Chapter 1: Introduction to Nanocatalysts

The global population could exceed 9 billion people by 2050.¹  This increase will require tremendous efforts to establish more sustainable production systems to ensure that there are sufficient materials and energy resources in the future. Sustainable chemistry is one of the key research areas dedicated to tackling the challenges on this road. Catalysis is essential to sustainable chemistry; catalysts can improve process economy and reduce the net emissions of greenhouse gases and the associated volume of waste (Figure 1.1). Sustainability concepts also need to be further applied in green chemistry protocols, particularly in the synthesis of modern catalysts and in analysis practices.

Catalysis has evolved into a mature research area, constituted by well-defined theories and concepts. Today in catalysis, researchers search for novel catalyst materials to improve the performance of existing ones or to develop catalysts for emerging processes. The best performing catalysts are those that provide high activity, selectivity, stability and efficiency at the same time (Figure 1.2). Their performance is considered in terms of the energetic, electronic, optical and photonic efficiencies. Bringing this knowledge into practice allows researchers to prepare tailor-made catalysts, for example, materials with a combination of certain properties that are purposefully made in a controlled way.^2–4  Preparing highly functionalized catalysts is not always straightforward and the development and optimization of catalysts still relies on high-throughput technology.^5–7 

Research fields in physics,^8–10  chemistry,^11–13  biology,^14–16  medicine^17–20  and particularly catalysis,^21–29  have taken tremendous advantage of the rapid developments in nanoscience and nanotechnology. Nanoparticles (NPs) can exhibit multiple morphologies, including nanospheres, nanosheets, nanoclusters, nanograins and nanofibers. They are defined as substances in the shape of spherical dots, rods, thin plates, or any irregular shape with a cross section of less than 100 nm.³⁰  The substances can be composites, compounds, alloys or elemental solids. Nanocatalysts stand at the boundary between homogeneous and heterogeneous catalysts, in the sense that in many cases they offer advantages from both, in terms of activity, selectivity, efficiency and re-usability.^28,31,32  They go beyond the limits of homogeneous and heterogeneous catalysts via nano-effects, which are not fully understood as yet. Catalytic nano-effects result from structural,^33–35  quantum size and electronic effects from nanopore confinement.^36–39  These phenomena cause a shift towards metals with a higher binding energy (BE) in volcano-shaped correlation curves between the catalytic activity and the dissociative BE of the reactants.^39–41 

Important structural changes in the size and morphology occur during the preparation of nanocatalysts, but also when they are applied under certain conditions (e.g. during the course of a chemical reaction). The structural changes are illustrated schematically in Figure 1.3. Materials have a higher specific surface when their particle size is reduced. To give an example, for a spherical dot of 1 µm across the volume, the surface-to-volume ratio is only 1%, but for a dot of 10 nm size it is 25%. The surface-to-volume ratio reaches 100% when the solid is ca. 1 nm (sections of three atomic shells or less). The particle size reduction is often assisted by an enrichment in the surface defects (crystal edges, corners and faces). The effect on the catalyst performance also depends on how the accessible active sites are fitted into the porous structure and onto the surface of the support. Tuning the porosity characteristics of the support is an essential task in the design of functional catalysts. Porosity analysis not only provides data on the specific surface area (m² g⁻¹), pore sizes (nm) and volumes (cm³ g⁻¹), but also on the pore morphology, topology and tortuosity (interconnection between the pores). Porous materials can be classified into macro, meso and microporous materials, described by established porosity analysis methods,^42–45  although in some cases these are not applied or interpreted correctly.⁴⁶  Analytical challenges still exist in the characterization of more advanced materials, such as hierarchical porous and nanoporous materials, owing to the effects of physical confinement (phase changes, condensation, etc.).⁴⁷ 

Researchers not only pursue catalysts with a high activity and selectivity, but also catalysts with a high stability, depending on the temperature, pH, solvent, and so forth. Catalysts can be deactivated in many ways due to surface aggregation (alloying), sintering, phase changes, leaching, and so forth. A wide range of analytical techniques are used to study these effects both on the surface layers, such as in bulk nanomaterials, including inductively coupled plasma optical emission spectroscopy (ICP-OES), Fourier transform infrared (FTIR) spectroscopy, Raman spectroscopy, powder X-rays diffraction (XRD) spectroscopy, X-ray photoelectron (XPS) spectroscopy, X-ray absorption spectroscopy (XAS), electron paramagnetic resonance (EPR), scanning electron microscopy (SEM), (high resolution-) transmission electron microscopy ((HR-)TEM), scanning transmission electron microscopy (STEM), atomic force microscopy (AFM), energy-dispersive X-ray spectroscopy (EDX), high-angle annular dark-field (HAADF) imaging, electron energy loss spectroscopy (EELS), extended X-ray absorption fine structure analysis (EXAFS), thermogravimetric analysis (TGA), and temperature-programmed desorption (TPD), reduction (TPR) and oxidation (TPO). To study the stability of a catalyst, these techniques should ideally be employed operando, or at least in situ, but many of them require conditions that are different from the reaction conditions (e.g. vacuum) and are not always sufficiently time-resolved to allow studying of fast catalytic cycles and structural changes.⁴⁸  In some cases spatial-time resolved techniques have been developed.^49,50 

Recent developments in nanotechnology have enabled researchers to study confinement effects in nanocatalysts, including changes in the physical properties (condensation, adsorption, mechanical strength, plasticity, melting, diffusivity, sintering and alloying ability), and the electronic and photonic properties.^51–55  These effects result from the increased surface area of NPs along with more abundant surface defects (compared to their bulk counterparts), which eventually alter the electronic micro-environment and the electric charge transfer.^2,4,56  These effects not only affect the properties of NPs that are hosted in small voids, but also those of organic molecules and in particular their molecular vibration modes, as recently demonstrated for propane.⁵⁷  Nanoreactors are often used to study confinement effects, for example with NPs hosted inside carbon nanotubes (CNTs) for which the sizes are controlled via etching and filling techniques in the nanoscale.^3,39,58  By selectively hosting NPs in confined spaces, the catalytic activities can be tuned (Figure 1.4 ). Zhang et al. succeeded in enhancing the catalytic activity of subnanometer titania clusters⁵⁹  and tuning the redox activity of metal clusters inside CNTs (as compared to their counterparts supported on the outer wall of CNTs).⁶⁰ 

Quantum effects only occur in very small dimensions. These effects are caused by changes in the electronic configurations of atoms in nanomaterials, which are significantly different from those of their bulk counterparts.^61–63  When the size of a nanocrystal (i.e. a single crystal nanoparticle) is smaller than the Broglie wavelength, discrete electronic energy levels arise, as electrons and holes are spatially confined giving rise to the formation of electric dipoles. The separation between the adjacent energy levels increases with decreasing particle size, resulting in changes in the spatial electron energy density configurations of the nanocrystals, which eventually change the electronic and optical properties. This is schematically illustrated in Figure 1.5a–d for bulk, quantum wells, quantum wires (or rods) and quantum dots (QDs) in semiconductor materials. These quantum materials are spatially confined in one, two and three dimensions, respectively.⁶⁴  To describe the confinement effects mathematically, at least in semiconductor nanocrystals, the Bloch wave functions of the bulk materials are multiplied with an envelope function to correct for the spatial confinement of the charge carriers (electrons and holes) and the exciton. The band gap of the QDs is the sum of the fundamental bulk band gap (E_g) and the confinement energy (E_conf) of both the electrons and holes. Quantum confinement effects lead to larger band gaps with decreasing size and to the formation of discrete energy levels with different quantum numbers at the band-edges of both the conduction and valence bands (Figure 1.5e). Another bottom-up method that can be used to describe QDs is the Linear Combination of Atomic Orbitals, in which stochastic orbital wave functions are constructed from the individual atomic orbitals.^64,65  Similar to those in molecules, atomic orbitals in QDs combine into a binding and anti-binding molecular orbital (with a lower and higher energy compared to the atomic orbital energy, respectively). The electrons occupy the molecular orbitals in such a way that the potential energy of the metal nanocluster is minimized. In bulk materials, a quasi-continuum of molecular orbital energy levels exists, analogous to the conduction and valence bands of the semiconductors. The molecular orbital energies start to present discrete levels when moving from the bulk to quantum size materials and eventually molecules. In QDs, the confinement of the electrons and holes results from the lower potential energy inside the dots compared to their outer side. Further insight into the physical and mathematical background of quantum size effects in nanomaterials has recently been provided in a clear and concise way.⁶⁵  The calculation methods are based on classic approaches, such as Finite-Difference Time Domain (FDTD) Maxwell solvers, which describe nano-effects rather poorly. Input from quantum mechanical methods is required to fit the experimental data well, such as stochastic Time Domain Density Functional Theory (TDDFT).⁶⁵  Considerable discrepancies are observed when particles become extremely nanosized. For instance, important spectral differences were observed between the calculated and observed absorption spectra of octahedral Au NPs at sizes less than 3.1 nm (Figure 1.6). Semiconductor nanocrystals are mostly used in electronic and optical applications, for example QDs exhibiting different optical bands lead to different size-dependent luminescence colours. Improved charge transfer can be obtained via construction of potential wells for electrons and holes in various architectures assembled using different synthesis methods, including colloidal, epitaxially vapour phase grown and laterally gated QDs.

The design of nanocatalysts comprises of several tasks, including: (1) how to build functionalized building blocks into a nanostructure; (2) how to control the allocation of active sites on and into the support matrix; (3) how to control the dispersion and distance between active sites; and (4) how to tune accessibility to the active site. Most of the nanocatalysts are prepared via immobilization onto a support (just like heterogeneous catalysts) or via encapsulation (sandwich structures and nanoreactors), as illustrated in Figure 1.7. These architectures were recently reviewed by Zhan et al.⁶⁶  In practice, nanocatalysts are synthesized either via bottom-up or top-down strategies, as reviewed by Sing and Tandon.⁶⁷  All of these strategies present advantages and drawbacks, and should be considered part of the designers’ toolbox. Not only can the morphology of the nanoparticle (NP) itself affect the reactivity, but also the morphology of the support material, depending on the type of metal NP-support interaction. This aspect is elegantly described in Chapter 3, in which cerium oxide and zinc oxide are evaluated as support materials to deposit various metal NPs and then tested in some important reactions. The stability of the nanocatalysts is also an important aspect. Supported metal NPs face deactivation effects related to their prolonged use under reaction conditions, such as the formation of coke, aggregation, coarsening and sintering. Ensuring that they are highly dispersed on the support surface is not always a straightforward task. Some synthesis methods can overcome this, as found in Ni NPs exsolved in perovskites for example.⁶⁸  Mechanistic insights show that the exsolution occurs via formation of sub-surface nucleates and that the particle-in-a-pit morphology could provide a unique thermal stability to the NPs. Numerous core–shell and york-shell architectures have been proposed to improve the stability of nanocatalysts.⁶⁹  Finally, in view of industrial applications, other properties such as the mechanic properties and recoverability are also important to the design of nanocatalysts. In recent years magnetic nanocatalysts have become extremely popular.^70,71 

The rational design of nanocatalysts is based on the design of metal NPs (Section 1.3.2) and on the design of the space (nanopores) surrounding these NPs (Section 1.3.3). After all, confinement effects not only affect the properties of the NPs, but also those of the reactants and products.

It is widely accepted that the key elementary steps of catalytic reactions are: (1) adsorption of the reactants; (2) diffusion of the intermediates; and (3) the desorption of products.^72–74  Conversion steps involve the transfer of electrons between the catalyst surface and reactive species. The catalytic properties of nanomaterials can be controlled by tuning the spatial and energy distribution of the valence electrons at the surface, which determine their reactivity (i.e. activation energy) in various reaction pathways, and thus the selectivity. In the molecular orbital theory, the adsorption and bonding strength of reactants on catalyst surfaces is related to the orbital symmetry and the spin state of the reactant and these depend on the match of energy levels between the reactant and the catalyst surface. The d-band centre theory implies that the density of the d-band valence electrons near the Fermi level is an important factor affecting the reactivity. DFT calculations suggest that the chemisorption, activation and dissociation energies of small molecules on metal surfaces can be correlated to the d-band centre of gravity of that metal. Although this has proven to be true in many catalytic reactions, caution must be used with theoretical calculations for more complex catalysts, such as core–shell NPs, as clearly demonstrated by Gorzkowski and Lewera with the use of bimetallic Pt-Pd nanocatalysts for the oxidation of formic acid.⁷⁵  Researchers have explored how to change the electronic properties of nanomaterials by altering their morphology, size and lattice framework. The sizes of NPs can be altered via nanotechniques including soft templating, etching, colloid synthesis and vapour phase epitaxial growth. The size and size distribution of NPs can be controlled to some extent by using stabilizing or capping agents, such as ligands (ethiols, phosphines and amines), surfactants (ammonium salts), polymers (polyvinyl alcohols, polyvinyl pyrrolidone, block co-polymers), dendrimers (polyamidoamine), polyoxoanions, and so forth.⁶⁷  The use of these agents can also alter the active site (morphology) and chemical environment (steric and electronic effects).

Why are Au bulk particles not catalytically active, and why are Au NPs are highly active? Understanding the answer to this question is believed to provide useful information for a more rational design of metal NPs. Descriptors for the rational design of metal NPs include intrinsic properties such as the surface chemical composition, lattice constant, atomic density, BE, and so forth.⁶⁷  More specific descriptors include the orbitalwise coordination number (to compare adsorption energies)^76–78  and the surface distortion (to compare electrocatalytic properties).⁷⁹  More advanced descriptors in bottom-up computational engineering strategies for the design of metallic nanoclusters include the smooth overlap of atomic positions (adsorption energy),⁸⁰  many-body tensor representation, atom-centred symmetry functions, surface tension (lattice contraction/expansion)⁸¹  and the surface electrostatic potential.^61,62  The catalytic activity of Au NPs can be described as a landscape with areas of positive electrostatic potential (σ-holes) at Au atoms with low-coordination owing to the overlap of singly occupied s-orbitals (Figure 1.8). In general, the decreasing binding affinity and catalytic activity in Au nanoclusters are in the order: corners>edges>facets. This method has been applied most frequently to noble metals, but could be translated to transition metal nanoclusters, to describe their Lewis acidity for instance.⁶² 

One elegant way to tune the size and size distribution of NPs is to selectively host them within porous nanomaterials, provided the latter can be produced in a controlled way in terms of the particle size, porosity and morphology. Various nanoreactors were introduced to demonstrate confinement effects, creating unique micro-environments in spaces with nanoscale dimensions (sub-µL volumes).^82,83  Supramolecular organization of small molecules and QDs in confined spaces can give rise to novel properties.⁸⁴  The controlled synthesis of confined spaces offers a unique opportunity to tune NPs that are hosted inside nanoreactors (either via confined crystal growth or through deposition after synthesis), as recently demonstrated in numerous studies: Pt NPs for the low-temperature oxidation of ethylene,⁸⁵  Ag NPs in CNTs for the hydrogenation of dimethyl oxalate,⁸⁶  ZnO NPs on nitrogen-rich carbon spheres⁸⁷  and Rh NPs on SBA-15.⁸⁸  Spatial confinement can also prevent undesired morphology changes to NPs such as sintering,⁸⁹  exfoliation/pulverization of silicon NPs during use in the anode materials of batteries⁹⁰  and agglomeration in MCM-22 of Pt clusters into large Pt NPs during CO oxidation in the water–gas shift (WGS) reaction.⁹¹  One aspect that still remains a challenge is the construction of hierarchical porous structures without collapse of the pores, to optimize the pore gradient for efficient mass transfer from the bulk phase to confined spaces (kinetic constrains),⁹²  particularly in electrocatalysis.^93–96 

Confinement effects can also be induced by the intimate positioning of 2D surfaces, either by neighbouring layers or by overlayers. Researchers have succeeded in synthesizing a variety of non-layered materials into 2D structures including graphene, silicene, hexagonal boron nitride, metals (Ag, Fe, Ru, etc.), metal oxides (NiO, TiO₂, CeO₂, etc.), metal chalcogenides (PbS, CuS, ZnSe, etc.), topological crystalline insulators (Pb_1−xSn_xSe, etc.) and organic–inorganic hybrid perovskites (CH₃NH₃PbI₃), among others. They can be either present in the form of free-standing atomic crystals, or they can be supported. Two concise reviews on the progress in this research field (chemistry under 2D covers) were published recently,^97,98  in terms of the synthesis methods, dimensions, applications and performances of the materials, but also from a more fundamental point of view. Gong and Bao edited a series of reviews on interfacial chemistry,⁹⁸  covering ceria catalyst model systems (for CO oxidation, WGS reaction, CO₂ hydrogenation and reforming),⁹⁹  metal oxides,¹⁰⁰  carbides¹⁰¹  and 2D surfaces such as graphene and boron nitride supported on metals and metal oxides.¹⁰²  Characteristic of these materials is their strong in-plane bonding but weak van der Waals-like interactions between neighbouring layers and/or overlayers. Single layers can be chemically functionalized and single atoms can be removed or substituted by heteroatoms via doping. Nitrogen-doped graphenes are among the most studied materials.^94,102  Single atoms and molecules (mainly gases such as H₂, O₂ and CO) can be intercalated between these 2D surfaces, which have already been studied in detail for graphene, but less often for other materials, such as 2D surfaces supported on metals or metal oxides, which have shown favourable results (as the energy needed to counter the Van der Waals forces between the overlayers to separate the layer from the support can be compensated by the strong adsorption of adsorbates on the support surface).¹⁰² 

Nowadays, the design of confined spaces and surfaces is a hot topic, not only in catalysis, but also in molecular separations, energy storage and medicinal applications. Nanoporous materials used for the design of confined spaces and surfaces are illustrated in Figure 1.9. Important descriptors of these porous nanomaterials include their crystallinity and the degree of long-range order, which not only determine their activity and selectivity, but also change their mechanical properties, which are important for their processability. These materials can be used as they are, but they are often further functionalized to tune the micro-environment in which the reactions or separations take place. They can also be used as templates to introduce desirable architectures into other materials. This means that they must be removed afterwards in some way.

Metal (oxide) NPs confined in mesoporous silicas and silicates have been studied for numerous catalytic applications^103,104  for instance in oxidations,^105–108  alkylations,¹⁰⁹  tandem reactions,^109,110  reformation,^111,112  desulfurization,¹¹³  but also in other research fields such as sensors,^114,115  drug delivery^116,117  and removal of pollutants.^118,119  In the synthesis of mesoporous silica particles control over the solubility, micellization and assembly of the ionic (cethyltrimethylammonium bromide, CTAB) and pluronic (P123) surfactants is important.¹²⁰  In addition to the CTAB : P123 ratio used, the presence of additives has important effects on the structure and morphology of silica NPs. Various methods have been applied to deposit NPs on and into the mesoporous structure of silicas and silicates. Caballero and co-workers succeeded in promoting strong NPs-support interactions by selectively hosting Ni NPs in the mesopores of SBA-15 using a deposition–precipitation (DP) method, which drastically reduced the amount of coke formation in dry methane reforming (DMR) as compared to Ni/SBA-15 in which most of the Ni was present in larger NPs on the surface located outside the mesopores (prepared by modified impregnation).¹¹¹  Other organic transformations require preferential deposition onto the surface outside the pores to increase the activity. Surface deposition of NPs can be promoted using novel deposition methods such as microwave irradiation,¹²¹  ball-milling^105,122  and continuous flow-based preparation.^106,107  SBA-15 (Santa Barbara Amorphous) and MCM-41 (Mobil Catalytic Material) are among the most studied silicates. Transition metals are deposited on and into the mesoporous silicas/silicates via direct synthesis methods or via post synthesis modifications. In general, they require high dispersion of the active sites and high stability, which can by promoted by inducing strong support-metal interactions, via the formation of a chemical bond (–O–M–O–Si–O–, M=Al, Fe, Cr, Zr, Co, Ni, Cu, etc.). Various characterization techniques are used to study the structure, morphology, acidity and metal coordination state of these materials, including porosity analysis, UV-Vis, FTIR and Raman spectroscopy, XPS, EXAFS, TEM and SEM. To really understand the difference in the catalytic performance observed among nanomaterials prepared via different synthesis methods, high-resolution methods are required such as HR-TEM, high resolution SEM and HR-STEM.¹²²  In this way, the interconnections between the hexagonally packed mesoporous channels and the metal NPs can be better observed, as demonstrated recently with SBA-15 using high resolution SEM in specific image modes.¹²³  It should be noted that classic methods used to study the chemical composition of surfaces such as EDX are not usually suitable for nano-sized materials, depending on the resolution limits. As a recent topic in this research area, Chapter 4 describes the preparation of metal-modified silicates, used for the production of fine chemical precursors.

Zeolites are crystalline aluminosilicates that were invented in the early 1960's, since then more than 100 types of zeolite have been developed and some of them have been used commercially as catalysts, ion-exchangers and adsorbents. The most common zeolite framework structures are zeolite A, Y, L and ZSM. A detailed text book on their properties and applications in catalysis has been published recently.¹²⁴  Most zeolites are strictly microporous, but more recently mesoporous zeolites have also been developed.^125,126  The technique involves connecting mesopores with micropores in a hierarchical architecture without the porous structure collapsing.^127,128  This is an important facet, because it enables better diffusion characteristics and widens the spectrum of substrates that can react inside the pores. It can also postpone or re-allocate the formation of coke on and into the zeolite structure.¹²⁹  One prominent example was demonstrated in the fluidized catalytic cracking (FCC) processes, which are performed in harsh conditions that are very demanding to the zeolite catalyst, especially in resistance to degradation by clotting, fragmentation and decrystallization. Vogt and Weckhuysen illustrated the importance of advances in the characterization techniques used to describe these phenomena in zeolites at operando.¹²⁸  Y zeolites are often used for FCC, they differ from X zeolites in the Si : Al ratio in their faujasite framework, which is an important parameter in the synthesis, structure and catalytic behaviour of zeolites.¹²⁹  Lutz described the effects of dealumination methods on the zeolite structure in a comprehensive way.¹³⁰  Y zeolite analogues can be prepared by post-synthetic removal of the framework Al, which emigrates and eventually turns into extra-framework aluminous species. Framework Al can be removed by acid extraction, isomorphous substitution and thermochemical treatment of NH₄Y, provided that the extra-framework Al is effectively washed out. During dealumination mesopores are formed, but only those that are hierarchically interconnected with the micro and macropores contribute to the transportation of larger molecules. In addition to the Si : Al ratio, the micro-environment of the framework Al (acid sites) is fundamentally important to the reactivity and stability of the zeolites.^131–134  In addition to the hierarchical porosity and tuning of the acid sites, recent advances have also described the synthesis of nanosized zeolites.^135,136  Li et al. reported the synthesis of zeolite L crystals with uniform sizes of less than 30 nm.¹³⁶  Confinement effects have been observed in nanosized zeolites¹³⁷  and nanosized zeolite crystals can provide a longer catalyst life time.¹³⁸  Zeolites are used in various applications such as in sensors for exhaust gas streams, detergent production, hydrocracking, dewaxing, isomerization, alkylation, the production/derivatization of aromatics, and more recently in biomass conversion and the production of fine chemicals.

The unique combination of organic and inorganic moieties, mostly arranged in a crystalline 3D structure, has afforded a new class of porous solids: reticular materials. Reticular materials have been postulated to be the most promising class of materials to achieve high activity, selectivity and energetic efficiency at the same time.³¹  Reticular materials include metal–organic frameworks (MOFs), covalent organic frameworks (COFs) and porous molecular materials, among others. They are heterogeneous materials constituted of molecular building units, which allow the integration of well-defined highly functionalized catalysts within the backbone of the architecture. COFs consist of well-defined 2D or 3D crystallites built from exclusively light elements (i.e. B, N and O) which are connected by strong covalent bonds to form rigid structures with pore sizes ranging between 0.7 and 2.7 nm, whereas MOFs are constructed by joining metal-containing units (secondary building units, metal ions or metal clusters) with organic linkers using strong bonds to create an open crystalline framework.¹³⁹  In general, COFs exhibit a higher thermal and solvent stability than MOFs. MOFs and COFs both exhibit considerable surface areas and tuneable pore metrics, allowing the effective diffusion of various reactants to the active sites. Research efforts have also been directed at improving their optoelectronic properties, in particular their photonic efficiency in terms of the bandgap adjustment and charge carrier mobility (e.g. photocatalytic CO₂ reduction).¹⁴⁰  Various up to date reviews have been published, mostly on the synthesis, characterisation and applications of MOFs^140–145  and COFs.^146–148  Progress has also been reported in their characterization using transmission electron microscopy (TEM), a technique that was initially considered to be inappropriate owing to their sensitivity to electron beams.¹⁴⁹  Although industrial applications of MOFs and COFs at present are quasi non-existent (mainly due to the elevated costs associated with their synthesis), several research groups have already prepared these materials at the pilot scale.¹⁵⁰  MOFs and COFs have potential applications in organic synthesis, catalysis, cancer therapy, sensors, electronics, separations, energy conversion, storage, and so forth. Various design strategies and post-synthesis modifications have been reported as a function of their application,^151–153  including the synthesis of nanocomposites using polymers to improve their recoverability.^154,155  A Cambridge structural database on MOFs has recently been created.¹⁵⁶  Prominent design aspects are assessed in Chapter 5, not only for MOFs but also for metal NPs hosted in MOFs.

In contrast to mesoporous silicas and silicates, zeolites and most of the reticular materials described above, carbon materials do not contain metals (only in trace amounts). Numerous carbon forms and anisotropic variants exist. They are versatile materials which can be chemically adapted via functionalization and doping. They can be used to prepare materials for numerous applications, such as batteries, supercapacitors, fuel cells, separations, catalysts and catalyst supports, carbon capture, gas storage, sensors, drug delivery, and so forth.¹⁵⁷  Carbon materials include activated carbons (AC), graphite and diamond. More recently numerous carbon nanoforms have been described, including graphene oxide (GO), graphene, fullerene, carbon nanotubes (CNTs), carbon nanohorns, and so forth.¹⁵⁸  They perform particularly well in photo and electrocatalytic applications as hole and electron transporting materials, as additives in perovskite layers in solar cells and as electrodes. The high compatibility of GO and reduced GO materials with perovskite materials is promising for the future of solar energy.^158,159  The state of the art in the field of electrocatalytic carbon nanomaterials – with graphene materials standing at the front – consists of preparing few-layered materials (thick layers and multilayer graphitic carbon or composites having considerably lower charge transfer performances).⁵⁶  Catalytic confinement effects have been demonstrated in various circumstances, especially in CNTs^38,39,86  and in (nitrogen doped) graphene QDs.^160–162  Synthesis routes are bottom-up (hydrothermal treatment, pyrolysis and stepwise organic synthesis) or top-down (cleavage of bulk carbon materials via acidic oxidation, electrochemical exfoliation, and solvothermal treatment, etc.). Top-down routes have some advantages such as a simple operation, abundant precursors, high water solubility and easy surface functionalization, but have a time-consuming and expensive cutting process, sophisticated separation processes and size-controlled production is difficult to achieve.^163,164  In the last decade, the preparation, characterization and application of nanostructured carbon materials have been extensively reported in the literature. In this book, the focus will be set on the synthesis of nanostructured carbons for applications in energy conversion and storage, and electrochemical sensing and organic transformations, as described in Chapter 6.

Nanocatalysis opens up many doors in sustainable chemistry. Novel nanomaterials have been developed for numerous applications (Figure 1.10), ranging from more established research areas such as refining, environmental remediation, and food processing to emerging research areas such as biorefinery processes, reforming, sensors, and energy conversion and storage. The present book presents recent topics in key application areas of nanocatalysts in Chapter 3 (fundamental processes), Chapter 4 (fine chemicals), Chapter 6 (energy storage and green chemistry), Chapter 7 (biorefinery), Chapters 9 (energy conversion and storage), Chapter 10 (reforming), Chapter 11 (environmental applications), and Chapter 12 (sensors).

Catalysis has an important position in the 12 green chemistry principles,¹⁶⁵  mainly because it reduces the unit production costs, the amount of waste and the energy use associated with a chemical process.⁷²  Advances in refining petroleum using nanocatalysts have been reviewed recently.¹⁶⁶  Nanocatalysts have also shown tremendous potential for the production of bulk chemicals starting from renewable compounds (biorefinery processes in Chapter 7 and CO₂ conversion in Chapter 8).¹⁶⁷  The most prominent organic transformations for the selective production of fine chemicals include hydrogenations, oxidations, alkylations, cyclizations, amidations and tandem synthesis. In addition to petroleum, fine chemicals can also be produced from biomass. One particular example is the production of terpenes and their oxides (e.g. α-pinene oxide). Chapter 4 demonstrates how Co NPs deposited on mesoporous silicates and zeolites can be used as efficient and robust catalysts for the selective oxidation of α-pinene oxide to campholenic aldehyde. In addition to reducing waste volumes and improving the economy of the processes, nanocatalysts also provide opportunities to improve the energy efficiency and reduce the carbon footprint of a process. Hydrogen (H₂) is considered to be one of the most sustainable energy carriers for the future.^168–170  At present, the vast majority of industrial H₂ production occurs via reforming reactions starting from methane, methanol, ethanol, acetic acid, and so forth. These processes require harsh conditions, associated not only with high energy use but also with significant catalytic deactivation effects. Nanocatalyst synthesis strategies to overcome these challenges are proposed in Chapter 10. Significantly lower energy use is associated with photocatalytic H₂ production, however considerably slower kinetics and lower yields are found. Important advances using plasmonic nanocatalysts are currently under way, as described in Chapters 9 and 11. Green chemistry principles should be applied to all of the processes associated with the life cycle of a production chain, including the synthesis of the nanocatalyst itself. In recent years, various green synthesis protocols for nanocatalysts have become popular.^{28,105,171–174}  Another important sustainability factor, in terms of resource management, is the use and recycling of metals. Transition metal oxides have been widely investigated as a substitute for noble metals.^175–177  Various predictive models based on the Sabatier principle (optimal BE), activity maps, d-band energy, coordination number and Slater orbitals have been used to explore transition metal catalysts,^178–180  in which computational chemistry has played an important role.¹⁸¹  Researchers have studied the confinement of transition metal NPs in nanoporous materials to give catalytic performances close to those of noble metals.^182–184  Carbon derived nanomaterials are interesting metal-free candidates which could take over the important role of Pt catalysts, not only in electrocatalysis, but also in various organic transformations (e.g. hydrogenations).^185–187  Chapter 6 presents some recent advances in this field.

Lignocellulose is by far the most abundant renewable biomass source on Earth, and consists mainly of the biopolymers cellulose, hemicellulose (carbohydrates) and lignin (which has a highly heterogeneous aromatic structure). Extractives and ashes make up the rest of the composition. In the past biorefinery research was mostly dedicated to the production of pulp and paper, in the last decade enormous efforts have been undertaken to develop novel scenarios, within current pulp production plants, or for use in independent biorefinery scenarios. Advances in catalysis are vital for the development of viable biorefinery scenarios. Figure 1.11 illustrates the role of nanocatalysts in the valorisation of biomass. The design of nanocatalysts for biomass conversion relies on the well-established knowledge base of catalytic petroleum refining, but intensive optimization is required, as biomass is mostly constituted of heterogeneous polymers with highly functionalized hydrophilic subdomains, which are not well matched with more the hydrophobic catalysts designed for petroleum refining. Research efforts in the last decade have mostly focused on lignocellulose and algae based biorefinery (e.g. for the production of second and third generation biofuels). Currently, at least for the large scale production of biofuels, and from a techno-economic point of view, projections for the near future are more favourable for lignocellosic biomass compared to algae biomass.^188,189  Nevertheless, important advances in different refinery scenarios are on the way for algae-based research as well.^190,191 

Lignocellulosic biomass is highly heterogeneous in structure and composition. This implicates considerable challenges related to the reactivity, selectivity and downstream processing, but it also provides a unique opportunity to produce a wide range of chemicals. Tailor-made catalyst materials with a high functionality and desirable properties are required for this purpose. They should also exhibit a good re-usability. Various up to date reviews on the design of nanocatalysts (mostly heterogeneous) have been published.^192–196  However, to the best of our knowledge, only a few works in the field of nanocatalyst design for non-pyrolytic lignocellulose valorisation have been published.^197,198  Chapter 7 aims to provide further insight into how nanocatalysts can lead to improved reactivity and/or selectivity in the lignocellulose biorefinery.

Important progress has been made in the field of lignocellulose biorefineries for the production of biofuels and chemicals, including pretreatments^199–204  and catalytic conversions.^205,206  Cellulose and hemicelluloses can potentially provide a wide range of compounds to replace chemicals that are currently derived from crude oil.²⁰⁷  Whereas cellulose consist of cellobiose units and C6 glucose monomers, hemicelluloses are constituted of various C5 carbohydrate monomers (with xylose being among the most abundant). The main disadvantage of their heterogeneity is the fact that C-5 sugars are not converted biochemically as efficiently as C-6 sugars (using whole cells or enzymes). Therefore, the chemocatalytic valorisation of hemicelluloses and their co-generated waste streams is an imminent research area. After (partial) depolymerisation, cellulose and hemicelluloses can be catalytically converted into hydroxymethylfurfural and furfural, respectively.^208,209  Alternative routes from cellulose include conversion to levulinic acid,^209,210  or levulinates via alcoholysis,^211–213  and to polyols (such as hexitol).²¹⁴  Recent advances in the catalytic valorisation of hemicelluloses suggest that conversion to levulinates, lactones and furans will be the key platform chemicals,^215–218  which could be further converted to a wide range of biofuels (or precursors) and chemicals via hydrogenation (or hydrogen transfer using alternative hydrogen donors), oxidation, amination, esterification, and so forth.

In the last decade the catalytic valorisation of lignin has been at the forefront of biorefinery research, because lignin is by far the richest and most abundant renewable source of aromatics. This research field covers three important subareas: (1) lignocellulose fractionation; (2) lignin depolymerization; and (3) upgrading to chemicals. The yield of chemicals produced from lignin and their value depends on the interplay between these three subareas, which unfortunately do not always work well with each other and are often rather poorly integrated into current and future biorefinery scenarios (which are mostly directed at carbohydrate valorisation). Various reviews have described advances in the field of catalytic lignin valorisation,^219–224  but the connection and interplay between these three subareas is often missing. This aspect has been discussed recently in a clear and concise way in an up to date study.²²⁴  The authors also addressed the fact that it was not until a couple of years ago that researchers started to realize that the structure of lignin itself determined the outcome more than the type of catalytic depolymerisation and upgrading methods used. The first challenge in biomass fractionation is to design a versatile process which is compatible with multiple biomass sources (feedstocks vary in space and time). To preserve the lignin reactivity towards depolymerisation, native β-O-4′ bonds in lignin should be retained and the formation of recalcitrant carbon–carbon bonds between reactive intermediates and lignin fragments should be avoided, as this compromises the value extracted from the carbohydrates.^224–228  Degraded lignins on the other hand, constitute a more challenging feedstock for the production of chemicals, and therefore could be better used as a solid fuel or in material applications. Another challenge in the interplay is to reduce the amount of different chemicals obtained to avoid high refinery costs, which could require extra conversion steps before the refining process can be started.

The catalytic conversion of CO₂, which is the most abundant greenhouse in the Earth's atmosphere, is one of the strategies that has been proposed to mitigate CO₂ emissions to reduce the potential for global warming.²²⁹  Other CO₂ conversion strategies include absorption (e.g. in amine solvents)²³⁰  and (bio)sequestration (e.g. using algae).^190,231  The conversion of CO₂ gas using catalysts has been demonstrated in various recent studies,^232–234  but to the best of our knowledge no review study on this topic has been published recently, not to mention the improvements achieved from using nanocatalysts. Catalytic CO₂ conversion pathways are based on: (1) conventional catalysts; (2) photocatalysts; (3) electrocatalysts; (4) photoelectrocatalysts; and (5) photothermocatalysts. These pathways are assessed in Chapter 8, with a focus on the use of different types of nanocatalysts, including nanospheres, nanosheets, nanorods, nanoclusters and nanofibers.

Reforming reactions can provide cost-effective ways of generating H₂. The main disadvantage of reforming, as compared to photocatalytic water splitting, is that CO and CO₂ products are generated, which need to be processed properly to minimize their impact on global warming. Although methane reforming and WGS are established processes, at present they are still studied intensively, particularly using nanocatalysts.^235–237  Important effects from the interaction with the support and from promotors have been demonstrated. These parameters, in particularly the size and dispersion of the NPs, are of primary importance to facing the harsh conditions present in dry and steam methane reforming, which typically causes catalyst deactivation due to coke formation and migration of the NPs outside the pores leading to them sintering into larger NPs. More recently, other hydrogen sources such as methanol,^238–240  ethanol,^241–243  acetic acid^244,245  and glycerol²⁴⁶  have also been studied for reforming to produce H₂ for fuel cells. The important aspects of nanocatalyst design for reforming reactions are discussed in Chapter 10.

In the last decade the development of novel 2D photocatalysts has grown exponentially.²⁴⁷  Ab initio simulations have made great contributions to the more rational design of photocatalysts. Water splitting is one of the most reported areas in the application of photocatalysts, as it could provide a sustainable technology for the generation of H₂. Photocatalytic systems in water splitting are denominated as artificial leaves, as they deal with the same thermodynamics as photosynthetic catalysts do in nature. Still, at present, more cost-effective methods exist for H₂ production such as natural gas reforming.²⁴⁸  In fact, reduction of the production cost of photocatalysts is one of the key challenges to improving the competitiveness of water splitting.²⁴⁹  Water splitting involves a redox reaction in which protons are reduced to H₂ and water is oxidized to O₂, as illustrated in Figure 1.12a. Photocatalytic water splitting can be considered as being electrocatalytic water splitting, in which the potential shifts are induced by a photo absorber. The development of photocatalysts requires tedious work on the design, because the overall process involves multiple steps, which occur in different times scales and with different spatial resolutions. These steps include: (1) photon absorption; (2) exciton separation; (3) charge carrier diffusion and transport; and (4) mass transfer (ion diffusion). Recently, some important reviews have described the progress in photocatalyst design for water splitting very well.^249–254  In general, 2D materials demonstrate the best performance as photocatalysts, because they are layered structures exhibiting only weak interactions in between the layers (Van der Waals and hydrogen bonding), they can be obtained from their bulk counterparts via exfoliation. Figure 1.12b shows the elements involved in inorganic 2D systems (p-block and d-block elements), and Table 1.1 lists the formation energies of some 2D materials (which are useful for engineering materials with optimized band gaps).

Takanabe evaluated the rational design and the performance of photocatalysts based on quantitative data and calculations.²⁵⁵  These data included: (1) the H₂ generation rate or the solar-to-H₂ conversion efficiency (% of H₂ energy to total solar irradiation energy); (2) exciton BE; (3) charge carrier lifetime, concentration, diffusion and transport; (4) the Fermi energy level, reduction potential, bandgap, dielectric constant and isoelectric point in water at the semiconductor interface; and (5) the electrocatalytic efficiency (e.g. by comparing overpotentials) and activity (photocurrent intensity, but also the metal-hydrogen, metal-oxygen and metal-oxyhydroxide bond strengths). Detailed information on how to assess the effect of these parameters has been provided recently.^255,256  Essential to the design of a good photocatalyst is the choice of an appropriate electrocatalyst, that is species that are optically transparent and show good compatibility with the semiconductor material at the interface (to reduce the effective bandgap and the overpotential). Qualitative characteristics should also be compared such as the effect of pH on the activity, the tolerance to acidic and alkaline media and the stability (‘self-healing’ capability). The charge carrier transport is driven by the potential gradients. Therefore, semiconductor–electrolyte interfaces need to be designed carefully as a function of the semiconductor Fermi-level and the electrolyte reduction potential. Important properties of the semiconductor materials are the electronic configuration (which in turn determines the densities of the energy states) and the absorption coefficient (which determines the film thickness). Semiconductor materials are designed based on their density of state (DOS), which is based on the Franck–Condon principle for photo absorption.^257,258  They should have narrow bandgaps with appropriate band structures (see Chapter 9).

Plasmonic photocatalysts have recently received a lot of attention. Significant improvements in various steps of the water splitting process have been demonstrated.^256,259  Their design is based on the integration of plasmonic NPs (mostly Au and Ag) in a semiconductor material. Various efforts have been made to describe the role of the surface plasmon resonance (SPR) effect in the water splitting process, but it was not until recently that further insight was provided.^{256,259–261}  SPR is a confinement effect attributed to nanostructures that exhibit local surface plasmon resonance (LSPR) regions to induce hot-electron injection, near-field effects and light scattering/trapping. LSPR is caused by the collective oscillations of the surface conduction band electrons excited by an oscillating electric field, typically a photon. LSPR effects can enhance the photocatalytic efficiency, provided that plasmonic hot holes are confined within the electrocatalytic sites at the interface with the semiconductor materials. Photocatalysts for environmental applications are described in Chapter 11, including the preparation methods used to construct plasmonic nanocomposites.

Sensors comprise a large group of devices used for the detection or (semi-)quantitative determination of target analytes and for the determination of physical parameters. The role of sensors in sustainable chemistry cannot be overlooked, as they may offer reliable, fast and in situ alternatives to established analytical determination methods, which are often tedious and time consuming, and have additional requirements for proper waste management. On the one hand, sensors (or at least their recognition elements) can be considered to be a particular research field of nanocatalysts. On the other hand, novel sensors can find potential applications in many different research areas, ranging from ‘basic’ parameters such as temperature and pH in physics to single molecule detection in biology. The fast progress of both nanoscience and nanotechnology has contributed enormously to the development of novel sensing devices and materials (selectivity, sensibility, miniaturization, etc.). In addition confinement effects can provide unprecedent properties; the concept of confined spaces is currently being explored in various single molecule detectors, which are especially important in biological applications (nucleic acids, peptides, proteins and other biomolecules), but also for chemicals, polymers, (micro)pollutants, and so forth.^262–264  Advanced examples include the detection of purine bases in DNA by Au NPs confined in MFI zeolites,²⁶⁵  the detection of H₂O₂ at 30 nM sensitivity level by CNTs-based electrochemical sensors,²⁶⁶  the conversion of the redox properties of various analytes to a detectable ionic current²⁶⁷  and the model-free dynamic observation of enzymatic activity in nanopores.²⁶⁸  Chapter 12 introduces the fundamental concepts of sensing and presents a case-study on Au NPs for the detection of bisphenol A with very low detection limits.