Perspective pubs.acs.org/JPCL
Progress in Heterogeneous Photocatalysis: From Classical Radical Chemistry to Engineering Nanomaterials and Solar Reactors Wey Yang Teoh,† Jason A. Scott,‡ and Rose Amal‡,* †
Clean Energy and Nanotechnology (CLEAN) Laboratory, School of Energy and Environment, City University of Hong Kong, Hong Kong, S.A.R. ‡ ARC Centre of Excellence for Functional Nanomaterials, School of Chemical Engineering, The University of New South Wales (UNSW), Sydney, Australia ABSTRACT: The field of heterogeneous photocatalysis has expanded rapidly in the last four decades, having undergone various evolutionary phases related to energy and the environment. The two most significant applications of photocatalysis are geared toward solar water splitting and the purification of air and water. Notably, the interdisciplinary nature of the field has increased significantly, incorporating semiconductor physics, surface sciences, photo and physical chemistry, materials science, and chemical engineering. Whereas this forms the basis on which the field continues to grow, adequate bridging of multidisciplinary knowledge remains essential. By recalling some of the classical fundamentals of photocatalysis, this Perspective provides contemporary views on heterogeneous photochemical conversion, encompassing charge transport characteristics, radical chemistry and organic degradation mechanisms, photocatalyst design, and photoreactor engineering.
Overview of Time Evolution of Photocatalysis. Research in “functional” heterogeneous photocatalysis has attracted interest from scientists and engineers alike, driven by its broad applications in photochemical conversions. The term functional designates photomediated redox reactions such as photocatalytic water splitting, environmental pollutant abatement, and organic synthesis. Ultimately, these applications were fuelled by the expectation of efficiently harnessing and utilizing freely available solar energy. Although studies on heterogeneous photocatalysis existed prior to the 1970s, appreciable interest in the area did not occur until the landmark paper by Fujishima and Honda in 19721 on the decomposition of water into H2 and O2 over single-crystal TiO2 (rutile) in a photoelectrochemical (PEC) cell (Figure 1a). Coinciding with the peak oil and energy crisis of the time, the work garnered attention as a promising technology for sustainable solar hydrogen generation as an alternative fuel. Subsequently, “miniaturization” of the PEC cell, in the form of individual particulates,2 was demonstrated through platinized TiO2 (Figure 1c). Incorporating organics (denoted as “R” in Figure 1b) such as carboxylic acids, alcohols, and saccharides as sacrificial electron donors into the photocatalytic process increased charge-separation efficiencies and gave higher H2 generation rates.3 In doing so, CO2 rather than O2 was generated. It soon became clear that heterogeneous photocatalysis could be extended to the photocatalytic remediation of organic and inorganic pollutants in the aqueous and gaseous phases.4 Despite easing energy concerns in the early 1980s, the subsequent wave of environmental remediation technologies continued to advocate the rapid development of functional photocatalysis.4−7 Extending to the late 1990s, extensive fundamental knowledge on photophysics and photochemistry was established, © 2012 American Chemical Society
including the understanding of radical chemistry and photoinduced charge transfer across photocatalytic surfaces, particularly that of TiO2.8 The mechanisms of organic photocatalytic oxidation as well as reductive dehalogenation9 and metallization10 became focal points of the research in conjunction with the interest in water and air pollution abatement. Although it was substrate-specific, organic degradation was predominantly centered on direct (hole-mediated) and indirect oxidation reactions (mediated by hydroxyl and superoxide radicals).7,11 This was complemented by identifying the key function of molecular oxygen as an efficient electron scavenger,12 prompting system aeration or oxygenation when assessing photocatalytic oxidation reactions. Further interfacing with engineering saw effective photon delivery to the photocatalyst surface through light harvesting and photoreactor design.13,14 Over the past decade, interest in photocatalysis has again been heightened by future energy (in)security. However, in addition to the peak oil production issues as in the 1970s, the urgency in developing renewable energy technologies has been driven to a large extent by the concerns of excessive fossil-fuel burning and net CO2 emissions.15 In this instance, photovoltaic (PV) conversion by excitonic (dyes, quantum dots, plasmonic) solar cells is a vanguard photocatalysis-related solution. Because the subject has been amply reviewed and updated recently,16−18 it will not be discussed here. Rather, the Perspective will focus exclusively on heterogeneous photochemical conversions such Received: January 15, 2012 Accepted: February 10, 2012 Published: February 10, 2012 629
dx.doi.org/10.1021/jz3000646 | J. Phys. Chem. Lett. 2012, 3, 629−639
The Journal of Physical Chemistry Letters
Perspective
Figure 1. Schematic diagram of (a) the photoelectrochemical system using a TiO2 single crystal as the photoanode; (b) the corresponding energy diagram of the system (which is similarly applicable to); and (c) the “miniaturizsed” photoelectrochemical system in the form of suspended platinized TiO2 particulates. R denotes general organic electron-donor molecule. Adapted with permission from ref 31.
Figure 2. Number of publications pertaining to photocatalysis on a decadal basis originating from the top 12 (overall) publishing nations (topic: photocatalysis/photocatalytic/photocatalyst, ISI Web of Science, 12/01/2012). Besides the traditional photocatalysis powerhouses of Japan and the United States, the research output from China has dominated over the past decade, whereas South Korea and India have also experienced substantial growth in the field.
is well-understood. Initially, electron−hole (e−−h+) pairs are generated upon bandgap excitation. Depending on the excitation lifetime relative to that of charge recombination a net fraction of “live” photocharges are present, which are trapped at defect sites or diffuse toward the photocatalyst surface. These surface holes and electrons can oxidize and reduce surface-adsorbed molecules, respectively, through interfacial charge transfer to produce radical species. These then participate in various reactions, as illustrated in Figure 3. The capacity for radical formation is governed by the intrinsic conduction (ECB) and valence (EVB) band potentials of the photocatalyst in relation to the redox potential of the surface reaction. Unless the excited high-energy photoelectrons, that is, hot carriers, can be extracted before their rapid relaxation to the conduction band edge, it is not possible to catalyze a reduction reaction where the reduction potential is more negative than the ECB. Likewise, it is difficult to catalyze an oxidative reaction where the reduction potential is more positive than the EVB. As such, it is not surprising to find some of the most oxidative photocatalysts, such as TiO2, ZnO, WO3 and BiVO4, to be
as organic oxidation as well as hydrogen evolution and water splitting. A significant portion of publications in the field over the past decade comprises photocatalyst syntheses and the various photon-mediated applications. Of the latest progress in photocatalysis, major advances have been driven by its integration with materials science and nanotechnology. This is accompanied by a demographic shift in photocatalysis research, as shown in Figure 2. Whereas material developments have recently forged ahead, they represent only one of the four key elements of a photocatalytic system. That is, designing a “functional” photocatalytic system requires a multiscale, integrated approach in terms of (1) understanding charge transport, (2) identifying the reaction mechanism, (3) designing an innovative photocatalyst, and (4) engineering photon delivery. This Perspective considers each of these four aspects as they are understood today, highlighting what is currently known as well as identifying key challenges which remain. Mechanistics of Heterogeneous Photocatalysis: The Radical Species. The mechanism behind semiconductor photocatalysis 630
dx.doi.org/10.1021/jz3000646 | J. Phys. Chem. Lett. 2012, 3, 629−639
The Journal of Physical Chemistry Letters
Perspective
Figure 3. Possible reaction pathways arising from the excitation of photocatalysts. Included are primary redox reactions and selected ensuing secondary radical reactions. R denotes the general organic electron donor molecule.
state in WO3 (+0.5 VNHE),27 alternative multiple electron-reduction mechanisms have been proposed: O2 + 2H++ 2e−→ H2O2, +0.695VNHE and O2 + 4H+ + 4e−→ 2H2O, +1.23VNHE,28 requiring the assistance of suitable cocatalysts such as Pt deposits.27 Secondary radical formation from the oxidation of organic substrates often increases the complexity of photocatalytic reaction mechanisms. For example, following from the holemediated decarboxylation reaction, highly reductive formate radicals (E0(CO2•−/CO2) = −1.9 VNHE) can be formed.29 These radicals are particularly useful for reducing highly electropositive metals such as Cd2+, when the ECB of the photocatalyst is less negative than that required for metallizing the cations.30 Similarly, the photocatalytic oxidation of alcohols may form reductive hydroxyalkoxyl radicals, for example, the α-hydroxymethyl radical from methanol oxidation (E0(•CH2OH/ CH2O) = −0.95 VNHE), which in turn injects an additional electron into the photocatalyst conduction band.31 In this current doubling effect, a single photon is capable of generating two photoelectrons, that is, one from intrinsic bandgap excitation and another through injection by the hydroxyalkoxyl radical. In this instance, the achievable apparent quantum yield can be >90% (theoretical maximum of 200%), as observed for photocatalytic hydrogen evolution in the presence of methanol as an electron donor.32 In gas-phase photocatalysis, the dominant photodegradation mechanism is not always straightforward. Surface hydroxyl groups or adsorbed water may contribute to the photodegradation process via hydroxyl radical formation. Although, if the relative humidity is too high, then photoactivity may decrease due to competitive adsorption between water molecules and the target pollutant.33 Nevertheless, the decreased presence of water (relative to aqueous-phase photocatalysis) implies that oxygen may play a greater part in gas-phase photocatalysis. Recent evidence has suggested that chemisorbed oxygen can: (i) be reduced upon UV illumination, where it remains either in
defined by their EVB being more positive than that required for hydroxyl radical (•OH) generation. Hydroxyl radicals have been identified as one of the most active and nonselective initiators of photocatalytic oxidation of organic substrates, particularly in relation to weakly adsorbing species such as alcohols and aromatics. Hydroxyl radicals are efficient H atom abstractors (e.g., ROH + •OH → RO•)19 and attack electron-rich moieties (e.g., C 6 H 6 + • OH → C6H5OH).20 Depending on the nature of the hydroxyl radicals, whether they are surface-bound (•OHads, E0ox > +1.6 VNHE, from the hole oxidation of hydroxyl moieties)21 or diffused species (•OHfree, E0ox = +2.72 VNHE, from the oxidation of hydroxyl ions and water molecules),22 they may have different oxidation potentials. In general, it is perceived that strongly surfaceconjugated molecules such as carboxylic acids (through covalent bonding with deprotonated hydroxyl moieties) are prone to direct oxidation or decarboxylation by the photo-Kolbe reaction. During photocatalysis, oxygen molecules often act as the oxidant (O2 + e− → O2•−). The effectiveness of molecular oxygen as electron scavenger was originally discussed by Gerischer and Heller12 and demonstrated in an accompanying study.23 Despite being a weaker radical (E0(O2/O2•−) = −0.33 VNHE) and possibly in lower concentrations than hydroxyl radicals, the role of O2•− as well as the subsequently formed reactive oxygen species, for example, singlet oxygen (1O2), hydroperoxyl radicals E0(O2/HO2•) = −0.05 VNHE, and hydrogen peroxide E0(O2/H2O2) = −0.33 VNHE, cannot be underestimated24 and may be complementary to the overall photocatalytic oxidation. Where the EVB of a photocatalyst is insufficient to catalyze the production of •OH, or similarly in the case of sub-bandgap excitation such as over anionic-doped TiO2 (N-, C-, and S-doped TiO2), superoxide radicals may dominate but are a weaker oxidant source.25,26 On the contrary, where the ECB of the photocatalytic material is insufficient for direct reduction of molecular oxygen, for example, the W 5d 631
dx.doi.org/10.1021/jz3000646 | J. Phys. Chem. Lett. 2012, 3, 629−639
The Journal of Physical Chemistry Letters
Perspective
exact anatase-rutile content, we observed differences in the photocatalytic activity of P25 and the samples prepared within our laboratory, the degree depending on the class of organic substrate.49 This was further traced to the materials intrinsic efficiencies for •OH generation and direct charge transfer, which governed the dominant degradation mechanism for each of the substrate functional groups: carboxylic acids − direct hole transfer; saccharides − efficient reductive pathway; and alcohols and aromatics − •OH-mediated oxidation.49 In a broader context, Ryu and Choi50 upon comparing the photocatalytic performance of 8 commercial variants of TiO2 (5 of which were pure anatase) across 19 different substrates, concluded “one photocatalyst does not fit all”. Multivariate statistical analysis, taking into account the key physicochemical properties of the photocatalyst in relation to the photocatalytic mechanistics, may provide a meaningful glimpse into the real performance of a photocatalyst.51
an undissociated state or dissociates to produce atomic oxygen species, or (ii) undergo hole-mediated desorption as molecular O2.34 However, the extent to which these adsorbed oxygen species exist in an ambient environment and their participation in the photodegradation mechanism is yet to be conclusively demonstrated.35 It should be apparent from the preceding discussion that on a molecular scale the photocatalytic mechanism is extremely intricate. A myriad of radical-based reactions are potentially available, each of which is governed by the local environment as well as the target species. Adding complexity to the system is the properties of the photocatalyst material itself, from its electronic faculties such as band potentials to its physicochemical attributes such as crystallinity or surface hydroxylation. These material aspects are explored in the following sections. Photocatalytic Materials and Substrate Specif icity. Prior to the Honda−Fujishima effect, interest in photocatalysis was centered on ZnO, which has similar ECB and EVB to TiO2. The material is, however, limited by intrinsic photocorrosion upon excitation in aqueous media by the photoinduced hole weakening of Zn2+−O2− bonds to produce O2 and soluble Zn2+.36 Hematite (α-Fe2O3) is another material that was originally thought to be an ideal photocatalytic material due to its low cost, abundance, and narrow bandgap for harnessing solar energy (EBG = 2.0 to 2.2 eV, excitation wavelength up to 620 nm).37 However, the material suffers from rapid charge recombination (lifetime
