 13 Views 83MB Size
Review pubs.acs.org/CR

Atomic-Scale Observations of Catalyst Structures under Reaction Conditions and during Catalysis Franklin (Feng) Tao*,†,‡ and Peter A. Crozier*,§ †

Department of Chemical and Petroleum Engineering, University of Kansas, Lawrence, Kansas 66045, United States Department of Chemistry, University of Kansas, Lawrence, Kansas 66045, United States § School of Engineering of Matter, Transport and Energy, Arizona State University, Tempe, Arizona 85287, United States ‡

ABSTRACT: Heterogeneous catalysis is a chemical process performed at a solid−gas or solid−liquid interface. Direct participation of catalyst atoms in this chemical process determines the significance of the surface structure of a catalyst in a fundamental understanding of such a chemical process at a molecular level. High-pressure scanning tunneling microscopy (HP-STM) and environmental transmission electron microscopy (ETEM) have been used to observe catalyst structure in the last few decades. In this review, instrumentation for the two in situ/operando techniques and scientific findings on catalyst structures under reaction conditions and during catalysis are discussed with the following objectives: (1) to present the fundamental aspects of in situ/operando studies of catalysts; (2) to interpret the observed restructurings of catalyst and evolution of catalyst structures; (3) to explore how HP-STM and ETEM can be synergistically used to reveal structural details under reaction conditions and during catalysis; and (4) to discuss the future challenges and prospects of atomic-scale observation of catalysts in understanding of heterogeneous catalysis. This Review focuses on the development of HP-STM and ETEM, the in situ/operando characterizations of catalyst structures with them, and the integration of the two structural analytical techniques for fundamentally understanding catalysis.

CONTENTS 1. Introduction 1.1. Model catalysts with low surface area versus nanoparticle catalysts with high surface area 1.2. Atomic resolution in situ/operando techniques characterizing catalyst structures: HPSTM and ETEM 1.3. Organization of this Review 2. Development of HP-STM and ETEM techniques 2.1. Development of high-pressure scanning tunneling microscopy 2.1.1. Overview of scanning tunneling microscopy 2.1.2. Creating a gaseous environment around a catalyst for in situ/operando studies using HP-STM 2.1.3. Controlling temperature of piezoelectric scanning tube and coarse approaching motor 2.1.4. Simultaneous qualitative and quantitative analysis of reactants and products 2.1.5. Instrumentation for visualization of surfaces under reaction conditions or during catalysis using high-pressure scanning tunneling microscopy 2.2. Development of environmental transmission electron microscopy 2.2.1. Creating reactive gas environments

© 2016 American Chemical Society

2.2.2. Environmental scanning transmission electron microscopy 2.2.3. Instrumentation for in situ characterization with other stimuli 2.2.4. Instrument-related detrimental effects during imaging under reactive gas conditions 3. New structures of catalyst surfaces revealed with HP-STM under reaction conditions or during catalysis 3.1. Pressure-dependent binding geometry of chemisorbed molecules on catalyst surfaces 3.1.1. Evolution from site-specific adsorption to non-site-specific adsorption with increase of pressure of reactant gas 3.1.2. Pressure-dependent change of site-specific adsorption of reactant molecules on model metal catalysts 3.2. Pressure-dependent restructuring of catalyst surfaces 3.2.1. Hex-Pt(100) in CO 3.2.2. Pt(110) in CO 3.2.3. Pt(557) and Pt(332) in CO 3.2.4. Cu(110) in H2 3.2.5. Pt(557) in O2 3.2.6. Ni(557) in CO

3488 3489

3490 3493 3493 3493 3493

3493

3494 3494

3494 3495 3496

3497 3498

3498

3499 3499

3499

3500 3502 3502 3503 3504 3507 3507 3508

Received: May 18, 2014 Published: March 9, 2016 3487

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews 4. Operando studies of catalyst surfaces during catalysis using HP-STM 5. Atomic-scale characterization of structural and chemical evolution in nanocatalysts with ETEM under reaction conditions 5.1. Catalyst preparation 5.2. Catalyst activation 5.3. Structural evolution in gas reactants 5.3.1. Supported metal nanoparticles 5.3.2. Reducible oxides 5.4. Deactivation and regeneration of catalysts 5.4.1. Metal nanoparticle sintering 5.4.2. Other deactivation mechanisms 5.4.3. Catalyst regeneration 5.5. In situ studies of catalyst under other reaction conditions and stimuli 5.5.1. Photocatalysis 5.5.2. Catalysis in a liquid 6. Operando studies of catalysts using ETEM 6.1. In situ studies of catalytic growth of nanomaterials 6.2. Gas-phase reactions 7. Integration of HP-STM and ETEM 7.1. Synergistic exploration of restructuring of metal catalyst at nanoscale 7.2. Determining the arrangements of atoms on the topmost atomic layer of a catalyst and adsorbates with ETEM and HP-STM 7.3. Packing and density of oxygen vacancies on topmost surface of reducible oxide 7.4. Understanding structural evolution of Cu nanoparticle supported on ZnO 7.5. Fundamental studies of bimetallic catalysts 8. Future challenges and prospects 8.1. Thermal drift 8.2. Achieving high resolution of images of catalyst surface at high temperature with HP-STM 8.3. Increasing pressure of reactant gases during in situ/operando studies 8.4. Fast and ultrafast imaging of catalyst 8.5. Chemical identity of adsorbates 8.6. Catalysts in a liquid phase 8.7. Correlative approaches Author Information Corresponding Authors Notes Biographies Acknowledgments References

Review

along with changes in pressure of reactants and temperature of a catalyst. To establish an intrinsic correlation between structure of a catalyst and its corresponding catalytic performance, it is crucial to characterize the surface (Figure 1c and d) and bulk structures (Figure 1a and b) of the catalyst at the atomic level under reaction conditions or during catalysis. As schematically shown in Figure 2, a catalytic reaction on a heterogeneous catalyst takes place at the interface between a catalyst particle and a mixture of reactants A and B. The tremendous effort of the surface science community in the last few decades revealed catalytic events that occur on a catalytic site typically consisting of a few atoms of the catalyst surface.1,2 The geometric and electronic structures of the catalytic site determine its catalytic performance to a large extent. A catalytically active surface may only be formed under a specific reaction condition or during catalysis and may not be stable either in air or under highvacuum conditions. Here the “reaction condition” is defined to be the condition where a catalyst is in the gas (or liquid) phase of a reactant or a gas phase of a pretreatment gas; “during catalysis” is defined to be the status in which a catalyst is in a gas (or liquid) phase of all reactants of a catalytic reaction. Thus, these studies in a gas or liquid phase are essential to mechanistically explore the surface processes involved in the catalytic reaction including adsorption, dissociation, and desorption of reactant molecules on an active catalyst, as well as coupling of intermediates. A simple thermodynamic argument can illustrate why, in general, the catalyst surface structure could change in the presence of reactants. The chemical potential of the gas phase around the catalyst can be expressed as a function of pressure, p, p total + kAT ln po where E is the standard by μA(gas)(T , p) = EA(gas)

3509

3510 3510 3512 3512 3512 3516 3517 3517 3519 3520 3520 3520 3521 3522 3522 3522 3524 3525

3526 3526

chemical potential. The chemical potential of the surface of a catalyst can be described as μcatalyst surface(T,p) = NMμM + NAμA(adsorbate), in which NM and NA are the number of exposed catalyst atoms on the surface (M = element) and the number of chemisorbed reactant molecules A, respectively, and μM and μB are the chemical potentials of the catalyst atoms M and adsorbed molecules A on surface, respectively. A change of pressure of the gas phase around the catalyst will change the chemical potential of the gas phase around the catalyst. A possible consequence of the change of pressure of the gas phase around the catalyst is a likely change of chemical potential of the surface because the gas phase around the catalyst is the source of the chemisorbed molecules or dissociated species of the reactant molecules. The likely change of the chemical potential of the catalyst surface can be realized by changing one or more of its four factors: NM, μM, NA, and μA. For example, the increase of pressure of reactant gas could increase the coverage of adsorbates, by which NA is increased. At a higher coverage, binding configuration of adsorbates could be different and thus μA could be varied. If the catalyst surface is restructured along with the increase of pressure of the reactant, the exposed number of catalyst atoms could be different and thus NM could be changed; the coordination environment of catalyst atoms of the catalyst surface upon a restructuring could be changed and thus μM could be very different. For example, when the pressure of CO is increased from 10−7 to 0.1−1 Torr,3 the step edges of Pt(557) surface restructures to a surface fully covered with nanoclusters with a size of ∼2 nm (Figure 3c). Compared to the original surface in CO at a low pressure (Figure 3a and b), the coverage of CO is increased nearly to 100% (Figure 3d); the binding configuration of CO is switched from bridge and on-top bindings to pure on-top binding.3 The formation of nanocluster surfaces

3526 3528 3529 3529

3529 3529 3530 3530 3530 3530 3531 3531 3531 3531 3531 3531

1. INTRODUCTION Heterogeneous catalysis is a chemical process performed at a solid−gas or a solid−liquid interface.1,2 The function of a catalyst is to direct a chemical reaction through a new pathway with a lower activation energy barrier. Heterogeneous catalysis is the cornerstone of many chemical and energy transformation processes in industries. The surface of a catalyst functions at a solid−gas or solid−liquid interface in a kinetically favorable pathway. Figure 1 schematically shows the structures of the bulk and surface of a catalyst nanoparticle loaded on a support. During catalysis, the structure of a catalyst may vary in a complex way 3488

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

the active catalyst sites are only located on the surface. Thus, in situ/operando studies of the evolution of the topmost subsurface, deep layers, and bulk structures are crucial in developing a fundamental understanding of structure−reactivity relations. Parts c and d of Figure 1 present the diverse packings of catalyst atoms of the topmost surface layer of a catalyst nanoparticle. It is extremely challenging to identify atomic lateral packing of the topmost surface of catalyst nanoparticles with a size of 1−10 nm (Figure 1d). To achieve a fundamental understanding of the interaction of molecules with the catalyst surface at the atomic level, model catalysts are often investigated. These model catalysts may be single crystals of metal, bimetallics, oxide, carbide, or sulfide with specific crystallographic faces or nanoclusters such as CeO2 nanoclusters (1−10 nm) supported on a flat crystal surface such as Au(111).30 Surface structures of such model catalysts are usually well-defined. These model catalysts eliminate the complexity of variations in particle size, shape, and irregular defect structures of high-surface-area systems. Figure 4 illustrates how single crystal model catalysts can be used to simulate the surface of a high-surface-area nanoparticle catalyst and to understand the potential structural evolution of its surface structure. A high-surface-area nanoparticle catalyst of metal M (Figure 4a) typically consists of nanoparticles with different sizes and shapes and even different compositions (e.g., if it is a bimetallic catalyst). One feature of nanoparticles is the large atomic fraction of undercoordinated atoms of the edges of a nanoparticle among all surface atoms of the nanoparticle. For example, atoms of the edges of a cuboctahedral metal nanoparticle (Figure 4c) have 7 nearest neighbors (NN). Here the number of nearest neighbors is equivalent to the coordination number (CN) of a metal atom M. The fraction of these types of catalyst atoms (NN = 7) strongly depends on the size of the nanoparticle (Figure 4d).31 Direct visualization of the packing of metal atoms of step edges and corners of such a nanoparticle with either high-pressure scanning tunneling microscopy (HP-STM) or transmission electron microscopy (TEM) is extremely challenging. An appropriate approach is to use a stepped surface of a metal single crystal catalyst (Figure 4e) as a model catalyst and visualize this vicinal surface with STM. For example, a stepped surface, Pt(557), has a large fraction (16.7%) of undercoordinated atoms (CN = 7) at its step edge; the step edge of Pt(557) can experimentally simulate the undercoordinated atoms at the edges of a catalyst nanoparticle (Figure 4a). Although usually only atoms on the topmost layers of the catalyst surface directly participate in a catalytic reaction, the subsurface and bulk are crucial for the formation of the topmost surface of a catalyst particle. In many cases, to form an active surface, a pretreatment must be performed during which a catalyst particle is typically oxidized, reduced, and partially or completely transformed to a new phase. For example, a reduction of catalyst precursor anchored on support like Al2O3 in H2 is typically necessary for the formation of a metallic catalyst. The active phase of a catalyst for Fischer−Tropsch synthesis, FexCy, is formed through the reduction of Fe2O3 at high temperature in the mixture of CO and H2.32 The carburization of Fe2O3 is a critical process for the formation of the active catalyst surface: iron carbide.33−35 Thus, in many cases a catalytically active surface is generated through significant involvement of subsurface and bulk changes. From this point of view, although model catalysts at different pressures and temperatures can approximate the surface of high-surface-area nanoparticle catalyst under

increases the fraction of undercoordinated Pt atoms (Figure 3e and f). Whether there is a change of catalyst surface structure along with the increase of pressure is largely related to the original surface structure of the catalyst before the pressure is increased.4,5 In situ and operando studies are two terms that are used frequently in the literature of catalysis and surface science. In this article, we focus mostly on the structure of the surface, subsurface, and bulk of catalyst particles in a gas phase of one or more reactants. For this article both in situ and operando studies mean the charactierzation of catalyst surface or bulk while the catalyst is in a gas (or liquid) phase of one or more reactants. For operando studies, catalytic activity or reaction rates are measured as well. A number of in situ/operando techniques have been used to track catalysts and their adsorbate or even intermediates in a gas phase of one reactant or a mixture of all reactants, although the catalyst surface chemistry in a liquid has not been explored. They can be categorized into two groups: (1) photon-in−photon-out and (2) electron-based analytical techniques. The first category includes the well-known spectroscopy techniques, such as X-ray absorption spectroscopy (XAS),6−11 Raman,12 infrared (IR) spectroscopy, 1 3 , 1 4 and sum frequency generation (SFG).15−23,2,24,25 In contrast to the photon-in−photon-out techniques, the application of electron-based surface analytic techniques to in situ/operando studies of catalysts has been quite challenging because scattering of electrons by molecules in the gaseous environment makes generation of electron beams of a constant energy or the collection of elastically scattered electrons difficult. However, in situ/operando studies using these electronbased techniques have been developed through sophisticated instrumentation in the past decade. One example is the ambientpressure photoelectron spectroscopy,26−29 in which elastically scattered photoelectrons traveling through a gas phase, with a thickness of the inelastic mean free path of phototelectrons or thinner, can be collected and their kinetic energies can be measured.26,27 It is noted that the sampling lateral dimension of these surface analytical techniques is typically in the range of tens of micrometers to a few millimeters. This average chemical information on surface or bulk of a catalyst is significant for understanding catalytic mechanisms because they well represent the structural and chemical information at large sampling lateral dimension. Other than these pieces of information at large lateral dimensions, local structural information at the atomic scale is crucial for a fundamental understanding of catalysis at a molecular level because a catalytic event is performed on a catalytic site consisting of one or several atoms. Thus, as schematically shown in Figure 1, in situ or operando studies of local structures of different sections of a catalyst particle at the atomic scale, including surface, subsurface, and bulk of a catalyst particle, are necessary. 1.1. Model catalysts with low surface area versus nanoparticle catalysts with high surface area

Because the active surface could be generated from a complicated structural evolution starting from catalyst precursors, information on the evolution of the subsurface and bulk of a catalyst particle during pretreatment is critical for understanding how catalytically active sites are formed. Figure 1b schematically illustrates a catalyst nanoparticle consisting of surface, subsurface, and bulk, which are closely correlated in generation of catalytic sites through structural evolution of a catalyst particle, although 3489

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 1. Schematics of structure of a catalyst nanoparticle. (a) Catalyst nanoparticle (gray cube) loaded on a support (light blue plate). (b) Surface (gray), subsurface (yellow), and bulk (green) of a catalyst nanoparticle supported on a support; the characterization techniques of structures of different sections are marked on the schematic. Scanning tunneling microscopy (STM) is the technique to visualize the topmost surface of a catalyst. Environmental transmission electron microscopy (ETEM) provides projected structural and chemical information on surface, subsurface, and bulk. (c) Schematic of a topmost surface of a catalyst nanoparticle consisting of different types of catalytic sites that can be identified with scanning tunneling microscopy because STM can visualize both the lateral and vertical arrangements of atoms on the topmost surface of a catalyst. (d) Schematic of arrangements of catalyst atoms of the topmost surface of a nanoparticle.

Figure 2. Schematics showing (a) a catalyst surface with low coverage of adsorbates in a mixture of reactants A and B with a low pressure p1 and (b) a catalyst surface with high coverage of adsorbates in a mixture of reactants A and B with a high pressure p2.

1.2. Atomic resolution in situ/operando techniques characterizing catalyst structures: HP-STM and ETEM

reaction conditions, it is certainly necessary to study subsurface and bulk structures of nanoparticles of a high-surface-area catalyst such as metal nanoparticles loaded on an inert support (Figure 4a). Environmental transmission electron microscopy (ETEM) in both broad-beam and focused-beam mode [scanning transmission electron microscopy (STEM)] has played a key role in elucidating the structures of high-surface-area nanoparticle catalysts. Atomic resolution imaging, spectroscopy, and diffraction have proven to be a powerful combination for understanding structures of catalyst nanoparticles.

A catalytic event is performed on a catalytic site consisting of one or more catalyst atoms or vacancies. Such a site is typically located on the topmost surface of a catalyst nanoparticle. Thus, the ultimate goal of in situ/operando studies of a catalyst is to identify arrangement of catalyst atoms, adsorbates, or even intermediates on the topmost surface and correlated this structural information with its corresponding catalytic performance. To establish structure−catalytic performance correlations to achieve a fundamental understanding of catalytic reaction mechanisms at a molecular level, the electronic and geometric 3490

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 3. STM images of Pt(557) in different pressures of CO at 25 °C. (a) STM image of Pt(557) in ultrahigh vacuum (UHV); (b) structural model of the Pt(557); (c) surface of Pt nanoclusters formed from Pt(557) surface in 0.1 Torr CO at 25 °C. (d) CO coverage on Pt(557) in CO at different pressures of CO gas at 25 °C. (e) Structural model of the formed surface consisting of Pt nanoparticles. (f) Surface structure of Pt nanoclusters with adsorbed CO optimized with density functional theory (DFT) calculations. Reproduced with permission from ref 3. Copyright 2010 AAAS.

Figure 4. Illustration of how a low-surface-area model catalyst can be used to experimentally simulate the surface of a high-surface-area nanoparticle catalyst. (a) TEM images of high-surface-area nanoparticle catalyst, Pt nanoparticles supported on γ-Al2O3 consisting of nanoparticles with different sizes and shapes. (b) Schematic of a cuboctahedral nanoparticle. (c) Packing of catalyst atoms of the topmost surface of an cuboctahedral nanoparticle; “nn” stands for nearest neighboring atoms. (d) Dependence of fractions of atoms with coordination numbers (CNs) of 4, 7, and 9 among all surface atoms on size of the octa−octahedral nanoparticles; CN stands for coordination number, which is equivalent to nn in (c). (e) Surface structure of (557) surface consisting of terraces with (111) surface structure and step with (100). (f) Illustration of the rows of atoms on and below a step edge. Reproduced with permission from ref 31. Copyright 1969 Elsevier, Inc.

restructuring of Pt(557) in gas of CO and Rh(110)-1 × 2 during CO oxidation in the mixture of CO and O2.3,36,37 As shown in Figure 5, the restructuring of Rh(110)-2 × 1 to Rh(110)-1 × 1 during CO oxidation in the mixture of CO and O2 was clearly observed with HP-STM. Compared to the approach of studying model catalyst surfaces with HP-STM, at present ETEM cannot easily image (1) adsorbates on surface of a catalyst particle, the lateral arrangement of catalyst atoms on different terraces of the topmost surface of the catalyst particle, and (2) the packing of atoms at step edges (Figure 1d). From this point of view, the approach of studying catalysts with HP-STM is complementary for ETEM in identifying diverse atomic

structures of species at surfaces, subsurface sites, and interfaces of catalysts must be determined at an atomic level. Currently, highpressure scanning tunneling microscopy (HP-STM) and environmental transmission electron microscopy (ETEM) are the two main in situ/operando analytical techniques identifying local catalyst structure. In particular, they provide in situ structural information on catalysts at the atomic scale under reaction conditions and during catalysis. HP-STM has been used to observe the lateral and vertical arrangements of catalyst atoms and adsorbates of the topmost surface of a model catalyst as shown in Figure 1c and d. For example, HP-STM has successfully identified the surface 3491

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 5. Restructuring of Rh(110)-1 × 2 during CO oxidation in the mixture of 0.08 Torr CO and 0.02 Torr O2. (a) STM of the topmost surface of nominal Rh(110)-1 × 2 in the mixture of 0.08 Torr CO and 0.02 Torr O2 at t = to; the surface consists of restructured Rh(110)-1 × 2 and Rh(110)-1 × 1 formed during CO oxidation. (b) STM images of the topmost surface of nominal Rh(110)-1 × 2 at t = to + 4 min. (c) Structural model of the atom arrangements of Rh(110)-1 × 2. Gray, green, and blue balls are atoms of the first, second, and third layers, respectively. (d) Structural model of the atoms of Rh(110)-1 × 1. Gray and green balls are atoms of the first and second layers, respectively.

Table 1. Complementary Functions of HP-STM and ETEM in In Situ Studies of Heterogeneous Catalysts entry

category

1

function in identification of catalyst structure

2

state of a catalyst

3

capability in identification of adsorbate/ dissociated species of reactant molecules lateral resolution vertical resolution surface sensitivity information depth time per image of 10 nm ×10 nm potential damage to topmost surface of a catalyst by electron beam potential damage to subsurface and bulk of a catalyst by electron beam influence of electron beam on adsorbates of reactants on the topmost surface of a catalyst

4 5 6 7 8 9 10 11

HP-STM

ETEM

lateral and vertical packings of catalyst atoms of the topmost surface at atomic level and location and size of adsorbates on a catalyst surface nanoclusters deposited on flat surface, single crystal model catalyst yes

projected atomic columns of subsurface and bulk and chemical information on subsurface and bulk regions at atomic scale high-surface-area nanoparticle catalyst with a thickness of a few to a few hundred nanometers possible but very challenging

0.1 Å 0.1 Å highest 1−2 atomic layers at least milliseconds no

0.1−1 Å 0.1−1 Å no subsurface and bulk nanosecond to seconds yes

no

possible

no

beam-induced desorption of adsorbates or dissociation of adsorbed molecules

terms of characterizing different aspects of catalyst structures. Although ETEM is not normally considered as a traditional surface-sensitive technique, it provides two-dimensional images representing projections of the three-dimensional atomic arrangement of the sample.40 A modern TEM has a large number of different imaging, diffraction, and spectral analyses modes, permitting detailed structural information to be obtained in both real and reciprocal spaces. With the recent development of aberration correction, much of this information can now be obtained at a sub-Ångstrom resolution.41−45 Tomographic methods have also been developed that allow three-dimensional imaging of nanostructures including catalysts, although its resolution is still far from atomic scale.46−49 Moreover, the availability of energy-dispersive X-ray spectroscopy and electron energy-loss spectroscopy allows for tracking chemical informa-

arrangements of the topmost surface of a catalyst including lateral arrangement of catalyst atoms on terraces and at the steps and adsorbates on the surface38,39 (Figure 1d). Compared to ETEM, HP-STM cannot imagine atoms of the subsurface and bulk of a model catalyst or a nanoparticle catalyst. In contrast to HP-STM, ETEM can provide structural and chemical information on subsurface and bulk of a catalyst nanoparticle in projection. Thus, ETEM is complementary for HP-STM in identifying chemical and structural information on a catalyst nanoparticle at an atomic scale. HP-STM can be used to study single crystal model catalysts and nanoclusters supported on a flat substrate. ETEM is mainly used to study high-surface-area nanoparticle catalysts. Although HP-STM and ETEM cannot study the same sample in most cases, HP-STM and ETEM are complementary in many ways in 3492

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 6. Chamber-in-chamber design of a high-pressure scanning tunneling microscope. (a) CAD drawing of a HP-STM system showing that the reaction cell is hung in the UHV chamber. (b) CAD drawing of a reaction cell. (c) Photo of a reactor with incanted STM body hung in UHV chamber. (d) STM body integrated on the lip of the reactor cell. Reproduced with permission from ref 28. Copyright 2012 Royal Society of Chemistry.

tion at an atomic scale. 50,51 Table 1 summarizes the complementary function of ETEM and HP-STM in in situ/ operando studies of heterogeneous catalysts.

scanning tube to correct a sample-tip distance at different locations of a model catalyst surface can be used to generate different contrasts in the STM image; mapping of the contrasts at different locations of the surface gives the corrugation of surface atoms. In the constant-height mode, the tip always remains at the same height. The different distances between tip and surface atoms give different tunneling currents. A plot of the tunneling currents at different locations gives the corrugation of surface atoms. In the constant-height mode, the contribution of local density of states to the tunneling current is much smaller. Thus, the local density of states is better reflected in the constantcurrent mode. STM is a powerful approach to study geometric structure and electronic states of catalyst atoms of the topmost surface and adsorbate on the surface. It has been widely used in most fields of physics, chemistry, and materials science66−68 2.1.2. Creating a gaseous environment around a catalyst for in situ/operando studies using HP-STM. A simple approach to perform in situ studies of catalyst surfaces used in early literature69 is a direct heating of a catalyst in a gasfilled UHV chamber where the STM parts are located. This simple approach has several disadvantages. During heating of a catalyst, the STM chamber wall receives heat from the hot catalyst through thermal conduction of reactant gas around the catalyst. The heated chamber wall desorbs its adsorbed molecules such as H2O and CO. Unfortunately, these molecules contaminate the catalyst surface significantly. The huge heating capacity of the gas filled in the UHV chamber limits the highest temperature to which the catalyst can be heated because all gas in the chamber is heated. In addition, it is almost impossible to perform catalysis in a flow mode because the volume of the reactant gas (5−15 L) around the catalyst is large. As the reactant gases with a large volume remain in the chamber, the diffusion of products is limited; consequently, product molecules likely readsorb on the surface. In addition, the gas between the STM parts (scanning tube, coarse approach motor, and other parts) and the UHV chamber will transduce low-frequency mechanical vibrations from the UHV chamber or pumps to the STM parts as high-pressure gas largely weakens the function of damping parts of the STM head (designed for UHV). An alternative method is the isolation of the gas environment of reactant gas(es) and/or product(s) from the UHV environment. A chamber-in-chamber design is one type of instrumentation for keeping gas in an internal cell without loss of STM function.70,71 The internal cell has a quite small size. It is called a reactor in some cases. Typically, it is hung in a UHV chamber (Figure 6a) and the gas volume of the cell (Figure 6b) is a few milliliters to a few tens of milliliters. The reaction cell consists of a sample stage (Figure 6d), gas inlet and outlet ports, a window for receiving radiation for heating the sample, and STM head

1.3. Organization of this Review

This Review first discusses the importance of studying catalyst structures in gas and liquid phase of reactants. In section 2 the development of HP-STM and ETEM techniques and the latest advances of instrumentation are briefly described. The surface structures of model catalysts under reaction conditions and during catalysis studied with HP-STM are reviewed in sections 3 and 4. Sections 5 and 6 discuss the in situ/operando studies of high-surface-area catalysts in chemical processes including activation, catalysis, and deactivation. Integration of HP-STM and ETEM for understanding of catalyst structures and mechanisms is discussed in section 7. Section 8 looks at the future challenges and prospects of HP-STM and ETEM.

2. DEVELOPMENT OF HP-STM AND ETEM TECHNIQUES 2.1. Development of high-pressure scanning tunneling microscopy

2.1.1. Overview of scanning tunneling microscopy. The first STM instrument was designed in 1981 by Binnig and Rohrer.52−60 In classical mechanics, electrons cannot cross an energy barrier if their total energy is less than the barrier height. However, in quantum mechanics, there is a finite probability that an electron can “tunnel” through a potential barrier. With the STM technique, tunneling happens between a conducting tip and the surface of a sample.61−64 The tunneling current is determined by a few terms including bias applied to tip or sample, tip−sample distance z, the local density of states ρs(k, E) of the sample and tip, and other factors.65 It can be expressed with a simplified formula: I ∝ Vρs(0, EF) e−2kz. As the tunneling current is exponentially dependent on the tip−surface distance, z, the corrugation of surface atoms can be tracked with a resolution of 0.1 Å or even higher. In addition, the tunneling current is proportional to the local density of states of the sample surface. STM can be used to identify differences in local density of states of different atoms, which is typically accomplished through a technique called scanning tunneling spectroscopy. Thus, atoms of different elements of a catalyst surface or of adsorbates can be readily distinguished, although it does not give the identity of each element of the surface atoms or adsorbates. STM can be operated in either constant-current or constantheight mode. In the constant-current mode, a comparison of the measured tunneling current to the set current generates a feedback to the Z-component of the scanning tube so that the tip−sample distance can be precisely tuned to keep the instant tunneling current at a set value. The applied voltage to the piezo 3493

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

was installed on the Cu substrate of the STM motor to heat the motor and scanning tube if the temperature is lower than the set cable. This instrumentation preserves a constant temperature of the scanning tube and motor of the STM room in the temperature range of 30−60 °C for collecting atom-resolved images when the sample is at a relatively high temperature of 25− 250 °C in a gaseous environment. 2.1.4. Simultaneous qualitative and quantitative analysis of reactants and products. One of the goals of in situ/operando studies is to establish an intrinsic correlation between surface structures under reaction condition and during catalysis and their corresponding catalytic performance such as activity, selectivity, deactivation, and poisoning resistance. To establish such a correlation, a simultaneous track of the formation of product molecules is necessary.73 A widely used technique for analyses of products is online mass spectrometry in which a small amount of gas in the reactor is leaked into a quadruple mass spectrometer through a capillary quartz tube.28 A simultaneous analysis of products during visualization of surface structure can offer two types of information on products. One is the gas composition in the reaction cell. It can identify whether products are formed at a catalytic condition. Thus, a correlation of the surface structure of a catalyst to the corresponding catalytic performance can be established. Another type of information is to track the evolution of partial pressure of a product as a function of the catalytic temperature or reaction time. By recording temperature-dependent partial pressures of this product and comparing them to their corresponding surface structure, a structure−reactivity correlation can be determined. 2.1.5. Instrumentation for visualization of surfaces under reaction conditions or during catalysis using highpressure scanning tunneling microscopy. Examination of the surface of metal single crystals using STM in the gas phase was started in the early 1990s.69,74 In early tests, gas was admitted to a large UHV chamber of a STM system and the catalyst was imaged in an environment of a reactant gas.69,75 In the last two decades, different designs of cell-type HP-STMs have been reported.70,76 These instrumentations are briefly discussed in the following paragraphs. A HP-STM with a very compact reactor with a volume of only 300 μL, in which only the tip and the sample were introduced, was reported.71 Most parts of the STM body such as the scanning piezoelectric tube were placed in UHV environments.71 Figure 8 is the section view of this HP-STM. A stick−slip motor was used to move the tip to the tunneling region of the sample in this HPSTM.71 A different high-pressure STM was reported in 2005 that can work in the pressure range from UHV to 1 bar.77 Figure 9 shows the CAD drawing of this HP-STM. The STM body, which is in a high-pressure cell with a relatively large volume of 1.5 L, has a high resonance frequency of 2.6 kHz. One of the features of this design is the small voltage required for the piezoelectric tube. By virtue of this, potential gas discharging of gas molecules induced by the voltage applied to the scanning piezoelectric tube was successfully avoided in this design,72,78 although such an electrical charging was hardly reported in in situ studies of HPSTM. Besides the above HP-STM, a chamber-in-chamber, HP-STM was designed by Tao et al. in 2008.78 The entire STM body was housed in the high-pressure reactor with volume for gases ≈ 19 mL. Unlike the previous HP-STMs, the high-pressure reactor of this design was placed in a UHV chamber by a compatible docking scaffold and mounting system (Figure 6a). While scanning, the reaction cell is isolated from the UHV chamber and

consisting of a tip, scanning tube, and coarse approaching motor.72 2.1.3. Controlling temperature of piezoelectric scanning tube and coarse approaching motor. Although the piezoelectric scanning tube and coarse approaching motor of STM can be baked to 120 °C for days, they only function well at a temperature below 70 °C. In addition, the temperature must remain constant for minimizing thermal drift. A higher temperature of the piezoelectric scanning tube and coarse approaching motor will make the preservation of a constant current more challenging, resulting in a larger fluctuation in temperature and a faster thermal drift. Thus, keeping the scanning tube and coarse approaching motor at a constant temperature (in the range of 25−60 °C) is the key for achieving atomic resolution during in situ/operando studies of model catalysts at a relatively high temperature. Separation of the reaction cell from the STM body is a crucial approach in control of temperature of the STM body.72 Figure 7

Figure 7. Schematic of the latest reaction cell HP-STM designed by Tao in collaboration with Specs Surface Nano Analysis. The main feature is the separation of the hot reaction cell and the cold STM body. Reproduced with permission from ref 72. Copyright 2013 American Institute of Physics.

schematically shows this separation. The reaction cell (marked with red in Figure 7) for catalysis at room temperature or a relatively high temperature is separated from the STM body (tip, scanning tube, and coarse approaching motor) at a temperature near room temperature. A shutter with an aperture is installed between the STM body and the reaction cell for tip approaching, scanning, and retracting. This aperture has a minimized diameter (typically 0.5 mm) to limit the flow of warm gas from the reaction cell to the STM room. In addition, a backfilling of gas of room temperature from the end of the STM room can further prevent the hot gas from diffusing to the STM room (Figure 7). Additionally, the capabilities of (1) cooling the STM body through the copper braid connected to both the Cu substrate of the STM body and an external dewar of liquid N2 and (2) heating of the substrate of the scanning tube and coarse approaching motor by the Zener diode both help to preserve a constant temperature of the STM body during scanning.72 The Cu braid will cool the STM body and thus further cool the scanning tube if the temperature of the coarse approaching motor and scanning tube is higher than the set temperature. A miniature Zener diode 3494

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 8. Schematic of structure of a HP-STM with a minimized reactor. (a) Whole reactor-like STM head. (b) Enlargement of (a) to show details of the minimized reactor. Reproduced with permission from ref 71. Copyright 1998 American Institute of Physics.

UHV.72 More importantly, atoms of the graphene layer on Ru(0001) can be identified when the sample is at 230 °C in 25 Torr N2 (Figure 12b). Other than the function of visualizing surface atoms at a relatively high temperature in the gas with a relatively high pressure, another advantage is the fast scanning of the Arahus STM head, which is significant for tracking the dynamics of surface structure of a model catalyst under reaction conditions or during catalysis. Other than the HP-STM used for in situ studies of a catalyst in a gaseous environment, liquid STM was developed by integrating an electrochemical cell with STM. It is mainly used to study the surface of the electrode in the solution of the electrolyte. Another important aspect of all instrumentations of HP-STM is the introduction of pure reactant gases. Impurity gases such as H2S, CO, and nickel carbonyl could largely poison the catalyst surface. Installation of a system of gas purification is necessary for ensuring that any observed surface structure is solely induced by reactant gas(es) instead of any sources of impurity such as a gas tank.

Figure 9. CAD drawing of a high-resolution HP-STM. The preparation chamber (left) is equipped with low-energy electron diffraction (LEED) (light gray), mass spectrometer (blue), ion gun for sputtering (light gray), an XPS system (yellow and reddish parts), and a gas inlet. The STM chamber (right) is pumped by a separate pumping system and can be sealed off by gate valves (green and reddish). The sample is moved between the two chambers by means of a transfer rod (left side of the preparation chamber) and two wobble sticks (red). The entire system rests on four pneumatic legs. Reproduced with permission from ref 77. Copyright 2005 American Institute of Physics.

2.2. Development of environmental transmission electron microscopy

The earliest motivation to control the gas environment inside an electron microscope dates back to efforts in the 1930s to examine samples in their hydrated state79 and for the study and control of contamination.79−82 ETEM was more extensively developed during the 1970s. This earlier history of the development of ETEM will not be given here. Rather, the basic approaches of ETEM will be summarized. A comprehensive review on ETEM and other in situ techniques from that period can be found in the book by Butler and Hale.83 One of the major goals of ETEM is to elucidate the structure− reactivity relations on high-surface-area nanoparticle catalysts. As mentioned before, an in situ/operando characterization using electron-based spectroscopy and microscopy has been challenging. Conventional TEM characterization is either “fresh” or “postmortem” analysis of catalysts in high vacuum. It may provide information that could be either misleading or difficult to interpret unambiguously. Ideally, TEM characterization should be performed under catalytic conditions. Moreover, it would be highly desirable to quantitatively measure the activity and

hung with three springs to minimize the mechanical vibration (Figure 6a). Very recently, a new high-temperature, ambient-pressure STM of catalyst studies was designed by Tao in collaboration with scientists at Specs Surface Nano Analysis using Aarhus STM head for achieving high-resolution structural details of surfaces of model catalysts at a relatively high temperature in a gaseous environment at near-ambient pressure (Figure 10).72 The first one is located in Tao’s group.72 After this successful design,72 one at Brookhaven, one at Florida, one at KAIST, two in China were installed. This design allows for visualization of a catalyst surface at the atomic scale at a temperature up to ∼230 °C and a pressure up to 25 Torr.5 As shown in Figure 11b, atomic details of graphene grown on Ru(0001) were visualized at 700 K in 3495

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 10. Reactor-like HP-STM consisting a reaction cell and a separated STM body. (a) CAD drawing of the whole system designed by Tao et al. (b) Picture of the HP-STM system. (c) Picture of the reaction cell assembled with STM body. Reproduced with permission from ref 72. Copyright 2013 American Institute of Physics.

2.2.1. Creating reactive gas environments. Similar to in situ studies using HP-STM, a local gas environment around a catalyst has to be created while maintaining a high vacuum throughout most of the TEM column and the electron source. There are three approaches that can currently be employed to satisfy this requirement: (a) Window method: gas or liquid is confined around the sample region by using thin electron transparent windows, e.g., thin amorphous carbon or SiN films. (b) Differential pumping: a pressure difference is maintained by installing small apertures above and below the sample area and using additional pumping. (c) Gas-injection system: an injection needle is placed near the sample surface and gas is allowed to flow from the tip of the needle into the sample area. The gas-injection approach has been used extensively in applications where localized chemical vapor depositions (CVDs) are required.85−87 This approach is advantageous because it introduces a relatively small volume of gas into the system and usually does not require extra pumping capacity. One disadvantage of this approach is that the pressure on the sample varies with the distance from the injection point, although it does not vary very much within a distance of several microns. Moreover, the pressure in the vicinity of the sample is only ∼0.1 Torr.87 To achieve information better representing a highsurface-area nanoparticle catalyst, it is often necessary to image catalyst nanoparticles dispersed on different locations of a TEM grid because of statistical variations in the catalyst powder. In the method using windows, thin electron transparent membranes are employed to separate the gaseous environment around the catalyst from the high vacuum of the rest of the microscope. This is usually accomplished via sample holders that incorporate windows along with a gas-inlet system, by which atomic resolution has been demonstrated.88,89 The window design has the advantage of being able to handle gas pressures up to 1 bar or higher, depending upon the strength and thickness of the window. Recent application of microelectromechanical systems (MEMS) technology has resulted in the development of windowed structures where the area of the thin window is rather small, as illustrated in Figure 13. This permits a much higher pressure to be maintained and minimizes the likelihood of membrane failure. The group at Delft first demonstrated atomic resolution of 0.18 nm at a pressure above 1 bar and a temperature of 500 °C.90 Recent results show that controlled-atmosphere experiments with windowed cells can be performed at a pressures

Figure 11. STM images of graphene grown on Ru(0001). Both images were collected when the sample was at 700 K in UHV. (a) 100 Å × 100 Å. The large spots form a Moiré pattern. The area marked with a rectangle box is O (2 × 2) formed at a step edge of Ru(0001). (b) 74 Å × 74 Å. The large, bright feature forms the Moiré pattern. Each bright spot in (b) is one carbon atom of graphene grown on Ru(0001). Reproduced with permission from ref 72. Copyright 2013 American Institute of Physics.

Figure 12. STM images of graphene grown on Ru(0001). These images were taken when sample was at 500 K in 25 Torr N2 gas. (a) Large-scale image of graphene on Ru(0001): 1700 Å × 1700 Å. Bright spots form Moiré pattern. (b) Atom-resolved image of graphene at 500 K in 25 Torr N2 gas. Reproduced with permission from ref 72. Copyright 2013 American Institute of Physics.

selectivity of the catalyst simultaneously so that a structure− catalytic performance relation can be established. In situ studies of the growth and type of nanotubes/wires relate directly to the catalytic activity and selectivity. However, for many gas-phase (or liquid-phase) reactions, the reactant and product gases are not visible with microscopy imaging and diffraction techniques. Recent works suggest that this limitation can be overcome, which has brought ETEM a step closer to a catalytic condition.84 3496

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 13. Schematic diagram of MEMS-based windowed cell reactor. (a) Side view. (b) Schematic of arrays of small holes (black) that are electrontransparent. Reproduced with permission from ref 90. Copyright 2008 Elsevier, Inc.

first atomic-resolution ETEM was accomplished with this approach.108 This approach was continuously developed throughout the 1990s and involved substantial modifications of the objective lens pole pieces.109−112 2.2.2. Environmental scanning transmission electron microscopy. Working in collaboration with scientists at Haldor Topsoe, FEI designed and commercialized a differential pumped system and demonstrated that this system was compatible with a Schottky field emission electron source of a Philips CM 300 FEG-TEM.113 The modified column was shown to permit cell pressures up to 10 mbar without adversely affecting the performance of the field emission gun. These modifications were incorporated into an FEI Tecnai F20 TEM/STEM, and the first commercial system was installed at Arizona State University in 2002.114 A photograph of this system is shown in Figure 14, and a schematic diagram of the differential pumping system is shown in Figure 15. The microscope column has been modified to add three sets of differential pumping apertures for three pumping stages between the gun valve and the viewing chamber. The first two sets of apertures are placed within the upper and lower objective lens pole pieces, and most of the gas leaking through this first set of apertures is pumped out using a turbomolecular pump. The second stage is between the condenser aperture and the selected area aperture and is pumped using a molecular drag pump. An additional ion pump is employed for the last stage before the gun valve. There is a control box to open or close various pneumatic valves, and thus the microscope can be readily switched between environmental mode and highvacuum mode. The gas inlet is controlled by a set of shut-off valves, and the pressure is regulated using a needle valve. The controlled-atmosphere approach can also be employed with the microscope operating in scanning transmission (STEM) mode. The first work on catalysts performed with a field emission environmental transition scanning electron microscope (ESTEM) at high pressures was conducted on CoRu bimetallic particles.115 It reported nanometer-resolution annular dark-field images (ADF) and electron energy loss spectroscopy (EELS) at 400 °C in 1 Torr of H2 and N2. This work was able to take advantage of the enhanced high-spatial-resolution spectroscopy available in a ESTEM as a result of the fine electron probe, which can be focused to a dimension of 10−2 Torr CO. The newly formed incommensurate Moiré lattice orients either along a high-symmetry direction of the substrate or along a high-symmetry direction of the adsorbate overlayer. These pressure-dependent binding sites of CO can be rationalized by taking into account the substrate−adsorbate interaction potential and adsorbate−adsorbate lateral repulsive potential. At a pressure lower than 10−2 Torr, 30° rotation of Moiré lattice by substrate can be understood as the maximum

desire to image catalysts with a resolution of 0.1 nm or better under reaction conditions or during catalysis. Recently a number of quantitative studies show that resolution loss can occur under high dose rates133 and high total beam currents.134 Bright et al. speculated that large beam currents largely lead to an enhancement in gas ionization and the positive ions migrate while the amount of electron scattering is increased.134

3. NEW STRUCTURES OF CATALYST SURFACES REVEALED WITH HP-STM UNDER REACTION CONDITIONS OR DURING CATALYSIS A catalytic event typically includes adsorption or dissociation of reactant molecules, diffusion of reactant molecules on a catalyst surface, migration of catalyst atoms (not always necessarily), and coupling of reactant molecules or its dissociated species to form product molecules on surfaces. Among these surface processes, molecular adsorption is of great importance. A tremendous amount of effort has been devoted in the last three decades to the investigation of the adsorption of molecules on metal or oxide surfaces using STM and other surface science techniques in ultrahigh vacuum through a vacuum surface science approach.1,2 Here the vacuum surface science approach is defined as a study of the surface of a sample in UHV (typically ≤3 × 10−10 Torr) upon both exposing to a reactant gas at a certain pressure (10−6−10−10 Torr) at a certain temperature for a certain time (1−1000 s) and then purging the reactant gas to restore its original pressure (typically ≤3 × 10−10 Torr). In contrast, with the in situ/ operando approach, characterization is performed while the catalyst is in a gaseous phase of reactants at a certain pressure. Numerous publications on atomic details of the catalyst surface and its adsorbates studied with the vacuum surface science approach can be found in the literature.1,2 However, the molecular packing of the adsorbate and the atomic packing of the catalyst surface in gas phase of reactants may be quite different from that of UHV, as discussed in the Introduction section. In the past decade, significant efforts have been made in exploring molecular packing of adsorbate and arrangement of catalysts atoms under reaction conditions and during catalysis. In a reactive gaseous environment, two types of pressuredependent evolutions of surface structure were revealed.3,33,73,105,135−139 One is the pressure-dependent binding sites of chemisorbed molecules on a catalyst surface while the catalyst is in a reactant gas. Another type is the pressuredependent restructuring of a catalyst surface in a gas phase of reactant, in which catalyst atoms on surface are largely reorganized. These pressure-dependent surface reactions of a model catalyst revealed by HP-STM are reviewed in this section. 3.1. Pressure-dependent binding geometry of chemisorbed molecules on catalyst surfaces

In the surface science vacuum approach, surface structures of chemisorbed molecules were identified by exposing a clean catalyst surface to a certain number of molecules of a reactant in a unit of Langmuir (1 × 10−6 Torr·s). After the exposure of a catalyst surface to a gas of reactant with a certain pressure for a certain time, the reactant gas, typically in the range of 10−9−10−6 Torr, is pumped. This approach restores the UHV environment before characterization of the binding geometry of the chemisorbed molecules of the surface of a model catalyst using STM. In most cases, the chemisorbed molecules are packed on one or two specific binding sites of the catalyst surface such as ontop, bridge, or hollow site for a (111) surface of a metal catalyst with a face-centered cubic (fcc) lattice. An image with atomic 3499

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

on a catalyst surface with a site-specific binding configuration different from that at low pressure or in UHV. This section will review the pressure-dependent, site-specific packing structures of reactant molecules. 3.1.2.1. CO on Rh(111). By using STM, a pressure-dependent adsorption behavior was identified on Rh(111) in CO.142 After exposing a Rh(111) surface to CO at 5 × 10−8 Torr, the adsorbed CO forms a (2 × 1) structure (Figure 18a and b). Although the adsorption sites of CO on Rh(111) cannot be resolved from the STM image due to the limited resolution, chemisorption at ontop sites at a coverage of ≤0.33 was proposed in previous studies based on low-energy diffraction pattern (LEED) and highresolution electron energy-loss spectroscopy (HREELS). As the pressure increases to ≥10−7 Torr, a periodicity of (√7 × √7)R19° is observed in the constant-current STM image (Figure 18c). Figure 17d shows two structural models of a (√7 × √7)R19° unit cell. In the gas phase of CO at a pressure of 10−6 Torr (Figure 18c), CO molecules exclusively adsorb on 3-fold hollow sites. However, CO molecules bond to both hollow and top sites when the CO pressure is increased to 10−5 Torr. This surprising evolution of surface structure along with the increase of CO pressure (Figure 18) demonstrates the influence of reactant pressure on molecular binding of reactants on a model catalyst. In this evolution, adsorption sites of the CO molecules change although the packing of metal atoms of the catalyst surface is preserved. When the CO pressure is increased to 5 Torr, a new structure with (2 × 2) periodicity (Figure 18e) is formed. As shown in Figure 18f, only one bright spot is present in the image of each unit cell, suggesting that the structure model contains one CO molecule on a top site and two CO molecules on hollow sites. This visualization of CO adsorbates on Rh(111) in gas of CO at different pressures using HP-STM clearly demonstrates the evolution of molecular adsorption from one type of specific site at low pressure (top site) to another type of specific site at a relatively higher pressure (both top and hollow sites). 3.1.2.2. NO on Rh(111). There is a pressure-dependent packing structure of NO molecules in the pressure range of 10−8−0.03 Torr.143 As the pressure of NO increases from 10−8 to 0.01 Torr, the (2 × 2)-3NO structure, which is stable at low pressure, transforms to the (3 × 3)-7NO structure at 1 atm. Figure 19 presents the dynamics of this transformation at 0.03 Torr as a function of time. White circles in parts a and b of Figure 19 mark the same defect at different times. As shown in Figure 19, the domain of the (3 × 3)-7NO structure grows with time. At the beginning (t = 0, Figure 19a) only a small area at the upper right corner of the image shows the (3 × 3)-7NO structure. However, the area of the (3 × 3)-7NO structure is doubled 55 s later (Figure 19b). In Figure 19c, the domain sizes of the (3 × 3)-7NO (right side) and (2 × 2)-3NO (left side) structures are nearly the same. The energy barrier between the (2 × 2)-3NO and the (3 × 3)-7NO structures was calculated to be ∼0.7 eV based on the dynamic data. As shown in the free energy diagram (Figure 20), the (2 × 2)-3NO is thermodynamically stable at low pressure of NO.143 However, (3 × 3)-7NO becomes thermodynamically favorable along with increase of NO pressure.143 On the basis of reported calculations,144 the transition temperature from (2 × 2)-3NO on Rh(111) to (3 × 3)-7NO is room temperature when the pressure of NO is ≥0.03 Torr. This calculation144 is consistent with the experimental observations with HP-STM (Figure 20).143 These observation with HP-STM at a relatively high pressure validated the predicted evolution of adsorbate structures in terms of packing pattern of adsorbates at different

Figure 16. STM images of 55 × 55 Å2 and corresponding ball models showing Moiré superstructures of CO on Pt(111) at room temperature. (a) Incommensurate structure at p = 10−2 Torr, It = 1.06 nA, Vt = 8.2 mV. (b) p = 720 Torr, It = 1.27 nA, Vt = 4.9 mV. The images a and b are aligned so that the bulk [110] direction is oriented along the x axis. (c) Incommensurate structure at 10−2 Torr. (d) Commensurate (√19 × √19) R23.4°-13CO structure at 720 Torr. Reproduced with permission from ref 4. Copyright 2004 American Chemical Society.

Figure 17. CO coverage θ on Pt(111) as a function of CO background pressure. The filled circles indicate values that were directly determined from the STM images, while the indirectly determined values are indicated by crosses. Reproduced with permission from ref 4. Copyright 2004 American Chemical Society.

occupation of the energetically favorable on-top sites. However, the weight of adsorbate−adsorbate lateral repulsion evidently increases along with an increase of gas pressure. At the pressure above 10−2 Torr, the lateral repulsion exceeds the energy barrier for switching from on-top to bridge sites; thus, the on-top site occupation tendency decreases, and the Moiré pattern no longer arranges along the high-symmetry direction. 3.1.2. Pressure-dependent change of site-specific adsorption of reactant molecules on model metal catalysts. As briefly mentioned at the beginning of section 3.1, molecules of reactant gas at a high pressure could chemisorb 3500

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 18. STM studies of Rh(111) in CO. (a) Large-area topographic image showing steps on the Rh(111) surface that are the result of emerging (111)-type dislocation planes. (b) Expanded view of a small terrace area showing the (2 × 1) structure of adsorbed CO. (c) Detailed view of the (√7 × √7)R19° periodicity, showing the content of the unit cell. (d) Schematic model of the adsorption sites of CO. (e) Topographic image (50 × 50 Å2) of the Rh(111) surface in 700 Torr CO. This image shows the (2 × 2) pattern observed in the CO pressure range between 5 and 700 Torr. (f) Structural model showing the (2 × 2) pattern of CO on Rh(111). Reproduced with permission from ref 142. Copyright 2000 Elsevier, Inc.

pressures of reactants by calculation of chemical potential of metals in reactant gases.144 3.1.2.3. NO on Pd(111). The (2 × 2)-3NO adsorbate structure was formed on Pd(111) over a wide range of pressures. On the basis of the surface free energy calculations (Figure 21),144 (2 × 2)-3NO on Pd(111) is thermodynamically stable at both low pressure and high pressure, suggesting a lack of a pressure dependence of adsorbate structures. This calculation is supported with the experimentally observed (2 × 2)-3NO structure at 720 Torr NO (Figure 22).144 Compared to the transformation from (2 × 2)-3NO formed on Rh(111) at low pressure to the (3 × 3)-7 NO formed in 720 Torr of NO,143 only (2 × 2)-3NO144 was observed on Pd(111) at 720 Torr (Figure 22). This difference in adsorbate structure at 720 Torr between Rh(111) and Pd(111) is likely related to the high binding energy of nitric oxide molecules on Rh(111) in contrast to Pd(111). The binding energies of NO in (3 × 3)-7 NO on Rh(111) and Pd(111) are 1.65 and 1.07 eV,144 respectively. 3.1.2.4. Rh(111) in CO and NO. The coadsorption of NO and CO on Rh(111) in a pressure range of 1 Torr were studied by using HP-STM.145 Upon adding 0.15 Torr of NO to an initial CO pressure of 0.5 Torr, bright spots appear and cover ∼1% of the image of the (2 × 2) structure to form a mixed (2 × 2)3(CO−NO) structure (Figure 23a). Because the adsorbed NO molecules appear much brighter than the nearby CO molecules at the experimental imaging condition, it is easy to distinguish them in STM images even at a relatively high pressure. As the partial pressure of NO increases, replacement of CO by NO takes place. Parts b1 and b2 of Figure 23 show sequential STM images taken with a time interval of 55 s. The arrows in these figures mark the area of one substitution. When the partial pressure of NO is 2−4 times greater than that of CO, the surface domains rich in NO separate from the mixed phase. This study demonstrates the complexity of adsorbate structures under reaction conditions and during catalysis when two or more reactants coexist in a gas phase around the catalyst. On a surface with coadsorbed molecules, replacement of one reactant with another one could occur spontaneously in a gas mixture of the two reactants. However, such a replacement typically could not

occur for a surface with coadsorbed reactant molecules in UHV,145 suggesting the complexity of the surface of catalysts in the mixture of all reactant gases of a catalytic reaction and the necessity of in situ studies. 3.1.2.5. O2 on Ru(0001). Chemisorption of oxygen atoms on Ru(0001) in UHV forms O(2 × 2) and O(2 × 1) structure with the coverage of 1/4 monolayer (ML) and 1/2 ML.146,147 Figure 24a is a representative image of Ru(0001) upon being exposed to O2 at 1 × 10−7 Torr for 20 s.77 Obviously, oxygen atoms pack into an O(2 × 1) phase in UHV. However, as shown in Figure 24b, chemisorbed oxygen atoms on the surface at the pressure of 200 mbar O2 in fact form a new phase of O(1 × 1).77 The O(1 × 1) ordered structure is formed on the entire Ru(0001) surface in 200 mbar O2. A DFT calculation suggests that the switch of O(2 × 2) or O(2 × 1) formed in high vacuum to O(1 × 1) at 200 mbar O2 is driven by the high binding energy of oxygen atoms to Ru.148 The formation of O(1 × 1) is a thermodynamically favorable process as the oxygen coverage of O(1 × 1) is twice that of O(2 × 1) at a relatively high pressure of O2. Here the distinctly different binding geometries of the adsorbed oxygen atoms in UHV and in O2 at a relatively high pressure clearly suggest a pressure gap in binding geometry of adsorbates on a surface. It illustrates the importance of in situ studies of a catalyst surface under reaction conditions and during catalysis from a fundamental understanding of a catalytic reaction point of view. Overall, when gas molecules adsorb on a metal surface at a high pressure, the adsorbate−adsorbate lateral repulsion plays an important role in the formation of a packing structure of adsorbates. As the coverage of adsorbates increases, the intermolecular repulsion could cause the adsorbate to transform from one site-specific binding to another site-specific binding that has a larger binding energy. Alternatively, the increased repulsion between adsorbates at a larger coverage could restructure the surface if catalyst atoms of the original surface are not closely compacted. The following section will demonstrate the restructuring of a catalyst surface in a gas phase with a relatively high pressure of reactants. 3501

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 20. Surface free energies for the (2 × 2)-3NO (red line) and (3 × 3)-7NO (blue line) adsorption structures on Rh(111). Temperature axis is shown for a NO pressure of 0.03 Torr. Reproduced with permission from ref 144. Copyright 2005 American Chemical Society.

Figure 21. Surface free energies for the (2 × 2)-3NO (red line) and (3× 3)-7NO (blue line) adsorption structures on Pd(111). Temperature axes are shown for NO pressures of 1 atm and 1 × 10−12 atm. Reproduced with permission from ref 144. Copyright 2005 American Chemical Society.

Figure 19. STM images of Rh(111) in 0.03 Torr of NO at 25 °C, showing the phase transition between a (2 × 2) and a (3 × 3) structure. In (A) the majority of the surface shows the (2 × 2) structure. (B and C) Domain boundary between the two phases moves across the image at a rate of ∼20 Å/min. Reproduced with permission from ref 143. Copyright 2001 American Physical Society.

the catalyst in a gas phase of the reactant. Compared to a closely packed surface such as Pt(111), an open surface such as Pt(110) or a surface with a high fraction of undercoordinated catalyst atoms such as Pt(557) tends to be restructured much more readily. In fact, these findings are significant because nanoparticle catalysts typically have a large fraction of undercoordinated catalyst atoms on their surfaces. 3.2.1. Hex-Pt(100) in CO. A hex-Pt(100) consists of a topmost surface layer with a quasi-hexagonal structure and all other subsurface layers of (100).149 Roughly speaking, the topmost surface layer is a pseudo-(111) atomic layer (pink and red balls) packed on the second layer which is (100) face (blue balls) as seen in Figure 25a. The unit cell of the hex-Pt(100) is (6 × 15) or (6 × 20) as shown in STM images (Figure 25c and d).36 In a CO gaseous environment of 10−6−10−5 Torr, the surface

3.2. Pressure-dependent restructuring of catalyst surfaces

In addition to a pressure-dependent switch of binding configuration of adsorbed molecules, the catalytic surface below the adsorbate layer could significantly restructure in order to release the increased intermolecular repulsion of the adsorbate layer at a high coverage of molecules. For example, strong adsorbate−adsorbate repulsion resulting from an increased surface coverage of reactant molecules at a relatively high pressure of a reactant could induce a reorganization of the catalyst atoms of the topmost surface or even subsurface layers of 3502

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 22. STM studies of Pd(111) in NO at a pressure of 720 Torr. (a) STM image of NO on Pd(111) at room temperature at 720 Torr of NO. (b) STM image of a clean Pd(111) surface. (c) Line scan showing the three NO molecules in the unit cell; atop adsorbed NO is clearly distinguished from 3fold hollow site adsorbed NO by their different apparent heights. (d) Ball model of the p(2 × 2)-3NO structure on Pd(111) (oxygen atoms are red, and nitrogen atoms are blue). The (2 × 2) unit cell and line scan direction corresponding to those in the STM image are indicated. Reproduced with permission from ref 144. Copyright 2005 American Chemical Society.

Figure 23. STM studies of surface structure of Rh(111) while Rh(1111) is in the gaseous mixture of CO and NO gas. (a) STM image of Rh(111) in equilibrium with a mixture of CO and NO. The brighter unit cells are due to NO adsorption on the top sites; neither CO nor NO can be seen when it is adsorbed on the hollow-sites. (b1 and b2) Pair of sequential STM images taken 55 s apart. These 200 Å × 115 Å images were taken on the same area of the surface in 0.50 Torr CO + 0.92 Torr NO (I = 260 pA, V = 50 mV). The top arrow shows one top site occupied by NO in the top image and by CO in the bottom one. The bottom arrow shows the opposite. (c) Schematic showing how a vacancy (boxed V) can diffuse across the unit cell and be substituted by a neighboring NO molecule, which is found in most of the hollow sites. Reproduced with permission from ref 145. Copyright 2002 American Chemical Society.

topography significantly changes and Pt nanoclusters are formed on Pt(100) substrate (Figure 26).36 Even at a low pressure of 5 × 10−9 Torr (Figure 27), the topmost pseudo (111) layer is restructured and small Pt nanoclusters are arranged along the [01−1] direction. When the CO pressure is increased to 10−5 Torr, a large number of islands with size of 0.5−3.5 nm and height of 0.23 nm are formed on the entire surface (Figure 26a). Each island consists of spots with high contrast. These spots are always packed into a square lattice (Figure 26b and c). Combining this experimental result with DFT calculations, a fan-out model was proposed in Figure 28b. In this model, CO molecules bond to small Pt nanoclusters through a tilted configuration, also called a fan-out configuration. In this model, the distance between two adjacent CO molecules in these nanoclusters in CO gas with a relatively high pressure is larger than the Pt−Pt distance in the bulk of Pt single crystals. Overall,

the increase of CO pressure results in breaking of the surface layer of hex-Pt(100) in a CO environment even with a pressure of only 10−6 Torr or lower. Compared to hex-Pt(100), there is no such restructuring on a closely compact surface Pt(111) and Rh(111) in CO at this pressure range. 3.2.2. Pt(110) in CO. Early studies performed in 1993 revealed the formation of a stripe-like surface structure of Pt(110) although there were no atom-resolved images available.150 Later on, atomic details of this surface in a CO environment were reported.151 After the introduction of CO, the (1 × 2) structure is lifted and a new structure forms. Parts a, b, and c of Figure 29 show the evolution of the Pt(110) surface structure as the CO pressure is increased. The channel-shapelike structure along [001] appears in Figure 29a, suggesting that the transformation of (1 × 2) to (1 × 1) occurs at the relatively low pressure of 10−5 Torr. The density of kinks largely increases as 3503

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

3.2.3. Pt(557) and Pt(332) in CO. In most cases a highsurface-area catalyst consists of nanoparticles that have high fractions of undercoordinated catalyst atoms. These undercoordinated atoms include atoms at step edges, corners, vacancies, and adatoms on surfaces of a nanoparticle (Figure 1c and d). From the point of similarity in the fraction of undercoordinated atoms on the catalyst surface, a stepped surface with a high fraction of undercoordinated catalyst atoms was considered as the model catalyst closest to an industrial nanoparticle catalyst for simulating the evolution of catalyst surface structures. One example is the concave Pt−Cu bimetallic nanoparticles (Figure 30) that are highly active in oxygen reduction on the cathode in low-temperature fuel cell technology.153 Clearly, the surfaces of these nanoparticles are stepped facets (Figure 30). On the basis of the correlation of single crystal model catalysts and high-surface-area nanoparticle catalysts reviewed in section 1.1, Pt(557) (Figure 31) can be used to simulate the surface of the high-surface-area nanoparticle catalysts (Figure 31). Pt(557) consists of parallel terraces with the same width of 1.3 nm along the [1−12] direction (Figure 31).3 The step height between two adjacent terraces is 0.21 nm. After 5 × 10−8 Torr of CO was introduced to a reaction cell, restructuring of the Pt(557) surface occurred. As shown in Figure 32b, the steps become meandering and some terraces are obviously wider than 1.3 nm, resulting in the increase of density of the step edge. As the coordination number of Pt atoms at step edges is only 7 for Pt(557), Pt−Pt interactions at the step edge are weakened due to the steric and electronic repulsion of two neighboring COs bonded to the two adjacent Pt atoms at a step edge. Thus, the mobility of Pt atoms at step edges is increased while Pt(557) remains in CO at 5 × 10−8 Torr. Along with the increase of CO pressure, the increased mobility offers opportunities for Pt atoms to recombine to form kink sites at 5 × 10−8 Torr and further to

Figure 24. STM studies of Ru(0001) exposed to O2. (a) STM image of the Ru(0001) surface after 20 s dosing with O2 at 10−7 mbar. The striped structure represents the O(2 × 1) phase; the image shows two rotational domains. On the large domain, the one fold period along the rows is resolved. (b) STM image recorded at 200 mbar of O2, showing the O(1 × 1) structure and dark and bright defect sites (A and B, respectively). Reproduced with permission from ref 77. Copyright 2005 American Institute of Physics.

the CO pressure increases for pressures larger than 10−2 Torr (Figure 29b). It is driven by the entropy effect determined by CO pressure. Moreover, the increase of surface energy of Pt(110) in the gas of CO resulting from formation of kinks can be compensated by the decrease of surface energy by the fact that more CO molecules are packed and thus more open surface area is passivated. At the pressure of 1 bar, the elongated islands of Pt atoms are observable (Figure 29c). High-resolution STM imaging (Figure 29d) revealed the formation of (1 × 2)-p2mgCO structure on Pt(110) at this pressure of CO, in which the chemisorbed CO overlayer covers the (1 × 1) substrate.153 Obviously, the Pt(110) surface exhibits significant restructuring in CO gaseous environment.

Figure 25. STM studies of clean hex-Pt(100). (a) Top-view of the quasi-hexagonal layer on the Pt(100) surface. (b) Side-view of the quasi-hexagonal layer on the Pt(100) surface. In (a) and (b), the Pt atoms of the unreconstructed (1 × 1) bulk plane are represented by light balls. (c) STM image showing parallel bands along the [011] direction spaced by 5 Pt atom distances. (d) High-resolution image of the hex-Pt(100) surface. (e) Line profile along A1−A2 in (c). Images in (c) and (d) were collected at a tunneling condition of 0.5 V and 0.3 nA. Reproduced with permission from ref 36. Copyright 2009 American Chemical Society. 3504

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 26. Surface structure of Pt(100) in CO at different pressures. (a) STM image (38 × 36 nm) of the Pt(100) surface after restructuring induced by adsorbed CO in equilibrium with the gas at 10−5 Torr. Numerous islands are formed by Pt atoms expelled from the hexagonal structure. The islands cover ∼45% of the area of the surface. (b) Atomically resolved image (10 × 10 nm) of the islands formed in 10−6 Torr of CO. (c) Schematic representation of the image, in which each dot corresponds to a maximum in the STM image. Reproduced with permission from ref 36. Copyright 2009 American Chemical Society.

Figure 27. Surface structure of hex-Pt(100) in low pressure of CO (5 × 10−9 Torr). (a) Large-scale image of Pt(100) in an environment of 5 × 10−9 Torr of CO. Step edges are marked with arrows. The surface reconstruction is only partially lifted, giving rise to clusters of Pt atoms aligned along the [01−1] direction. (b) High-resolution image showing areas of lifted and unlifted atoms in the same terrace. (a) Large-scale image of Pt(100) in an environment of 5 × 10−9 Torr of CO. Step edges are marked with arrows. The surface reconstruction is only partially lifted, giving rise to clusters of Pt atoms aligned along the [01−1] direction. (b) High-resolution image showing areas of lifted and unlifted atoms in the same terrace. Reproduced with permission from ref 36. Copyright 2009 American Chemical Society.

Figure 28. DFT calculations for the adsorbed CO on hex-Pt(100) in CO. (a) Starting geometry of CO and Pt atoms forming a (6 × 3) cell before relaxation. The Pt atoms and CO molecules are separated by 0.37 nm as the maxima in the STM images. (b) After relaxation to minimize energy in the DFT calculation, two (3 × 3) clusters are formed. The average Pt−Pt distance in the new clusters is 0.275 nm, as in the bulk, while the average O−O distance from the CO molecules is 0.37−0.41 nm, which matches the experimental findings. The distance between the nearest oxygen atoms of the two (3 × 3) clusters is 0.32 nm. This relaxed geometry is the stable one. In (a) and (b), the dark blue circles represent platinum atoms in the slab layers, whereas the light blue circles represent Pt atoms at the surface. The red and gray circles represent oxygen and carbon atoms, respectively. Reproduced with permission from ref 36. Copyright 2009 American Chemical Society.

form nanoclusters at the step edge in CO gas at pressures of ≥0.1 Torr (Figure 32c and d). As the coverage increases, the lateral repulsion is also increased and each terrace is eventually broken into deeper kink sites at 5 × 10−8 Torr CO. The equilibrium structure at 1 Torr is triangle-like nanoclusters fully covering the surface. For such a nanoclustercovered surface, the fraction of Pt atoms at the edge of the triangular nanocluster is ∼50% (Figure 32g), which is much larger than the 16% of Pt(557) in CO at 5 × 10−8 Torr (Figure 32b). Ambient-pressure X-ray photoelectron spectroscopy (APXPS) measurements showed that the coverage of CO on such a nanocluster-covered surface is 100%. The 100% coverage of CO on the broken Pt(557) surface, also called a nanocluster-covered surface, results from the high fraction of undercoordinated Pt atoms at the edge of these nanoclusters (∼2.5 nm). It suggests that the increased repulsion of lateral CO molecules bonded to undercoordinated Pt atoms at step edge induces the formation of nanoclusters through breaking the step edge because Pt(557) is an open surface compared to Pt(111). From a thermodynamics point of view, a CO molecule at an edge of triangle nanoclusters (Figure 32g) can release the repulsion between two adjacent CO

molecules by tilting adjacent CO molecules to the open side of the triangular clusters. Thus, the increased surface energy due to repulsion of CO molecules at high coverage is partially decreased by the fan-out effect of CO adsorbates at the edge of triangular nanoclusters. In addition, the decrease of surface energy resulting from passivating more Pt atoms of the surface through adsorption of more CO molecules at 0.1−1 Torr partially compensates the increase of repulsion. The formation of nanoclusters was supported by DFT calculations.3 Among these three potential surface structures of Pt(557) with 100% coverage of CO molecules, double-step (Figure 32e), parallelo3505

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 31. Vicinal surface of Pt(557). Reproduced with permission from ref 3. Copyright 2010 AAAS.

Figure 29. STM images of Pt(110) in CO at different pressures: (a) 10−7 mbar CO (1000 × 1000 Å2); (b) 10−2 mbar CO (1000 × 1000 Å2); (c) 1000 mbar CO (900 × 900 Å2); (d) 1000 mbar CO (28 Å × 28 Å). All images were obtained at 373 K. Reproduced with permission from ref 152. Copyright 2003 American Institute of Physics.

Figure 30. Schematic showing vicinal surface of the Pt−Cu CNC nanoparticle.

gram (Figure 32f), and triangular nanoclusters (Figure 32g) were considered in DFT calculation;3 the surface consisting of triangular nanoclusters (Figure 32g) has the lowest energy. A similar restructuring was observed on Pt(332) in CO. The restructured Pt(557) surface consisting of nanoclusters is restored to a stepped surface consisting of kink sites at the edge (Figure 33a) once 1 Torr CO is purged to 1 × 10−8 Torr CO. The surface in CO with a pressure of 10−8 Torr CO is very similar to the stepped surface with kink sites formed while a clean Pt(557) is directly exposed to CO gaseous environment of 5 × 10−8 Torr CO. Obviously, this restructuring is reversible. It clearly illustrates how the pressure of reactant molecules could restructure the packing of catalyst surface atoms.3 Once the gas reservoir of adsorbates is removed, the surface energy of the catalyst is increased due to the decrease of CO coverage at a low pressure (5 × 10−8 Torr). To decrease the surface energy, these undercoordinated Pt atoms prefer to reorganize to form terraces, by which the average coordinating number of all Pt atoms is increased. In other words, the fraction of Pt atoms at edge sites is

Figure 32. Surface structure of Pt(557) in different pressures of CO: (a) in ultrahigh vacuum with a background pressure of 1 × 10−10 Torr; (b) under ∼5 × 10−8 Torr of CO; (c) under 1 Torr of CO. Images are 40 nm by 50 nm in size. (d) Enlarged view of (c) showing the roughly triangular shape of the nanoclusters formed at 1 Torr. Two of the clusters are marked with red lines. (e) Optimized structure of terrace with doubleheight step. (f) Optimized parallelogram. (g) Optimized triangular nanocluster. Reproduced with permission from ref 3. Copyright 2010 AAAS.

obviously decreased. Therefore, this reorganization makes the system return to a stepped surface because a stepped surface (Figure 33a) has a smaller fraction of undercoordinated Pt atoms in contrast to a nanocluster-covered surface (Figure 33b). If the stepped surface is exposed and remains in a gas of CO with a pressure about 1 Torr, the terraced structure is broken again. A very similar restructuring was clearly observed on Pt(332) in gas of CO with a pressure of 1 Torr. These explorations demonstrate 3506

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

structure of Cu(110) in H2 at 1 bar.156 No surface reconstruction of Cu(110) in H2 at ∼10−8 Torr was revealed. However, as the pressure increases, the equilibrium coverage of hydrogen increases accordingly. The high coverage of adsorbates attained at high pressure of H2 induces surface reconstruction of Cu(110) from 1 × 1 to 2 × 1. Figure 34 presents the surface structure of Cu(110) at different pressures. Nonetheless, the surface structure of (1 × 1) is restored if the pressure of H2 goes back to 10−9 bar. Extensive investigations of Cu(110) surface at various pressures revealed that the critical pressure of this structural transition is in a pressure range of 1.5−15 Torr. The pressure-driven surface restructuring is supported by the restructuring of Pt nanoparticles revealed with in situ X-ray absorption spectroscopy (XAS).157 A recent in situ study using XAS revealed that both surface and bulk of Pt nanoparticle catalyst supported on SiO2 restructure in the gas phase of H2 at 20 bar.157 This is another example illustrating that in situ studies of catalyst surface structure are necessary because the surface structures of some catalysts such as Pt(557), Pt(332), and Cu(110) could be observed only at a relatively high pressure of reactants. Compared to the reversible restructuring of Cu(110) surface in H2 explored with HP-STM, ETEM studies of Cu/ZnO also suggest that H2O vapor at Torr pressure range can induce restructuring of the Cu surface of Cu nanoparticles supported on ZnO, which will be reviewed in sections 5 and 7. 3.2.5. Pt(557) in O2. Similar to CO on Pt(557), the coverage of oxygen atoms on Pt(557) is dependent on the pressure of O2. The measured oxygen coverage on Pt(557) is shown in Figure 35.37 Along with the increase of O2 pressure, the coverage of oxygen atoms is increased. It reaches 0.92 for Pt(557) in O2 at 1 Torr and decreases when O2 is purged. HP-STM studies show that Pt(557) is broken into nanoclusters at a relatively high pressure of O2,37 similar to Pt(557) in CO. AP-XPS studies show that these nanoclusters are platinum nanoclusters covered with oxygen atoms. Atomic fractions of Pt atoms of platinum oxide in all Pt atoms were measured with AP-XPS. The low photon energy (hν = 350 eV) and high photon energy (hν = 475 eV) allow for measuring this fraction of Pt(oxide)/Pt(total) in a shallow region with a mean free path of 0.4 nm and a relatively deep region with a mean free path of 0.6 nm. The decrease of the fraction, Pt(oxide)/Pt(total), in a larger detection volume suggests that platinum oxide is mainly on the topmost surface of Pt(557).37 In fact, the chemical state of the platinum oxide is the chemisorbed oxygen atoms on Pt(557) instead of platinum oxide.

Figure 33. STM images of Pt(557) at room temperature in 10−8 Torr CO (a) and 1 Torr CO (b). Reproduced with permission from ref 3. Copyright 2010 AAAS.

the hypothesis discussed in section 1 that the pressure of reactant could drive a surface restructuring. In the expression of chemical potential of a catalyst surface (section 1.1), μM is the term to describe the chemical potential of the catalyst atom on the surface. Obviously, μM of a Pt atom at the edge of a nanocluster is very different from μM of a Pt atom of a terrace. Clearly, this reversible pressure-dependent restructuring elegantly demonstrates an evolution of pressure-dependent surface structure. It is a demonstration of the existence of a pressure gap for some catalyst systems. More importantly, the reversible restructuring at low pressure and high pressure clearly shows that surface structure of some catalysts under a reaction condition or during catalysis could be observed only when the catalyst is in the gas of the reactant at a certain pressure. It further suggests that an in-situ/operando characterization of surface structure of catalyst under reaction conditions or during catalysis is necessary, at least for some catalyst systems such as Pt(557) in CO. In addition to the restructuring of Pt(557) in CO, most recently the surface of Cu(111) can restructure to Cu nanoclusters in CO at Torr pressure range as well.155 The formed Cu nanoclusters are active for H2O dissociation. 3.2.4. Cu(110) in H2. The adsorption of hydrogen atoms on Cu(110) is fundamentally important because supported Cu catalyst is the catalyst of two important industrial catalytic reactions including methanol synthesis and water gas shift. The HP-STM experiments have been conducted to reveal the surface

Figure 34. Sequence of atom-resolved STM images recorded at 298 K in the HP-STM depicting the Cu(110) surface (A) before H2 exposure, (B) during exposure to 1 bar of H2 (the image is recorded 1 h 50 min after exposure), and (C) after a time lapse of ∼1500 s directly after evacuating the HP cell. The unit cell of the (1 × 2) missing-row structure is indicated in (A) (2.55 × 7.22 Å2). All figures are drawn on the same scale (46 × 45 Å2). Reproduced with permission from ref 156. Copyright 2001 American Physical Society. 3507

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

reactant. So far, restructuring driven by reactant gases at a certain pressure has been demonstrated on Pt(557), Pt(332), Pt(110), and Cu(110) at room temperature in CO, H2, and O2. Parallel to the restructuring explored with HP-STM, ETEM has revealed that gas of H2O vapor at 0.3 Torr can restructure the Cu nanoparticle surface at 220 °C.138 Similar to the restructuring of Pt(557) revealed with HP-STM observation and confirmed with AP-XPS, a restructuring of Pt nanoparticles in gas of O2 at 0.013 Torr was revealed with ETEM studies. 158 The restructuring of these metal nanoparticles in O2 will be discussed in sections 5 and 7 3.2.6. Ni(557) in CO. Nickel is an important commercial catalyst for hydrocarbon reforming and many other reactions of C−H activation in chemical and energy transformations.1 In addition, it is even active for the decomposition of ammonia.159,160 A fundamental understanding of the evolution of surface structures under reaction conditions or during catalysis is critical. Step edges of a clean Ni(557) in 3 × 10−10 Torr CO were identified (Figure 36b).5 However, the step edge becomes rougher while Ni(557) is in the gas environment of CO with a pressure as low as 10−6 Torr. Segments of some terraces become wider by a factor of 3 compared to clean Ni(557). Clearly, the step edge of Ni(557) is significantly restructured in CO at a pressure of 10−6 Torr. In contrast to the structural evolution of step edge of Pt(557) in 1 Torr CO (Figure 32), the restructuring of Ni(557) (Figure 36e) is more dramatic. The restructuring of the Ni(557) surface at 25 °C in 1 Torr CO was supported by the evolution of coverage of chemisorbed CO molecules on Ni(557) at different pressures of CO measured with ambient-pressure X-ray photoelectron spectroscopy at room temperature.5

Figure 35. Changes of surface coverage of atomic oxygen (black line) and the fraction of Pt of Pt oxide in all Pt atoms formed on Pt(557) (red lines) as a function of the pressure of oxygen gas phase. Two different photon energies were used to check the potential depth-dependent fraction of platinum oxide. Clearly, the coverage of oxygen atoms increases along the pressure of oxygen gas and reaches the saturation value about 1 Torr. At higher pressure, the fraction of platinum oxide is higher. The same evolution of surface fraction of platinum oxide as surface coverage of oxygen atoms suggests that the gas-phase pressure plays an important role in the formation of platinum oxide on Pt(557). Reproduced with permission from ref 37. Copyright 2012 American Chemical Society.

If O2 is purged and then a pressure of O2 at 10−8 Torr is reached, these nanoclusters are restructured to their original stepped surface, similar to the one formed in O2 at a pressure of 10−8 Torr. The switch between a surface of nanoclusters at 1 Torr of O2 and a surface of terraces at a pressure of 10−8 Torr of O2 is reversible. This is another example showing restructuring of the surfaces of model catalysts driven by the change of pressure of a

Figure 36. Restructuring of Ni(557) in CO gas. (a) Structural model of Ni(557). (b) STM image of Ni(557) under UHV with a base pressure of 3 × 10−10 Torr. (c) STM image of Ni(557) in CO gas at 1 × 10−8 Torr. (d) STM image of Ni(557) in CO gas at 1 × 10−6 Torr. (e) STM image of Ni(557) at 1 Torr CO. All images have a size of 24 nm × 24 nm. The inset parts marked with black boxes in (c) and (e) are the enlargements of a part of (c) and (e), respectively. Reproduced with permission from ref 5. Copyright 2013 American Chemical Society. 3508

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 37. (Upper panel) Mass spectrometer signals of O2, CO, and CO2, measured directly from the reactor cell. The mass spectra were recorded simultaneously with the STM movie. Labels (A−H) correspond to the STM images in the lower panel. Rlow and Rhigh denote the low and high CO2production rate branches. Pth indicates the threshold value of the CO pressure at which the rate switched from Rlow to Rhigh and the surface morphology changed simultaneously from smooth to rough. (Lower panel) STM images of 210 × 210 nm2 from an STM movie of a Pt(110) surface at a temperature of 425 K in a 3.0 mL/min flow of mixtures of CO and/or O2 at 0.5 bar (65 s/image). The images were differentiated to enhance the contrast. Images (A, B, E, F, and H) show flat terraces separated by steps of the Pt lattice. This corresponds to the metallic, CO-covered surface. Image (C) shows the change of the surface during the step in activity (the image was built up from bottom to top). The bright scan lines and the change in slope are the result of extra thermal drift due to the increased, exothermic reaction. Images (D) and (G) show the rough surface consisting of protrusions, with heights of 0.2−0.4 nm, and pits (see inset). These are formed during the high activity stage. There is a direct correspondence between the buildup of the roughness and the total amount of produced CO2. Reproduced with permission from ref 161. Copyright 2002 American Physical Society.

4. OPERANDO STUDIES OF CATALYST SURFACES DURING CATALYSIS USING HP-STM An operando study of CO oxidation on Pt(110) with HP-STM and online mass spectrometry has been performed. The pressure of the reactant gases is ∼1 bar. A quadruple mass spectrometer was used to analyze the composition of gas in the reactor under different catalytic conditions including different pressure ratios of CO to O2 at different reaction temperatures while HP-STM was being used to identify the active surface phase of Pt(110) for CO oxidation under a semirealistic industrial catalysis condition.161,162 A correlation between the surface structure and the corresponding catalytic performances in terms of activity was established simultaneously. The upper panel of Figure 37 shows the catalytic activity of CO oxidation in terms of the partial pressure of product CO2 and reactants CO and O2 at different periods marked as A−H in this figure. The corresponding surface structures at these periods are shown in the lower panel of Figure 37. As shown in the upper panel of Figure 37, at period A the Pt surface was covered with CO molecules in CO gas before O2 was introduced and this initial surface was metallic. At 500 s or so, O2 was introduced and reacted with CO (period B in Figure 37). As there was excess CO in the reactor, the production rate of CO2 was determined by the O2 pressure. When the CO pressure was below a certain value,

the Pt(110) surface was oxidized (period C in Figure 37). Surprisingly, while the CO pressure is decreased to a threshold pressure, the production rate of CO2 is increased largely (panel D in Figure 37). The active phase at this period was suggested to be platinum oxide. It further indicates that PtOx is more active than a metallic Pt surface. When CO pressure was increased, more O2 was consumed. The proposed platinum oxide was reduced to metallic Pt as shown in period E of Figure 37. When O2 pressure was increased, more CO was consumed. At the periods of F and G, the partial pressure of O2 is higher than that for CO definitely; in this period, the surface was suggested to be platinum oxide as shown in the HP-STM image collected at period G in the lower panel of Figure 37. As shown in the lower panel of Figure 37, in periods C and D the progressive formation of a rough surface of Pt(110) indicates that CO molecules react with surface lattice oxygen atoms to form CO2 molecules. Under this condition, catalysis includes an oxidizing step (filling oxygen vacancies) and a simultaneous reducing step (surface lattice oxygen atoms reacts with CO molecules). When the pressure of CO is increased again (period E of Figure 37), the proposed surface platinum oxide is reduced to metal and it gives a CO-covered metal surface. The reaction rate was obviously decreased because the partial pressure of the formed CO2 was decreased. The surface at period E in the lower 3509

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

panel of Figure 37 became relatively flat when CO partial pressure was increased to higher than the threshold pressure of CO (Pth in Figure 37). The active phase is a metallic Pt surface that exhibits a lower catalytic activity in contrast to that for the platinum oxide on a surface formed in O2-rich environment. These operando studies using HP-STM rationalized the similary catalytic activity for CO oxidation on Pt(110). Without in situ/operando studies using HP-STM, these different active surfaces of model catalyst Pt(110) formed under different catalytic conditions cannot be identified and the correlation in Figure 37 cannot be revealed. It demonstrates that operando studies using HP-STM along with the measurements of the corresponding catalytic performances can establish structure− reactivity correlations. With such intrinsic correlation to be established through operando studies, catalytic mechanisms may be understood at an atomic level.

workers; they resolved lattice planes of a CeO2 catalyst in a reactive gas.89,108 In many ways this study opened the door to modern ETEM of catalyst where atomic resolution is now routinely accomplished. Exploration of the structure of the catalyst during preparation, activity, catalysis, deactivation, and regeneration under different conditions is a critical part of exploring correlations between structure and performance. By integrating the capabilities of ETEM with the fundamental studies of catalyst surface structure with HP-STM, a deep understanding of a catalytic reaction at the atomic level can be achieved. ETEM is an ideal tool for exploring these changes down to the atomic level using imaging, diffraction, and spectroscopy. The following subsections reviewed the in situ studies of preparation, activation, active form, and deactivation of catalysts with ETEM in the last ten years. For reviews of early work, the readers can refer to earlier literature.106,173−180 5.1. Catalyst preparation

5. ATOMIC-SCALE CHARACTERIZATION OF STRUCTURAL AND CHEMICAL EVOLUTION IN NANOCATALYSTS WITH ETEM UNDER REACTION CONDITIONS Catalytic nanomaterials undergo an evolutionary life cycle starting from the preparation process, moving through activation and operation, and concluding with deactivation and possible regeneration. The phase transformations taking place during these processes are often driven by interactions with the surrounding gas/liquid of reactants of a catalytic reaction or by exposure to an external stimuli such as heat, light, or electric fields. The structural changes are strongly correlated with catalytic performance. For example, the phase changes taking place in a material during catalyst preparation or activation are accompanied by gross changes in the structure and composition bulk-phase transformations. Catalyst evolution under steadystate conditions is often associated with more subtle changes that may involve shape changes or surface reconstruction. Deactivation is another gradual change that can be associated with sintering, poisoning, burying of surface structure, or phase transformation. Regeneration is the process of restoring the catalyst to its origin active state. Each of these processes can be investigated by ETEM, although the number of studies is still somewhat limited and there is certainly need for much more work. The value of the ETEM approach for characterizing highsurface-area nanoparticle catalysts was recognized by Baker. He was the first to extensively apply the technique to the studies of heterogeneous catalysts in a series of early works performed on filamentous carbon growth on transition metal nanoparticles− not too dissimilar to modern work on carbon nanotubes−but atomic resolution TEM was not yet available.163−165 Baker investigated the effect of Fe-based catalyst for coal conversion using graphite as a model system.166,167 He was the first to explore morphological changes of metal-supported catalysts, which included spreading of metal particles under oxygen,168 changes in sulfide particles in H2,169 and strong metal support interactions.170,171 Gai was also an early pioneer who used ETEM to explore the evolution of defects in oxide catalysts on exposure to reducing and oxidizing gases.172 She demonstrated that a variety of defects including dislocations and crystallographic shear planes can form in MoO3 on exposure to reducing gases. Discussion of this and other early work in ETEM of catalysts can be found in earlier reviews.173,174 The first atomic resolution in situ studies using ETEM was achieved by Parkinson and co-

Understanding new and existing methodologies for catalyst preparation is important for developing new catalysts.181−183 Many methods have been developed for fabricating well-defined nanoparticle catalysts using a wide range of chemical and physical methods.183−191 For example, a traditional preparation method192 typically involves impregnating a high-surface-area support with a salt solution and then thermally decomposing the salt via calcination and a reduction step to yield a dispersion of metal nanoparticles on the support. Studies employing physicochemical macroscopic characterization techniques including ex situ transmission electron microscopy (TEM) have been carried out over decades on the effect of varying preparation conditions on the final metal dispersion.189,191,193−196,182,189,195 ETEM has been employed to follow the nanoscale processes occurring during catalyst preparation. For the impregnation approach, ETEM allows for tracking the spatial distributions of the initial salt, the oxide, and the final metal to be correlated from the same region of the support. From such observations, the role of the initial salt dispersion, diffusion, and chemical decomposition during the thermal treatments can be evaluated directly with imaging techniques. For instance, Li et al. investigated the evolution of the Ni phase during reduction of a nickel salt on TiO2 support.197,198 They showed that, for high metal loading, the spatial distribution of the metal phase on the support is not significantly altered during reduction. The exact shape of the metal particle depended on whether particle nucleation occurred on anatase or rutile. Evidence for simultaneous diffusion of TiO2 onto the surface of the growing Ni particles during reduction in CO (Figure 38)198 was found. This is an example of a strong metal−support interaction in which reduced titania species diffuse onto the metal surface similar to that demonstrated in the pioneering work of Tauster and colleagues.199−201 A challenge in in situ/operando studies of the catalyst preparation results from the inherent complexity of highsurface-area support in the nanostructure and composition. For example, TiO2 and γ-alumina, which are important supports for industrial catalysts, show local variations in structure including different surface facets and many surface defects. The various facets and defects of the support may have promotional effects for catalysis, but the structural complexity greatly complicates the interpretation of the catalytic performance. In addition, the complexity of the support may obscure information about the fundamental mechanisms leading to morphological, chemical, and structural changes taking place during catalyst preparation. To address this issue, Banerjee undertook an investigation of the 3510

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

from 200 to 500 °C leads to the formation of metal particles with internal voids. The voids were not stable above 500 °C, and a reduction at higher temperatures gave solid metal particles. The reducibility of cobalt oxide on 12 wt % Co/α-Al2O3 with and without 0.5 wt % Re promoter205 was investigated. Through imaging and EELS of this catalyst precursor in 2.5 Torr of H2, it was found that the initial Co3O4 started to transform to CoO at 180 °C and Co metal started to appear at 260 °C with complete reduction at 360 °C. In this case, the Re appeared to be uniformly distributed over the support and there was no difference in the reducibility between the promoted and nonpromoted catalysts. A similar result was also found for a CoOx/SiO2 composite system.206 Other than the studies of the preparation of supported singlemetal catalysts, the preparation of supported bimetallic nanoparticle catalysts was studied with ETEM. The reduction of metal oxides and the formation of bimetallic catalysts were tracked during the preparation and activation. For instance, the preparation of bimetallic catalysts including a Fischer−Tropsch catalyst (20 wt % Co + 2 wt % Ru)/γ-Al2O3115 was investigated. During reduction, small particles nucleate from the metal precursor and grow mainly by coalescence. By comparing the Ru-promoted Co catalyst with pure cobalt nanoparticle catalysts, in situ STEM EELS showed that, in gentle reducing conditions (1 Torr of 10% H2 and 90% N2) at 400 °C, small metallic Ru and CoRu metal particles coexist with larger CoOx particles, demonstrating that Ru enhanced the reduction of CoOx. The morphological and nanoscale compositional changes during reduction of (3 wt % Ni + 1 wt % Cu)/TiO2207 in the preparation of Ni−Cu bimetallic catalyst were investigated with ESTEM. Particles nucleated from the nitrate precursors and grew via both Ostwald ripening and particle migration and coalescence. Cu/TiO2 and Ni/TiO2 nanoparticles can be reduced to metal at 300 and 400 °C, respectively. Particles containing both Cu and Ni supported on TiO2 can be reduced to a metallic state in a 1.5 Torr mixture of 20% H2 and 80% N2 at 300 °C. As shown in Figure 40, the energy-loss spectra recorded from points marked with 1 and 2 show that point 1 contains Ni and Cu whereas point 2 contains only Ni. The white lines at the front of the Ni L23 edge showed that the Ni associated with the Cu is metallic whereas the Ni without association with Cu at point 2 is still oxide. This in situ study demonstrated the

Figure 38. Bright-field ETEM image showing simultaneous nucleation of Ni particle on rutile and rutile cluster on the Ni particle recorded during reduction of Ni precursor at 350 °C in 0.2 Torr of CO. The inset is part of a background-subtracted EELS recorded from the cluster showing the presence of the Ti L23 edge at 455 eV, confirming that the cluster is TiO2. Reproduced with permission from ref 198. Copyright 2006 Elsevier, Inc.

preparation of model Ni catalysts during both calcination and reduction steps on a support consisting of SiO2 spheres.202 The simple geometry of this support is ideal for interpreting changes in structure and morphology of the metal species during the calcination and reduction steps. By using this support, it was easier to track the diffusion of metal during reduction. This study202 demonstrated that, as shown in Figure 39b, finer metal dispersions were obtained by eliminating the calcination step and performing a direct reduction in H2. In addition, silica spheres were also employed as a support to investigate the effect of reduction temperature on particle morphology for the PdO to Pd reaction.203,204 In this case, reduction in the temperature range

Figure 39. In situ Z-contrast image (i.e., ESTEM image) showing change in morphology that takes place during in situ direct reduction of a well-dispersed precursor on silica support. (a) Initial distribution of precursor on silica support. Faint precursor rings are visible on the support marked B. (b) Image recorded after 3 h exposure to 1 Torr of H2 at 400 °C showing a fine dispersion of Ni metal particles. Reproduced with permission from ref 202. Copyright 2012 American Chemical Society.

Figure 40. In situ Z-contrast image showing a bimetallic CuNi particle on a TiO2 support in H2 at 300 °C at an intermediate stage in reduction. Energy-loss spectra recorded from points marked 1 and 2 show that point 1 contains Ni and Cu whereas point 2 contains only Ni. The white lines at the front of the Ni L23 edge show that the Ni associate with the Cu is metallic whereas the Ni at point 2 is still oxide. Reproduced with permission from ref 207. Copyright 2009 Elsevier, Inc. 3511

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

promotional effect of Cu in reducing Ni to form Cu−Ni bimetallic particles. BaO is the catalyst for storing NOx. Its preparation from precursor Ba(NO3)2 and morphological changes of Ba(NO3)2/ α-Al2O3 during annealing208 were tracked with ETEM. The nitrate was observed to decompose to BaO in the annealing of Ba(NO3)2/α-Al2O3 in the mixture of N2 and O2, although the overall triangular shape of the original Ba(NO3)2 particles at micrometer scale was basically preserved during the annealing. An obvious change during this annealing is that the volume of particles shrunk by ∼30%. In addition, ETEM studies show that triangular aggregates of BaO consists of many small crystallites with different orientations. The combination of in situ imaging, electron diffraction, and spectroscopy of the ETEM is able to provide nanoscale or even atomic-scale information about the structure, composition, and phase transformation at different stages of catalyst preparation. A major advantage of the ETEM approach is that it explicitly reveals the evolution of the structure and composition at an atomic scale during preparation. Even on a highly controlled model catalyst, considerable heterogeneity in structure and composition may exist at the nanometer scale during the intermediate stage of the catalyst preparation such as calcination and reduction.

Figure 41. O−K-edge energy-loss spectra of reoxidized catalyst, CuO and Cu2O references (lower traces), and a ZnO reference (upper). The four catalyst sample spectra were acquired on the same sample area with electron beam flux increasing from the spectrum marked “low” to the spectrum marked “high”. Reproduced with permission from ref 209. Copyright from 2006 Union of Crystallography.

5.2. Catalyst activation

42a convert to a porous Ni metal structure during reduction due to the volume shrinkage associated with the oxide to metal phase transformation as shown in Figure 42b. Interestingly, as shown in Figure 42c during reoxidation, NiO forms voidlike structures due to the outward transport of Ni cations along grain boundaries and dislocations.213,214 This leads to an overall expansion in the NiO phase compared to the original NiO. This explanation results in strain in the adjacent YSZ grains and eventually leads to anode failure. The process of activation of Fischer−Tropsch catalyst in which iron oxide converts to iron carbide was studied with ETEM.34,215,216 The catalyst precursor, iron oxide, could be activated with pure CO or the reactant gases of Fischer−Tropsch synthesis, i.e., a mixture of CO and H2. Annealing in CO gas at 20 Torr reduced the hematite to metallic iron, accompanied by a severe sintering, and finally formed iron carbide, which is the active phase of a catalyst of Fischer−Tropsch synthesis (confirmed with other techniques). Upon a prolonged reaction, these newly formed crystallites of iron carbide exhibited a size in the 5−30 nm. These in situ studies of activation processes illustrate the value of ETEM to identify phase transformations leading to catalyst activation.

Catalyst activation may be thought of as a process taking place inside a reactor that alters the surface and/or bulk structure of catalyst nanoparticles to form an active surface. In some cases, it is hard to differentiate the preparation and the activation of a catalyst. For a metal-supported catalyst, catalyst activation may simply be a repetition of a reduction of surface oxide formed during exposure to air or a more complex phase transformation in a reactive gas. In either case, studies of the catalyst activation with ETEM may provide important information on the formation of active surface phases of catalysts. Activation of 4 wt % Cu/ZnO catalysts prepared using coimpregnation techniques was studied with ETEM and X-ray diffraction (XRD).209 In the as-prepared catalyst, Cu substituted into the ZnO lattice. In situ EELS and XRD showed the formation of metallic Cu species on the surface of the catalyst during reduction in H2 at 1 Torr at 277 °C (Figure 41). In situ XRD suggested the reoxidation of metal Cu to Cu2+ during calcination in air. In situ EELS did not conclusively suggest the formation of copper oxide, although it is spectulated that an intermediate phase of CuO supported on ZnO was formed. Recent work shows that Cu−Zn bimetallic metal clusters can transform into polycrystalline clusters consisting of separate CuO and ZnO nanocrystals. Upon subsequent reduction in H2, the CuO converts into metallic Cu with ZnO nanocrystal covering their surface.210 This illustrates the complexity of the phase transformations that may take place in complex nanoparticles systems during activation. ETEM has also been used to track the activation of electrocatalysts. The as-prepared anode on the high-temperature solid oxide fuel cell often consists of NiO and yttria-stabilized zirconia (YSZ). The anode must be activated by reducing the NiO to metallic Ni so that it acts as both a current collector and a catalyst. In addition, the anode is prone to cracking under redox cycling. ETEM was used to follow the evolution of the anode under both reducing and oxidizing conditions.211,212 Figure 42 shows the evolution of the anode during a reduction and oxidation cycle. The initially dense NiO grains shown in Figure

5.3. Structural evolution in gas reactants

Majority of ETEM studies falls under the broad category of performing in situ imaging, spectroscopy, and diffraction on catalysts under near reaction conditions. In these experiments, the active catalyst is exposed to conditions that are similar to those present inside a reactor. For in situ ETEM, the volume of catalyst present in the microscope is extremely small and conversions could be very small and even negligible. Certainly there is considerable scientific value in observing the atomic-level changes taking place in the catalyst on exposure to gases that may be present inside the reactor. However, it is important to give careful consideration to the gas composition employed in an ETEM so that meaningful comparisons with ex situ reactor data can be made.217 5.3.1. Supported metal nanoparticles. Metal nanoparticles loaded on high surface area supports is an important 3512

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 42. In situ reduction and reoxidation of a Ni/YSZ fuel cell anode. (a) Initial structure composed of NiO/YSZ; (b) fully reduced anode showing porous Ni metal regions; (c) reoxidized anode showing formation of porous NiO regions and bend contour in YSZ indicative of strain. Reproduced with permission from ref 211. Copyright 2010 Elsevier, Inc.

category of catalysts. Au, Pt, Pd, Ag, Ni, Cu, and NiRu have been investigated recently with ETEM. 5.3.1.1. Au in CO and O2. Au nanoparticles have received considerable attention because of their high activity for CO oxidation as first demonstrated over 20 years ago.218 Despite tremendous efforts made in the last 20 years, the catalytic mechanism of the prototype reaction, CO oxidation on Au nanoparticles, is still being debated.219 The best Au catalysts are usually supported on reducible oxides like TiO2, CeO2, and FeOx. Giorgio and co-workers have conducted several studies on the shape of 1−8 nm Au nanoparticles on MgO, TiO2, and carbon supports on exposure to 1.5 Torr of O2, H2, and CO + O2 at room temperature.88,220 Exposure to H2 resulted in the particles assuming truncated octahedron shapes bounded by (111) and (100) surfaces, whereas exposure to O2 resulted in the particle assuming a more rounded shape. This was interpreted in terms of oxygen atoms chemisorbing at edge and defect sites. These shapes were present regardless of the type of employed substrates. By performing a detailed analysis of the shape of these nanoparticles, they were able to determine the interfacial energy for Au on TiO2 and MgO. They find a strong metal−support interation with the TiO2 (adhesion energy 0.98 J/m2) and a much weaker adhesion energy with MgO (0.5 J/m2). Takeda’s group has investigated the Au particles on CeO2 at room temperature in the presence of CO or/and O2.221,222 They defined a morphology index to obtain a more quantitative indication of particle shape and performed systematic studies on the particle shape as a function of CO and O2 partial pressures (Figure 43). Interestingly, even very low concentrations of CO mixed with O2 can stabilize the (111) and (100) facets of Au nanoparticles, presumably because of the much larger adsorption energy of CO compared to O2. For CO oxidation conditions they found that the predominantly faceted particles are present with slightly rounded corners. Interestingly, they also found that Au nanoparticles supported on TiC show faceted shapes that are independent of gas composition. The same group examined the morphology of Au nanoparticles in 1% vol CO in air (0.34 Torr) at room temperature using aberration-corrected ETEM.222 They found that there is an outward relaxation of ∼0.5 Å of the top (100) layer in the presence of CO (Figure 44). This reconstruction allows CO to adsorb on (100) facets of Au nanoparticles with high coverage. The FEI Titan ETEM can be operated at 80 kV to increase the sensitivity to light atoms. With this mode the authors were able to observe a contrast on the

Figure 43. Variation in shape of Au particle supported on CeO2 at room temperature as a function of O2 and CO partial pressures. Reproduced with permission from ref 221. Copyright 2011 Wiley-VCH.

Figure 44. Differences in the surface structure of Au in (a) vacuum, showing predominantly a bulk-terminated structure, and (b) the presence of CO, showing outward relaxation of the top (100) layer. The two lower figures for (a) and (b) are higher magnification images from the boxed regions showing changes in plane spacing as a result of changes in the reactive gas. Reproduced with permission from ref 222. Copyright 2012 AAAS.

surface of the Au nanoparticles that appears to be consistent with direct imaging of CO adsorbates (see section 7.2 and Figure 63). In addition, they observed a stepwise lateral displacement of the 3513

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 45. Differences in the shape of Cu particle on ZnO in H2 (left), H2 and H2O (middle), and H2 and CO (right). Reproduced with permission from ref 138. Copyright 2002 AAAS.

The faceting suggested by ETEM studies has not been investigated on Pt model catalysts with HP-STM. It is expected that a HP-STM study could provide more clear information on how the surface structure of Pt nanoparticles evolve with temperature in the mixture of CO and O2 if a model catalyst of Pt nanoparticles is used. 5.3.1.3. Cu on ZnO in H2 and H2O. There has been considerable discussion about the nature of the interaction between Cu and ZnO under conditions relevant to this reaction.227 Hansen and co-workers addressed this issue by investigating the shape of Cu particles supported on ZnO in different gas atmospheres.138 They observed reversible shape transformation in Cu particles at 220 °C when the reduction strength of reactant gas around Cu/ZnO catalyst nanoparticles (Figure 45) was varied. The Cu nanoparticles restructured when pure H2 was changed to the mixture of H2 and H2O. The fraction of (110) facets increased in a mixture of H2 and H2O and decreased when the gas was changed to H2 and CO. This evolution suggests that H2O molecules can stabilize the (110) facet. A Wulff construction analysis was employed to estimate changes in adhesion energy associated with the change of morphology of Cu nanoparticles supported on ZnO. They were also able to estimate the thermodynamics (Gibbs free energies) of surface facetting. Thus, by analyzing ETEM data, the thermodynamics (Gibbs free energies) of surface faceting can be estimated. In a stronger reducing atmosphere (mixture of H2 and CO), more oxygen vacancies are created on the surface of the ZnO support and the Cu particle spreads.227 In situ studies using EELS showed subtle changes of the Cu L23 near-edge fine structure under strong reducing conditions. The change of the Cu L23 near-edge fine structure was assumed to be associated with tensile strain in the flattened Cu particles, consistent with the formation of a ZnCu alloy layer at the Cu/ZnO interface.228 The correlation between catalyst structure and reactivity has been explored explicitly by investigating the relationship between enhanced activity of methanol synthesis and catalyst pretreatment in different mixtures of CO and H2.229 5.3.1.4. Ni and Ni−Ru on SiO2 for partial oxidation of CH4. Evolution of Ni catalysts during oxidation of methane was tracked with ETEM, by which a correlation of catalyst structure

Au nanoparticles on ceria even at room temperature, which suggests these Au nanoparticles are weakly bonded to an Oterminated surface.223 Au nanoparticles on ceria nanorods during CO oxidation at room temperature and water gas shift at 573 K were studied with ETEM.224 The Au nanoparticles appear faceted in both O2 and CO at 573 K, and only a subtle change in particle shape occurs when the gas composition is changed. Presumably, the residence time for O on Au at high temperature is so short that the coverage is not high enough to trigger the shape transformation observed at room temperature. However, the interface between the Au and ceria reconstructs when CO is introduced. Isotope-labeling experiments suggested that CO oxidation occurs at the Au/CeO2 interface with CO adsorbing on the nearby Au atoms and atomic oxygen for CO oxidation coming from the ceria. 5.3.1.2. Pt in CO and O2. Pt nanoparticles supported on both carbon or ceria can assume shapes close to the Wulff form, consisting of mostly (111) surfaces truncated with (100) in vacuum, N2, and H2 at room temperature.225,226 Small Pt nanoparticles of 4 nm supported on CeO2 become rounded when exposed to O2,226 suggesting a restructuring of the Pt nanoparticle. This is consistent with the observation of restructuring of vicinal Pt(557) surface in O2 revealed with HP-STM,37which is reviewed in section 3.2.5. Compared to the lack of restructuring of larger Pt nanoparticle of 20 nm, the restructuring of small Pt nanoparticles may suggest the role of undercoordinated Pt atoms on surface restructuring.225,226 The restructuring of Pt nanoparticles of 4 nm but lack of restructuring of Pt nanoparticles of 20 nm is similar to the difference identified with HP-STM, in which narrow terraces of Pt(557) (1.3 nm) are readily restructured in CO at ≥0.1 Torr but not the wide terrace of Pt(111) in CO at 1 bar. Both ETEM and HP-STM revealed the importance of size in terms of fraction of undercoordinated atoms in restructuring. More complementary functions of HPSTM studies of model catalyst and ETEM studies of highsurface-area catalyst will be discussed in section 7. The 4 nm particles in mixtures of CO and O2 at room temperature are rounded. However, they become progressively faceted with heating to 100 °C. This is understandable because the coverage of CO and O2 at 100 °C could be quite low due to the much shorter residence time at a relatively high temperature. 3514

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 46. In situ ETEM images and electron diffraction patterns of Ni/SiO2: (a) in the presence of 1 Torr of H2 at 400 °C (condition A); (b) from the same region in the presence of 1 Torr of mixture of CH4 and O2 in 2:1 ratio at 400 °C (condition B). Reproduced with permission from ref 217. Copyright 2011 Wiley-VCH.

with ex situ catalytic activity was established.217,230 ETEM was used to track the phase changes taking place on the 2.5 wt % Ni/ SiO2 and 7.5 wt % Ni/SiO2 catalysts. During the thermal rampup to 350 °C in 1 Torr of mixture of 2CH4 + O2, the initial metallic Ni particles were oxidized to NiO void structures due to the Kirkendall mechanism (Figure 46). The NiO particles catalyze complete oxidation of CH4, forming catalytic products CO2 and H2O. The finding of reoxidation of metallic Ni particles to NiO during ramp-up at an early stage of catalysis revealed that the prereduction step of NiO particles in hydrogen is unnecessary. As the temperature continues to rise, the oxygen in the reactor gets converted and the atmosphere becomes reducing, resulting in reduction of NiO to metallic Ni at ∼700 °C. At this point reforming begins and the products change from CO2 and H2O to CO and H2. The reduction mechanism of NiO to Ni is the reverse of the oxidation process. The formed metallic Ni on the surface diffuses to subsurface sites (see Figure 47) and delays the onset of syngas production until the NiO particle is almost completely reduced. The complete changes in catalytic activity in terms of consumption of CH4 and corresponding schematic of catalyst structure derived from ETEM are shown in Figure 48. For this simple model system, there is a very direct and simple correlation between catalyst structure and reactivity. The major changes in catalytic activity that appear at 400 and 750 °C correlate very precisely with structural transformations within the catalysts. The behavior of the bimetallic NiRu/SiO2 is particularly interesting for the partial oxidation of methane. Using a combination of imaging and in situ STEM EELS, it was shown

Figure 47. In situ high-resolution ETEM images along with Fourier transforms of Ni/SiO2 catalyst in the presence of a few Torr of CH4 showing formation of subsurface Ni metal domains in a NiO nanoparticle. Reproduced with permission from ref 217. Copyright 2011 Wiley-VCH.

that initially homogeneous NiRu particles undergo two distinct changes during ramp-up in mixtures of CH4 and O2.231 At ∼350 °C, Ni diffuses to the surface and oxidizes, giving rise to a NiO/ Ru core−shell structure that performs complete oxidation of CH4. At ∼550 °C, this shell of NiO is reduced to metallic Ni metal and then remixes with the underlying metallic Ru core to form Ni−Ru bimetallic nanoparticles. However, the high spatial 3515

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

change in the structure of Ru particles suggests that the Ba may act as an electronic promoter for the reaction. 5.3.2. Reducible oxides. Reducible oxides represent another important category of industrial catalysts. Reducible oxides undergo changes in catalytic atmospheres often associated with Mars van Krevelen-type mechanisms in which oxygen atoms for the reaction come from lattice oxygen and the resulting oxygen vacancies are refilled by dissociation of molecular O2. Consequently, there has been significant interest in exploring the changes that take place in oxide nanoparticles under redox conditions. While catalysis takes place on the surface, the generation or filling of oxygen vacancies on the surface could vary the subsurface or bulk structure of the catalyst. 5.3.2.1. Cerium-based oxides. Ceria and doped ceria play significant roles in numerous oxidative catalysis processes in chemical transformation and environmental remediation. For example, they play important roles in three-way automotive catalysis and have potential applications in intermediatetemperature solid oxide fuel cells.233−238 The capability of ceria to generate atomic oxygen for oxidative catalysis is evidenced by the direct observation of soot gasification at the soot−ceria interface.239 The rate of gasification of soot in contact with ceria in O2 was much greater than soot in contact with inert oxide, alumina. Due to the reducibility of Ce4+ to Ce3+, oxygen vacancies can be generated readily in a reducing environment such as H2 or CO. The high activity of oxygen vacancies can readily dissociate molecular oxygen or reactants with a relatively low activation barrier; thus, oxygen vacancies can be readily refilled in an oxidative environment such as O2. This reversible process of releasing oxygen atoms and refilling oxygen vacancies makes ceria active for numerous reactions. The capability of generating oxygen vacancies can be tuned by doping ceria with other elements. For example, doping Zr cations can increase oxygen storage capacity at low temperatures (e.g., refs 238 and 240−243) and enhance thermal stability at high temperature. In situ electron energy-loss spectroscopy (EELS) in the ETEM has been employed to determine the local oxidation state of the cerium in subsurface and bulk of ceria NPs and thus measure the density of oxygen vacancies in cerium-based oxides.244−246 As shown in Figure 49, the two peaks in the EELS spectra, Ce M5 and Ce M4, are known as white lines and they undergo a reversal in intensity when the Ce oxidation state changes. Their relative intensity can be used to explore the evolution of oxidation states and generation of oxygen vacancies in bulk ceria because the identification of oxygen vacancies in bulk is challenging for STM and XPS. This method has been used to investigate the reversible redox activity of individual CexZr1−xO2 nanoparticles.247 It has also been employed to study restructuring and reversible oxygen vacancy ordering under reducing conditions.247,248 Figure 50 shows a redox cycle for a CeO2 nanoparticle where the sample is heated and cooled in 0.5 Torr of H2. Compared to the EELS spectrum of ceria at 245 °C in 0.5 Torr H2 (Figure 50c), the Ce M5 and Ce M4 white lines (the lower panel of Figure 50f) reverse at 693 °C, which shows an abrupt change in the oxidation state. Simultaneously, superlattice fringes appear in the atomic resolution image (Figure 50d) and superlattice spots caused by ordering of oxygen vacancies appear in the electron diffraction pattern (Figure 50e). The reaction is completely reversible and the structure vanishes on cooling to 600 °C as the ceria reoxidizes. The phase transformation is associated with a doubling of the lattice parameter and the formation of a bodycentered cubic structure.

Figure 48. Structure−reactivity relation for activation of Ni/SiO2 for partial oxidation of methane during temperature ramp-up in flowing CH4 and O2.

resolution ESTEM EELS shows that the bimetallic particle is not compositionally uniform with regions rich in either Ru or Ni. This is not surprising given the phase diagram for this system because Ni is fcc and Ru is hexagonal close-packed (hcp) under these conditions. This heterogeneous metallic nanoparticle performs partial oxidation of methane to produce syn gas. Ni is also active for many CO-based catalytic reactions such as Fischer−Tropsch synthesis. The structure of Ni catalyst in CO was studied with ETEM.197 The surface morphology of Ni nanoparticles (NPs) supported on TiO2 in 0.2 Torr of 5% CO/ N2 was investigated when the Ni/TiO2 was at 350 °C.197 At an early stage of the exposure to CO, (111) is the preferential exposed facet due to its low surface energy.91 With longer CO exposure, the Ni (111) facet shrunk because it was not thermodynamically favored in the CO atmosphere. On the basis of the vaporization of Ni(CO)n formed on Au/Ni(111) in CO revealed with HP-STM, the shrinkage of Ni nanoparticles could result from the formation of vapor of nickel carbonyl in the presence of CO. Upon shrinkage, the surface was crystallized at 350 °C with an exposed (100). 5.3.1.5. Other metal catalyst systems. Pd is active for many important reactions of chemical transformations and environmental remediation including oxidative and reductive catalysis. Structures of Pd catalysts in O2 or H2 were studied with ETEM.88,91 In terms of Pd NPs/TiO2 in O2, its shape was changed in contrast to UHV. Wulff−Kaishev analysis of the Pd NPs was performed. The extracted adhesion energies between Pd and TiO2 in vacuum and O2 are 0.7 and 0.4 J/m2, respectively.88 The decrease of adhesion energy likely results in the change of shape of Pd nanoparticles/TiO2 when exposed to O2. By exposing Pd nanoparticles in H2 with a pressure of 4 bar, Pd nanoparticles transform to palladium hydride in H2 through a nucleation revealed in electron diffraction and imaging.91 Ru is well-known for its high activity in ammonia synthesis. Structural evolutions of Ru catalyst NPs in N2 and H2 were studied.232 Both the catalysts consisting of Ba-promoted and unpromoted Ru particles supported on BN were investigated. Upon exposure of the as-prepared catalyst to a mixture of H2 and N2 with a pressure of 4 Torr at 550 °C, Ru NPs smaller than 1 nm diffused across the BN surface and then sintered, forming clean Ru NPs of 2 nm. The Ba promoter enhances the activity for ammonia synthesis by more than an order of magnitude. In the presence of promoter, the crystalline BaOx phase partially covered the Ru nanoparticle surfaces. The lack of a significant 3516

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

to the strong metal−support interaction. This phenomenon was clearly demonstrated in the Pt/TiO2 system with ETEM. Clean Pt nanoparticles without any oxide shell are supported on rutile titania support at 300 °C in gas of H2.174 Surprisingly, a few monolayers of reduced titania covered the surface of a 10 nm Pt particle when the temperature was raised to 450 °C. In some cases, covering the entire metal particle with a layer of support oxide may deactivates the catalyst, although the uncovered surface can still participate in the catalysis. On the other hand, it could enhance the catalysis slightly because the formation of a thin partially covered oxide shell is an extension of the metal− oxide interface. A similar phenomenon was observed on Ni supported on CeO2 in a methane atmosphere by using ETEM.250 Partial coverage of reducing oxide on the surface of metal nanoparticles could play an important role in many catalyst systems because such coverage would extend the interfacial zones where Mars van Krevellen processes can occur. Further work is necessary to explore the relevance of this phenomenon to real catalysts. Another example of nonlocal metal−oxide−gas interactions is related to the spillover phenomenon, in which molecules dissociate on metal particles and the dissociated fragments spill to the oxide support and undergo a further reaction on the surface of the oxide. In situ EELS was used to characterize a hydrogen spillover in nickel-loaded praseodymium-doped ceria.251 A preferential reduction of ceria in the vicinity of the Ni particle at temperatures much lower than the reduction temperature of praseodymium-doped ceria was observed. The experimentally measured spatial extent of the reduction zone was found to be ∼20 nm. 5.4. Deactivation and regeneration of catalysts

Catalysts can be deactivated by poisoning, fouling, sintering, and phase transformations.252,253 Understanding the mechanism of deactivation is significant for developing catalysts with high durability and finding methods for regenerating activity. 5.4.1. Metal nanoparticle sintering. Sintering likely results in deactivation for many supported metal catalysts working at high temperature. Typically, the surface area of catalyst nanoparticles is largely decreased along with the formation of large catalyst particles; thus, catalytic activity gradually decreases. For catalysis at a temperature significantly lower than the melting point of the catalyst, particle sintering occurs either through Ostwald ripening or particle migration and coalescence. In the last two decades there has been considerable interest in developing a fundamental understanding of the sintering process at the atomic level in order to develop strategies to prevent sintering.254 ETEM has been used to investigate sintering processing of supported metal nanoparticle catalysts under different reactive gas environments. Liu et al. investigated the sintering of Pd on an industrial α-Al2O3 support in 0.5 Torr of H2O vapor at 700 °C over a period of 8 h.255,256 It is important to control electron beam effects so that the observed sintering phenomena solely result from interaction of gas and heat with the sample. To reduce the exposure of the catalyst particles to electron beam, the electron beam was turned on only once every hour for rapid image recording. The mean particle size increased from 7 to 10 nm, and image sequences showed that small particles shrunk and larger particles grew, demonstrating that Ostwald ripening was the dominant mechanism in this case. Sintering of Pt particles supported on thin amorphous Al2O3 and SiO2 films257−259 was investigated. These systems are

Figure 49. EELS study of Rh(1%)/Ce0.8Pr0.2O2−x catalyst. Spectra recorded under H2 pressure (4 mbar) at the indicated temperatures. The peak positions for some of the characteristic Ce and Pr species are marked by the solid and dashed vertical lines, respectively. Reproduced with permission from ref 246. Copyright 2003 Royal Society of Chemistry.

5.3.2.2. Vanadium oxide in selective oxidation. Evolution of extended defects in vanadyl pyrophosphate catalysts for selective oxidation was investigated with ETEM.173,249 In a reducing atmosphere of n-butane at 400 °C, oxygen vacancies form in the surface region, resulting in strain between the surface region and the underlying bulk layer. To accommodate the strain, the crystal responds by the formation of extended defects at subsurface sites via a glide shear mechanism. This process can continue for a period of days and give rise to highly defective structures, which may strongly influence the catalytic activity. These defects sites may play an important role in activation of alkane, especially for dehydrogenation. 5.3.2.3. Nonlocal metal−oxide−gas interactions. The catalytic performances of many metal/oxide bifunctional catalysts rely on the interactions between metal nanoparticles and a reducible oxide support. In these metal/oxide catalysts, the interaction may be highly localized to the interface between the metal and support. Alternatively, the interaction between the metal and support199−201 may also give rise to structural changes beyond the interfacial contact area, which can be investigated with ETEM. One of the structural changes beyond the interface is the formation of partial oxide shells on metal nanoparticles due 3517

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 50. Dynamic changes in atomic-level structure characterized by high-resolution imaging (upper) and electron diffraction (middle) and oxidation state change determined from EELS (lower) at (a−c) 245, (d−f) 693, and (g−i) 600 °C in 0.5 Torr of H2 along the [21−1] zone axis from the same individual ceria nanoparticle during a redox cycle. After reduction (d), arrows in the high-resolution image and extra spots along the (220) plane in the electron diffraction pattern show the superstructure formed during reduction. The reversal of the Ce M5/M4 white lines indicates the reduction and reoxidation from Ce4+ to Ce3+ and from Ce3+ to Ce4+ during the heating and cooling, respectively. Reproduced with permission from ref 248. Copyright 2009 American Chemical Society.

Figure 51. Pt nanoparticles supported on SiO2/Si3N4 acquired at 560 Pa O2 at 550 °C. Panels (a) and (b) were acquired after 30 min, and panels (c) and (d) were acquired after 10 h. At medium magnifications (c) and (d) the appearance of anomalously large particles is evident. Reproduced with permission from ref 260. Copyright 2012 American Chemical Society.

Ostwald ripening over a 7 h period. Ostwald ripening involves atoms leaving small nanoparticles which may diffuse across the support and attach to larger particles. The concentration of diffusing atoms on the support immediately adjacent to the small particle is generally assumed to be larger than the mean

considerably simpler than industrial supports and are well-suited to determine fundamental mechanisms and parameters associated with sintering models. For Pt nanoparticles on Al2O3, experiments conducted in 10 mbar of air at a temperature of 650 °C provided clear evidence for coarsening taking place via 3518

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

concentration of adatoms on the support. In the simpler form of ripening theories, a mean field approximation is made; it assumes that the concentration of diffusing atoms on the support is constant beyond some critical distance away from the particle. Simonsen used a mean field theory to model the rate of change of the distribution with time and found that this gave good agreement with experiments but that there were local deviations in certain regions of the support.259 Sinterings of Pt particles on SiO2 amorphous film at 500 °C in O2 and Pd on an amorphous carbon film at 600 °C in H2260 were investigated. In both cases, sintering proceeded via Ostwald ripening and much of the distribution could be described by a mean field theory that assumes that the concentration of diffusing atoms on the support is constant beyond some critical distance away from the particle. In addition, these authors also found an anomalous growth model of large particles, which was not predicted by the mean field theory. Figure 51 shows a sequence of images taken from the same area of the Pt/SiO2 sample. In Figure 51, the large particles (Figure 51b) were not formed from nonuniformity in the metal distribution. In fact, they were formed in the area very similar to any typical areas in Figure 51a. Thus, the mechanism of the formation of large Pt nanoparticles in this case is not clear at this moment but it also occurred in the Pd sample. Ostwald ripening was the sintering mechanism for Pt on amorphous planar SiO2 annealed in air at 650 °C, and the data were well described with a mean field theory but with local variations in behavior.259,257 The local variations were associated with nonuniform particle aggregates consisting of particles of irregular size and distribution as shown in Figure 52a. Figure 52b shows the change in size of each particle during the sintering. A kinetic model was developed. It took the local mass distribution into account using a Voronoi weighting scheme and was able to make a reasonable prediction of the particle-size evolution starting from the initial particle configuration shown in Figure 52c. ETEM showed that the sintering of Ni/MgAl2O4 at 750 °C in a mixture of H2 and H2O at a pressure of a few Torrs proceeded mostly via Ostwald ripening.261,262 Challa et al. analyzed the evolution of the system to extract the energetic parameter for a mean-field ripening model, allowing them to predict ripening behavior over a short time period. ETEM work on sintering seems to conclusively show that Ostwald ripening is the primary mechanism for sintering under oxidizing and reducing conditions on a variety of metal/support systems. This mechanism can apparently lead to long tails in the distribution of particle size, which may be caused by local but subtle variations in the particle size and shape. Further research is required to investigate the effect of local fluctuations in the initial particle-size distribution and the evolution of the particle-size distribution with more sophisticated ripening models including gas-phase transport. 5.4.2. Other deactivation mechanisms. Another type of deactivation of catalysts is the carbon deposition in which carbon is built up on the surface of a catalyst and essentially blocks the access of reactant molecules to active sites. Different types of carbon depositions such as pyrolytic carbon, encapsulating carbon (gum), and whisker carbon, depending on the reaction condition, were reported in the literature. Compared to the large number of in situ ETEM studies of the growth of carbon materials including carbon nanotubes, carbon fibers, and graphene (see section 6.1), there are few in situ studies on carbon deposition and deactivation. Although carbon deposition

Figure 52. (a) TEM image of area containing Pt nanoparticles on SiO2 support from sintering experiments. Nanoparticles are labeled 1−9, and their size is plotted as a function of time in (b). (c) Predicted time evolution of particle radii using a kinetic model that takes the local mass variations into account. Reproduced with permission from ref 259. Copyright 2011 Elsevier, Inc.

is a well-known source of deactivation of catalysts, fundamental understanding of this process is very limited. More effort in investigating the carbon deposition by using ETEM is definitely necessary. Coke formation on a Pt/MgO system263 was studied with ETEM and ex situ aberration-corrected TEM. In situ measurements were obtained after exposing the catalyst to isobutene at 475 °C, and ex situ observations were obtained after exposing the catalyst to mixtures of ethane and hydrogen at 600 °C. Graphene sheets preferentially formed at the steps on the Pt nanoparticles. For Pt nanoparticles greater than 6 nm, the surfaces of 3519

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 53. Photodecomposition of Cu2O on exposure to light and H2O. (a) TEM image of initial particle. (b) TEM image of final Cu particles. (c and d) Electron diffraction. (e) Electron energy-loss spectra. Reproduced with permission from ref 222. Copyright 2012 Institute of Physics.

annealing the catalyst in air at 350 °C. The Pd particles catalytically gasified the green oil and moved through the organic matrix, generating tunnels and tracks. Pd particles that had chance collisions during this process coalesced, resulting in a coarsening of the metal particle distribution. This observation was able to explain why the Pd particles underwent significant coarsening at the relatively low temperatures used for the catalyst regeneration.

nanoparticles were completely covered with graphene layers. For Pt nanoparticles smaller than 6 nm, the graphene formed into tubes or spilled over onto the adjacent support. The sizedependent behavior of the carbon formation was attributed to an accommodation of the strain energy in the growing carbon sheet.263 This fundamental understanding about coking could help the design of new catalysts for either enhancing or suppressing carbon formation. Loss of activity can also be associated with phase segregation in bimetallic nanoparticle. There has not been much work conducted in this area by ETEM, but a recent work showed that Pd segregation to the surface during CO oxidation is the likely cause of the loss in catalytic activity in AuPd alloy catalysts.264 5.4.3. Catalyst regeneration. Similar to carbon deposition, catalyst regeneration has received relatively little attention. ETEM was used to investigate the sintering process that takes place during regeneration of Pd/Al2O3 by removing hydrocarbon deposits (so-called green oil) that build up on the catalyst during the catalysis of hydrogenation.255,256 The green oil formation in the used catalyst resulted in the Pd metal NPs being lifted off the Al2O3 support, and essentially the Pd particles were mostly embedded in the organic medium. The regeneration involved

5.5. In situ studies of catalyst under other reaction conditions and stimuli

5.5.1. Photocatalysis. Photocatalysis, where light irradiation provides most of the energy for a reaction, is becoming increasingly important for environmental remediation and removal of organic waste.265,266 Photocatalysis also has great potential for sustainable energy development, especially in the area of solar fuels.267−269 To understand structure−reactivity correlations of the photocatalyst in the process of catalytic reaction by ETEM, it is necessary to introduce light as a stimulus. This is usually achieved by installation of an optical fiber to introduce photons to the sample in the ETEM. Some work has been conducted on photocatalytic reactions inside the electron microscope. An 3520

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 54. Changes in surface of anatase nanoparticle under gas-phase photocatalytic water-splitting conditions. (a) Typical anatase particle before exposure to reaction conditions. (b) Typical anatase particle after exposure to light and water showing formation of disordered surface layer. The inset show the energy-loss spectra taken from the surface (red) and center (blue) of the nanoparticle indicating Ti4+ in the center and Ti3+ in the disorder layer. Reproduced with permission from ref 121. Copyright 2013 American Chemical Society.

Figure 55. Image sequence separated by 0.5 s time intervals of a growing carbon nanofiber using a Ni catalyst in 2.1 mbar of 50/50 mixture of CH4/H2 at 536 °C. The scale bar is 5 nm, and drawings are included to guide the eye. Reproduced with permission from ref 280. Copyright 2004 Nature Publishing Group.

optical fiber was attached to a high-vacuum TEM and the photodecomposition of polyhydrocarbon compounds on a TiO2 film was studied.270 In addition, photocatalytic nucleation of Au nanoparticles on TiO2 NPs under UV irradiation271 was investigated. Light-illumination sample holders can also be designed for ETEM.119 These holders were employed to investigate the photodecomposition of the photocatalyst, Cu2O, to Cu under UV irradiation in a H2O environment as shown in Figure 53.272,273 The formation of metallic Cu NPs was characterized with electron diffraction (Figure 53c) and EELS (Figure 53e).273 A fiber-based photoexcitation system for ETEM, which permits control of temperature, gas environment, and light, was developed by the Crozier group.120 This photoexcitation system has been used to investigate changes taking place on the

surface of anatase nanocrystals during photoinduced splitting of water molecules in vapor.121 As shown in Figure 54, the top one or two monolayers of TiO2 nanoparticles amorphize under exposure to 1 Torr of H2O at 150 °C while being exposed to UV light. The formation of an amorphous layer of TiO2 results from heavy hydroxylation of the surface due to the adsorption and dissociation of H2O onto oxygen vacancies produced on the anatase surface during photon illumination. More recent work has focused on cocatalyst photocorrsion and its role in catalyst deactivation.274 5.5.2. Catalysis in a liquid. Rather few works have been conducted in the electron microscope on liquid-phase catalysis. A liquid cell was designed by Gai et al. that allowed liquid to flow directly over the surface of a catalyst while making TEM observations.275,276 With this liquid cell, Co-promoted and Au3521

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

are deposited onto the surface of a semiconductor. The sample is then heated until a semiconductor−metal eutectic liquid drop is formed. This liquid drop is the catalyst for this reaction. The reactant (e.g., silane or germane) is introduced, and it adsorbs and dissociates on the surface of the eutectic. This Si or Ge dissolves in the eutectic drop, causing supersaturation; then the growth materials precipitate at the bottom of the droplet, resulting in the formation of a semiconducting nanowire. The catalytic eutectic particle can either be a liquid, which follows a vapor−liquid−solid (VLS) mechanism, or be a solid, which gives a vapor−solid−solid (VSS) mechanism. The first application of operando ETEM to image the vapor− liquid−solid growth process in semiconductors was made by Wu and Yang.292 Many other groups have employed ETEM to learn about the fundamental steps taking place during nanowire growth.293−295

promoted Ru/TiO2 catalysts were directly imaged when they were catalyzing the formation of hexamethylenediamines from liquid adiponitrile. It is found that the catalyst maintains its structural integrity during product formation. In addition, water splitting on Au NPs on TiO2 in an ionic liquid was successfully performed in a liquid cell.271 Recently significant advances in development of liquid cell for TEM studies of nanoparticle growth in liquid have been made and atom-resolved images have been obtained.102−105,277−279 It is expected that these newly designed liquid cells will enable in situ TEM studies of heterogeneous catalyst under liquid reaction conditions.

6. OPERANDO STUDIES OF CATALYSTS USING ETEM 6.1. In situ studies of catalytic growth of nanomaterials

ETEM has been used to observe catalytic growth of nanowires and nanotubes. These nanostructures are grown on catalysts from gas-phase precursors. Typically, a gaseous precursor was introduced to the electron microscope and catalytically decomposed onto catalyst in the microscope. The nanowires and nanotubes were directly observed and recorded in real time during growth. Local measurement of growth kinetics can be carried out and then correlated with local catalyst structure, which provides the operando feature of this type of experiment. The groups at Haldor Topsoe and Arizona State University were the first to use in situ electron microscopy to study the nucleation and growth of carbon nanotubes (CNTs).280,281 In these experiments, Ni, Co, or Mo catalyst nanoparticles were loaded onto oxide supports and then exposed to reactant gases that contained either methane or acetylene. Carbon nanotubes or nanofibers were observed to grow from the metal particles. The growth kinetics studies showed that the resulting number of walls and bends in the tubes depends on gaseous precursor, temperature, and pressure.282,283 The changes in shape of Ni nanoparticles during growth of carbon nanotubes were correlated with the formation of the graphene sheets on Ni catalyst. The formation of carbon nanotubes was assisted by a dynamic formation of monatomic step edges of the nickel nanoparticle surface, as shown in Figure 55. Their observations were consistent with a growth mechanism involving surface diffusion of carbon and the restructuring of nickel catalyst nanoparticles. The growth of carbon nanotubes on an iron catalyst supported on SiO2 was investigated with ETEM.284 It was found that a fluctuating cementite phase (FeC3) was the active catalyst for the grown of carbon nanotubes.284 A further study showed that addition of Mo can suppress the formation of inactive silicate phases and the active phase of the catalyst is M23C6 where M consists of both Mo and Fe. This is consistent with the report from Sharma and co-workers that FeC3 was the active phase for carbon nanotube growth, although metallic Fe has also been proposed as the active phase for the growth.285 Ni nanoparticles doped with Au were also active for carbon nanotube formation; there was some evidence for the formation of NiC3.286 There was no evidence presented for carbide formation in FeCu bimetallic systems.287 The role of water vapor in the growth and gasification of nanotubes was investigated under a variety of conditions,288 and the deactivation mechanism that limits the total tube length was studied.289 In addition, localized oxidation of nanotubes has also been investigated with ETEM.290 Operando ETEM has also been extensively employed to study the dynamics of semiconducting nanowires growth.291 In this approach, catalyst precursors in the form of metal nanoparticles

6.2. Gas-phase reactions

Continued improvements in gas-analysis techniques now allow operando TEM to be performed for gas-phase catalytic reactions. Most ETEMs are equipped with a residual gas analyzer that is employed to measure the composition of the vacuum in the column and the reaction cell. Giorgio et al. used such a setup to confirm CO conversion to CO2 over an Au catalyst in experiments performed in a windowed cell reactor in a TEM.220 They were able to correlate the CO2 signal with the corresponding shape of the Au nanoparticles, although no quantitative conversion or activity data was presented. In the larger cells associated with differentially pumped ETEMs, the quantification of the mass spectrometry data is somewhat challenging because the mass spectrometer must operate under high-vacuum conditions; careful consideration should be given to the gas-diffusion pathways from the reaction cell to the spectrometer inlet. Even if no gas is admitted into the reaction cell, the mass spectrometer will show many peaks associated with the background gas of the reaction cell. With the open architecture and high surface areas of the differentially pumped ETEM, care must be exercised to ensure that the observed catalytic products result from catalysis on the surface of the catalyst and not other components of the reaction cell such as the objective lens pole pieces, the liner tube, or the cell walls. A solution to this background problem is to dramatically increase the amount of the catalyst loaded in the TEM. An operando pellet formed from pyrex or quartz wool allows high catalyst loadings in the TEM so that high catalytic conversion can be achieved.296 Control experiments with and without the catalyst powder should be performed to confirm that the detected products are predominantly formed from the catalytic reaction on the loaded catalyst powder. In addition to mass spectrometry, electron energy-loss spectroscopy (EELS) can be used to quantitatively measure catalytic products formed from a catalyst. It has been demonstrated that EELS in the TEM can be used to quantify the gas composition in the reaction cell.297 EELS is not as sensitive as mass spectrometry but has the advantage that it directly measures the composition of the gas in the reaction cell because the electron beam directly passes through the gas in the reaction cell. Standard methods for quantifying the inner-shell edges in the energy-loss spectra can be applied to the analysis of gas composition. In addition, gases as light as H2 can be quantified with the low-loss region of the spectra as well. A comparison of EELS and mass spectrometry for ETEM applications has been published.298 3522

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

geometries and the amount of catalyst present. For this reaction, the difference in reactant pressure in the ETEM (a few Torr) and the quartz tube reactor (∼760 Torr) seems not to affect the evaluation of catalytic performance. It is noted that it is challenging to use the current design of ETEM reaction cells to determine reaction kinetics. For example, the current differentially pumped systems behave very differently from standard plug flow reactors. Measurement of the spatial variation in gas composition and the residence time suggests that the gas environment around the pellet is more similar to a continuously stirred tank reactor.296 Despite these limitations, the quantitative determination of products allows relative changes in reaction kinetics to be determined. The capability of measuring relative kinetics is illustrated in Figure 58 for CO oxidation over a Ru catalyst.298 The catalyst is initially in the form of RuO2 and is heated progressively to 500 °C and then cooled to room temperature in a flowing mixture of CO and O2. As shown in Figure 58, at 300 °C during ramping-up, the molar fraction of CO2 in the mixture of the gases of this reaction cell is 0.04 whereas it is 0.3 at 300 °C during the ramping-down process. Because the reactor geometry remains constant, the change in CO conversion is a consequence of an increase in catalyst activity. This demonstrates that changes in reaction kinetics can be measured with this approach. A remarkable demonstation of the potential power of operando ETEM was the recent work investigating a wellknown oscillation in the conversion of CO to CO2 that occurs over Pt.299,300 The operando studies were conducted in a MEMS-based windowed cell at a pressure of 1 atm. A residual gas analyzer (RGA) was employed to monitor CO, O2, and CO2 with a 0.1 s dwell time and 1.8 s total spectrum acquisition time. This did not define the temporal resolution of the technique, however, which was determined by space-time broadening of the gas in the tubing connecting the nanoreactor to the RGA. This broadening was found to be on the order of 20 s, with a delay of 5 s between gas leaving the nanoreactor and being detected by the RGA. Fast oscillations in the reaction rate (which were still detected by monitoring the heater power used to maintain a constant nanoreactor temperature) could therefore not be observed using the RGA (Figure 59a), but oscillations with a period of about 30 s were clearly resolved (Figure 59b). These authors were able to correlate the variations in the gas composition in the cell with atomic-level images that showed oscillations in the faceting of the Pt nanoparticles. This is the first operando ETEM experiment

One example of analysis of gas products using EELS is the detection and quantification of products of CO oxidation performed on a Ru-based catalyst loaded in the ETEM.84 The product of CO oxidation, CO2 was detected by the appearance of a second π* peak at the onset of the carbon K-edge in the EELS spectrum. This π* peak of CO2 is 3 eV higher in energy than the π* peak of CO, making it easy to be detected by EELS. Figure 56

Figure 56. Background-subtracted energy-loss spectra acquired at different temperatures during CO oxidation on Ru/SiO2 catalyst in an ETEM reactor. The small peak indicated with the red arrow is the π* peak of the carbon K-edge of CO2 and grows steadily with increasing temperature, proving that catalysis is taking place in the ETEM reaction cell. Reproduced with permission from ref 84. Copyright 2012 American Chemical Society.

shows the change in the energy-loss spectra as a function of the catalysis temperature while the mixture of CO and O2 with a total pressure of 1 Torr is flowing through the ETEM reaction cell. At 150 °C a second peak appears and grows as the catalyst is activated for transforming CO to CO2 in the cell. The spectra can be quantified to determine the conversion of CO. Parts a and b of Figure 57 show conversion of CO as a function of catalysis temperature determined by operando TEM and determined from an ex situ quartz tube reactor, respectively. Their overall shapes of the two conversion curves are very similar to the same light-off temperature of ∼150 °C. The absolute values of the conversions are different because of differences in the reactor

Figure 57. Plot showing the CO conversion during CO oxidation reaction on a Ru/SiO2 performed in (left) ex situ quartz tube reactor and (right) measured from in situ energy-loss spectroscopy. Reproduced with permission from ref 84. Copyright 2012 American Chemical Society. 3523

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 58. Plot showing results from an operando experiment in which the Ru−RuO2 catalyst was heated from room temperature to 460 °C and then cooled back to 100 °C. The hysteresis shows that the catalyst is relatively inactive on the ramp up and much more active on the ramp down. Inset highresolution images were taken at different points in the heating and cooling cycle. The initial RuO2 converts to hcp Ru metal at ∼500 °C, giving a much more active form of the catalyst. Reproduced with permission from ref 298. Copyright 2014 Cambridge University Press.

Figure 59. Atomic-scale visualization of the dynamic refacetting of a Pt nanoparticle during the oscillatory CO oxidation. (a−e) TEM images show the more spherical shape (a,c,e) and the more facetted shape (b,d), during the oscillatory reaction. Fast Fourier transforms included as insets in c−e reveal a lattice spacing corresponding to the Pt(111) lattice planes. Reproduced with permission from ref 299. Copyright 2014 Nature Publishing Group.

HP-STM has been well used to characterize surface structures on both (1) single-crystal model catalysts and (2) metal or oxide nanoclusters grown on a flat surface; in these HP-STM studies, the size of the nanoclusters can be as small as 1 nm. Thus, from the size point of view, both ETEM and HP-STM can work in a similar regime of size of catalyst particles. It is noted that ETEM and HP-STM cannot work on the same sample in most cases because of their different principles of imaging. HP-STM cannot be used to study the nanoparticles with weakly adsorbed H2O or carbon-contained species and nonrigid capping layer of surface molecules even if they can be loaded on a flat support such as Au(111) or graphite surface because of the lack of a rigid interface between catalyst nanoparticles and the flat surface. For STM studies, clean, naked 1−10 nm nanoclusters are typically grown on a flat support in UHV through e-beam evaporation or different sputtering-deposition techniques followed by annealing in UHV. An example of atom-resolved STM imaging of Pd

that correlates real-time changes in nanocatalyst structure with activity.

7. INTEGRATION OF HP-STM AND ETEM Both HP-STM and ETEM are powerful approaches for characterizing the atomic structure of catalysts under reaction condition and during catalysis. HP-STM can provide information on the lateral packing of catalyst atoms and defects of the topmost surface as well as the spatial location and size of adorbates on the surface of a single-crystal model catalyst or nanoclusters grown on a flat substrate. ETEM can provide projection information on structure and composition of highsurface-area nanoparticle catalysts. Integration of the complementary information from both techniques can potentially provide a much deeper understanding of the complex processes taking place on the surface of a catalyst. 3524

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 60. STM studies of metal nanocluster model catalyst supported on Al2O3 thin film on NiAl. These Pd nanoclusters were grown on a flat support. They are nanocluster model catalysts that can be well characterized with STM. (a) STM image of a large scale of Pd nanoclusters with a size of 2−8 nm; the range of size is very similar to that for a high-surface-area metal nanoparticle catalyst. (b) and (c) Atom-resolved surfaces of Pd nanocluster model catalysts. Reproduced with permission from ref 301. Copyright 1999 American Physical Society.

Figure 61. ETEM images of Pt nanoparticles supported on CeO2 and the corresponding three-dimensional model. (a) In high vacuum at 25 °C. (b) 3D structural model of image (a). (c) In 1 mbar CO at 25 °C. (d) In mixture of CO and air with a pressure of 1 mbar at 25 °C. (e) 3D structural model of image (d). Reproduced with permission from ref 226. Copyright 2011 Japan Society of Applied Physics.

nanoparticles is shown in Figure 60.301 From a size point of view, the sizes of catalyst particles studied with STM and TEM are definitely in the same range. The lateral packing of catalyst atoms and adsorbates on the topmost surface of the catalyst revealed with HP-STM will be complementary to the structural and chemical information on high-surface-area nanoparticle catalysts revealed with ETEM.

sample region of the ETEM, the surface morphology of Pt nanoparticles in vacuum (Figure 61a) changed to a round shape, as shown in Figure 61c. The clear observation of the round shape of the same nanoparticle in CO suggested that the Pt nanoparticles are restructured in CO at Torr pressure. The restructuring is understandable because the CO molecules in the gas phase of CO around the Pt nanoparticles directly interact with the surface of the nanoparticle. However, the image in Figure 61c cannot provide any indications of the lateral arrangements of Pt atoms on the surfaces and whether defects (adatom, step edge, and vacancies, as schematically shown in Figure 1c and d) are present. HP-STM can provide information on the lateral arrangement of atoms, defect, and adsorbate on the topmost surface. In fact, surface structures of (111), (110), and high Miller index surfaces such as (557) and (332) of the Pt model catalysts were studied with HP-STM in the literature.3 Although HP-STM cannot image nanoparticles synthesized through wet chemistry methods due to the deconvolution of a STM tip, surfaces of model catalysts with large terraces (>10 nm) such as Pt(111) single crystal or with parallel terraces (>1 nm) such as Pt(557) have been successfully used to experimentally simulate how these

7.1. Synergistic exploration of restructuring of metal catalyst at nanoscale

How the pressure of a reactant gas could restructure catalyst particles has attracted significant attention as it is fundamentally important for understanding the surface process performed on a catalyst. ETEM studies of Pt nanoparticles supported on CeO2 in CO or O2 exhibit a reactant-driven restructuring. There is no restructuring for Pt NPs in vacuum (Figure 61a) under a beam voltage of 200 kV and a dosage of 4.0 A/cm2.302 The TEM image simulation indicates that the exposed facets are (100) and (111) (Figure 61b). As the contrast in TEM images results from projection of atomic columns, whether the exposed surface is a high Miller-index surface or not cannot be readily judged by the ETEM images. For instance, when CO was introduced to the 3525

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

on the packing of adsorbed CO molecules on the surface of the Au nanoparticles. HP-STM studies of Au nanoclusters supported on CeO2 surface can be used to visualize catalyst atoms of the topmost surface and the arrangement of adsorbed molecules on the surface. This complementary information on Au nanoclusters to be visualized with HP-STM allows a complete description of the structure of the Au nanoparticle catalyst at atom scale.

surfaces could respond to a reactant gas at a certain pressure. High-pressure STM studies (Figure 15) show that Pt(111), the most closely packed surface, does not restructure in CO gas with a pressure ≥20 Torr.4 Compared to the close-packed (111) surface consisting of Pt atoms with CN (Pt−Pt) = 9, open surfaces such as Pt(110) consisting of Pt atoms with CN(Pt−Pt) = 7−8 and vicinal surfaces such as Pt(557) consisting of Pt atoms with CN(Pt−Pt) = 9 on terrace and CN(Pt−Pt) = 7 restructure into homogeneously distributed nanoclusters on the surface (Figure 32) in a gaseous environment of CO.3,4,161 Due to the projection nature of a single ETEM image, the restructuring on the topmost surface layers cannot be easily addressed. Even if 3D reconstruction techniques are employed, it would be challenging because the electron beam may affect atomic arrangement of the columns of atoms at the surface (marked with a red line in Figure 62). Thus, a marriage of HPSTM and ETEM would be ideal for fundamental studies on the surface structure of catalysts in the gas environment of a reactant.

7.3. Packing and density of oxygen vacancies on topmost surface of reducible oxide

Reducible oxides are one of the most important categories or components of heterogeneous catalysts. They typically provide oxygen vacancies for adsorption or dissociation of reactant molecules, participate in the charge transfer at a metal−oxide interface, and offer oxygen atoms for oxidation reactions. How the oxygen vacancies of a reducible oxide such as CeO2 and TiO2 participate in a catalytic reaction at the atomic scale still remains unclear even for a prototype reaction such as CO oxidation. As shown in Figure 50, the generation of oxygen vacancies in CeO2 during reduction in H2 can be detected by the abrupt change of Ce M5 white lines in in situ EELS. Unfortunately, these changes do not offer any structural information on oxygen vacancies on the topmost surface of a catalyst such as the locations, distribution, and density of surface oxygen vacancies. The generation, distribution, and chemical environment of oxygen vacancies on the surface and subsurface of reducible oxides can be studied with STM. For example, as shown in Figure 64, oxygen vacancies of the CeO2 (111) surface can be clearly identified with STM.303,304 Oxygen vacancies exist in the form of single, dimer, and chain clusters on the topmost surface layer (Figure 64), which are clearly identified with STM because oxygen vacancies appear with dark contrast compared to oxygen atoms on the surface. In addition, statistical analysis can offer information on the density of oxygen vacancies (Figure 64c). In addition, by varying the bias of the tunneling condition, oxygen vacancies at subsurface sites (blue triangles in Figure 64a and b) can be identified with STM. Thus, there is no reason that the oxygen vacancies of the topmost layer and subsurface of CeO2 in a gas phase cannot be identified with HP-STM. It is expected that the distribution and density of oxygen vacancies in the surface region of reducible oxides in reactive gases can be well studied with HP-STM. It will provide information on the packing and density of oxygen vacancies on the topmost surface of reducible oxide nanoparticles, which is complementary to the information from ETEM.

Figure 62. Illustration of the projected image achieved in TEM (a) and 3D structural model. The edge of a projected image was marked with a set of red lines; each red line represents a facet. The surface of a nanoparticle consists of many difference facets. It is extremely challenging to achieve the information on the packing of metal atoms on the topmost layer of each facet of a metal nanoparticle consisting of different facets, although this piece of information is crucial for identifying active sites of the catalyst surface and adsorbates on this surface. Reproduced with permission from ref 226. Copyright 2011 Japan Society of Applied Physics.

7.2. Determining the arrangements of atoms on the topmost atomic layer of a catalyst and adsorbates with ETEM and HP-STM

7.4. Understanding structural evolution of Cu nanoparticle supported on ZnO

It is reported that the (100)-1 × 1 surface of Au nanoparticles supported on CeO2 reconstructs into a hexagonal structure on exposure to CO and O2 (Figure 44). ETEM observation showed two structural evolutions during gas exposure. One is the increase of interplanar distance of Au atoms along [001] (Figure 44, part c versus d); the other is the decrease of interatomic distances of atoms on topmost layer along [010] (Figure 44, part c versus d). DFT calculation suggests that the topmost layer of the surface forms a hexagonal structure (Figure 63e). On the basis of the measured distance of two projected CO molecules (Figure 63a− c), a half coverage of CO on the newly formed hexagonal layer was proposed in Figure 63e and f. Solid evidence for the increase of interatomic distance between the topmost and second topmost layers along the [001] direction and the decrease of distance between two adjacent Au atoms along the [010] direction of the topmost layer is given. However, there is no direct observation of the lateral packing of Au atoms on the topmost surface of Au nanoparticles and no information

Cu NPs supported on ZnO are important industrial catalysts of methanol synthesis from CO and H2. The change of the surface morphology of Cu NPs supported on ZnO in different reactant gases was observed with ETEM.138,305 As shown in Figure 45, the surface of Cu nanoparticles in 1.5 mbar of H2 mainly consists of facets (100) and (111). Notably, it is changed to a surface consisting of (110), (100), and (111) while the gas around the catalyst is changed to a mixture of 1.12 mbar H2 and 0.38 mbar H2O. The fraction of (110) in the mixture of H2 and H2O is increased, which is due to the introduction of the oxidizing gas, H2O. The fraction of (110) facets is decreased in a reducing gas environment such as a mixture of 1.42 mbar H2 and 0.08 mbar CO. This ETEM study suggests that the Cu(110) could be stabilized by adsorption of H2O molecules. Through Wulff construction, the surface energies of (100), (110), and (111) in different reactants were extracted138,306 (Table 2). This 3526

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Figure 63. Aberration-corrected ETEM images and structural model of Au-hex reconstructed surface of Au NP in vacuum and the mixture of 1% CO and air with a total pressure 1 mbar at 25 °C. (A) Image in a vacuum. (B) Image of the mixture of 1% CO and air with a total pressure of 1 mbar at 25 °C; this image was taken using 80 keV electrons using an under-defocus condition. (C) Observed image in the rectangular region in (B) at higher magnification. (D) Simulated image based on an energetically favorable model. The model in (E) plan view along the [001] direction of crystalline gold and (F) cross-sectional view along the [110] direction of crystalline gold to show the undulating topmost Au layer. Reproduced with permission from ref 317. Copyright 2015 Nature Publishing Group.

Figure 64. STM studies of oxygen vacancies of the topmost layer and the subsurface of ceria through filled-state (A) and empty-state (B) images of single vacancies and related structural models (left, surface vacancy; right, subsurface vacancy; characteristic O rim atoms in blue) and (C) calculated density of states (DOS) and simulated filled-state STM images (bias −3.0 V). Ce 4f gap states, displayed as unshaded curves, do not contribute to the STM images because of their strong spatial localization. Reproduced with permission from ref 304. Copyright 2005 AAAS.

suggested that the adsorption of H2O on the (110) surface of Cu/ZnO changes the ordering of surface energy of (110) and (111). Similar to the UHV environment, (110) of Cu nanoparticles in H2 has a higher surface energy than (111) in H2. However, (110) of the Cu nanoparticles in H2O has a lower surface energy than (111) of Cu nanoparticles in H2O.

To identify the lateral arrangement of Cu atoms of the topmost surface of Cu nanoparticles in water vapor, the most appropriate technique is HP-STM because it can visualize the detailed surface structure of Cu nanoclusters at the atomic scale. For example, it would be very important to use HP-STM to investigate the (110) surface of Cu in the presence of water 3527

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Table 2. Relative Surface Free Energies (γhkl and γ*) and Work of Adhesion (W) for a Cu NP Supported on ZnO in Different Gases of Reactants; Values in Parentheses Are Surface Free Energies Calculated for Cu Single Crystals in Vacuum γ100 γ110 γ111 γ* W

H2

H2/H2O

H2/CO

1.08 ± 0.03 (1.10) 1.11 ± 0.02 (1.15) 1 0.09 ± 0.05 0.91 ± 0.05

0.61 ± 0.03 0.57 ± 0.03 0.59 ± 0.02 0.09 ± 0.02 0.91 ± 0.02

1.08 ± 0.03 1.11 ± 0.02 1 −0.54 ± 0.05 1.54 ± 0.05

7.5. Fundamental studies of bimetallic catalysts

Bimetallic catalyst such as Ni−Cu, Ru−Cu, Os−Cu, Pt−Ir, and Pt−Ru has been one of the most important categories of catalysts.307,308 The added guest metal can tune catalytic performance through electronic effects, geometrical effects, or bifunctional effects. Electronic effects can be found in most bimetallic catalysts because the coordination of the guest metal to the host metal typically changes the d-band center and thus modifies the adsorption energies of reactants or intermediates and thus affects catalytic activity. Geometrical effects are realized by changing the fractions of specific sites (such as monomers, dimers, or trimers of atoms of host metal) among all available sites; in most cases the fraction of sites of monomers and dimers of host metal atoms are increased while the fraction of trimer sites is decreased by increasing the concentration of the guest metal.309 The change of fractions of different sites could tune catalytic selectivity as different reaction pathways may require a specific type of binding sites. A bifunctional effect of a bimetallic catalyst results from the binding of reactants A and B to the guest and host metal atoms, respectively. In many cases, these bimetallic catalysts are randomly mixed alloys instead of intermetallics. Thus, atomic packing of host and guest metals on the topmost surface of an alloy is typically different from that of subsurface and bulk.310 One example of the challenge in distinguishing the packing or distribution of the guest metal on the topmost surface of

vapor. In fact, HP-STM has already been used to address a similar catalyst system. As shown in Figure 34, HP-STM studies revealed that the Cu(110) surface is restructured to Cu(110)-2 × 1 in H2 gas with a pressure higher than 20 mbar at 25 °C. As HP-STM can offer atomic details of a catalyst surface up to ≥230 °C in ∼33 mbar gas,5,72 it is expected that HP-STM studies of surfaces of the model catalysts including Cu(110), Cu(100), and Cu(111) single crystals and Cu nanoclusters in H2, a mixture of H2 and H2O, and a mixture of H2 and CO can provide atom-resolved images of the topmost surface of Cu catalyst in different mixtures of reactant gases, by which an intrinsic nature of the Cu catalysts of methanol synthesis may be revealed.

Figure 65. Correlation between catalytic activity in formation of vinyl acetate and the fraction of a pair of Pd monomers. (a) Acetyloxylation of ethylene to form vinyl acetate. (b) Optimized distance between two adsorption sites (Pd atoms) for an effective coupling of the two reactant molecules to form a vinyl acetate and the distance of a pair of Pd monomers on Au(100); the distance between two Pd monomers is reasonable for adsorption of the two reactant molecules and coupling for formation of product molecules. (c) Catalytic activity in terms of turnover frequency of formation of vinyl acetate per second on each Pd sites as a function of Pd coverage in terms of atomic fraction of Pd atoms of the topmost layer of the (100) surface of Au−Pd alloy. (d) Fraction of the pair of Pd monomers as a function of Pd coverage on the topmost layer of the (100) surface of Au−Pd alloy; the pair of Pd monomers is marked with a red oval. Reproduced with permission from ref 309. Copyright 2005 AAAS. 3528

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

temperature changes.311 This makes it possible to perform pulse experiments in ETEM because the temperature fluctuations will not result in large sample drift when the gas composition is changed or the thermal conductivity of the gas is changed.

bimetallic nanoparticles is the Au−Pd system. As shown in Figure 65a, the optimized distance for an effective coupling between acetic acid and ethylene is ∼3.3 Å.309 The measured turnover frequency (TOF) for formation of vinyl acetate quickly increases along the increase of Pd coverage, reaches the maximum rate at the coverage of Pd of 0.10, and then slowly decreases along with the increase of surface coverage of Pd atoms (Figure 65c). This trend was explained by the dependence of fractions of Pd monomers on the Pd coverages as shown in Figure 65d. For the atomic lateral arrangement on the topmost surface of the Au−Pd catalyst schematically shown in the inset of Figure 65d, STM is the most appropriate technique to identify the distribution of guest metal atoms Pd on surface of the Au−Pd alloy catalyst because Au and Pd atoms on a surface can be readily distinguished in STM.309 HP-STM studies can qualitatively identify the lateral packing of Pd atoms on the surface of the bimetallic catalyst, including Pd on a terrace, step edge, and even adatom and vacancy, and quantitatively measure the fractions of the different types of sites (monomer, dimer, trimer, pair of monomers). The information on packing of catalyst atoms on the topmost surface of Au−Pd nanoclusters to be provided by HP-STM and is key to achieve before an intrinsic correlation between catalytic performance and the corresponding surface structure of catalysts at the atomic scale can be built. Definitely, ETEM can provide the overall structure and compositional heterogeneity of high-surface Au−Pd bimetallic nanoparticle catalyst. The combination of the two techniques would provide a deeper understanding of the surface of the bimetallic catalyst.

8.2. Achieving high resolution of images of catalyst surface at high temperature with HP-STM

In addition to the thermal drift, other instrumentation challenges must be addressed in order to apply STM techniques to visualize the surface of a model catalyst with atomic details at a temperature higher than the current limit, 230 °C. A challenging issue is how to keep the STM body in the temperature range of 25−60 °C (closer to 25 °C is better). A smaller separator with an aperture between the STM reactor and the STM body will minimize the diffusion of hot gas from the reactor to the STM body. However, the smallest hole size is the diameter of the tip, typically 0.2 mm. In addition, a better cooling and heating loop with a faster response to any variation of temperature of the STM body is necessary to maintain the temperature of the STM body at a set temperature. If the reactant gas is oxidative, additional challenges must be addressed. In the case of a tungsten tip, oxidization of this tip will form a layer of amorphous tungsten oxide, which makes the surface of the STM tip nonconductive. Even for a Pt−Ir alloy tip, its surface can be readily oxidized at a temperature of 100 °C in O2. Using an Au tip can avoid the problem of oxidation, but the softness of gold makes it relatively difficult to achieve atomresolved images. Alternatively, Au-based alloys or other oxidation-resistant alloys could be used to make an STM tip. Another choice is to use an STM tip made of a conductive oxide or an STM tip coated with a layer of well-crystallized conductive oxide through atomic layer deposition. In addition, using a carbon nanotube as an STM tip could be another choice because it is highly conductive and more resistant to oxidative gases at a high temperature. If this is a choice, it is necessary to develop a repeatable protocol to immobilize carbon nanotubes to a tip holder. For instance, growth of a carbon nanotube on a tungsten tip through CVD could be a realistic approach.

8. FUTURE CHALLENGES AND PROSPECTS HP-STM and ETEM have been applied to the studies of catalyst structures for a couple of decades. New structures and related new chemistries of catalysts under reaction conditions have been revealed. However, significant limitations remain for both techniques. It is expected that the ongoing development of new instrumentation of HP-STM and ETEM will reduce these limitations in the near future. What follows is a discussion of these limitations, challenges, and potential strategies to move each technique forward. 8.1. Thermal drift

8.3. Increasing pressure of reactant gases during in situ/operando studies

Many heterogeneous catalytic reactions are performed at temperatures in the range 100−800 °C. However, the reaction conditions at high temperature have given rise to a challenge in achieving atomic-scale resolution of images. For HP-STM, thermal drift mainly results from the difference of temperature between a catalyst and the STM tip. This makes it difficult to achieve atom-resolved images at a high temperature. The current maximum operational temperature of HP-STM is 230 °C at a pressure of 25 Torr if an atom-resolved image is required.72 The temperature of a catalyst in 25 Torr definitely can be >230 °C if atomic resolution of STM images is not required. In addition, placing the tip close to the sample surface by 20−100 nm during the heating process of the catalyst before imaging could largely reduce the time to reach an approximate thermal equilibrium with the sample.72,78 Compared to the challenging issue of thermal drift in in situ studies using HP-STM, ETEM has less concern with how rising sample temperature could decrease the resolution of images. In situ ETEM studies using MEMS-based windowed cells allow nanocatalysts to be imaged with atomic resolution at elevated temperatures. The MEMS-based sample holders are also very stable with a very small amount of sample drift during

The ultimate goal of fundamental studies of catalysis is to design better catalysts with higher selectivity and activity for a green, least-energy-consumption, and low-cost production of chemical and fuel feedstocks beneficial to our society. Thus, the long-term target of in situ studies is to simulate important catalytic reactions under the reaction conditions relevant to industry. This means that experimental simulation in a research lab should be done at a pressure of 1 bar or higher. On the basis of the imaging principle of HP-STM, there is no limitation for the working pressure as long as the reaction cell can accommodate the high pressure of the reactants and products. From an instrumentation point of view, there are challenges when the pressure is increased. One is a possible gas discharging at a higher pressure. The voltage applied to piezo/materials of the coarse approach motor and scanning tube could result in discharge in the pressure range of 10−100 mbar although it is a rare incident in studies of HP-STM. One solution is to mix the reaction gas (10−100 mbar) with an inert gas with a pressure of 990−900 mbar to bring up the total pressure to 1 bar. Another instrumentation challenge is the sealing of the reaction cell at a higher pressure in the range of 1−10 bar. It is necessary to preserve the UHV environment of the STM chamber while the 3529

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

reaction cell is filled with gas(es) at a pressure higher than 1 bar. The ongoing collaborative design of the Tao group and Specs Surface Nano Analysis GmbH is expected to solve these issues. In terms of in situ studies using ETEM, the limitation for working at a high pressure is the enhanced inelastic scattering of the electron beam by an increased density of reactant gas(es) around the catalyst. To minimize inelastic scattering, a higher working pressure must be accompanied by a decrease of the thickness of the gas path length or an increase in the electron mean free path by using a higher accelerating voltage. Both of these options are expected to be explored in future developments.

ever, such a combination of STM/STS technique is extremely challenging for in situ studies of catalyst surfaces at relatively high temperatures in gas or liquid phase. The chemical identification of adsorbates on the surface of model catalysts or nanocluster catalyst has to be done with in situ electronic or vibrational spectroscopies such as in situ IR. From an instrumentation point of view, it is quite easy to incorporate in situ IR into an HP-STM. In ETEM, a high-spatial-resolution EELS can be readily employed to probe the composition of adsorbates provided that electron beam damage effects can be controlled. The energy resolution of conventional EELS in the electron microcopy is typically on the order of 1 or 0.3 eV if a system is equipped with cold-field emission sources. Commercially available ETEMs are now equipped with monochromators that offer sub-100 meV energy resolution. The recent development of the monochromator has now achieved an energy resolution of ∼10 meV, and vibrational spectroscopy has now been demonstrated.316 The delocalized nature of the interaction allows the vibrational modes to be detected without placing the electron beam directly on the catalyst, which will successfully avoid electron beam damage. In principle, it should be possible to combine ultrahigh-energyresolution EELS with an environmental cell to achieve spatial atomic resolution and energy resolution of tens of meV so that vibrational excitations of adsorbates on a catalyst may be explored.

8.4. Fast and ultrafast imaging of catalyst

Time-resolved in situ/operando studies are important for observing the dynamic structural changes of catalysts under reaction conditions. Definitely, fast or ultrafast in situ imaging techniques are still at an early stage of the development. Development of ultrafast imaging technique and integration with in situ studies is challenging but necessary and significant for fundamental studies of catalytic mechanism. Nanosecond or even femtosecond imaging is the ideal temporal resolution for tracking dynamics of catalyst structures under reaction conditions and during catalysis. For HP-STM, the fastest scan is currently at the level of one frame per millisecond. Absolutely, there is a long way to go to reach the time scale of catalyst dynamics. The barriers are the robustness and rigidness of the STM body and the electronic communication between the STM body and data acquisition. For dynamic studies of a catalyst in ETEM, a high recording sensitivity is required and must also be coupled with faster detector readout rates and shorter readout times. Most current detectors operate at close to TV rates with detection quantum efficiencies (DQEs) significantly smaller than unity. The development of direct detection camera systems promises to revolutionize in situ TEM by offering readout rates of 1000 frames per second (fps) with correspondingly high DQEs.312 While 103 fps will provide new insights into dynamic processes taking place in catalysts, it is still slower than the resident time or life of molecules or intermediates. The goal of so-called dynamic transmission electron microscopy (DTEM) is to image materials with atomic resolution and ultrafast time resolution.313 The current temporal resolution is 10−9 s or better under a UHV condition. It should be possible to achieve atomic resolution imaging with a temporal resolution of ∼10−6 s in future instrumentations. These techniques can be combined with spectroscopy so that chemical information (at least from low-loss EELS) with high spatial and temporal resolution is feasible.

8.6. Catalysts in a liquid phase

As discussed in section 3, there is no technical barrier to imaging the surface of a catalyst in a liquid with STM because the molecular layers of a liquid between the STM tip and the catalyst surface equivalently contribute to the contrast of STM images at different locations. The main application of in situ studies of liquid using STM is to track the surface of model catalysts in liquid under an electrochemical condition. Minimization of the etching rate of the STM tip by an acidic solution during an in situ electrochemical experiment is still a challenging issue, although coating the tip with inert metal may be a solution. In situ TEM experiments in liquid are now being performed by several groups around the world.99−105,277−279 At present the published work has focused mostly on particle nucleation, battery work, and life science applications. Experiments on liquid-phase catalytic reactions are certain to be performed in the near future because more than 2/3 of industrial catalytic reactions are in fact performed at solid−liquid interfaces. The challenge of in situ studies of a catalyst in a liquid phase using ETEM is how to minimize the inelastic scattering of electron beam and beam damage effects. To achieve atom-resolved images of a catalyst in liquid, the strategy is to keep the liquid layer as thin as possible. In addition, the ongoing development has made the addition of electrical biasing in liquid possible; it could open up opportunities for investigating electrocatalytic processes of a model fuel cell and chargeable battery under working conditions.

8.5. Chemical identity of adsorbates

Identification of adsorbates on the catalyst surface is a key component in fundamental studies of catalysis. One of the main functions of HP-STM is the identification of adsorbates. The location and size of adsorbates can be visualized at the atomic scale because the lateral and vertical resolutions of HP-STM techniques can be as high as 1 pm. Even if there are free gas molecules surrounding the adsorbates on a catalyst surface, these adsorbates can be readily identified with STM. For example, the chemisorbed CO molecules on Pt(111) in 1 Torr of CO were clearly visualized (Figure 16). However, it is challenging to chemically identify the adsorbate with STM techniques, although the identity of some adsorbates at cryogenic temperatures could be achieved through a combination of scanning tunneling microscopy and scanning tunneling spectroscopy.314,315 How-

8.7. Correlative approaches

While the primary focus of this Review has been on structural observation with in situ/operando microscopy of catalysts, there is growing recognition for the need to increase the number of correlative approaches that can be performed simultaneously. Operando methods represent one class of such measurements where catalytic activity and selectivity are correlated with its structure, obtained usually from one characterization technique. A recent U.S. Department of Energy workshop on in situ methodologies placed emphasis on the need for more 3530

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

sophisticated approaches to correlative analysis. For catalysis research, bringing together a wide range of operando imaging and spectroscopy techniques simultaneously will dramatically improve our ability to establish structure−reactivity relations. Novel operando cells are being constructive for correlative microscopy. For example, correlated operando studies using ETEM and X-ray absorption spectroscopy have shown complex ensemble behavior of Pt nanoparticles during ethylene hydrogenation.317 The trend in combining sophisticated atomicresolution catalyst characterization with other characterization tools under operando conditions will dramatically accelerate our understanding of the structure−functionality relations in heterogeneous catalysts.

Catalysis Society. He has organized numerous workshops, schools, and symposia on electron microscopy and has published more than 140 archival journal and book articles. He is on the editorial boards of Micron and Microscopy Today and is a member of the scientific advisory boards for the Molecular Foundry, Lawrence Berkeley Laboratory, and the Center for Nanophase Materials Sciences, Oak Ridge National Laboratory. In 2011 he was elected a Fellow of the Microscopy Society of America. Crozier is currently an Associate Professor in the School for Engineering of Matter, Transport and Energy at Arizona State University.

ACKNOWLEDGMENTS F.T. acknowledges financial support from the NSF Career Award NSF-CHE-14162121, Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy, under Grant No. DESC0014561, and the Integrated Mesoscale Architectures for Sustainable Catalysis (IMASC), Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award No. DESC0012573, and National Science Foundation under the grant No. NSF-OIA-1539105 and NSF-CBET-1264798. F.T. acknowledges help from Longhui Nie for requesting licenses for using figures from publishers and from Luan Nguyen and Shiran Zhang for preparing a couple of structural models. P.A.C. acknowledges financial support as follows: Sections of this work have been supported from National Science Foundation under Award Nos. CTS-0306688 (ceria redox studies), CBET 0553445 (NiCu bimetallic catalysts), and CBET 1134464 (EELS of gases, RGA analysis, operando EELS) and U.S. Department of Energy, Office of Science, Basic Energy Science, under Award Nos. DEFG02-07ER46442 (hydrogen spillover) and DE-SC0004954 (photocatalysis, monochromated, and vibrational EELS).

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest. Biographies Franklin (Feng) Tao is a tenured Miller associate professor of chemical and petroleum engineering at Department of Chemical and Petroleum Engineering and of chemistry at Department of Chemistry at University of Kansas. He received a Ph.D. from Princeton University with a following postdoc research at University of CaliforniaBerkeley and Lawrence Berkeley National Laboratory. His research interests include heterogeneous catalysis for chemical transformations of small molecules to chemical and fuel feedstocks and for environmental remediation, application of catalysis to energy transformation, synthesis and catalysis of single-site metal and oxide catalysts, development of in situ and operando characterization instruments/techniques for fundamental studies of catalyst surfaces under reaction conditions and during catalysis, and in situ and operando studies of structure and chemistry of catalysts with ambient-pressure X-ray photoelectron spectroscopy, ambient/high-pressure, high-temperature scanning tunneling microscopy, and environmental transmission electron spectroscopy. He published over 100 research articles. He is an elected fellow of Royal Society of Chemistry. He received Eugene P. Wigner Fellowship (2010), finalist prize of Gerhard Ertl Young investigator award (2011), Paul Holloway award of AVS (2012), NSF-career award, and Miller research award (2014). He is on the advisory editorial boards or editorial boards of several journals including Chemical Society Reviews, Catalysis Science & Technology, and Science China Materials. He is married to Hong Peng and has a daughter, Lauren P. Tao, and a son, Steven P. Tao.

REFERENCES (1) Ertl, G.; Knözinger, H.; Weitkamp, J. Handbook of Heterogeneous Catalysis; Wiley VCH: Weinheim, Germany, 1997. (2) Somorjai, G. A.; Li, Y. M. Introduction to Surface Chemistry and Catalysis; Wiley VCH: Weinheim, Germany, 2010. (3) Tao, F.; Dag, S.; Wang, L. W.; Liu, Z.; Butcher, D. R.; Bluhm, H.; Salmeron, M.; Somorjai, G. A. Break-Up of Stepped Platinum Catalyst Surfaces by High CO Coverage. Science 2010, 327, 850−853. (4) Longwitz, S. R.; Schnadt, J.; Vestergaard, E. K.; Vang, R. T.; Stensgaard, I.; Brune, H.; Besenbacher, F. High-Coverage Structures of Carbon Monoxide Adsorbed on Pt(111) Studied by High-Pressure Scanning Tunneling Microscopy. J. Phys. Chem. B 2004, 108, 14497− 14502. (5) Nguyen, L.; Cheng, F.; Zhang, S.; Tao, F. Visualization of Surfaces of Pt and Ni Model Catalysts in Reactive Environments Using Ambient Pressure High Temperature Scanning Tunneling Microscopy and Understanding the Restructurings of Surfaces of Model Metal Catalysts under Reaction Conditions at Near Ambient Pressure. J. Phys. Chem. C 2013, 117, 971−977. (6) Frenkel, A. I.; Hills, C. W.; Nuzzo, R. G. A View from The Inside: Complexity in the Atomic Scale Ordering of Supported Metal Nanoparticles. J. Phys. Chem. B 2001, 105, 12689−12703. (7) Bare, S. R.; Ressler, T. Characterization of Catalysts in Reactive Atmospheres by X-ray Absorption Spectroscopy. Adv. Catal. 2009, 52, 339−465. (8) Alexeev, O. S.; Li, F.; Amiridis, M. D.; Gates, B. C. Effects of Adsorbates on Supported Platinum and Iridium Clusters: Characterization in Reactive Atmospheres and during Alkene Hydrogenation Catalysis by X-ray Absorption Spectroscopy. J. Phys. Chem. B 2005, 109, 2338−2349.

Peter A. Crozier obtained a B.S. and Ph.D. in physics from the University of Glasgow with specialization in scanning transmission electron microscopy and spectroscopy. He worked as a postdoctoral fellow on electron energy loss spectroscopy at the University of Alberta with Ray Egerton. He then went onto to study microscopy of surfaces under UHV conditions with John Cowley at Arizona State University (ASU). He was Director of the Industrial Associates Program at ASU for 15 years and worked with many chemical and petrochemical companies on characterization problems of industrial catalysts. He has extensive experience in developing and applying advanced transmission electron microscopy techniques to nanomaterials including semiconductors, catalytic materials, and atmospheric aerosols. He has 15 years of experience in developing and applying the technique of in situ environmental transmission electron microscopy to problems in catalytic materials. He is a member of the Microscopy Society of America, the Materials Research Society, and the North American 3531

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(9) Argo, A. M.; Gates, B. C. MgO-Supported Rh6 and Ir6: Structural Characterization during the Catalysis of Ethene Hydrogenation. J. Phys. Chem. B 2003, 107, 5519−5528. (10) Argo, A. M.; Odzak, J. F.; Goellner, J. F.; Lai, F. S.; Xiao, F. S.; Gates, B. C. Catalysis by Oxide-Supported Clusters of Iridium and Rhodium:Hydrogenation of Ethene, Propene, and Toluene. J. Phys. Chem. B 2006, 110, 1775−1786. (11) Singh, J.; Lamberti, C.; van Bokhoven, J. A. Advanced X-ray Absorption and Emission Spectroscopy: In Situ Catalytic Studies. Chem. Soc. Rev. 2010, 39, 4754−4766. (12) Bañares, M. A.; Mestl, G. Structural Characterization of Operating Catalysts by Raman Spectroscopy. Adv. Catal. 2009, 52, 43−128. (13) Ryczkowski, J. IR Spectroscopy in Catalysis. Catal. Today 2001, 68, 263−381. (14) Vimont, A.; Thibault-Starzyk, F.; Daturi, M. Analysing and Understanding the Active Site by IR Spectroscopy. Chem. Soc. Rev. 2010, 39, 4928−4950. (15) Yang, M.; Somorjai, G. A. Adsorption and Reactions of C6 Hydrocarbons at High Pressures on Pt(111) Single-Crystal Surfaces Studied by Sum Frequency Generation Vibrational Spectroscopy:Mechanisms of Isomerization and Dehydrocyclization of n-Hexane. J. Am. Chem. Soc. 2004, 126, 7698−7708. (16) Chen, P.; Kung, K. Y.; Shen, Y. R.; Somorjai, G. A. Sum Frequency Generation Spectroscopic Study of CO/Ethylene Coadsorption on the Pt(111) Surface and CO Poisoning of Catalytic Ethylene Hydrogenation. Surf. Sci. 2001, 494, 289−297. (17) Cremer, P. S.; Su, X.; Shen, Y. R.; Somorjai, G. A. Ethylene Hydrogenation on Pt(111) Monitored In Situ at High Pressures Using Sum Frequency Generation. J. Am. Chem. Soc. 1996, 118, 2942−2949. (18) Su, X.; Cremer, P. S.; Shen, Y. R.; Somorjai, G. A. High-Pressure CO Oxidation on Pt(111) Monitored with Infrared-Visible Sum Frequency Generation (SFG). J. Am. Chem. Soc. 1997, 119, 3994−4000. (19) Cremer, P. S.; Su, X.; Somorjai, G. A.; Shen, Y. R. High Pressure Catalytic Processes Studied by Infrared-Visible Sum Frequency Generation. J. Mol. Catal. A: Chem. 1998, 131, 225−241. (20) Su, X.; Cremer, P. S.; Shen, Y. R.; Somorjai, G. A. Pressure Dependence (10−10−700 Torr) of the Vibrational Spectra of Adsorbed CO on Pt(111) Studied by Sum Frequency Generation. Phys. Rev. Lett. 1996, 77, 3858. (21) Shen, Y. R. The Principles of Nonlinear Optics; Wiley: New York, 1984. (22) Shen, Y. R. Surface Properties Probed by Second-Harmonic and Sum-Frequency Generation. Nature 1989, 337, 519−525. (23) Zhu, X. D.; Suhr, H.; Shen, Y. R. Surface Vibrational Spectroscopy by Infrared-Visible Sum Frequency Generation. Phys. Rev. B: Condens. Matter Mater. Phys. 1987, 35, 3047. (24) Somorjai, G. A.; McCrea, K. R. Sum Frequency Generation: Surface Vibrational Spectroscopy Studies of Catalytic Reactions on Metal Single-Crystal Surfaces. Adv. Catal. 2000, 45, 385−438. (25) Somorjai, G. A.; York, R. L.; Butcher, D.; Park, J. Y. The Evolution of Model Catalytic Systems; Studies of Structure, Bonding and Dynamics from Single Crystal Metal Surfaces to Nanoparticles, and from Low Pressure (< 10−3 Torr) to High Pressure (> 10−3 Torr) to Liquid Interfaces. Phys. Chem. Chem. Phys. 2007, 9, 3500−3513. (26) Ogletree, D. F.; Bluhm, H.; Lebedev, G.; Fadley, C. S.; Hussain, Z.; Salmeron, M. A Differentially Pumped Electrostatic Lens System for Photoemission Studies in the Millibar Range. Rev. Sci. Instrum. 2002, 73, 3872−3877. (27) Salmeron, M.; Schlögl, R. Ambient Pressure Photoelectron Spectroscopy: A New Tool for Surface Science and Nanotechnology. Surf. Sci. Rep. 2008, 63, 169−199. (28) Tao, F. Design of an In-House Ambient Pressure AP-XPS Using a Bench-Top X-Ray Source and the Surface Chemistry of Ceria under Reaction Conditions. Chem. Commun. 2012, 48, 3812−3814. (29) Mangolini, F.; Ahlund, J.; Wabiszewski, G. E.; Adiga, V. P.; Egberts, P.; Streller, F.; Backlund, K.; Karlsson, P. G.; Wannberg, B.; Carpick, R. W. Angle-Resolved Environmental X-Ray Photoelectron Spectroscopy: A New Laboratory Setup for Photoemission Studies at Pressures up to 0.4 Torr. Rev. Sci. Instrum. 2012, 83, 093112.

(30) Rodriguez, J. A.; Ma, S.; Liu, P.; Hrbek, J.; Evans, J.; Perez, M. Activity of CeOx and TiOx Nanoparticles Grown on Au(111) in the Water-Gas Shift Reaction. Science 2007, 318, 1757−1760. (31) Van Hardeveld, R.; Hartog, F. Statistics of Surface Atoms and Surface Sites on Metal Crystals. Surf. Sci. 1969, 15, 189−230. (32) de Smit, E.; Weckhuysen, B. M. The Renaissance of Iron-Based Fischer−Tropsch Synthesis: on the Multifaceted Catalyst Deactivation Behaviour. Chem. Soc. Rev. 2008, 37, 2758−2781. (33) 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.; et al. Nanoscale Chemical Imaging of a Working Catalyst by Scanning Transmission X-Ray Microscopy. Nature 2008, 456, 222−225. (34) Janbroers, S.; Crozier, P. A.; Zandbergen, H. W.; Kooyman, P. J. A Model Study on the Carburization Process of Iron-Based Fischer− Tropsch Catalysts Using In Situ TEM-EELS. Appl. Catal., B 2011, 102, 521−527. (35) Sun, B.; Qiao, M. H.; Fan, K. N. A.; Ulrich, J.; Tao, F. Fischer− Tropsch Synthesis over Molecular Sieve Supported Catalysts. ChemCatChem 2011, 3, 542−550. (36) Tao, F.; Dag, S.; Wang, L. W.; Liu, Z.; Butcher, D. R.; Salmeron, M.; Somorjai, G. A. Restructuring of Hex-Pt(100) under CO Gas Environments: Formation of 2-D Nanoclusters. Nano Lett. 2009, 9, 2167−2171. (37) Zhu, Z. W.; Tao, F.; Zheng, F.; Chang, R.; Li, Y. M.; Heinke, L.; Liu, Z.; Salmeron, M.; Somorjai, G. A. Formation of Nanometer-Sized Surface Platinum Oxide Clusters on a Stepped Pt(557) Single Crystal Surface Induced by Oxygen: A High-Pressure STM and AmbientPressure XPS Study. Nano Lett. 2012, 12, 1491−1497. (38) Vang, R. T.; Laegsgaard, E.; Besenbacher, F. Bridging the Pressure Gap in Model Systems for Heterogeneous Catalysis with High-Pressure Scanning Tunneling Microscopy. Phys. Chem. Chem. Phys. 2007, 9, 3460−3469. (39) Lauritsen, J. V.; Besenbacher, F. Model Catalyst Surfaces Investigated by Scanning Tunneling Microscopy. Adv. Catal. 2006, 50, 97−147. (40) Williams, D. B.; Carter, C. B. Transmission Electron MicroscopyA Textbook for Materials Science; Plenum Press: New York, 1996. (41) Dellby, N.; Krivanek, O. L.; Nellist, P. D.; Batson, P. E.; Lupini, A. R. Progress in Aberration-Corrected Scanning Transmission Electron Microscopy. Microscopy 2001, 50, 177−185. (42) Pennycook, S. J. Seeing the Atoms More Clearly: STEM Imaging from the Crewe Era to Today. Ultramicroscopy 2012, 123, 28−37. (43) Haider, M.; Rose, H.; Uhlemann, S.; Schwan, E.; Kabius, B.; Urban, K. A Spherical-Aberration-Corrected 200 KV Transmission Electron Microscope. Ultramicroscopy 1998, 75, 53−60. (44) Rose, H. Correction of Aberrations, a Promising Means for Improving the Spatial and Energy Resolution of Energy-Filtering Electron-Microscopes. Ultramicroscopy 1994, 56, 11−25. (45) Smith, D. J. Development of Aberration-Corrected Electron Microscopy. Microsc. Microanal. 2008, 14, 2−15. (46) de Jong, K. P.; van den Oetelaar, L. C. A.; Vogt, E. T. C.; Eijsbouts, S.; Koster, A. J.; Friedrich, H.; de Jongh, P. E. High-Resolution Electron Tomography Study of an Industrial Ni−Mo/γ-Al2O3 Hydrotreating Catalyst. J. Phys. Chem. B 2006, 110, 10209−10212. (47) Jinschek, J. R.; Batenburg, K. J.; Calderon, H. A.; Kilaas, R.; Radmilovic, V.; Kisielowski, C. 3-D Reconstruction of the Atomic Positions in a Simulated Gold Nanocrystal Based on Discrete Tomography: Prospects of Atomic Resolution Electron Tomography. Ultramicroscopy 2008, 108, 589−604. (48) Li, Z. Y.; Young, N. P.; DiVece, M.; Palomba, S.; Palmer, R. E.; Bleloch, A. L.; Curley, B. C.; Johnston, R. L.; Jiang, J.; Yuan, J. ThreeDimensional Atomic-Scale Structure of Size-Selected Gold Nanoclusters. Nature 2008, 451, 46−48. (49) Midgley, P. A.; Weyland, M. 3D Electron Microscopy in the Physical Sciences: the Development of Z-Contrast and EFTEM Tomography. Ultramicroscopy 2003, 96, 413−431. (50) Muller, D. A.; Kourkoutis, L. F.; Murfitt, M.; Song, J. H.; Hwang, H. Y.; Silcox, J.; Dellby, N.; Krivanek, O. L. Atomic-Scale Chemical 3532

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Imaging of Composition and Bonding by Aberration-Corrected Microscopy. Science 2008, 319, 1073−1076. (51) Watanabe, M.; Ackland, D. W.; Burrows, A.; Kiely, C. J.; Williams, D. B.; Krivanek, O. L.; Dellby, N.; Murfitt, M. F.; Szilagyi, Z. Improvements in the X-Ray Analytical Capabilities of a Scanning Transmission Electron Microscope by Spherical-Aberration Correction. Microsc. Microanal. 2006, 12, 515−526. (52) Binnig, G.; Rohrer, H.; Gerber, Ch.; Weibel, E. Tunneling Through a Controllable Vacuum Gap. Appl. Phys. Lett. 1982, 40, 178− 180. (53) Binnig, G.; Rohrer, H. Scanning Tunneling Microscopy - from Birth to Adolescence. Angew. Chem., Int. Ed. Engl. 1987, 26, 606−614. (54) Binnig, G.; Rohrer, H. Scanning Tunneling Microscopy - from Birth to Adolescence. Rev. Mod. Phys. 1987, 59, 615−625. (55) Marti, O.; Binnig, G.; Rohrer, H.; Salemink, H. Low-Temperature Scanning Tunneling Microscope. Surf. Sci. 1987, 181, 230−234. (56) Binnig, G.; Fuchs, H.; Gerber, C.; Rohrer, H.; Stoll, E.; Tosatti, E. Energy-Dependent State-Density Corrugation of a Graphite Surface as Seen by Scanning Tunneling Microscopy. Europhys. Lett. 1986, 1, 31− 36. (57) Binnig, G.; Rohrer, H. The Scanning Tunneling Microscope. Sci. Am. 1985, 253, 50−50. (58) Binnig, G.; Rohrer, H. Scanning Tunneling Microscopy. Surf. Sci. 1985, 152-153, 17−26. (59) Binnig, G.; Rohrer, H. Scanning Tunneling Microscopy. IBM J. Res. Dev. 1986, 30, 355−369. (60) Gerber, C.; Binnig, G.; Fuchs, H.; Marti, O.; Rohrer, H. Scanning Tunneling Microscope Combined with a Scanning Electron-Microscope. Rev. Sci. Instrum. 1986, 57, 221−224. (61) Holm, R. The Electric Tunnel Effect across Thin Insulator Films in Contacts. J. Appl. Phys. 1951, 22, 569−574. (62) Bell, R. P. The Tunnel Effect Correction for Parabolic Potential Barriers. Trans. Faraday Soc. 1959, 55, 1−4. (63) Chynoweth, A. G.; Logan, R. A.; Wolff, P. A. Effect of Landau Levels upon Tunnel Currents in Indium Antimonide. Phys. Rev. Lett. 1960, 5, 548−550. (64) Vul, B. M. Impact Ionization and Tunnel Effect in Semiconductors. Soviet Physics - Solid State 1961, 2, 2631−2636. (65) Chen, C. J. Introduction to Scanning Tunneling Microscopy; Oxford University Press: 1993. (66) Hla, S.-W.; Bartels, L.; Meyer, G.; Rieder, K.-H. Inducing All Steps of a Chemical Reaction with the Scanning Tunneling Microscope Tip: Towards Single Molecule Engineering. Phys. Rev. Lett. 2000, 85, 2777. (67) Zhang, Z.; Lagally, M. G. Atomistic Processes in the Early Stages of Thin-Film Growth. Science 1997, 276, 377−383. (68) Chen, X.; Fu, Y.-S.; Ji, S.-H.; Zhang, T.; Cheng, P.; Ma, X.-C.; Zou, X.-L.; Duan, W.-H.; Jia, J.-F.; Xue, Q.-K. Probing Superexchange Interaction in Molecular Magnets by Spin-Flip Spectroscopy and Microscopy. Phys. Rev. Lett. 2008, 101, 197208. (69) McIntyre, B. J.; Salmeron, M.; Somorjai, G. A. In Situ Scanning Tunneling Microscopy Study of Platinum (110) in A Reactor Cell at High Pressures and Temperatures. J. Vac. Sci. Technol., A 1993, 11, 1964−1968. (70) Laegsgaard, E.; Ö sterlund, L.; Thostrup, P.; Rasmussen, P. B.; Stensgaard, I.; Besenbacher, F. A High-Pressure Scanning Tunneling Microscope. Rev. Sci. Instrum. 2001, 72, 3537−3542. (71) Rasmussen, P. B.; Hendriksen, B. L. M.; Zeijlemaker, H.; Ficke, H. G.; Frenken, J. W. M. The ″Reactor STM″: A Scanning Tunneling Microscope for Investigation of Catalytic Surfaces at Semi-Industrial Reaction Conditions. Rev. Sci. Instrum. 1998, 69, 3879−3884. (72) Tao, F.; Nguyen, L.; Zhang, S. R. Design of a New Reactor-Like High Temperature Near Ambient Pressure Scanning Tunneling Microscope for Catalysis Studies. Rev. Sci. Instrum. 2013, 84, 034101. (73) Tao, F.; Salmeron, M. In Situ Studies of Chemistry and Structure of Materials in Reactive Environments. Science 2011, 331, 171−174. (74) McIntyre, B. J.; Salmeron, M. B.; Somorjai, G. A. A Scanning Tunneling Microscope that Operates at High Pressures and High Temperatures (430 K) and during Catalytic Reactions. Catal. Lett. 1992, 14, 263−269.

(75) McIntyre, B. J.; Salmeron, M.; Somorjai, G. A. A Variable Pressure/Temperature Scanning Tunneling Microscope for Surface Science and Catalysis Studies. Rev. Sci. Instrum. 1993, 64, 687−691. (76) Kolmakov, A.; Goodman, D. W. In Situ Scanning Tunneling Microscopy of Individual Supported Metal Clusters at Reactive Gas Pressures From 10−8 to 104 Pa. Rev. Sci. Instrum. 2003, 74, 2444−2450. (77) Rößler, M.; Geng, P.; Wintterlin, J. A High-Pressure Scanning Tunneling Microscope for Studying Heterogeneous Catalysis. Rev. Sci. Instrum. 2005, 76, 023705. (78) Tao, F.; Tang, D.; Salmeron, M.; Somorjai, G. A. A New Scanning Tunneling Microscope Reactor Used for High-Pressure and HighTemperature Catalysis Studies. Rev. Sci. Instrum. 2008, 79, 084101. (79) Marton, L. La Microscopie Electronique les Object Biologiques. Bull. Acad. R. Belg. Cl. Sci. 1935, 21, 553−554. (80) Watson, J. H. L. Specimen Contamination in Electron Microscopes. J. Appl. Phys. 1948, 19, 110−111. (81) Watson, J. H. L. An Effect of Electron Bombardment upon Carbon Black. J. Appl. Phys. 1947, 18, 153−161. (82) Ardenne, M.; Beischer, D. Untersuchung von Metalloxyd-rauchen mit dem Universal-Elektronenmikroskop. Z. Elektrochem. 1940, 46, 270−277. (83) Butler, P.; Hale, K. Dynamic Experiments in the Electron Microscope. Practical Methods in Electron Microscopy; Elsevier: New York, 1981; pp 239−308. (84) Chenna, S.; Crozier, P. A. Operando TEM: A Technique for Detection of Catalysis Using Electron Energy-Loss Spectroscopy in Transmission Electron Microscope. ACS Catal. 2012, 2, 2395−2402. (85) Kohlmann, K. T.; Thiemann, M.; Brünger, W. H. E-Beam Induced X-Ray Mask Repair with Optimized Gas Nozzle Geometry. Microelectron. Eng. 1991, 13, 279−282. (86) Matsui, S.; Ichihashi, T. In Situ Observation on Electron-BeamInduced Chemical Vapor Deposition by Transmission Electron Microscopy. Appl. Phys. Lett. 1988, 53, 842−844. (87) Kishita, K.; Sakai, H.; Tanaka, H.; Saka, H.; Kuroda, K.; Sakamoto, M.; Watabe, A.; Kamino, T. Development of an Analytical Environmental TEM System and Its Application. J. Electron Microsc. 2009, 58, 331−339. (88) Giorgio, S.; Sao Joao, S.; Nitsche, S.; Chaudanson, D.; Sitja, G.; Henry, C. R. Environmental Electron Microscopy (ETEM) for Catalysts with a Closed E-Cell with Carbon Windows. Ultramicroscopy 2006, 106, 503−507. (89) Parkinson, G. M. High Resolution In Situ Controlled Atmosphere Transmission Electron Microscopy (CATEM) of Heterogeneous Catalysts. Catal. Lett. 1989, 2, 303−307. (90) Creemer, J. F.; Helveg, S.; Hoveling, G. H.; Ullmann, S.; Molenbroek, A. M.; Sarro, P. M.; Zandbergen, H. W. Atomic-Scale Electron Microscopy at Ambient Pressure. Ultramicroscopy 2008, 108, 993−998. (91) Alan, T.; Yokosawa, T.; Gaspar, J.; Pandraud, G.; Paul, O.; Creemer, F.; Sarro, P. M.; Zandbergen, H. W. Micro-Fabricated Channel with Ultra-Thin yet Ultra-Strong Windows Enables Electron Microscopy under 4-bar Pressure. Appl. Phys. Lett. 2012, 100, 081903. (92) Yokosawa, T.; Alan, T.; Pandraud, G.; Dam, B.; Zandbergen, H. In-situ TEM on (de)Hydrogenation of Pd at 0.5−4.5 bar Hydrogen Pressure and 20−400 °C. Ultramicroscopy 2012, 112, 47−52. (93) de Jonge, N.; Bigelow, W. C.; Veith, G. M. Atmospheric Pressure Scanning Transmission Electron Microscopy. Nano Lett. 2010, 10, 1028−1031. (94) Xin, H. L. L.; Niu, K. Y.; Alsem, D. H.; Zheng, H. M. In Situ TEM Study of Catalytic Nanoparticle Reactions in Atmospheric Pressure Gas Environment. Microsc. Microanal. 2013, 19, 1558−1568. (95) Mehraeen, S.; McKeown, J. T.; Deshmukh, P. V.; Evans, J. E.; Abellan, P.; Xu, P. H.; Reed, B. W.; Taheri, M. L.; Fischione, P. E.; Browning, N. D. A (S)TEM Gas Cell Holder with Localized Laser Heating for In Situ Experiments. Microsc. Microanal. 2013, 19, 470−478. (96) Vendelbo, S. B.; Kooyman, P. J.; Creemer, J. F.; Morana, B.; Mele, L.; Dona, P.; Nelissen, B. J.; Helveg, S. Method for Local Temperature Measurement in a Nanoreactor for In Situ High-Resolution Electron Microscopy. Ultramicroscopy 2013, 133, 72−79. 3533

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(97) Picher, M.; Mazzucco, S.; Blankenship, S.; Sharma, R. Vibrational and Optical Spectroscopies Integrated with Environmental Transmission Electron Microscopy. Ultramicroscopy 2015, 150, 10−15. (98) Mecklenburg, M.; Hubbard, W. A.; White, E. R.; Dhall, R.; Cronin, S. B.; Aloni, S.; Regan, B. C. Nanoscale Temperature Mapping in Operating Microelectronic Devices. Science 2015, 347, 629−632. (99) de Jonge, N.; Peckys, D. B.; Kremers, G. J.; Piston, D. W. Electron Microscopy of Whole Cells in Liquid with Nanometer Resolution. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 2159−2164. (100) Ring, E. A.; de Jonge, N. Video-Frequency Scanning Transmission Electron Microscopy of Moving Gold Nanoparticles in Liquid. Micron 2012, 43, 1078−1084. (101) Inayoshi, Y.; Minoda, H.; Arai, Y.; Nagayama, K. Direct Observation of Biological Molecules in Liquid by Environmental PhasePlate Transmission Electron Microscopy. Micron 2012, 43, 1091−1098. (102) Evans, J. E.; Jungjohann, K. L.; Wong, P. C. K.; Chiu, P.-L.; Dutrow, G. H.; Arslan, I.; Browning, N. D. Visualizing Macromolecular Complexes with In Situ Liquid Scanning Transmission Electron Microscopy. Micron 2012, 43, 1085−1090. (103) Zheng, H. M.; Smith, R. K.; Jun, Y. W.; Kisielowski, C.; Dahmen, U.; Alivisatos, A. P. Observation of Single Colloidal Platinum Nanocrystal Growth Trajectories. Science 2009, 324, 1309−1312. (104) Liao, H. G.; Cui, L. K.; Whitelam, S.; Zheng, H. M. Real-Time Imaging of Pt3Fe Nanorod Growth in Solution. Science 2012, 336, 1011−1014. (105) Yuk, J. M.; Park, J.; Ercius, P.; Kim, K.; Hellebusch, D. J.; Crommie, M. F.; Lee, J. Y.; Zettl, A.; Alivisatos, A. P. High-Resolution EM of Colloidal Nanocrystal Growth Using Graphene Liquid Cells. Science 2012, 336, 61−64. (106) Sharma, R.; Crozier, P. A. Environmental Transmission Electron Microscopy in Nanotechnology. Handbook of Microscopy for Nanotechnology; Springer: New York, 2005; pp 531−563. (107) Swann, P. R.; Tighe, N. J. Proc. 5th Eur. Reg. Cong Electron Microscopy; Institute of Physics, Royal Microscopal Society: Manchester, U.K., 1972; pp 436−437. (108) Doole, R. C.; Parkinson, G. M.; Stead, J. M. High Resolution Gas Reaction Cell for JEM 4000. In Inst. Phys. Conf. Ser.; Institute of Physics Publishing: Bristol, U.K., 1991; pp 157−160. (109) Sharma, R.; Weiss, K. Development of a TEM to Study In Situ Structural and Chemical Changes at an Atomic Level During Gas-Solid Interactions at Elevated Temperatures. Microsc. Res. Tech. 1998, 42, 270−280. (110) Lee, T. C.; Dewald, D. K.; Eades, J. A.; Robertson, I. M.; Birnbaum, H. K. An Environmental Cell Transmission Electron Microscope. Rev. Sci. Instrum. 1991, 62, 1438−1444. (111) Boyes, E. D.; Gai, P. L. Environmental High Resolution Electron Microscopy and Applications to Chemical Science. Ultramicroscopy 1997, 67, 219−232. (112) Yao, N.; Spinnler, G. E.; Kemp, R. A.; Guthrie, D. C.; Cates, R. D.; Bolinger, C. M. Environmental-Cell TEM Studies of Catalyst Particle Behavior. In Proceeding of 49th Ann. Meeting of Microscopy Society of America; San Francisco Press, Inc.: San Francisco, 1991; pp 1028− 1029. (113) Hansen, L.; Wagner, J. B. In-situ reduction of a methanol catalyst in a new FEG TEM. In Proceeding of 12th European Congress on Electron Microscopy; Czechoslovak Society for Electron Microscopy: Brno, Czech Republic, 2000; pp 537−538. (114) Sharma, R.; Crozier, P. A.; Marx, R.; Weiss, K. An Environmental Transmission Electron Microscope for In-Situ Observation of Chemical Processes at the Nanometer Level. Microsc. Microanal. 2003, 9, 912− 913. (115) Li, P.; Liu, J.; Nag, N.; Crozier, P. A. In Situ Synthesis and Characterization of Ru Promoted Co/Al2O3 Fischer−Tropsch Catalysts. Appl. Catal., A 2006, 307, 212−221. (116) Allard, L. F.; Overbury, S. H.; Bigelow, W. C.; Katz, M. B.; Nackashi, D. P.; Damiano, J. Novel MEMS-Based Gas-Cell/Heating Specimen Holder Provides Advanced Imaging Capabilities for In Situ Reaction Studies. Microsc. Microanal. 2012, 18, 656−666.

(117) Gai, P. L.; Lari, L.; Ward, M. R.; Boyes, E. D. Visualisation of Single Atom Dynamics and Their Role in Nanocatalysts under Controlled Reaction Environments. Chem. Phys. Lett. 2014, 592, 355−359. (118) Yaguchi, T.; Suzuki, M.; Watabe, A.; Nagakubo, Y.; Ueda, K.; Kamino, T. Development of a High Temperature-Atmospheric Pressure Environmental Cell for High-Resolution TEM. J. Electron Microsc. 2011, 60, 217−225. (119) Cavalca, F.; Laursen, A. B.; Kardynal, B. E.; Dunin-Borkowski, R. E.; Dahl, S.; Wagner, J. B.; Hansen, T. W. In Situ Transmission Electron Microscopy of Light-Induced Photocatalytic Reactions. Nanotechnology 2012, 23, 075705. (120) Miller, B. K.; Crozier, P. A. System for In Situ UV-Visible Illumination of ETEM Samples. Microsc. Microanal. 2013, 19, 461−469. (121) Zhang, L.; Miller, B. K.; Crozier, P. A. Atomic Level Observation of Surface Amorphization in Anatase Nanocrystals During Light Irradiation in Water Vapor. Nano Lett. 2013, 13, 679−684. (122) Egerton, R. F.; Li, P.; Malac, M. Radiation Damage in The TEM and SEM. Micron 2004, 35, 399−409. (123) Egerton, R. F.; Wang, F.; Crozier, P. A. Beam-Induced Damage to Thin Specimens in An Intense Electron Probe. Microsc. Microanal. 2006, 12, 65−71. (124) Meyer, J. C.; Girit, C. O.; Crommie, M. F.; Zettl, A. Hydrocarbon Lithography on Graphene Membranes. Appl. Phys. Lett. 2008, 92, 123110. (125) Mitsuishi, K.; Shimojo, M.; Han, M.; Furuya, K. Electron-BeamInduced Deposition Using a Subnanometer Sized Probe of High-Energy Electrons. Appl. Phys. Lett. 2003, 83, 2064−2066. (126) van Dorp, W. F.; Hagen, C. W. A Critical Literature Review of Focused Electron Beam Induced Deposition. J. Appl. Phys. 2008, 104, 081301. (127) Crozier, P. A.; Tolle, J.; Kouvetakis, J.; Ritter, C. Synthesis of Uniform GaN Quantum Dot Arrays via Electron Nanolithography of D2GaN3. Appl. Phys. Lett. 2004, 84, 3441−3443. (128) van Dorp, W. F.; van Someren, B.; Hagen, C. W.; Kruit, P.; Crozier, P. A. Approaching the Resolution Limit of Nanometer-Scale Electron Beam Induced Deposition. Nano Lett. 2005, 5, 1303−1307. (129) Crozier, P. A. Nanoscale Oxide Patterning with Electron-SolidGas Reactions. Nano Lett. 2007, 7, 2395−2398. (130) Kuwauchi, Y.; Yoshida, H.; Akita, T.; Haruta, M.; Takeda, S. Intrinsic Catalytic Structure of Gold Nanoparticles Supported on TiO2. Angew. Chem., Int. Ed. 2012, 51, 7729−7733. (131) Hansen, T. W.; Wagner, J. B. Environmental Transmission Electron Microscopy in an Aberration-Corrected Environment. Microsc. Microanal. 2012, 18, 684−690. (132) Suzuki, M.; Yaguchi, T.; Zhang, X. F. High-Resolution Environmental Transmission Electron Microscopy: Modeling and Experimental Verification. Microscopy 2013, 62, 437−450. (133) Jinschek, J. R.; Helveg, S. Image Resolution and Sensitivity in an Environmental Transmission Electron Microscope. Micron 2012, 43, 1156−1168. (134) Bright, A. N.; Yoshida, K.; Tanaka, N. Influence of Total Beam Current on HRTEM Image Resolution in Differentially Pumped ETEM with Nitrogen Gas. Ultramicroscopy 2013, 124, 46−51. (135) Oppenheim, I. C.; Trevor, D. J.; Chidsey, C. E. D.; Trevor, P. L.; Sieradzki, K. In Situ Scanning Tunneling Microscopy of Corrosion of Silver-Gold Alloys. Science 1991, 254, 687−689. (136) Kim, J. S.; LaGrange, T.; Reed, B. W.; Taheri, M. L.; Armstrong, M. R.; King, W. E.; Browning, N. D.; Campbell, G. H. Imaging of Transient Structures Using Nanosecond In Situ TEM. Science 2008, 321, 1472−1475. (137) Tao, F.; Grass, M. E.; Zhang, Y. W.; Butcher, D. R.; Renzas, J. R.; Liu, Z.; Chung, J. Y.; Mun, B. S.; Salmeron, M.; Somorjai, G. A. ReactionDriven Restructuring of Rh-Pd and Pt-Pd Core-Shell Nanoparticles. Science 2008, 322, 932−934. (138) Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.; Clausen, B. S.; Topsoe, H. Atom-Resolved Imaging of Dynamic Shape Changes in Supported Copper Nanocrystals. Science 2002, 295, 2053− 2055. 3534

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(139) Behrens, M.; Studt, F.; Kasatkin, I.; Kuhl, S.; Havecker, M.; Abild-Pedersen, F.; Zander, S.; Girgsdies, F.; Kurr, P.; Kniep, B. L.; et al. The Active Site of Methanol Synthesis over Cu/ZnO/Al2O3 Industrial Catalysts. Science 2012, 336, 893−897. (140) Altman, M. S. CO Prefers the Aisle Seat. Science 2010, 327, 789− 790. (141) Kruse Vestergaard, E.; Thostrup, P.; An, T.; Lægsgaard, E.; Stensgaard, I.; Hammer, B.; Besenbacher, F. Comment on “High Pressure Adsorbate Structures Studied by Scanning Tunneling Microscopy: CO on Pt(111) in Equilibrium with the Gas Phase”. Phys. Rev. Lett. 2002, 88, 259601. (142) Cernota, P.; Rider, K.; Yoon, H. A.; Salmeron, M.; Somorjai, G. Dense Structures Formed by CO on Rh(111) Studied by Scanning Tunneling Microscopy. Surf. Sci. 2000, 445, 249. (143) Rider, K. B.; Hwang, K. S.; Salmeron, M.; Somorjai, G. A. Structure and Dynamics of Dense Monolayers of NO Adsorbed on Rh(111) in Equilibrium with the Gas Phase in the Torr Pressure Range. Phys. Rev. Lett. 2001, 86, 4330. (144) Vang, R. T.; Wang, J. G.; Knudsen, J.; Schnadt, J.; Laegsgaard, E.; Stensgaard, I.; Besenbacher, F. The Adsorption Structure of NO on Pd(III) at High Pressures Studied by STM and DFT. J. Phys. Chem. B 2005, 109, 14262−14265. (145) Rider, K. B.; Hwang, K. S.; Salmeron, M.; Somorjai, G. A. HighPressure (1 Torr) Scanning Tunneling Microscopy (STM) Study of the Coadsorption and Exchange of CO and NO on the Rh(111) Crystal Face. J. Am. Chem. Soc. 2002, 124, 5588−5593. (146) Madey, T. E.; Albert Engelhardt, H.; Menzel, D. Adsorption of Oxygen and Oxidation of CO on the Ruthenium (001) Surface. Surf. Sci. 1975, 48, 304−328. (147) Parrott, S. L.; Praline, G.; Koel, B. E.; White, J. M.; Taylor, T. N. Oxygen Chemisorption on a Stepped Ru (∼001) Crystal. J. Chem. Phys. 1979, 71, 3352−3354. (148) Stampfl, C.; Scheffler, M. Theoretical Study of O Adlayers on Ru(0001). Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 2868− 2872. (149) Hagstrom, S.; Lyon, H. B.; Somorjai, G. A. Surface Structures on the Clean Platinum (100) Surface. Phys. Rev. Lett. 1965, 15, 491. (150) McIntyre, B. J.; Salmeron, M.; Somorjai, G. A. A Variable Pressure Temperature Scanning Tunneling Microscope for Surface Scientific and Catalysis Studies. Rev. Sci. Instrum. 1993, 64, 687−691. (151) Thostrup, P.; Christoffersen, E.; Lorensen, H. T.; Jacobsen, K. W.; Besenbacher, F.; Nørskov, J. K. Adsorption-Induced Step Formation. Phys. Rev. Lett. 2001, 87, 126102. (152) Thostrup, P.; Vestergaard, E. K.; An, T.; Lægsgaard, E.; Besenbacher, F. CO-Induced Restructuring of Pt(110)-(1 × 2): Bridging the Pressure Gap with High-Pressure Scanning Tunneling Microscopy. J. Chem. Phys. 2003, 118, 3724−3730. (153) Hofmann, P.; Bare, S. R.; King, D. A. Surface Phase Transitions in CO Chemisorption on Pt{110}. Surf. Sci. 1982, 117, 245−256. (154) Yin, A. X.; Min, X. Q.; Zhu, W.; Liu, W. C.; Zhang, Y. W.; Yan, C. H. Pt-Cu and Pt-Pd-Cu Concave Nanocubes with High-Index Facets and Superior Electrocatalytic Activity. Chem. - Eur. J. 2012, 18, 777−782. (155) Eren, B.; Zherebetskyy, D.; Patera, L. L.; Wu, C. H.; Bluhm, H.; Africh, C.; Wang, L. W.; Somorjai, G. A.; Salmeron, M. Activation of Cu(111) surface by decomposition into nanoclusters driven by CO adsorption. Science 2016, 351, 475−478. (156) Ö sterlund, L.; Rasmussen, P. B.; Thostrup, P.; Lægsgaard, E.; Stensgaard, I.; Besenbacher, F. Bridging the Pressure Gap in Surface Scientific at the Atomic Level: H/Cu(110). Phys. Rev. Lett. 2001, 86, 460−463. (157) Mistry, H.; Behafarid, F.; Bare, S. R.; Cuenya, B. R. PressureDependent Effect of Hydrogen Adsorption on Structural and Electronic Properties of Pt/gamma-Al2O3 Nanoparticles. ChemCatChem 2014, 6, 348−352. (158) Takeda, S.; Kuwauchi, Y.; Yoshida, H. Environmental Transmission Electron Microscopy for Catalyst Materials Using a Spherical Aberration Corrector. Ultramicroscopy 2015, 151, 178−190.

(159) Tamaru, K.; Tanaka, K. I.; Fukasaku, S.; Ishida, S. Decomposition of Ammonia on a Nickel Catalyst. Trans. Faraday Soc. 1965, 61, 765−772. (160) Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous Catalysis; VCH-Wiley: 1997. (161) Hendriksen, B. L. M.; Frenken, J. W. M. Co Oxidation on Pt(110): Scanning Tunneling Microscopy inside a High-Pressure Flow Reactor. Phys. Rev. Lett. 2002, 89, 046101. (162) Ackermann, M. D.; Pedersen, T. M.; Hendriksen, B. L. M.; Robach, O.; Bobaru, S. C.; Popa, I.; Quiros, C.; Kim, H.; Hammer, B.; Ferrer, S.; et al. Structure and Reactivity of Surface Oxides on Pt(110) During Catalytic CO Oxidation. Phys. Rev. Lett. 2005, 95, 255505. (163) Baker, R. T. K.; Harris, P. S.; Thomas, R. B.; Waite, R. J. Formation of Filamentous Carbon From Iron, Cobalt and Chromium Catalyzed Decomposition of Acetylene. J. Catal. 1973, 30, 86−95. (164) Baker, R. T. K.; Barber, M. A.; Harris, P. S.; Feates, F. S.; Waite, R. J. Nucleation and Growth of Carbon Deposits from the Nickel Catalyzed Decomposition of Acetylene. J. Catal. 1972, 26, 51−62. (165) Baker, R. T. K.; Chludzinski, J. J., Jr. Filamentous Carbon Growth on Nickel-Iron Surfaces: The Effect of Various Oxide Additives. J. Catal. 1980, 64, 464−478. (166) Baker, R. T. K.; Rodriguez, N. M. A Review of the Use of In situ Electron Microscopy Techniques for the Study of Iron-Based Catalysts for Coal Conversion Processes. Energy Fuels 1994, 8, 330−340. (167) Baker, R. T. K.; Dudash, N. S.; Lund, C. R. F.; Chludzinski, J. J., Jr. Comparison of The Catalytic Influence of Nickel and Cu-Ni Alloys on the Graphite-Steam Reaction. Fuel 1985, 64, 1151−1156. (168) Derouane, E. G.; Chludzinski, J. J.; Baker, R. T. K. Direct Observation of Wetting and Spreading of Copper Particles on Magnesium Oxide. J. Catal. 1984, 85, 187−196. (169) Upton, B. H.; Chen, C. C.; Rodriguez, N. M.; Baker, R. T. K. In Situ Electron Microscopy Studies of the Interaction of Iron, Cobalt, and Nickel with Molybdenum Disulfide Single Crystals in Hydrogen. J. Catal. 1993, 141, 171−190. (170) Simoens, A. J.; Baker, R. T. K.; Dwyer, D. J.; Lund, C. R. F.; Madon, R. J. A Study of the Nickel-Titanium Oxide Interaction. J. Catal. 1984, 86, 359−372. (171) Dumesic, J. A.; Stevenson, S. A.; Sherwood, R. D.; Baker, R. T. K. Migration of Nickel and Titanium Oxide Species as Studied by In Situ Scanning Transmission Electron Microscopy. J. Catal. 1986, 99, 79−87. (172) Gai, P. L. Dynamic Studies of Metal Oxide Catalysts: MoO3. Philos. Mag. A 1981, 43, 841−855. (173) Gai, P. L. Environmental High Resolution Electron Microscopy of Gas-Catalyst Reactions. Top. Catal. 1999, 8, 97−113. (174) Gai, P. L. Developments in In Situ Environmental Cell HighResolution Electron Microscopy and Applications to Catalysis. Top. Catal. 2002, 21, 161−173. (175) Gai, P. L.; Boyes, E. D. Environmental High Resolution Electron Microscopy in Materials Science. In In Situ Microscopy in Materials Research; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1997; pp 123−146. (176) Gai, P. L.; Boyes, E. D. Advances in Atomic Resolution In Situ Environmental Transmission Electron Microscopy and 1Å Aberration Corrected In Situ Electron Microscopy. Microsc. Res. Tech. 2009, 72, 153−164. (177) Gai, P. L.; Boyes, E. D.; Helveg, S.; Hansen, P. L.; Giorgio, S.; Henry, C. R. Atomic-Resolution Environmental Transmission Electron Microscopy for Probing Gas-Solid Reactions in Heterogeneous Catalysis. MRS Bull. 2007, 32, 1044−1050. (178) Gai, P. L.; Sharma, R.; Ross, F. M. Environmental (S)TEM Studies of Gas-Liquid-Solid Interactions under Reaction Conditions. MRS Bull. 2008, 33, 107−114. (179) Sharma, R. Design and Applications of Environmental Cell Transmission Electron Microscope for In Situ Observations of GasSolid Reactions. Microsc. Microanal. 2001, 7, 494−506. (180) Sharma, R. An Environmental Transmission Electron Microscope for In Situ Synthesis and Characterization of Nanomaterials. J. Mater. Res. 2005, 20, 1695−1707. 3535

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(181) Ertl, H., Knozinger, J., Weitland, G., Eds. Preparation of Solid Catalysts; Wiley-VCH: New York, 1999. (182) De Jong, K. P. Assembly of Solid Catalysts. CatTech 1998, 3, 87− 87. (183) Moser, W. R., Ed. Advanced Catalysts and Nanostructured MaterialsModern Synthetic Methods; Academic Press: New York, 1996. (184) Schwarz, J. A.; Contescu, C.; Contescu, A. Methods for Preparation of Catalytic Materials. Chem. Rev. 1995, 95, 477−510. (185) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.; Yeung, L. K. Dendrimer Encapsulated Metal Nanoparticles: Synthesis, Characterization, and Application to Catalysis. Acc. Chem. Res. 2001, 34, 181−190. (186) Gates, B. C. Supported Nanostructured Catalysts: Metal Complexes and Metal Clusters. Adv. Chem. Eng. 2001, 27, 49−77. (187) Lafaye, G.; Williams, C. T.; Amiridis, M. D. Synthesis and Microscopic Characterization of Dendrimer-Derived Ru/Al2O3 Catalysts. Catal. Lett. 2004, 96, 43−47. (188) Menard, L. D.; Xu, F. T.; Nuzzo, R. G.; Yang, J. C. Preparation of TiO2-Supported Au Nanoparticle Catalysts from a Au 13 Cluster Precursor: Ligand Removal Using Ozone Exposure versus a Rapid Thermal Treatment. J. Catal. 2006, 243, 64−73. (189) Sietsma, J. R. A.; Friedrich, H.; Broersma, A.; Versluijs-Helder, M.; van Dillen, A. J.; de Jongh, P. E.; de Jong, K. P. How Nitric Oxide Affects the Decomposition of Supported Nickel Nitrate to Arrive at Highly Dispersed Catalysts. J. Catal. 2008, 260, 227−235. (190) Sobal, N. S.; Ebels, U.; Mohwald, H.; Giersig, M. Synthesis of Core-Shell PtCo Nanocrystals. J. Phys. Chem. B 2003, 107, 7351−7354. (191) Marceau, E.; Che, M.; Cejka, J.; Zukal, A. Nickel Nitrate vs. Acetate: Influence of the Precursor on the Structure and Reducibility of Ni/MCM-41 and Ni/Al-MCM-41 Catalyst. ChemCatChem 2010, 2, 413−422. (192) Pinna, F. Supported Metal Catalysts Preparation. Catal. Today 1998, 41, 129−137. (193) Bartholomew, C. H.; Farrauto, R. J. Chemistry of NickelAlumina Catalysts. J. Catal. 1976, 45, 41−53. (194) Louis, C.; Cheng, Z. X.; Che, M. Characterization of Ni/SiO2 Catalyst during Impregnation and further Thermal Activation Treatment Leading to Metal Particles. J. Phys. Chem. 1993, 97, 5703−5712. (195) Sietsma, J. R. A.; Meeldijk, J. D.; Versluijs-Helder, M.; Broersma, A.; van Dillen, A. J.; de Jongh, P. E.; de Jong, K. P. Ordered Mesoporous Silica to Study the Preparation of Ni/SiO2 Ex Nitrate Catalysts: Impregnation, Drying, and Thermal Treatments. Chem. Mater. 2008, 20, 2921−2931. (196) Yates, D. J. C.; Sinfelt, J. H.; Taylor, W. F. Catalysis over Supported Metals.I. Kinetics of Ethane Hydrogenolysis over Nickel Surfaces of Known Area. J. Am. Chem. Soc. 1964, 86, 2996−3001. (197) Li, P.; Liu, J.; Nag, N.; Crozier, P. A. Atomic-Scale Study of InSitu Metal Nanoparticle Synthesis in a Ni/TiO2 System. J. Phys. Chem. B 2005, 109, 13883−13890. (198) Li, P.; Liu, J.; Nag, N.; Crozier, P. A. Dynamic Nucleation and Growth of Ni Nanoparticles on High Surface Area Titania. Surf. Sci. 2006, 600, 693−702. (199) Tauster, S. J. Strong Metal-Support Interactions. Acc. Chem. Res. 1987, 20, 389−394. (200) Tauster, S. J.; Fung, S. C.; Baker, R. T. K.; Horsley, J. A. Strong Interaction in Supported-Metal Catalysts. Science 1981, 211, 1121− 1125. (201) Tauster, S. J.; Fung, S. C.; Garten, R. L. Strong Metal-Support Interactions - Group 8 Noble Metals Supported on TiO2. J. Am. Chem. Soc. 1978, 100, 170−175. (202) Banerjee, R.; Crozier, P. A. In Situ Synthesis and Nanoscale Evolution of Model Supported Metal Catalysts: Ni on Silica. J. Phys. Chem. C 2012, 116, 11486−11495. (203) Crozier, P. A.; Sharma, R.; Datye, A. K. Oxidation and Reduction of Small Palladium Particles on Silica. Microsc. Microanal. 1998, 4, 278− 285. (204) Crozier, P. A.; Datye, A. K. Direct Observation of Reduction of PdO to Pd Metal by In Situ Electron Microscopy. Stud. Surf. Sci. Catal. 2000, 130, 3119−3124.

(205) Dehghan, R.; Hansen, T. W.; Wagner, J. B.; Holmen, A.; Rytter, E.; Borg, O.; Walmsley, J. C. In-Situ Reduction of Promoted Cobalt Oxide Supported on Alumina by Environmental Transmission Electron Microscopy. Catal. Lett. 2011, 141, 754−761. (206) Xin, H. L. L.; Pach, E. A.; Diaz, R. E.; Stach, E. A.; Salmeron, M.; Zheng, H. M. Revealing Correlation of Valence State with Nanoporous Structure in Cobalt Catalyst Nanoparticles by In Situ Environmental TEM. ACS Nano 2012, 6, 4241−4247. (207) Li, P.; Liu, J.; Nag, N.; Crozier, P. A. In Situ Preparation of NiCu/TiO2 Bimetallic Catalysts. J. Catal. 2009, 262, 73−82. (208) Wang, C. M.; Kwak, J. H.; Kim, D. H.; Szanyi, J.; Sharma, R.; Thevuthasan, S.; Peden, C. H. F. Morphological Evolution of Ba(NO3)2 Supported on Alpha-Al2O3 (0001): An In Situ TEM Study. J. Phys. Chem. B 2006, 110, 11878−11883. (209) Andreasen, J. W.; Rasmussen, F. B.; Helveg, S.; Molenbroek, A.; Stahl, K.; Nielsen, M. M.; Feidenhans’l, R. Activation of a Cu/ZnO Catalyst for Methanol Synthesis. J. Appl. Crystallogr. 2006, 39, 209−221. (210) Holse, C.; Elkjaer, C. F.; Nierhoff, A.; Sehested, J.; Chorkendorff, I.; Helveg, S.; Nielsen, J. H. Dynamic Behavior of CuZn Nanoparticles under Oxidizing and Reducing Conditions. J. Phys. Chem. C 2015, 119, 2804−2812. (211) Jeangros, Q.; Faes, A.; Wagner, J. B.; Hansen, T. W.; Aschauer, U.; Van herle, J.; Hessler-Wyser, A.; Dunin-Borkowski, R. E. In Situ Redox Cycle of a Nickel-YSZ Fuel Cell Anode in an Environmental Transmission Electron Microscope. Acta Mater. 2010, 58, 4578−4589. (212) Jeangros, Q.; Hansen, T. W.; Wagner, J. B.; Damsgaard, C. D.; Dunin-Borkowski, R. E.; Hebert, C.; Van Herle, J.; Hessler-Wyser, A. Reduction of Nickel Oxide Particles by Hydrogen Studied in an Environmental TEM. J. Mater. Sci. 2013, 48, 2893−2907. (213) Atkinson, A.; Moon, D. P.; Smart, D. W.; Taylor, R. I. Tracer Diffusion Studies in NiO Bicrystals and Polycrystals. J. Mater. Sci. 1986, 21, 1747−1757. (214) Atkinson, A.; Taylor, R. I. The Self Diffusion of Ni in NiO and Its Relevance to the Oxidation of Ni. J. Mater. Sci. 1978, 13, 427−432. (215) Xiong, H. F.; Motchelaho, M. A.; Moyo, M.; Jewell, L. L.; Coville, N. J. Effect of Group I alkali metal promoters on Fe/CNT catalysts in Fischer−Tropsch synthesis. Fuel 2015, 150, 687−696. (216) Shroff, M. D.; Kalakkad, D. S.; Coulter, K. E.; Kohler, S. D.; Harrington, M. S.; Jackson, N. B.; Sault, A. G.; Datye, A. K. Activation of Precipitated Iron Fischer−Tropsch Synthesis Catalysts. J. Catal. 1995, 156, 185−207. (217) Chenna, S.; Banerjee, R.; Crozier, P. A. Atomic-Scale Observation of the Ni Activation Process for Partial Oxidation of Methane Using In Situ Environmental TEM. ChemCatChem 2011, 3, 1051−1059. (218) Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. Gold Catalysts Prepared by Coprecipitation for Low-Temperature Oxidation of Hydrogen and of Carbon-Monoxide. J. Catal. 1989, 115, 301−309. (219) Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings, G. J. Identification of Active Gold Nanoclusters on Iron Oxide Supports for CO Oxidation. Science 2008, 321, 1331−1335. (220) Giorgio, S.; Cabie, M.; Henry, C. R. Dynamic Observations of Au Catalysts by Environmental Electron Microscopy. Gold Bull. 2008, 41, 167−173. (221) Uchiyama, T.; Yoshida, H.; Kuwauchi, Y.; Ichikawa, S.; Shimada, S.; Haruta, M.; Takeda, S. Systematic Morphology Changes of Gold Nanoparticles Supported on CeO2 during CO Oxidation. Angew. Chem., Int. Ed. 2011, 50, 10157−10160. (222) Yoshida, H.; Kuwauchi, Y.; Jinschek, J. R.; Sun, K. J.; Tanaka, S.; Kohyama, M.; Shimada, S.; Haruta, M.; Takeda, S. Visualizing Gas Molecules Interacting with Supported Nanoparticulate Catalysts at Reaction Conditions. Science 2012, 335, 317−319. (223) Kuwauchi, Y.; Takeda, S.; Yoshida, H.; Sun, K.; Haruta, M.; Kohno, H. Stepwise Displacement of Catalytically Active Gold Nanoparticles on Cerium Oxide. Nano Lett. 2013, 13, 3073−3077. (224) Ta, N.; Liu, J.; Chenna, S.; Crozier, P. A.; Li, Y.; Chen, A.; Shen, W. Stabilized Gold Nanoparticles on Ceria Nanorods by Strong Interfacial Anchoring. J. Am. Chem. Soc. 2012, 134, 20585−20588. 3536

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

Pathway to Low-Temperature Reduction. Chem. Mater. 2005, 17, 1157−1166. (244) Sharma, R.; Crozier, P. A. In Situ Electron Microscopy of CeO2 and CeO2-ZrO2 Reduction. Electron Microsc. Anal. 1999, 61, 569−572. (245) Sharma, R.; Crozier, P. A.; Kang, Z. C.; Eyring, L. Observation of Dynamic Nanostructural and Nanochemical Changes in Ceria-Based Catalysts during In Situ Reduction. Philos. Mag. 2004, 84, 2731−2747. (246) Lopez-Cartes, C.; Bernal, S.; Calvino, J. J.; Cauqui, M. A.; Blanco, G.; Perez-Omil, J. A.; Pintado, J. M.; Helveg, S.; Hansen, P. L. In Situ Transmission Electron Microscopy Investigation of Ce(IV) and Pr(IV) Reducibility in a Rh (1%)/Ce0.8Pr0.2O2‑X Catalyst. Chem. Commun. 2003, 644−645. (247) Crozier, P. A.; Wang, R.; Sharma, R. In Situ Environmental TEM Studies of Dynamic Changes in Cerium-Based Oxide Nanoparticles during Redox Processes. Ultramicroscopy 2008, 108, 1432−1440. (248) Wang, R.; Crozier, P. A.; Sharma, R. Structural Transformation in Ceria Nanoparticles during Redox Processes. J. Phys. Chem. C 2009, 113, 5700−5704. (249) Gai, P. L.; Kourtakis, K. Solid-State Defect Mechanism in Vanadyl Pyrophosphate Catalysts - Implications for Selective Oxidation. Science 1995, 267, 661−663. (250) Cavendish, R. Structural Analysis of Nickel Doped Cerium Oxide Catalysts for Fuel Reforming in Solid Oxide Fuel Cells. Ph.D Dissertation, Arizona State University, Tempe, AZ, 2012. (251) Sharma, V.; Crozier, P. A.; Sharma, R.; Adams, J. B. Direct Observation of Hydrogen Spillover in Ni-Loaded Pr-Doped Ceria. Catal. Today 2012, 180, 2−8. (252) Bartholomew, C. H. Mechanism of Catalyst Deactivation. Appl. Catal., A 2001, 212, 17−60. (253) Forzatti, P.; Lietti, L. Catalyst Deactivation. Catal. Today 1999, 52, 165−181. (254) Datye, A. K.; Xu, Q.; Kharas, K. C.; McCarty, J. M. Particle Size Distributions in Heterogeneous Catalysts: What Do They Tell Us about the Sintering Mechanism? Catal. Today 2006, 111, 59−67. (255) Liu, R.-J.; Crozier, P. A.; Smith, C. M.; Hucul, D. A.; Blackson, J.; Salaita, G. A New Contact Sintering Mechanism Associated with Regeneration of Pd/Al2O3 Hydrogenation Catalyst. Appl. Catal., A 2005, 282, 111−121. (256) Liu, R. J.; Crozier, P. A.; Smith, C. M.; Hucul, D. A.; Blackson, J.; Salaita, G. In-Situ Electron Microscopy Studies of the Sintering of Palladium Nanoparticles on Alumina during Catalyst Regeneration Processes. Microsc. Microanal. 2004, 10, 77−85. (257) Simonsen, S. B.; Chorkendorff, I.; Dahl, S.; Skoglundh, M.; Meinander, K.; Jensen, T. N.; Lauritsen, J. V.; Helveg, S. Effect of Particle Morphology on the Ripening of Supported Pt Nanoparticles. J. Phys. Chem. C 2012, 116, 5646−5653. (258) Simonsen, S. B.; Chorkendorff, I.; Dahl, S.; Skoglundh, M.; Sehested, J.; Helveg, S. Direct Observations of Oxygen-Induced Platinum Nanoparticle Ripening Studied by In Situ TEM. J. Am. Chem. Soc. 2010, 132, 7968−7975. (259) Simonsen, S. B.; Chorkendorff, I.; Dahl, S.; Skoglundh, M.; Sehested, J.; Helveg, S. Ostwald Ripening in a Pt/SiO2 Model Catalyst Studied by In Situ TEM. J. Catal. 2011, 281, 147−155. (260) Benavidez, A. D.; Kovarik, L.; Genc, A.; Agrawal, N.; Larsson, E. M.; Hansen, T. W.; Karim, A. M.; Datye, A. K. Environmental Transmission Electron Microscopy Study of the Origins of Anomalous Particle Size Distributions in Supported Metal Catalysts. ACS Catal. 2012, 2, 2349−2356. (261) Challa, S. R.; Delariva, A. T.; Hansen, T. W.; Helveg, S.; Sehested, J.; Hansen, P. L.; Garzon, F.; Datye, A. K. Relating Rates of Catalyst Sintering to the Disappearance of Individual Nanoparticles during Ostwald Ripening. J. Am. Chem. Soc. 2011, 133, 20672−20675. (262) DeLaRiva, A. T.; Hansen, T. W.; Challa, S. R.; Datye, A. K. In Situ Transmission Electron Microscopy of Catalyst Sintering. J. Catal. 2013, 308, 291−305. (263) Peng, Z. M.; Somodi, F.; Helveg, S.; Kisielowski, C.; Specht, P.; Bell, A. T. High-Resolution In Situ and Ex Situ TEM Studies on Graphene Formation and Growth on Pt Nanoparticles. J. Catal. 2012, 286, 22−29.

(225) Cabie, M.; Giorgio, S.; Henry, C. R.; Axet, M. R.; Philippot, K.; Chaudret, B. Direct Observation of the Reversible Changes of the Morphology of Pt Nanoparticles under Gas Environment. J. Phys. Chem. C 2010, 114, 2160−2163. (226) Yoshida, H.; Matsuura, K.; Kuwauchi, Y.; Kohno, H.; Shimada, S.; Haruta, M.; Takeda, S. Temperature-Dependent Change in Shape of Platinum Nanoparticles Supported on CeO2 during Catalytic Reactions. Appl. Phys. Express 2011, 4, 065001. (227) Ovesen, C. V.; Clausen, B. S.; Schiotz, J.; Stoltze, P.; Topsoe, H.; Norskov, J. K. Kinetic Implications of Dynamical Changes in Catalyst Morphology during Methanol Synthesis over Cu/ZnO Catalysts. J. Catal. 1997, 168, 133−142. (228) Wagner, J. B.; Hansen, P. L.; Molenbroek, A. M.; Topsoe, H.; Clausen, B. S.; Helveg, S. In Situ Electron Energy Loss Spectroscopy Studies of Gas-Dependent Metal-Support Interactions in Cu/ZnO Catalysts. J. Phys. Chem. B 2003, 107, 7753−7758. (229) Vesborg, P. C. K.; Chorkendorff, I.; Knudsen, I.; Balmes, O.; Nerlov, J.; Molenbroek, A. M.; Clausen, B. S.; Helveg, S. Transient Behavior of Cu/ZnO-Based Methanol Synthesis Catalysts. J. Catal. 2009, 262, 65−72. (230) Chenna, S.; Crozier, P. A. In Situ Environmental Transmission Electron Microscopy to Determine Transformation Pathways in Supported Ni Nanoparticles. Micron 2012, 43, 1188−1194. (231) Chenna, S.; Crozier, P. A. In Situ Environmental TEM Studies to Understand the Effect of Ru Promotion on Supported Ni Catalysts. Microsc. Microanal. 2011, 17, 1602−1603. (232) Hansen, T. W.; Wagner, J. B.; Hansen, P. L.; Dahl, S.; Topsoe, H.; Jacobson, C. J. H. Atomic-Resolution In Situ Transmission Electron Microscopy of a Promoter of a Heterogeneous Catalyst. Science 2001, 294, 1508−1510. (233) Trovarelli, A., Fornasiero, P, Eds. Catalysis by Ceria and Related Materials; Imperial College Press: London, 2002. (234) Mogensen, M.; Lybye, D.; Bonanos, N.; Hendriksen, P. V.; Poulsen, F. W. Factors Controlling the Oxide Ion Conductivity of Fluorite and Perovskite Structured Oxides. Solid State Ionics 2004, 174, 279−286. (235) Gaudillere, C.; Vernoux, P.; Mirodatos, C.; Caboche, G.; Farrusseng, D. Screening of Ceria-Based Catalysts for Internal Methane Reforming in Low Temperature SOFC. Catal. Today 2010, 157, 263− 269. (236) Mogensen, D.; Grunwaldt, J. D.; Hendriksen, P. V.; DamJohansen, K.; Nielsen, J. U. Internal Steam Reforming in Solid Oxide Fuel Cells: Status and Opportunities of Kinetic Studies and Their Impact on Modelling. J. Power Sources 2011, 196, 25−38. (237) Bernal, S.; Blanco, G.; Calvino, J. J.; Hernandez, J. C.; PerezOmil, J. A.; Pintado, J. M.; Yeste, M. R. Some Recent Results on the Correlation of Nano-Structural and Redox Propertiesin Ceria-Zirconia Mixed Oxides. J. Alloys Compd. 2008, 451, 521−525. (238) Yeste, M. P.; Hernandez, J. C.; Bernal, S.; Blanco, G.; Calvino, J. J.; Perez-Omil, J. A.; Pintado, J. M. Comparative Study of the Reducibility under H2 and CO of Two Thermally Aged Ce0.62Zr0.38O2 Mixed Oxide Samples. Catal. Today 2009, 141, 409−414. (239) Simonsen, S. B.; Dahl, S.; Johnson, E.; Helveg, S. CeriaCatalyzed Soot Oxidation Studied by Environmental Transmission Electron Microscopy. J. Catal. 2008, 255, 1−5. (240) Fornasiero, P.; Kaspar, J.; Sergo, V.; Graziani, M. Redox Behavior of High-Surface-Area Rh-, Pt-, and Pd-Loaded Ce0.5Zr0.5O2 Mixed Oxide. J. Catal. 1999, 182, 56−69. (241) Otsuka-Yao-Matsuo, S.; Omata, T.; Izu, N.; Kishimoto, H. Oxygen Release Behavior of CeZrO4 Powders and Appearance of New Compounds k and t*. J. Solid State Chem. 1998, 138, 47−54. (242) Fornasiero, P.; Balducci, G.; Di Monte, R.; Kaspar, J.; Sergo, V.; Gubitosa, G.; Ferrero, A.; Graziani, M. Modification of the Redox Behaviour of CeO2 Induced by Structural Doping with ZrO2. J. Catal. 1996, 164, 173−183. (243) Montini, T.; Hickey, N.; Fornasiero, P.; Graziani, M.; Banares, M. A.; Martinez-Huerta, M. V.; Alessandri, I.; Depero, L. E. Variations in the Extent of Pyrochlore-Type Cation Ordering in Ce2Zr2O8: A t ’-k 3537

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(264) Delannoy, L.; Giorgio, S.; Mattei, J. G.; Henry, C. R.; El Kolli, N.; Methivier, C.; Louis, C. Surface Segregation of Pd from TiO2-Supported AuPd Nanoalloys under CO Oxidation Conditions Observed In Situ by ETEM and DRIFTS. ChemCatChem 2013, 5, 2707−2716. (265) Fujishima, A.; Zhang, X. T.; Tryk, D. A. TiO2 Photocatalysis and Related Surface Phenomena. Surf. Sci. Rep. 2008, 63, 515−582. (266) Kamat, P. V. Photophysical, Photochemical and Photocatalytic Aspects of Metal Nanoparticles. J. Phys. Chem. B 2002, 106, 7729−7744. (267) Kudo, A.; Miseki, Y. Heterogeneous Photocatalyst Materials for Water Splitting. Chem. Soc. Rev. 2009, 38, 253−278. (268) Osterloh, F. E. Inorganic Materials as Catalysts for Photochemical Splitting of Water. Chem. Mater. 2008, 20, 35−54. (269) Navarro Yerga, R. M.; Alvarez Galvan, M. C.; del Valle, F.; Villoria de la Mano, J. A.; Fierro, J. L. G. Water Splitting on Semiconductor Catalysts under Visible-Light Irradiation. ChemSusChem 2009, 2, 471−485. (270) Yoshida, K.; Yamasaki, J.; Tanaka, N. In Situ High-Resolution Transmission Electron Microscopy Observation of Photodecomposition Process of Poly-Hydrocarbons on Catalytic TiO2 films. Appl. Phys. Lett. 2004, 84, 2542−2544. (271) Yoshida, K.; Nozaki, T.; Hirayama, T.; Tanaka, N. In Situ HighResolution Transmission Electron Microscopy of Photocatalytic Reactions by Excited Electrons in Ion Liquid. J. Electron Microsc. 2007, 56, 177−180. (272) Cavalca, F.; Laursen, A. B.; Wagner, J. B.; Damsgaard, C. D.; Chorkendorff, I.; Hansen, T. W. Light-Induced Reduction of Cuprous Oxide in an Environmental Transmission Electron Microscope. ChemCatChem 2013, 5, 2667−2672. (273) Cavalca, F.; Laursen, A. B.; Kardynal, B. E.; Dunin-Borkowski, R. E.; Dahl, S.; Wagner, J. B.; Hansen, T. W. In Situ Transmission Electron Microscopy of Light-Induced Photocatalytic Reactions. Nanotechnology 2012, 23, 075705. (274) Liu, Q.; Zhang, L.; Crozier, P. A. Structure-Reactivity Relationships of Ni-NiO Core-Shell Co-Catalysts on Ta2O5 for Solar Hydrogen Production. Appl. Catal., B 2015, 172-173, 58−64. (275) Gai, P. L. Development of Wet Environmental TEM (WetETEM) for In Situ Studies of Liquid-Catalyst Reactions on the Nanoscale. Microsc. Microanal. 2002, 8, 21−28. (276) Gai, P. L.; Kourtakis, K.; Boyes, E. D. In Situ Nanoscale Wet Imaging of the Heterogeneous Catalyzation of Nitriles in a Solution Phase: Novel Hydrogenation Chemistry through Nanocatalysts on Nanosupports. Catal. Lett. 2005, 102, 1−7. (277) de Jonge, N.; Ross, F. M. Electron Microscopy of Specimens in Liquid. Nat. Nanotechnol. 2011, 6, 695−704. (278) Li, D.; Nielsen, M. H.; Lee, J. R. I.; Frandsen, C.; Banfield, J. F.; De Yoreo, J. J. Direction-Specific Interactions Control Crystal Growth by Oriented Attachment. Science 2012, 336, 1014−1018. (279) White, E. R.; Mecklenburg, M.; Singer, S. B.; Aloni, S.; Regan, B. C. Imaging Nanobubbles in Water with Scanning Transmission Electron Microscopy. Appl. Phys. Express 2011, 4, 055201. (280) Helveg, S.; Lopez-Cartes, C.; Sehested, J.; Hansen, P. L.; Clausen, B. S.; Rostrup-Nielsen, J. R.; Abild-Pedersen, F.; Norskov, J. K. Atomic-Scale Imaging of Carbon Nanofibre Growth. Nature 2004, 427, 426−429. (281) Sharma, R.; Iqbal, Z. In Situ Observations of Carbon Nanotube Formation using Environmental Transmission Electron Microscopy. Appl. Phys. Lett. 2004, 84, 990−992. (282) Sharma, R.; Rez, P.; Treacy, M. M. J.; Stuart, S. J. In Situ Observation of the Growth Mechanisms of Carbon Nanotubes under Diverse Reaction Conditions. J. Electron Microsc. 2005, 54, 231−237. (283) Sharma, R.; Rez, P.; Brown, M.; Du, G. H.; Treacy, M. M. J. Dynamic Observations of the Effect of Pressure and Temperature Conditions on the Selective Synthesis of Carbon Nanotubes. Nanotechnology 2007, 18, 125602. (284) Yoshida, H.; Takeda, S.; Uchiyama, T.; Kohno, H.; Homma, Y. Atomic-Scale In-Situ Observation of Carbon Nanotube Growth from Solid State Iron Carbide Nanoparticles. Nano Lett. 2008, 8, 2082−2086. (285) Hofmann, S.; Blume, R.; Wirth, C. T.; Cantoro, M.; Sharma, R.; Ducati, C.; Havecker, M.; Zafeiratos, S.; Schnoerch, P.; Oestereich, A.;

et al. State of Transition Metal Catalysts during Carbon Nanotube Growth. J. Phys. Chem. C 2009, 113, 1648−1656. (286) Sharma, R.; Chee, S. W.; Herzing, A.; Miranda, R.; Rez, P. Evaluation of the Role of Au in Improving Catalytic Activity of Ni Nanoparticles for the Formation of One-Dimensional Carbon Nanostructures. Nano Lett. 2011, 11, 2464−2471. (287) He, M. S.; Liu, B. L.; Chernov, A. I.; Obraztsova, E. D.; Kauppi, I.; Jiang, H.; Anoshkin, I.; Cavalca, F.; Hansen, T. W.; Wagner, J. B.; et al. Growth Mechanism of Single-Walled Carbon Nanotubes on IronCopper Catalyst and Chirality Studies by Electron Diffraction. Chem. Mater. 2012, 24, 1796−1801. (288) Feng, X. F.; Chee, S. W.; Sharma, R.; Liu, K.; Xie, X.; Li, Q. Q.; Fan, S. S.; Jiang, K. L. In Situ TEM Observation of the Gasification and Growth of Carbon Nanotubes using Iron Catalysts. Nano Res. 2011, 4, 767−779. (289) Kim, S. M.; Pint, C. L.; Amama, P. B.; Hauge, R. H.; Maruyama, B.; Stach, E. A. Catalyst and Catalyst Support Morphology Evolution in Single-Walled Carbon Nanotube Supergrowth: Growth Deceleration and Termination. J. Mater. Res. 2010, 25, 1875−1885. (290) Koh, A. L.; Gidcumb, E.; Zhou, O.; Sinclair, R. Observations of Carbon Nanotube Oxidation in an Aberration-Corrected Environmental Transmission Electron Microscope. ACS Nano 2013, 7, 2566− 2572. (291) Ross, F. M. Controlling Nanowire Structures through Real Time Growth Studies. Rep. Prog. Phys. 2010, 73, 114501. (292) Wu, Y. Y.; Yang, P. D. Direct Observation of Vapor-Liquid-Solid Nanowire Growth. J. Am. Chem. Soc. 2001, 123, 3165−3166. (293) Madras, P.; Dailey, E.; Drucker, J. Kinetically Induced Kinking of Vapor-Liquid-Solid Grown Epitaxial Si Nanowires. Nano Lett. 2009, 9, 3826−3830. (294) Ross, F. M.; Tersoff, J.; Reuter, M. C. Sawtooth Faceting in Silicon Nanowires. Phys. Rev. Lett. 2005, 95, 4. (295) Stach, E. A.; Pauzauskie, P. J.; Kuykendall, T.; Goldberger, J.; He, R. R.; Yang, P. D. Watching GaN Nanowires Grow. Nano Lett. 2003, 3, 867−869. (296) Miller, B. K.; Barker, T.; Crozier, P. A. Novel Sample Preparation for Operando TEM of Catalysts. Ultramicroscopy 2015, 156, 18−22. (297) Crozier, P. A.; Chenna, S. In Situ Analysis of Gas Composition by Electron Energy-Loss Spectroscopy for Environmental Transmission Electron Microscopy. Ultramicroscopy 2011, 111, 177−185. (298) Miller, B. K.; Crozier, P. A. Operando TEM of Ru/RuO2 Catalyst Performing CO Oxidation. Microsc. Microanal. 2014, 20, 815. (299) Vendelbo, S. B.; Elkjær, C. F.; Falsig, H.; Puspitasari, I.; Dona, P.; Mele, L.; Morana, B.; Nelissen, B. J.; van Rijn, R.; Creemer, J. F.; et al. Visualization of Oscillatory Behaviour of Pt Nanoparticles Catalysing CO Oxidation. Nat. Mater. 2014, 13, 884−890. (300) Imbihl, R.; Ertl, G. Oscillatory Kinetics in Heterogeneous Catalysis. Chem. Rev. 1995, 95, 697−733. (301) Hansen, K. H.; Worren, T.; Stempel, S.; Laegsgaard, E.; Baumer, M.; Freund, H. J.; Besenbacher, F.; Stensgaard, I. Palladium Nanocrystals on Al2O3: Structure and Adhesion Energy. Phys. Rev. Lett. 1999, 83, 4120−4123. (302) Yoshida, H.; Matsuura, K.; Kuwauchi, Y.; Kohno, H.; Shimada, S.; Haruta, M.; Takeda, S. Temperature-Dependent Change in Shape of Platinum Nanoparticles Supported on CeO2 during Catalytic Reactions. Appl. Phys. Express 2011, 4, 065001. (303) Campbell, C. T.; Peden, C. H. F. Chemistry - Oxygen Vacancies and Catalysis on Ceria Surfaces. Science 2005, 309, 713−714. (304) Esch, F.; Fabris, S.; Zhou, L.; Montini, T.; Africh, C.; Fornasiero, P.; Comelli, G.; Rosei, R. Electron Localization Determines Defect Formation on Ceria Substrates. Science 2005, 309, 752−755. (305) Behrens, M.; Studt, F.; Kasatkin, I.; Kuehl, S.; Haevecker, M.; Abild-Pedersen, F.; Zander, S.; Girgsdies, F.; Kurr, P.; Kniep, B.-L.; et al. The Active Site of Methanol Synthesis over Cu/ZnO/Al2O3 Industrial Catalysts. Science 2012, 336, 893−897. (306) Wulff, G. On the Question of Speed of Growth and Dissolution of Crystal Surfaces. Z. Krystallogr. 1901, 34, 449−530. (307) Sinfelt, J. H. Bimetallic Catalysts: Discoveries, Concepts and Applications, Vol. 7; ExxonMobile: Annandale, NJ, 1983. 3538

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539

Chemical Reviews

Review

(308) Tao, F. Synthesis, Catalysis, Surface Chemistry and Structure of Bimetallic Nanocatalysts. Chem. Soc. Rev. 2012, 41, 7977−7979. (309) Chen, M. S.; Kumar, D.; Yi, C. W.; Goodman, D. W. The Promotional Effect of Gold in Catalysis by Palladium-Gold. Science 2005, 310, 291−293. (310) Zeng, S.; Nguyen, L.; Cheng, F.; Liu, L.; Yu, Y.; Tao, F. Surface Structure and Chemistry of Pt/Cu/Pt(111) Near Surface Alloy Model Catalyst in CO. Appl. Surf. Sci. 2014, 320, 225−230. (311) Allard, L. F.; Bigelow, W. C.; Jose-Yacaman, M.; Nackashi, D. P.; Damiano, J.; Mick, S. E. A New MEMS-Based System for Ultra-HighResolution Imaging at Elevated Temperatures. Microsc. Res. Tech. 2009, 72, 208−215. (312) Panciera, F.; Chou, Y. C.; Reuter, M. C.; Zakharov, D.; Stach, E. A.; Hofmann, S.; Ross, F. M. Synthesis of Nanostructures in Nanowires Using Sequential Catalyst Reactions. Nat. Mater. 2015, 14, 820−825. (313) LaGrange, T.; Reed, B. W.; Santala, M. K.; McKeown, J. T.; Kulovits, A.; Wiezorek, J. r. M. K.; Nikolova, L.; Rosei, F.; Siwick, B. J.; Campbell, G. H. Approaches for Ultrafast Imaging of Transient Materials Processes in the Transmission Electron Microscope. Micron 2012, 43, 1108−1120. (314) Qiu, X. H.; Nazin, G. V.; Ho, W. Vibrationally Resolved Fluorescence Excited with Submolecular Precision. Science 2003, 299, 542−546. (315) Stipe, B. C.; Rezaei, M. A.; Ho, W. Single-Molecule Vibrational Spectroscopy and Microscopy. Science 1998, 280, 1732−1735. (316) Krivanek, O. L.; Lovejoy, T. C.; Dellby, N.; Aoki, T.; Carpenter, R. W.; Rez, P.; Soignard, E.; Zhu, J.; Batson, P. E.; Lagos, M. J.; et al. Vibrational Spectroscopy in the Electron Microscope. Nature 2014, 514, 209−212. (317) Li, Y.; Zakharov, D.; Zhao, S.; Tappero, R.; Jung, U.; Elsen, A.; Baumann, P.; Nuzzo, R. G.; Stach, E. A.; Frenkel, A. I. Complex Structural Dynamics of Nanocatalysts Revealed in Operando Conditions by Correlated Imaging and Spectroscopy Probes. Nat. Commun. 2015, 6, 7583.

3539

DOI: 10.1021/cr5002657 Chem. Rev. 2016, 116, 3487−3539