A photocatalytic reaction involves charge separation and transfer under photo-irradiation, and the photogenerated charge carriers (holes and electrons) are responsible for the photocatalytic activity of the catalyst. The active centres in a single-site photocatalyst are the isolated and spatially separated sites that may interact with reactants after photo-irradiation. Generally, single-site photocatalysts perform better than other types of photocatalysts owing to the presence of the efficient active centres. A porous structure can provide more single sites and special passages for charge transport. Thus, the introduction of a porous structure into a photocatalyst may result in markedly enhanced photocatalytic reactivity, providing a promising strategy for the design and fabrication of novel photocatalysts with high performances. In this review, we summarize the developments in single-site photocatalysts, particularly those with a porous structure, such as metal-incorporated zeolites, metal–organic frameworks and porous semiconductor photocatalysts. The synthesis, structures and catalytic performances of these single-site photocatalysts have been described, and characterization and reaction mechanisms for single-site photocatalysts have also been detailed. Finally, we point out the significance of study on single-site photocatalysts with a porous structure.
Over 80 years ago, Taylor (1925) introduced the concept of catalytically active sites. The active sites with the highest catalytic activity are atoms situated at special sites (peaks, fissures and other crystalline discontinuities). Usually, high catalytic activity is attributed to the preferential adsorption on a catalyst surface taking place at the atomic level. With the development of material science and characterization technology, the concept of single-site heterogeneous catalysts (SSHCs) was proposed by Thomas and co-workers (Thomas 1988, 1999; Thomas & Raja 2001; Thomas et al. 2005). Single-site catalysts are a class of solid catalysts in which one or a small number of isolated catalytic species are firmly and uniformly anchored to a support (generally porous materials; Thomas & Raja 2001; Yu et al. 2002).
‘Single site’, the catalytically active centre, may consist of one or more atoms. Such single sites are spatially isolated from one another, without spectroscopic or other cross talk between such sites. Each site has the same energy of interaction between it and a reactant as every other single site; and each such site is structurally well characterized. Since the concept of the single-site catalysts was proposed, a great deal of research has been carried out in this field. The investigation of single-site catalysts facilitates the determination of the kinetics and mechanism of catalytic turnover—both experimentally and computationally—and makes the energetics of various intermediates accessible.
Since Fujishima & Honda (1972) discovered the phenomenon of ultraviolet (UV) light-induced photocatalytic splitting of water on a TiO2 electrode, considerable attention has been directed towards the investigation of photocatalysis. A photocatalytic reaction is a chemical process initiated by photons absorbed by a photocatalyst (usually a semiconductor material). Separated electrons and holes are generated in the photocatalyst. These charge carriers can move to the surface of the photocatalyst and react with reactants such as water, pollutants and organic molecules. On the basis of the mechanism for photocatalytic reactions, an efficient photocatalyst must bear the following merits: (i) its electronic band gap is comparable to the energy of the incident light; (ii) it can produce separable charges; and (iii) it is good in transporting charge carriers so that the migration of photon-generated charges to the catalyst surface is facilitated (Fujishima et al. 2008). Owing to the enhancement in electron–hole separation, a space charge layer at the material interface may improve the efficiency of photocatalysts with dual- or multi-components (Fujishima & Honda 1972). SSHCs composed of individual isolated active sites and an open structure, such as mesoporous silica, have proved to be efficient in the separation of the photo-induced charges because of the existence of the interfaces (Osterloh 2008). Photocatalysts with the features of SSHCs are called single-site photocatalysts (SSPCs; Anpo & Thomas 2006). These photocatalysts have been widely used in the photodecomposition of pollutant molecules, generation of hydrogen by water splitting and the photocatalytic synthesis of organic species.
Porous materials, ranging from zeolites to coordination polymers, offer considerable internal surface areas for the separation, manipulation and catalytic transformation of guest molecules (Stein 2003; Kitagawa et al. 2004; Bradshaw et al. 2005). Photocatalysts with a porous structure may exhibit excellent catalytic performance (Schattka et al. 2002; Sreethawong et al. 2005). Therefore, considerable efforts have been devoted to the investigation of SSPCs with various porous structures. In this review, we summarize the recent developments in SSPCs with a porous structure and their photocatalytic properties. We also touch upon the mechanisms for the photocatalytic reactions and the rational design of new porous single-site photocatalysts.
2. Individual isolated metal ions in zeolites as photocatalytically active sites
Metal cations including transition metal cations can be introduced into the framework or the micropores of zeolites by various methods, such as the hydrothermal and ion-exchange approaches. The introduction of these metal cations remarkably modifies the catalytic and photocatalytic activities of zeolites. For example, zeolites doped with tetrahedrally coordinated Ti4+ atoms exhibit high photocatalytic activity under UV light (Yamashita & Anpo 2003; Anpo 2004). The high photocatalytic activity is due to the formation of Ti3+ active sites. The same phenomenon is also observed in ZSM-5 and zeolite Y containing two coordinated Cu+ cations. These Cu+-containing zeolites show high photocatalytic activities in the conversion of NO to N2 and O2 (Anpo et al. 1994; Matsuoka & Anpo 2003). Titanium oxides anchored within micropores of zeolites can photocatalytically reduce CO2 with H2O (Anpo et al. 1997). It is believed that, under photo-irradiation, the single-site metal atoms within the micropores or the framework of zeolites facilitate the formation of the charge transfer excited state and then improve the photocatalytic performance of the material.
Activation of C–H bonds in saturated hydrocarbons under mild conditions is a great challenge. Recently, Chen and co-workers reported a Zn+-modified ZSM-5 material that shows substantial photocatalytic activity for the dehydrogenative coupling of methane to ethane and hydrogen under UV and sunlight irradiation at room temperature (Li et al. 2011). First, a protonated Zn2+–ZSM-5− zeolite was prepared through the interaction of the zeolite with metallic zinc vapour. After the irradiation under UV light for 1 h, Zn+ species formed as a result of the photo-induced one-electron transfer from the zeolite framework to the 4s orbital of the Zn2+ cation in the as-prepared Zn2+–ZSM-5−. The 4s electron of the Zn+ cation in the resulting (Zn+, Zn2+)–ZSM-5− migrates back to the zeolite framework under thermal treatment in the absence of UV irradiation (figure 1). The energy needed to power the Zn+-modified ZSM-5 photocatalytic system is much lower than those systems reported previously (Choudhary et al. 1997; Arakawa et al. 2001; Zheng et al. 2008). In a related research system, electron transfer from a zeolite framework to guest organic molecules has been unambiguously observed (Li, J. R. et al. 2009; Li, L. et al. 2009), demonstrating that electron transfer in zeolite structures is not unusual. These results suggest that the design and synthesis of highly photoactive SSPCs with a zeolitic feature is quite feasible.
Metal cations have also been introduced into mesoporous and macroporous structures. For example, titanium has been loaded into various mesoporous materials, such as MCM-41, MCM-48 and hexagonal mesoporous silica. These Ti-containing mesoporous materials exhibit catalytic activities in the reduction of CO2 with H2O (Anpo et al. 1998), hydroxylation of benzene (Guo et al. 2003), decomposition of NO (Zhang et al. 2000; Hu et al. 2006), partial oxidation of alkenes (Yamashita et al. 2003) and decomposition of water (Liu et al. 2003).
In brief, the charge transfer excited state of the metal ions formed under irradiation played a crucial role for photocatalytic activity of the microporous or mesoporous systems with individual isolated metal ions or metal clusters.
3. Photocatalytically active sites in uranyl-containing metal–organic compounds
Metal–organic frameworks (MOFs) are a new type of porous materials with crystalline frameworks. The crystal structure of an MOF is an extended three-dimensional network constructed by metal ions or small discrete clusters through the linkage of multi-dentate organic molecules (Li et al. 1999; James 2003). MOFs exhibit many unique properties, such as large pore size, high surface area, good selectivity for small molecules and optical or magnetic responses to the inclusion of guests (Rowsell & Yaghi 2004). Therefore, MOFs have many potential applications in the fields of, for example, gas storage (Zhao et al. 2004), compound separation (; Li, L. et al. 2009), chemical sensing (Lu & Hupp 2010), biomedical imaging (Liu et al. 2011) and heterogeneous catalysis (Song et al. 2010).
The MOF-based catalysts are mainly classified into three types: MOFs with metal-active sites, MOFs with reactive functional groups and MOFs as host matrices or nanometric reaction cavities (Corma et al. 2010). The catalytic activities observed for MOFs with metal-active sites are directly related to the metallic components, either as isolated metal centres or as clusters connected through the organic linkers. Thus, MOFs with metal-active sites can be regarded as single-site catalysts.
Uranyl units may also be linked by organic molecules to form uranyl–organic framework (UOF) compounds. As a particular family of MOF materials, UOFs have attracted much attention in recent years because of their rich structural features, as well as their unusual physical and chemical properties (Wang & Chen 2011). The photocatalytic behaviour of uranyl compounds was observed in 1974 (Burrows & Kemp 1974). Since then, the preparation and investigation of the photocatalytic properties of uranyl compounds have received considerable interest. Nevertheless, the synthesis of three-dimensional framework uranyl–organic compounds with a porous structure has proved rather challenging, and in contrast two- and one-dimensional uranyl–organic materials are easily available. These low-dimensional uranyl compounds exhibit equally interesting photocatalytic properties.
Recently, a metal–organic extended structure [Ag(bipy)(UO2)(bdc)1.5] (structure 1 in figure 2) with a two-dimensional network was prepared using 2,2′-bipyridyl (bipy) and 1,4-benzenedicarboxylate (bdc) as mixed ligands (Yu et al. 2005). In the structure of 1, the uranyl units are connected by bridging bdc ligands to produce chains, which are then cross-linked by additional bdc groups to form a two-dimensional network decorated with [Ag(bipy)]+ subunits. This water-insoluble MOF material exhibits UV- and visible-light photocatalytic activities superior to those of commercial TiO2 (Degussa P-25) for the degradation of rhodamine B (RhB). Through the same strategy, another Ag-U-organic compound [Ag2(phen)2UO2(btec)] (structure 2 in figure 2) in the presence of 1,10-phenanthroline (phen) and 1,2,4,5-benzenetetracarboxylate (btec) was obtained (Yu et al. 2005). This uranyl–organic extended structure also exhibits photocatalytic activity for the RhB degradation. Structure 2 consists of two-dimensional layers with Ag-UO8-Ag trinuclear cores as the building blocks, which are linked by bridging btec ligands. The less spacious interlayer region and fully coordinated U centres in structure 2 are unfavourable for the access of the dye molecules. As expected, compound 2 is less efficient than 1 in degradation of RhB under both UV- and visible-light irradiations.
The effect of uranyl groups that act as photocatalytic active centres in different uranyl–organic extended structures has been further demonstrated by other uranyl-containing compounds. A uranium–nickel–organic compound [Ni2(H2O)2(QA)2(bipy)2U5O14(H2O)2(OAc)2]⋅2H2O (structure 3 in figure 2; HOAc=acetic acid, bipy=4,4′-bipyridine, H2QA=quinolinic acid) with micropores was prepared through a hydrothermal method (Yu et al. 2004). This compound exhibits photocatalytic activity for the degradation of methyl blue (MB). Through a solution volatilization route, two new uranyl-NDC compounds, UO2(NDC)[(CH3)2SO)]2 and [UO2(NDC)(CH2OH)2] (H2NDC=1,4-naphthalene dicarboxylate) with one-dimensional chain structures were synthesized (Xia et al. 2010). Both compounds possess new structural features, and exhibit high efficiency in degradation of RhB under the irradiation of UV and/or visible light.
Given the fact that Ag+ is sensitive to light, it is necessary to clarify whether or not the Ag species contribute to the photocatalytic activities of the Ag+-containing compounds. Recently, two uranyl–organic compounds, (UO2)8(NDC)12(4,4′-bipyH2)3(4,40-bipyH)3 (structure 4 in figure 2) and (UO2)3O[Ag(2,2′-bipy)2]2-(NDC)3 (structure 5 in figure 2) with and without Ag+ have been prepared (Liao et al. 2008). Both 4 and 5 have a two-dimensional layered structure with similar accessibility to active uranyl centres and similar uranium content. The comparable photocatalytic activities for these two compounds under both UV- and visible-light irradiations indicate that the uranyl units are the active centres responsible for the photocatalytic properties, whereas the Ag+ species are of less importance.
The hydrogen abstraction and electron transfer photodegradation mechanism is generally accepted for the photocatalytic reaction involving uranyl compounds (Burrows & Kemp 1974). The charge-transfer transition within the uranyl group leads to the absorption in the UV and the visible regions (Volkovich et al. 2001). Upon photoexcitation, the charge transfer actually takes place from oxygen to uranium to form an excited state *UO2+2. This excited state can abstract electrons from organic molecules (such as RhB) with an appropriate orientation, resulting in intermediates and protons. The excited electrons can be captured by electronegative substances, such as O2 in the solution, generating highly active peroxide anions. The peroxide anions further oxidize and decompose the organic intermediates, leading to complete degradation of the organic substances (figure 2). Consequently, the uranyl species play the role of active centres in the photocatalytic reactions.
4. Photocatalytically active species in semiconductors with porous structures
Photocatalytic properties of semiconductor materials have received much attention since TiO2 materials were successfully used for photochemical conversion (Fujishima & Honda 1972). The photocatalytic activity of a semiconductor is attributed to the generation of electron–hole pairs under irradiation. The electrons generated may move to the surface of the semiconductor, leading to the formation of highly active OH* and O−2 radicals (Parker et al. 2007). These radicals can react further with the substrate to achieve the decomposition of the reagent. Hwang and co-workers employed an in situ solid-state nuclear magnetic resonance technique to investigate the surface-bound intermediates in the photocatalysis oxidation of ethanol (Hwang & Raftery 1999). Two intermediates, surface Ti-ethoxide species and surface hydrogen-bonded ethanol, were found. Among these two intermediates, the surface Ti-ethoxide species resulting from the reaction between surface hydroxyl groups and ethanol are highly active.
Porous materials usually have high surface areas, abundant surface states and readily accessed channels (Grosso et al. 2004; Martinez-Ferrero et al. 2007; Pauporte & Rathousky 2007). Thus, it is expected that more active sites which are essential for catalytic reactions can be generated by the introduction of a porous structure to a semiconductor. It has been proved that semiconductor photocatalysts with porous structure show improved photochemical performances (Matsushita et al. 1998; Schattka et al. 2002; Anandan & Yoon 2003; Lu et al. 2008).
Recently, we developed a template-free preparation method for the fabrication of highly active porous titania or metal-doped porous titania (Zou et al. 2008). First, a non-stoichiometric titanium-based heterometal alkoxide (M-TG; TG=titanium glycolates, M=Fe, Mn) was synthesized in ethylene glycol. Using the M-TG as a single-source precursor, crystalline porous M-TiO2 photocatalysts have been obtained through a simple thermal treatment process. The dopant heteroatoms are incorporated into the TiO2 lattice at the Ti4+ positions. These porous M-TiO2 materials show high photocatalytic activities in degradation of phenol under UV irradiation. The high crystallinity, large surface area and appropriate transition metal doping are all beneficial for the enhancement of the photocatalytic performance of the porous TiO2 material. Particularly, charge trapping in the transition metal ion centre is conducive to separation of the photogenerated electron–hole pairs, leading to enhancement of photocatalytic activity for the M-TiO2 materials (figure 3). Compared with Mn-TiO2, Fe-TiO2 shows higher photocatalytic activity because Fe3+ may suppress the recombination of photogenerated electron–hole pairs more efficiently. In addition, the M-TiO2 materials can be easily separated and recovered from the reaction systems after photocatalytic reactions, facilitating the recyclability of the photocatalysts. The successful formation of non-stoichiometric heterometal alkoxides and their conversion into porous oxide materials open the way not only to the synthesis of new metal alkoxides but also to heterometal oxides with useful functions (Zou et al. 2008).
Recently, mesoporous TiO2 spheres with a large surface area and rich surface hydroxyl groups were prepared through a light-driven synthetic strategy (Zou et al. 2011). A titanium alkoxide, titanium glycolate, which is photoactive towards UV light, is employed as a precursor. Under UV irradiation, mesoporous TiO2 spheres (TiO2-UV) are generated through the decomposition of the organic components in the titanium glycolate spheres. The resulting TiO2-UV spheres with abundant surface hydroxyl groups can transform urea to graphitic carbon nitride under mild conditions. Although investigations on the photocatalytic performance are still underway, the formation of mesoporous TiO2 under UV irradiation provides a new strategy for the fabrication of inorganic functional materials with desirable morphology and properties.
Porous semiconductor materials can also be used for the fabrication of a variety of composite catalysts. For example, a metal–semiconductor composite with Pt clusters supported by ZnO nanocages was fabricated recently (Zeng et al. 2008). This composite material exhibited high photocatalytic activity in the photodegradation of methyl orange. The high photocatalytic activity is attributed to the presence of Pt clusters, which act as the reaction centres. The Pt clusters offer abundant nanoscale Schottky contacts in the Pt–ZnO metal–semiconductor interfaces which improve the separation efficiency of the photogenerated charges. Pt/TiO2 nanoarchitecture, a metal–mesoporous semiconductor photocatalyst, was also prepared (Wang et al. 2005a,b). In the Pt/TiO2 nanoarchitecture, platinum clusters with diameters smaller than 5 nm are well dispersed in a cubic mesoporous crystalline anatase. The platinum clusters together with the mesoporous structure provide efficient and easily accessible active sites for the photocatalytic redox reactions. The metal–semiconductor nanoheterojunction can also promote the separation of photogenerated charge carriers owing to the existence of the metal–semiconductor interfaces, guaranteeing high catalytic activity in the oxidation of carbon monoxide.
5. In situ characterization of single-site photocatalysts
The spatial separation of active sites also makes it feasible to characterize and identify the nature and the interactions of the active sites with reactants at the atomic level by techniques such as X-ray absorption (XRA), Fourier transform infrared (FT-IR), ultraviolet–visible absorption (UV-Vis), electron paramagnetic resonance spectroscopies, photoluminescence and microcalorimetry. In order to achieve the ultimate goal, in which the structure of catalysts is targeted, the greatest possible precision is demanded in determining the structure of the catalyst in general and of the active site in particular. Therefore, the in situ characterization methods are an ineluctable need for the catalysts as a prerequisite to engineering active sites (Thomas 1997). XRA spectroscopy and X-ray diffraction (XRD) are ideal tools for such in situ investigations. When solids possess three-dimensional surfaces, as most micro- and mesoporous materials do, XRA yields information pertaining to the immediate atomic environment of the absorbing atom (which is selected to be at, or close to, the active site). From XRD, one can obtain information about the overall structural integrity and other aspects of long-range order within the phase in question (Couves et al. 1991).
The detailed local atomic environment of the single sites in catalysts may be determined by in situ X-ray absorption fine structure (XAFS; including the pre-edge, near-edge and extended-edge fine structures) spectroscopy. For example, from XAFS spectroscopy, we can see the feasibility of quantitatively tracking the Ti(IV) site from its calcined state to its active state in the catalyst Ti/MCM-41 (Maschmeyer et al. 1995; Oldroyd et al. 1996). This is readily understood in view of the contrasting accessibility of the Ti(IV) sites in the respective catalysts. Moreover, using a joint XAFS and computational study, we could also directly probe the environment of the reactive centre. And we could arrive at the pathway and associated energetics for the mechanism of the catalytic reaction. Thomas and co-workers investigated the epoxidation mechanism of alkene by alkyl hydroperoxides. The experimental results and calculations using density functional theory showed that the original four-coordinated Ti active site passes through a six-coordinated state in the epoxidation process (Thomas et al. 2002; Williams & Thomas 2005).
Besides XRA and XRD, some other techniques are also helpful in the characterization of SSPCs. Applying various molecular spectroscopic techniques such as photoluminescence, UV-Vis, XAFS (X-ray absorption near-edge structure and Fourier transform extended X-ray absorption fine structure) and FT-IR spectroscopy, Anpo and co-workers deduced the photocatalytic decomposition mechanism of NO into N2 and O2 on Cu+-SSPC prepared using ZSM-5 zeolite as a host (Anpo et al. 1994; Matsuoka & Anpo 2003). Information on the electron transfer and details of the individual steps involved for the conversion of NO to its component parts were obtained.
In addition, the electron microscopy technique can give visual images of the solid catalysts. In the last decade, significant advances have been made in the deployment of electron microscopy. New instrumental and methodological concepts have appeared constantly. For instance, high-resolution electron microscopy (HREM) and four-dimensional electron microscopy reach picometre spatial resolution coupled with femtosecond time resolution. They can provide the possibility for in situ stages of the solid–gas and solid–liquid studies (Thomas & Midgley 2010). The merits of HREM include being able to image individual metal atoms and metal cluster catalysts. Therefore, electron microscopic techniques can be used to elucidate the nature of complex selective oxidation catalysts (Grasselli et al. 2006; Pyrz et al. 2008), clarify the nature of and design superior supported oxide and solid acid catalysts, and determine the structure of a complex zeolite catalyst (Serra et al. 2004; Baerlocher et al. 2007).
There are many other advanced characteristic techniques that have been applied in the investigation of SSPCs, promoting the development of these fascinating materials.
A catalyst should meet the following criteria: high stability, low cost, non-toxicity and good performance, especially with regard to facilitating separation of products from reactants and recyclability. Through molecular design, SSPCs can satisfy these criteria well. SSPCs with a porous structure usually exhibit interesting optical, electronic, thermal and structural properties, depending on their pore size, shape and structure. The presence of porosity in SSPCs also benefits their photocatalytic activities. Because of the high specific surface area and rich pores, porous structures provide hosts for the formation of spatially isolated and readily accessed single sites, guaranteeing high catalytic activities. Moreover, photocatalytic reactions can also take advantage of the unique features, such as the shape selectivities, of porous materials.
To tackle the issues on ‘energy’ and ‘environment’, green chemistry has received more and more attention. As a process of green chemistry, photocatalytic reaction will play an important role in environmental protection and in the search for renewable and clean energy. SSPCs with a porous structure are highly promising for the design and fabrication of novel photocatalysts with high activities in the decomposition of pollutant molecules, water splitting and synthesis of organic species.
We are grateful to the National Basic Research Programme of China (grant no. 2011CB808703) and the National Natural Science Foundation of China (grant no. 91022019) for financial support.
One contribution of 14 to a Special feature ‘Recent advances in single-site heterogeneous catalysis’.
- Received November 22, 2011.
- Accepted February 10, 2012.
- This journal is © 2012 The Royal Society