Review pubs.acs.org/CR
Hollow Nano- and Microstructures as Catalysts Gonzalo Prieto, Harun Tüysüz, Nicolas Duyckaerts, Johannes Knossalla, Guang-Hui Wang, and Ferdi Schüth* Department of Heterogeneous Catalysis, Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mülheim an der Ruhr, Germany ABSTRACT: Catalysis is at the core of almost every established and emerging chemical process and also plays a central role in the quest for novel technologies for the sustainable production and conversion of energy. Particularly since the early 2000s, a great surge of interest exists in the design and application of micro- and nanometer-sized materials with hollow interiors as solid catalysts. This review provides an updated and critical survey of the ever-expanding material architectures and applications of hollow structures in all branches of catalysis, including bio-, electro-, and photocatalysis. First, the main synthesis strategies toward hollow materials are succinctly summarized, with emphasis on the (regioselective) incorporation of various types of catalytic functionalities within their different subunits. The principles underlying the scientific and technological interest in hollow materials as solid catalysts, or catalyst carriers, are then comprehensively reviewed. Aspects covered include the stabilization of catalysts by encapsulation, the introduction of molecular sieving or stimuli-responsive “auxiliary” functionalities, as well as the single-particle, spatial compartmentalization of various catalytic functions to create multifunctional (bio)catalysts. Examples are also given on the applications which hollow structures find in the emerging fields of electro- and photocatalysis, particularly in the context of the sustainable production of chemical energy carriers. Finally, a critical perspective is provided on the plausible evolution lines for this thriving scientific field, as well as the main practical challenges relevant to the reproducible and scalable synthesis and utilization of hollow micro- and nanostructures as solid catalysts.
CONTENTS 1. Introduction 2. Synthesis Strategies toward Hollow Nano- and Microstructures 2.1. Wet Chemistry Synthesis Routes 2.1.1. Soft-Templating Methods 2.1.2. Hard-Templating Methods 2.1.3. Template-Free Methods 2.2. Spray Synthesis Routes 2.2.1. Spray Drying 2.2.2. Flame Spray Pyrolysis 2.2.3. Spray Pyrolysis 2.3. Catalytic Functions in Hollow Materials 3. Application of Hollow Nano- and Microstructures as Solid Catalysts 3.1. Mass Transport Phenomena in Hollow Catalysts 3.2. Catalyst Stabilization via Encapsulation in Hollow Structures 3.2.1. Stabilization of Encapsulated Nanoparticles against Sintering 3.2.2. Stabilization of Encapsulated Nanoparticles against Leaching and Aggregation in Liquid Phase Reactions 3.3. Size-Selective and Poison-Resistant Catalysis via Molecular Sieving 3.3.1. Encapsulation by Amorphous Shells © 2016 American Chemical Society
3.3.2. Encapsulation by Crystalline, Molecular Sieve Shells 3.4. Void-Confinement Effects in Catalysis 3.5. Spatial Compartmentalization of Multiple Catalytic Functionalities 3.6. Stimuli Responsive Catalysts 4. Application of Hollow Nano- and Microstructures as Biocatalysts 4.1. Stabilization of Enzymes via Encapsulation 4.2. Substrate Size-Selective Biocatalysis 4.3. Compartmentalization of Enzymes for Cascade Biocatalysis 5. Application of Hollow Nano- and Microstructures As Electro- And Photocatalysts 5.1. Advantages of a Hollow Structure in Electroand Photocatalysts 5.2. Hollow Nano- And Microstructures As Electrocatalysts 5.3. Hollow Nano- and Microstructures as Photocatalysts 6. Concluding Remarks and Outlook Author Information Corresponding Author Notes
14057 14059 14060 14060 14064 14067 14068 14068 14069 14069 14070 14071 14072 14073 14074
14076 14077 14077
14078 14081 14083 14088 14090 14090 14092 14094 14095 14095 14097 14100 14103 14105 14105 14105
Received: June 14, 2016 Published: October 7, 2016 14056
DOI: 10.1021/acs.chemrev.6b00374 Chem. Rev. 2016, 116, 14056−14119
Chemical Reviews Biographies Acknowledgments References
Review
14105 14106 14106
1. INTRODUCTION Catalysis is a core field, both in chemical research and the chemical and related industries. While a very high level of control and understanding has been achieved in molecular catalysis, the majority of chemical processes in industry rely on less well-understood solid catalysts. Solid catalysts can be used as neat compounds, such as in the case of zeolites, but in most cases they are multicomponent systems, or they are composed of a species, which carries the primary catalytic functionality, supported on a high surface area solid. The active component can consist of a variety of different solids, such as metal nanoparticles, metal oxides or sulfides, or molecular species (i.e., metal complexes). Also a great variety of support materials has been used, including metal and nonmetal oxides or carbons. While it is true that the level of understanding in heterogeneous catalysis is substantially lower than for molecular catalysts, a high degree of sophistication has been achieved in the synthesis of complex solid catalysts. It is nowadays possible to control the shape of metal nanoparticles with high precision for many different types of materials,1,2 alloys can be created with a variety of different compositions,3 hierarchical materials are accessible in which control of feature sizes over a range of different length scales is possible,4−6 and catalytic functionality can be placed at will at different positions of support materials.7−9 The advances in the control of morphologies of different materials over the last years are tremendous. One particular class of morphologies, namely hollow structures, has attracted particular attention.10 This is due to the range of materials, which has become accessible in the form of hollow structures, but also to their potential advantages in catalytic applications. Figure 1 illustrates this development. Up to the mid-2000’s, there was limited interest in catalysis using hollow structures, not so much because they were not interesting but more due to the lack of synthetic methodologies to produce such structures. Since then there has been exponential increase in the number of publications relating to hollow structures in catalysis, as detailed in the different categories depicted in Figure 1. The interest in hollow structures in catalysis is due to a number of possible key advantages. (i) They effectively isolate catalytic species, which can, for instance, lead to stabilization of particles against sintering or be used for the combination of incompatible catalytic functions within the same material. (ii) They can enable cascade reactions by placing catalytic functionality in sequentially localized compartments. (iii) If pore sizes in the shells are precisely controlled, hollow structures can be used to improve selectivity of catalytic reactions by molecular sieving or strong differences of diffusivities in nanopores. In addition to these central points, there are a number of other features, which are of interest in catalytic applications of hollow structures. Hollow structures are in general suitable for improving mass transfer as compared to the same mass of bulk supported catalysts, leading, however, to an increased catalyst volume, since the shells can be made very thin so that they do not produce a severe limitation to transport of reagents and products. If the hydrophobicity/hydrophilicity is well-adjusted, hollow structures may enrich reagents in the voids, which can
Figure 1. Time-evolution of the scientific interest in hollow materials as catalysts as illustrated with the number of publications over the last 30 recent years. Data included only for complete years, hence not covering 2016. These are based on publications listed in Scopus which contain the words “hollow” and (“electro-, “photo-, “bio)catalyst”, as indicated in the legend of the plot, within the title, keywords, and abstract.
lead to increased reaction rates. In several cases, it is also claimed that they have increased concentrations of low coordination sites with favorable catalytic properties and in photocatalysis hollow structures are assumed to improve scattering and thus the absorption of light. These points will be addressed and highlighted in more detail in the course of this review. Hollow structures, in general, and hollow catalysts, in particular, can exist in different configurations according to the type and number of structural features of which they consist. Many, if not most, hollow structures are more or less spherical, but this is not a necessary condition; basically, any shape may fulfill the specific function at which it is targeted. Also, size monodispersity is often realized, but it is not mandatory for many applications. In addition, different nomenclatures can be found in the literature to describe each conceptual architecture. Figure 2 summarizes schematically some of the most common material configurations and is used to establish the nomenclature adopted in this review. In the simplest case, the material consists of single-shell hollow particles, but it can also display multiple, nested shells organized in a Matryoshka-type arrangement, leading to multishelled or onion-like hollow particles, the latter term being often reserved for cases where the number of shells is larger than two or three. Either single or multishell hollow particles can contain an individual, single particle of various possible compositions in the central compartment, resulting in so-called yolk-(multi)shell particles. When the cargo material consists of multiple particles, materials are typically denoted as rattle-type particles. More complex multishell hollow structures may combine a single yolk particle in the central compartment and a rattle-type cargo within the intershell spacing. Particularly interesting are multishell hollow materials in which the yolk and/or rattle-type fillings have different chemical composition in the different intraparticle compartments. Principally, various other situations can be conceived, and synthetic methodologies 14057
DOI: 10.1021/acs.chemrev.6b00374 Chem. Rev. 2016, 116, 14056−14119
Chemical Reviews
Review
Figure 2. Schematic representation of various possible configurations of synthetic hollow materials. For convenience, a general spherical morphology is adopted in all cases. Red and green balls represent typically metal (oxide) nanoparticles. (a) Single-shell hollow particle, (b) multishelled or onionlike hollow particle, (c) yolk−shell particle, (d) rattle-type particle, and (e) yolk multishelled particle with additional intershell nanoparticle cargo.
have reached a level of maturity that almost any structure is experimentally accessible. These synthetic pathways, including soft- and hardtemplating, template-free methods, and spraying methods, will be covered in the first part of this review, followed by a short overview of methods to integrate catalytic functionality. The second part of the article focuses on the potential of hollow structures in order to impart specific functionality to different types of catalysts. These include the stabilization of catalytic species, the creation of size selectivity, compartmentalization of different functionalities, void confinement effects, and the generation of stimuli-responsive catalysts. The final two sections address the functions of hollow structures in biocatalysis and in electro- and photocatalysis. The review concludes with a critical appraisal of the use of hollow structures in catalysis in the light of the studies published up to now. The review focuses on materials, in which a species, often nanoparticle(s) is encapsulated in a hollow void, and where the encapsulated species does not snugly fit into the shell, but leaves additional void space, as opposed to core−shell structures, in which the core is directly covered by the shell. Different sample notations can be found in the scientific literature to denote both types of materials. For instance, I@hS or I, @S have been proposed to denote yolk−shell particles, in which the “h” (which stands for hollow) in the first case, and the “blank” following the comma in the second case, denote the void between the encapsulated species I and the hollow shell material S. Hence a gold-nanoparticle incorporated as a yolk into a ZrO2 hollow shell would be indicated as Au@hZrO2 or Au, @ZrO2, respectively, while the core−shell analogue would be denoted as Au@ZrO2. However, the latter nomenclature is also rather frequently applied in the literature for hollow structures. For simplicity, and since this review discusses almost exclusively hollow materials, we have adopted the most extended I@S notation. Whenever core−shell structures are described (e.g., as reference materials or precursors for hollow derivatives), this will be clearly indicated. One aspect with respect to the characterization of the yolk− shell structures should be mentioned before synthetic strategies are discussed. In most publications, proof for the existence of the yolk−shell structures relies almost exclusively on electron microscopy observation, coupled with the specific performance in catalytic reactions, which is interpreted in terms of the yolk− shell structures. Electron microscopy analysis, however, has the disadvantage that it only provides local information, and there is often uncertainty with respect to the selectivity of the synthesis. This is to some extent alleviated, if large areas are
covered in overview images, but such images are not often provided, so that it is in most cases unclear how selective are the syntheses really. In our own work,11,12 we show overview TEM images with over a hundred particles, demonstrating the perfection of the synthesis. For this protocol, we have for many different batches checked the perfection of the synthesis, and typically around 95% of the particles have the target morphology with a gold particle encapsulated in a zirconia shell, while of the remainder, about half of the shells are empty and half contain more than one encapsulated particle. Also in other publications, electron microscopy overviews are presented,13−15 and this information may also be in their Supporting Information.16 It would be helpful if researchers in this field would generally provide overview images in addition to high-magnification images showing the details of the structure. Finally, in some studies, particle size distributions of the encapsulated metal particles are reported based on the analysis of several hundred particles, such as, for instance, in the papers from the Farrusseng group on zeolite-encapsulated nanoparticles.17,18 This can be considered as implicit proof for a homogeneous quality of the samples. It would be ideal, if bulk analytical techniques, which provide a notably more reliable sampling, could provide information on the perfection of encapsulation. However, in analyzing these techniques one quickly realizes the limitations of most bulk techniques in this respect because in many cases just a mechanical mixture of the core particles and empty shells would give exactly the same analytical response. The only relatively generally applicable tools seem to be surface analytical methods with a limited analysis depth of only a few nanometers. When such methods are used for the analysis of the core particles, the signal from them should be strongly reduced or not present at all. Prominent examples for such methods are X-ray photoelectron spectroscopy (XPS) or secondary ion mass spectrometry (SIMS). For both techniques, the escape depth of the species carrying the signal (electrons in the case of XPS, ions in the case of SIMS) is only on the order of nanometers, depending on the exact nature of the material. A photoelectron generated in the encapsulated species will not be able to penetrate the shell, unless there are straight escape pathways through the shell, and the same holds for secondary ions. However, in spite of the general usefulness of the methods, they have not been used often. Figure 3 shows Au 4f XPS spectra gold particles encapsulated in zirconia (this is the type of particles reported in ref11) and for the same material after crushing the shells, both taken under the same conditions. As one can clearly see, the gold signal is barely visible for the encapsulated gold particles because the electrons can only pass 14058
DOI: 10.1021/acs.chemrev.6b00374 Chem. Rev. 2016, 116, 14056−14119
Chemical Reviews
Review
and nanomaterials, the reader is referred to excellent previous reviews.21−23 For the sake of discussion, the synthesis routes have been grouped into two categories. On the one hand, wet chemistry routes encompass synthesis methods in which the development of the hollow material is carried out in a liquid medium, starting from a solution, a liquid-in-liquid emulsion, or a solid-in-liquid suspension. These methods allow a high degree of control over several synthesis parameters and, thereby, over the ultimate structure of the product. As main downsides, however, one should mention the facts that they usually operate in highly diluted (often colloidal) media and, therefore, involve the processing of large volumes and are typically carried out in batches. It is challenging to extend them to operate in a continuous mode without compromising the quality of the product. On the other hand, spray techniques can be operated continuously, starting from precursors in the liquid or the gas phase. However, this goes generally at the expense of the degree of control over the isotropy of the synthesis media and the uniformity of the product at the nano- and micrometer length scales. The yolk−shell systems described in publications are often the result of rather intricate and elaborate synthetic protocols. The target is typically not the generation of industrial scale amounts of materials, since the catalysts are mostly designed for model studies. This, however, means that rather low amounts are generated by most protocols. The typical scale, at which the target materials are produced, is the 100 mg scale. In many cases, exact amounts synthesized and/or yields cannot be extracted from published data, since a substantial fraction of the syntheses involves leaching steps, and while amounts of reagents are typically reported, weights of the final products are not. There are a few notable exceptions, where higher amounts of samples were synthesized or scalability of the syntheses has been reported, although “higher amounts” or “scalability” often have to be taken with a grain of salt, since reactions typically proceed at high dilution, as mentioned before, so that even liter-scale syntheses only provide subgram amounts of sample. Some liquid phase processes are reported to be scalable, but it seems to be most straightforward in spray pyrolysis or related, intrinsically continuous methods. Gonzalez et al.16 described the synthesis of shape- and sizecontrolled hollow particles and yolk−shell or multishell particles by the Kirkendall effect. The authors report that the synthesis could be scaled from the ten milliliter scale to the liter scale without problems. Estimating from the description given in the supporting information of that publication, this would result in material amounts on the order of one gram. This example already highlights a problem with the synthesis of yolk−shell materials: they typically proceed at rather high dilution, resulting in relatively small amounts of sample even if high solution volumes are used. Higher amounts were reported by Fang et al.13 for the synthesis of alumosilicate shells, where 2 g per batch were produced. For the synthesis of yolk−shell materials, however, the batch sizes were appreciably lower, judging from the information in the supporting information. Multigram scale syntheses of Fe3O4@PANI (PANI:polyaniline) were reported by Jeong et al.,24 and the authors claim the synthesis to be scalable. Another procedure from the field of battery materials is the synthesis of encapsulated sulfur particles described in ref 15. For the production of material on the gram scale, 2 L of reaction volume were required, highlighting again
Figure 3. XPS spectra of gold particles encapsulated in ZrO2 hollow shells (purple) and of the same materials after crushing the hollow shells (yellow). Samples are described in ref 11. C. Weidenthaler is gratefully acknowledged for the measurements.
the shells under fortunate circumstances, while a clear gold signal is seen for the materials where the shells had been crushed. A similar approach, albeit without analyzing a crushed comparison sample, was followed by Zeng et al.,19 who encapsulated Co3O4 in a MOF. In the photoelectron spectra, the signal corresponding to cobalt was reduced to very small intensity after encapsulation of the Co3O4 nanoparticles in the MOF. Evidence obtained by all other methods is more circumstantial. Sometimes a pronounced cavitation effect in nitrogen sorption with a wide and almost rectangular hysteresis can be taken as evidence for a void with only small apertures (typically
