Personal Account
THE CHEMICAL RECORD
Perspectives on Heterogeneous Photochemistry Cynthia M. Friend Harvard University, Department of Chemistry and Chemical Biology, 12 Oxford St., Cambridge, MA 02138 (USA) E-mail: [email protected]
Received: April 11, 2014 Publised online: August 18, 2014
ABSTRACT: Heterogeneous photochemistry has a potentially important role in production of energy, in environmental remediation and in sustainable production of chemicals. Photochemical efficiency depends on both materials properties and the desired chemical reaction that is promoted through creation of an excited state. A detailed understanding of the interplay between materials properties and reactivity requires a molecular-scale approach that determines the elementary steps in the overall process. This personal account summarizes the role of defects in determining the photochemical and thermal reactions on rutile titania, a model for semiconductor metal oxide photocatalysts that defects, e.g., Ti interstitials present in the subsurface region, and O adatoms on the surface, have a substantial impact on the efficiency for photochemical conversion through modification of molecular binding and also through likely modification of charge carrier dynamics. Design of materials must include engineering of the optical and electronic properties of the semiconductor photocatalyst, and understanding of the key photochemical steps involved in specific processes to ensure proper alignment of their electronic states with the band structure of the material. Thus, fundamental surface science studies and development of time-dependent theoretical methods that map out the reaction mechanism for photochemical processes on materials with controlled composition and structure are critical. DOI 10.1002/tcr.201402037 Keywords: defects, oxidation, photochemistry, surface chemistry, titania
Introduction Some of the most pressing technological issues facing the world today revolve around energy and the environment. There is a desire to shift the production of chemicals and fuels to renewable resources to reduce dependence on petrochemical resources; additionally, removal of noxious molecules from the atmosphere is important for improving air quality. Solutions for both of these problems require efficient transformations of
944 www.tcr.wiley-vch.de
organic molecules, and heterogeneous photochemistry provides one avenue for driving these key processes. The design and implementation of photochemistry to address these pressing problems related to energy and the environment requires a detailed understanding of the interplay between materials properties and chemical bonding and reactivity. Herein, selected examples are used to illustrate how the
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
powerful tools of surface chemistry are used to establish a molecular approach to understanding heterogeneous photochemical processes. Fundamental studies of titania are an illustrative case because the chemical and structural properties of titania have been widely studied, as discussed in several excellent reviews.[1–3] Fundamental studies of surface chemistry under ultrahigh vacuum and associated theoretical models establish guiding principles for photocatalysis as has been discussed in several reviews.[2,4–6] For example, the mechanism for photo-oxidation of methanol to methyl formate was determined using surface science tools,[7] providing insight into the analogous vapour phase process over a supported catalyst at higher pressures.[8] The molecular level insight into surface photochemistry provided by surface science studies provides the basis for understanding photocatalysis despite the vastly different reaction conditions, especially since most catalytic processes operate at low steady-state concentrations of reactants on the surface. In this account, the roles of defects and dopants in charge carrier recombination, bonding of molecular intermediates to the surface, treatment of excited states of the semiconductormolecule system, and coupling to thermal processes are discussed. Although titania is used to illustrate these effects, similar factors may play a role in other metal oxide photocatalysts in general.
Overview of Heterogeneous Photochemistry on Semiconductors Principles Excitation of electron-hole pairs in a semiconductor material is the first step in photocatalysis. The energy of the band gap in the semiconductor is critical because it determines the photon energy required for excitation of charge carriers. The alignment of the band edges with electronic states in molecules and
Cynthia M. Friend is the T. W. Richards Professor of Chemistry and Director of the Rowland Institute at Harvard University. She joined the Harvard faculty in 1982 after one year of postdoctoral research at Stanford University. Friend earned a doctorate in Physical Chemistry from the University of California (UC) Berkeley (1981) and a B.S. degree from UC Davis (1977). She has been widely honoured for her research in surface chemistry and catalysis. Friend is a Fellow of the American Association for the Advancement of Science and the American Chemical Society.
Chem. Rec. 2014, 14, 944–951
Fig. 1. Qualitative illustration of principles of photocatalysis applied to water splitting from reference [10]. This schematic representation does not take into account the interaction of the electron-hole pair following excitation which alters the energies of these charge carriers.
intermediates on the surface also determines the types of reactions that the charge carriers (electrons and holes) can promote because charge carriers must have sufficient energy to drive reactions and also be able to interact with specific states of molecules that lead to oxidation or reduction.[9–12] For example, water splitting to O2 and H2 is an important reaction that has the potential to produce clean fuels. The minimum energy required for this reaction is 1.23 eV (Figure 1); thus, a minimum requirement for a photocatalyst for this process is a material with a band gap of 1.23 eV or greater. The optical band gap of semiconductor materials provides a guide to which materials might be viable photocatalysts; however, the actual energies of the electrons in the conduction band and holes in the valence band will be different due to Coulombic interaction (Figure 2).[13,14] Hence, the additional energy required to create an excitonic pair of electrons and holes must be considered. In addition, the conduction and valence bands of a material must also have a proper alignment with specific states in surface-bound intermediates; in other words, there must be electronic states in a molecule or molecular intermediate that can accept or donate an electron for the desired reaction to occur. For example, there must be states of methoxy that overlap with the top of the valence band of TiO2 in order for photo-oxidation to occur (Figure 3). Hence, the electronic band gap and band alignment are key factors in determining photochemical efficacy of a material for a specific reaction. The same principles apply to other photochemical reactions; however, the requirements for band alignment and the overall energetics will be different for every specific reaction. To fully
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 945
THE CHEMICAL RECORD
Energy
Conduction band
Eexc
Egap
Eop Valence band Density of states
Fig. 2. Qualitative illustration of the effect of the difference between the fundamental and optical gap due to exciton binding. The difference is related to the electron and hole effective masses and is typically less than 0.1 eV.
TiO2
bridging site in the photo-dissociation of water using 400 nm light.[15,16] In this case, water photo-dissociation occurs; however, the overall production of O2 and H2 requires subsequent steps that could be either thermal or photochemical. Furthermore, the efficiency for the water photo-dissociation depends on the alignment of the energy of the excited state in the molecule and the top of the valence band, where the holes are made to drive the O–H bond scission. More generally, the photo-active species may be a new intermediate formed from an earlier thermal step. Thus, the electronic structure of both the photo-active molecular species on the surface and the material to which it is bound are important because the excitation of electron hole pairs in the material must be aligned with excited states that lead to photoreaction in the molecule. Thus, designing efficient and selective photochemical processes requires not only an understanding of overall energetics, but also understanding of the nature of key intermediates and of the dynamics of excited states. Below, the role of defects in determining the nature of surface bound species is discussed.
Energy
conduction band
Methoxy occupied orbitals
Overlap of states
Materials Design
TiO2 valence band
Density of states Fig. 3. Qualitative illustration of the hybridization of the highest occupied orbitals of methoxy with the top of the valence band of titania which allows for transfer of a hole in the valence band to the molecule so as to induce photo-oxidation.
understand the requirements for promoting a reaction with light, a detailed molecular scale mechanism is required because the electronic states of different molecular intermediates will be aligned differently with the bands of the semiconductor.[12] Understanding the elementary steps that lead to an overall reaction is a critical aspect of photochemistry on surfaces that is not widely considered in designing new materials for photochemistry. The example above for water only considers the overall reaction energetics and not the elementary steps in the overall process. Understanding elementary steps is critical because the energetics for both photochemical and thermal steps depend on the intermediates for these steps and their bonding to the surface. Recent studies using scanning tunnelling microscopy (STM) were used to demonstrate that a proton is transferred to a bridging oxygen on TiO2 to yield one OH group bound on the Ti five-coordinate rows and one at a
946 www.tcr.wiley-vch.de
There has been a major recent effort to design new materials for photochemical reactions. To date, transition metal oxides are among the most common materials used for photocatalysis and within this class, titania is most commonly used. Titania has the advantage that it is easy to prepare, is earth abundant and chemically versatile; however, it has an optical band gap in the ultraviolet (3.1 eV for rutile), making it rather inefficient for solar energy conversion. Titania also has some activity for water splitting which has made it a popular material to investigate.[17,18] Because there is great interest in using solar energy for photochemical transformations, there is a drive to develop materials with excitations that overlap the solar spectrum, which has the highest photon irradiance between ∼1–3 eV. Therefore, there is a major effort to develop photocatalysts with appropriate optical and chemical properties to promote photochemical processes using visible light predominant in the solar spectrum.[19,20] To this end, other transition metal oxides and a variety of semiconductors with smaller band gaps (Figure 4) as well as hybrid thereof are being designed[20,21] and tested[10,22–24] for a variety of reactions. For example, hybrids of TiO2 and CdSe have been demonstrated to absorb visible light via the CdSe leading to charge injection into the titania.[25] Another approach is to sensitize semiconductor materials, such as titania, with metallic structures. Recent work has demonstrated the potential value of using SPR excitation of Ag or Au nanoparticles as photocatalysts.[26–29] Recent studies of ethylene oxidation using Ag nanoparticles as photocatalysts[30,31]
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
Fig. 4. Schematic of the variation of the optical band gap and band alignment using water splitting as a reference. Figure from reference [10] by Maeda and Domen.
are an excellent example of the value of fundamental studies that provide insight into mechanism. There are a range of studies of SPR-based photocatalysts for various reactions, including water splitting,[28,29] the decomposition of organic dyes and other VOCs,[26,32] and methanol oxidation,[33] suggesting wide applicability to this approach; however, the effect of SPR excitation on specific elementary steps has not been determined for most of these cases. A full description of these effects must describe the time-dependent behaviour of these collective excitations. Most of the endeavours to design new photocatalysts have all focused on engineering the optical properties of these materials; however, a molecular scale understanding of chemical reactivity is required to determine how materials can be modified to achieve more efficient photocatalysis. As motivated above in the discussion of water photo-dissociation, the bonding of photo-active species and alignment of electronic states that lead to bond scission with the valence and conduction bands for hole- and electron-mediated processes, respectively, are an important part of designing photocatalysts for specific reactions. Therefore, the bonding and thermal reactivity of specific molecules on a particular photocatalyst must be considered. Using the case of reduced rutile TiO2(110) as a case study, the value of fundamental studies of reactivity in understanding key factors that control photochemical reaction are revealed.
Defects and Photochemistry Titania is one of the most widely investigated materials for photocatalysis even though it has a large band gap of ∼3 eV in the ultraviolet. There has been a concerted effort to modify the electronic and optical properties of TiO2 through doping with the goal of creating charge carriers and inducing photochemical events using visible light. Doping with other transition metals, e.g. Ni, Fe, W,[20] and with other anions, e.g.
Chem. Rec. 2014, 14, 944–951
nitride,[34,35] is one approach used. Most commonly, creation of defects in the titania through reduction is employed.[1] Defects in titania are created through thermal, electron- or photoinduced reduction. Materials with a high defect density can also be created using a variety of methods, including laserassisted oxidation.[36] Doping of titania with defects significantly alters the electronic and optical properties by introducing states into the gap.[1] In titania, the change in optical properties manifests itself with a dark blue colour with increasing degree of reduction. Oxygen deficiency in titania and other transition metal oxides creates electronic states in the band gap that lower the energy for electron-hole pair excitation. In the case of titania, this leads to absorption of visible light. Typical defects that occur in TiO2 are oxygen vacancies on the surface, e.g. bridging oxygen vacancies on the rutile TiO2(110) surface, and subsurface defects such as Ti interstitials. The deficiency of O both on the surface and in the bulk lead to excess electron density near these defects,[9,11] creating charge differences that can lead to site-specific reactivity. Exposure of reduced surfaces to gaseous O2 also creates oxygen adatoms on the surfaces of reduced titania, filling in most of the surface oxygen vacancies and compensating for the charge due to the excess Ti interstitials.[15] All of these vacancies can play a role in both thermal and photochemical processes on the surface. While the potential role of surface O vacancies is clearly established, only recently has the effect of subsurface, interstitial defects been demonstrated to be important in thermal reactions on titania. In photochemical processes, defects can affect the lifetimes of charge carriers. For example, O vacancies are thought to trap charge carriers leading to an increase in the rate of electron-hole pair recombination, thus decreasing the efficiency of heterogeneous photochemistry. Because photons penetrate into the bulk of a photocatalyst, electron-hole pairs are created even in the bulk of the solid. The absorption depth is estimated to be ∼100 nm for light with an energy of 3.4.eV, for example.[17] In order to promote photochemical processes, these charge carriers must remain separated and must migrate to the surface where reactants are bound. Photocatalyst architectures can be designed to optimize photon absorption and survival of charge carriers as they migrate to the surface—for example, by using nanowires. However, control of materials composition and structure at an atomic level is important for the best efficiency because of the need to control chemical reactivity. Understanding the elementary steps that lead to an overall photochemical reaction on a semiconductor surface is an important and often overlooked factor in determining the overall efficiency. Specifically, photochemical transformations depend on the binding of reactants on the surface because bonding determines their ground and electronically excited states, including how they evolve in time. Defects play an
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 947
THE CHEMICAL RECORD
important role in determining the binding of molecular species to the surface and, therefore, changing the nature of ensuing photochemical events. Many studies have demonstrated the importance of bridging oxygen vacancies on the surface in determining the rates of thermal reactions. For example, water dissociates to two OH groups on a bridging oxygen vacancy.[19] Similarly, methanol reacts to form CH3O bound to bridging oxygen vacancies transferring the H to a neighbouring bridging O.[21] Here, the roles of O adatoms and subsurface defects are discussed in more detail. Oxygen Adatoms Oxygen adatoms on the surface affect chemical reactivity on titania by leading to the formation of new intermediates. O adatoms are formed from exposure to TiO2 that is partially reduced. O2 reacts to fill surface vacancies and O adatoms form on the Ti five-coordinate rows in concert.[4,37] Thus, the net effect is to neutralize excess charge on the surface and to change the nature of potential reaction sites.[15,16] Oxygen adatoms promote the formation of specific intermediates that are formed thermally prior to photochemical activation. The new intermediates have very different photochemical efficiencies for reactions. These new intermediates have a different electronic structure, leading to changes in their alignment relative to the states in titania. As a result, the photochemical behaviour will almost certainly be different than the original reactant molecule. There are several well-documented examples of changes in photochemical efficiency due to the presence of O adatoms on titania. In the case of methanol, O adatoms on titania promote O–H bond breaking to form methoxy on TiO2(110). The O adatoms are all removed as water below 350 K.[7,18,38] Methoxy formed in this manner is photo-oxidized to formaldehyde irradiation with ultraviolet light (Figure 5). Notably, the rate of methoxy photo-oxidation is ∼50 times faster than methanol on a reduced surface.[7,20,38,39] This difference has been rationalized in terms of better level alignment of methoxy with the titania valence band than methanol.[12,20] Similar reactions are observed for ethanol on titania[10,22–24,40] and Ru-modified TiO2[23,25]—acetaldehyde is formed. For dense layers of methoxy, a second photochemical pathway develops form the photo-oxidation of formaldehyde.
Fig. 5. Schematic of initial photochemical event in the photo-oxidation of methoxy formed from methanol on TiO2(110). Adapted from reference [7].
948 www.tcr.wiley-vch.de
The ultimate product is methyl formate that is formed from reaction with a nearby methoxy (Figure 6). The fact that methyl formate is produced shows that methoxy is photooxidized more slowly than formaldehyde. If methoxy reacted faster, it would not be available for the coupling process. The mechanism for both the primary and secondary oxidation of methoxy was established based on fundamental studies using scanning tunnelling microscopy (STM) and isotopic labelling in conjunction with both photon-stimulated and thermal desorption measurements.[7] The efficiency of the photo-oxidation and the sequential nature of formaldehyde is illustrated by the photon-stimulated desorption of the products during irradiation (Figure 7). The example of methoxy photo-oxidation demonstrates how O adatoms affect the efficiency of photooxidation by promoting the formation of active surface intermediates with electronic structure that aligns better with the semiconductor states of titania. This case also demonstrates the value of fundamental studies in determining the pathways for photo-oxidation at a molecular level. A second example of the effect of O adatoms on photochemistry via formation of a new photo-active species is the case of ketone photo-oxidation on titania. Ketones are not photochemically active on reduced titania and they are relatively weakly bound to the surface.[2,41] Butyrophenone, for example, is highly mobile on the surface even at low temperature,
Fig. 6. Schematic of secondary photo-oxidation of formaldehyde and coupling with methoxy on titania to form methyl formate. Adapted from reference [4].
Fig. 7. Production of formaldehyde (red) and methyl formate (blue) as a function of irraditation time from photo-oxidation of methoxy on TiO2(110). The sharp rise in formaldehdye decays as the reactions deplete the surface of methoxy. Figure from reference [4].
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
Fig. 8. Titanium interstitials are required for formation of a benzaldehyde dimer (diolate) as illustrated in the schematic on the top right. The diolate is observed using scanning tunnelling microscopy (upper left). Density functional theory calculations show that the formation of the diolate is not favoured on a stoichiometric surface— interstitials are required. Corresponding simulations of the STM images are shown below the experimental image on the left.
forming a 1-D gas along the Ti five-coordinate rows.[41,43] Oxygen adatoms present on the surface react with ketones to form a surface complex that is photo-active.[2,42,43] This specific case clearly demonstrates the importance of O adatoms in determining the photochemical activity for organic oxidation reactions by formation of a distinct surface intermediate. Clearly, O adatoms play a major role in determining the types of intermediates that form on the titania surface which, in turn, affects photochemical efficiency. These results indicate that the presence of O2 during photo-oxidation plays a direct role in the formation of photo-active species. Indeed, the photo-oxidation pathways for trimethyl acetate on TiO2(110) strongly depend on the pressure of O2 in the gas phase during photo-reaction.[2,44] The initial photo-reaction is hole-mediated C–C bond breaking to yield CO2 and a t-butyl radical for all O2 pressures investigated. The ensuing reactions are altered by variation in the pressure of O2 because the distribution of the acetate and the oxygen on the surface is affected. Subsurface Defects While the roles of defects such as vacancies and adatoms in determining the nature of surface intermediates are clear and intuitive, less obvious is the fact that subsurface defects, e.g. Ti interstitials, are also important in determining reactive intermediates on the surface as illustrated by the case of the thermal reaction of benzaldehyde to yield stilbene and surface oxygen
Chem. Rec. 2014, 14, 944–951
on reduced TiO2(110). Interstitial Ti promotes the dimerization of benzaldehyde on reduced TiO2(110) as shown using scanning tunnelling microscopy (STM) and density functional theory (DFT) (Figure 8).[45] More recently, we used infrared spectroscopy to identify the diolate intermediate and to demonstrate a direct correlation of the amount formed and the degree of bulk reduction of the titania, providing further evidence for the key role of subsurface defects in promoting the formation of this intermediate.[46] The origin of this effect is that Ti interstitials present in the subsurface region donate charge to Ti4+ on the surface, leading to partial reduction. This increased electron density on the surface titanium ions weakens the C=O bonds and promotes C–C bond formation. Oxidation of the surface to neutralize this charge prevents the formation of the diolate.[47] While the photochemical activity of this species has not been investigated, it is clear that subsurface defects have a role on the bonding of species adsorbed on the surface. Hence, the bonding and possibly the very nature of photo-active intermediates present on a surface will be altered by subsurface defects.
Molecular Structure and Photo-oxidation Due to the importance of alignment of the electronic states of molecular species and the bands of the titania, the molecular structure of homologous sets of molecules also affects
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 949
THE CHEMICAL RECORD
Fig. 9. The differences in the rate of photo-decompostion of benzoate and trimethylacetate on reduced TiO2 is evident from STM images taken after illuminaton. Benzoate (top) does not undergo any detectable photo-reaction whereas trimethyl acetate (bottom) decomposes after exposure to light for only a few minutes.
photochemical efficiency. For example, there is a vast difference in the rate of photo-decomposition of benzoate and trimethyl acetate on reduced TiO2(110) (Figure 9).[2,48] While the origin of this effect is not understood in detail, it is interesting to note that the efficiency for photo-decomposition of these carboxylates correlates with the ease of C–C bond breaking. Specifically, the resonance stabilization of the carboxylate group via interaction with the phenyl ring most likely affects photochemical efficiency. Theoretical treatment of this effect would provide more insight into the large differences in photo-reaction efficiency.
Theoretical Treatment of Surface Photochemistry Theoretical treatment of photochemical events has the potential to provide deep insight into the design of more efficient processes. For example, the vast difference in the photochemical efficiency for decomposition of trimethyl acetate and benzoate could be probed by theory. In order to properly treat photochemical processes on surfaces, time dependent methods that include a description of electronically excited states and their evolution are required. Unfortunately, most theoretical studies invoked to describe photochemical processes have focused entirely on the bonding in the ground state.[42,49] There are few examples of the application of timedependent theoretical methods to surface photochemistry. For example, the photo-dissociation of methanol and phenol on anatase TiO2.[50] This study revealed a charge transfer from
950 www.tcr.wiley-vch.de
phenol to titania so as to create a molecular radical cation that subsequently reacts with oxygen and/or water to lead to the overall photo-decomposition reaction. The degree of charge transfer was greater for phenol than for methanol, which is consistent with experimental results cited in the paper.[50] Nonadiabatic molecular dynamics simulations have also been used to model electrons created through photon absorption on TiO2.[51] These hydrated (so-called “wet”) electrons are highly delocalized and require the presence of both water and titania to exist in this state. This example illustrates the importance of electron-phonon coupling and the need to consider strong interaction of the adsorbate and surface in modelling photo-induced processes. Nevertheless, ground state calculations can provide insight into photochemical mechanisms and efficiency. For example, comparison of the valence structure of methanol and methoxy on titania probed the level alignment with titania and concluded that methoxy will react more efficiently than methanol,[12,34,35] consistent with experimental observations.[1,38] These studies were coupled with time-dependent experimental measurements (two-photon photo-emission) that provide added insight about the temporal evolution of electronic states.[12] Further development of accurate time-dependent methods for treating excited states coupled with an understanding of the key intermediates that lead to photochemical reactions remains a challenge in the development of more efficient photochemistry on semiconductors. Theoretical studies have the potential to establish guiding principles in designing materials for specific photochemical reactions.
Outlook and Summary Development of effective strategies for efficient photochemical processes requires an understanding of the elementary steps that lead to the overall reaction. Specifically, the identification of photo-active intermediates on the surface and a description of excited state dynamics are an important and generally overlooked factor in heterogeneous photochemistry. While design of materials with appropriate optical properties is important, alignment of the electronic states of key reactive intermediates with states in the material is also critical. Thus, fundamental surface science studies that map out the reaction mechanism for photochemical processes on materials with controlled composition and structure play a critical role in understanding heterogeneous photochemistry. A more detailed understanding of charge carrier dynamics in photocatalytic materials, while not discussed in detail here, is also required. To this end, the development of new theoretical and experimental tools that probe the time evolution of excited states in conjunction with
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
existing methods that elucidate molecular bonding, structure and reactivity are needed as a complement to band-gap engineering of materials.
[24] [25]
Acknowledgements
[27]
The support of the U.S. National Science Foundation under grants DMR-0934480 and CHE-0956653 is gratefully acknowledged. I also thankfully acknowledge the assistance of Mr. Dmitry Vinichenko in preparing several figures.
[28]
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
[17] [18] [19] [20] [21] [22] [23]
[29] [30] [31] [32]
J. Tao, T. Luttrell, M. Batzill, Nat. Chem. 2011, 3, 296–300. M. A. Henderson, I. Lyubinetsky, Chem. Rev. 2013, 113, 4428–4455. C. L. Pang, R. Lindsay, G. Thornton, Chem. Soc. Rev. 2008, 37, 2328–2353. A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515–582. C. Sousa, S. Tosoni, F. Illas, Chem. Rev. 2013, 113, 4456– 4495. M. A. Henderson, Surf. Sci. Rep. 2011, 185–297. K. R. Phillips, S. C. Jensen, M. Baron, S.-C. Li, C. M. Friend, J. Am. Chem. Soc. 2013, 135, 574–577. H. Kominami, H. Sugahara, K. Hashimoto, Catal. Commun. 2010, 11, 426–429. C. Di Valentin, G. Pacchioni, A. Selloni, J. Phys. Chem. C 2009, 113, 20543–20552. K. Maeda, K. Domen, J. Phys. Chem. C 2007, 111, 7851– 7861. E. Finazzi, C. Di Valentin, G. Pacchioni, A. Selloni, J. Chem. Phys. 2008, 129, 154113. A. Migani, D. J. Mowbray, A. Iacomino, J. Zhao, H. Petek, A. Rubio, J. Am. Chem. Soc. 2013, 135, 11429–11432. L. Chiodo, J. M. Garcia-Lastra, A. Iacomino, S. Ossicini, J. Zhao, H. Petek, A. Rubio, Phys. Rev. B 2010, 82, 045207. W. Kang, M. S. Hybertsen, Phys. Rev. B 2010, 82, 085203. Z.-T. Wang, N. Aaron Deskins, I. Lyubinetsky, J. Phys. Chem. Lett. 2012, 3, 102–106. S. Tan, H. Feng, Y. Ji, Y. Wang, J. Zhao, A. Zhao, B. Wang, Y. Luo, J. Yang, J. G. Hou, J. Am. Chem. Soc. 2012, 134, 9978– 9985. L. Li, P. A. Salvador, G. S. Rohrer, Nanoscale 2014, 6, 24–42. A. Fujishima, K. Honda, Nature 1972, 238, 37–38. M. A. Henderson, Surf. Sci. 1996, 355, 151–166. P. Liao, E. A. Carter, Chem. Soc. Rev. 2013, 42, 2401–2422. Z. Zhang, O. Bondarchuk, J. M. White, B. D. Kay, Z. Dohnálek, J. Am. Chem. Soc. 2006, 128, 4198–4199. W.-T. Chen, V. Jovic, D. Sun-Waterhouse, H. Idriss, G. Waterhouse, Int. J. Hydrogen Energy 2013, 38, 15036–15048. S. Kundu, A. B. Vidal, M. A. Nadeem, S. D. Senanayake, H. Idriss, P. Liu, J. A. Rodriguez, D. Stacchiola, J. Phys. Chem. C 2013, 117, 11149–11158.
Chem. Rec. 2014, 14, 944–951
[26]
[33]
[34] [35] [36] [37]
[38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51]
P. V. Kamat, Acc. Chem. Res. 2012, 45, 1906–1915. I. Robel, V. Subramanian, M. Kuno, P. V. Kamat, J. Am. Chem. Soc. 2006, 128, 2385–2393. X. Zhang, Y. L. Chen, R.-S. Liu, D. P. Tsai, Rep. Prog. Phys. 2013, 76, 046401–046442. S. Linic, P. Christopher, H. Xin, A. Marimuthu, Acc. Chem. Res. 2013, 46, 1890–1899. S. Linic, P. Christopher, D. B. Ingram, Nat. Mater. 2011, 10, 911–921. W. Hou, S. B. Cronin, Adv. Funct. Mater. 2012, 23, 1612– 1619. P. P. Christopher, H. H. Xin, A. A. Marimuthu, S. S. Linic, Nat. Mater. 2012, 11, 1044–1050. S. Linic, P. Christopher, US Patent 20,130,122,396 2013. X. Chen, H.-Y. Zhu, J.-C. Zhao, Z.-F. Zheng, X.-P. Gao, Angew. Chem. Int. Ed. 2008, 47, 5353–5356. R. Sellappan, M. G. Nielsen, F. González-Posada, P. C. K. Vesborg, I. Chorkendorff, D. Chakarov, J. Catal. 2013, 307, 214–221. J. Tao, M. Yang, J. Chai, J. Pan, Y. P. Feng, J. Phys. Chem. C 2014, 118, 994–1000. S. S. Soni, M. J. Henderson, J.-F. Bardeau, A. Gibaud, Adv. Mater. 2008, 20, 1493–1498. E. C. Landis, K. C. Phillips, E. Mazur, C. M. Friend, J. Appl. Phys. 2012, 112, 063108–063108-5. A. C. Papageorgiou, N. S. Beglitis, C. L. Pang, G. Teobaldi, G. Cabailh, Q. Chen, A. J. Fisher, W. A. Hofer, G. Thornton, Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 2391–2396. M. Shen, M. A. Henderson, J. Phys. Chem. Lett. 2011, 2, 2707–2710. W. Yang, Z. Ren, Z. Ma, D. Dai, T. K. Minton, J. Phys. Chem. C 2013, 117, 5293–5300. P. M. Jayaweera, E. L. Quah, H. Idriss, J. Phys. Chem. C Nanomater Interfaces 2007, 111, 1764–1769. S. C. Jensen, A. Shank, R. J. Madix, C. M. Friend, ACS Nano 2012, 6, 2925–2930. M. A. Henderson, N. A. Deskins, R. T. Zehr, M. Dupuis, J. Catal. 2011, 279, 205–212. S. C. Jensen, K. R. Phillips, M. Baron, E. C. Landis, C. M. Friend, Phys. Chem. Chem. Phys. 2013, 15, 5193–5201. M. Henderson, J. M. White, H. Uetsuka, H. Onishi, J. Catal. 2006, 238, 153–164. L. Benz, J. Haubrich, S. C. Jensen, C. M. Friend, ACS Nano 2011, 5, 834–843. P. Clawin, C. M. Friend, K. Al-Shamery, Chem. Eur. J. 2014, DOI 10.1002/chem.201402102. S. C. Jensen, J. Haubrich, A. Shank, C. M. Friend, Chem. Eur. J. 2011, 17, 8309–8312. E. C. Landis, S. C. Jensen, K. R. Phillips, C. M. Friend, J. Phys. Chem. C 2012, 116, 21508–21513. Q. Guo, C. Xu, Z. Ren, W. Yang, Z. Ma, D. Dai, H. Fan, T. K. Minton, X. Yang, J. Am. Chem. Soc. 2012, 134, 13366–13373. S. Suzuki, T. Tsuneda, K. Hirao, J. Chem. Phys. 2012, 136, 024706. S. A. Fischer, W. R. Duncan, O. V. Prezhdo, J. Am. Chem. Soc. 2009, 131, 15483–15491.
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 951
THE CHEMICAL RECORD
Perspectives on Heterogeneous Photochemistry Cynthia M. Friend Harvard University, Department of Chemistry and Chemical Biology, 12 Oxford St., Cambridge, MA 02138 (USA) E-mail: [email protected]
Received: April 11, 2014 Publised online: August 18, 2014
ABSTRACT: Heterogeneous photochemistry has a potentially important role in production of energy, in environmental remediation and in sustainable production of chemicals. Photochemical efficiency depends on both materials properties and the desired chemical reaction that is promoted through creation of an excited state. A detailed understanding of the interplay between materials properties and reactivity requires a molecular-scale approach that determines the elementary steps in the overall process. This personal account summarizes the role of defects in determining the photochemical and thermal reactions on rutile titania, a model for semiconductor metal oxide photocatalysts that defects, e.g., Ti interstitials present in the subsurface region, and O adatoms on the surface, have a substantial impact on the efficiency for photochemical conversion through modification of molecular binding and also through likely modification of charge carrier dynamics. Design of materials must include engineering of the optical and electronic properties of the semiconductor photocatalyst, and understanding of the key photochemical steps involved in specific processes to ensure proper alignment of their electronic states with the band structure of the material. Thus, fundamental surface science studies and development of time-dependent theoretical methods that map out the reaction mechanism for photochemical processes on materials with controlled composition and structure are critical. DOI 10.1002/tcr.201402037 Keywords: defects, oxidation, photochemistry, surface chemistry, titania
Introduction Some of the most pressing technological issues facing the world today revolve around energy and the environment. There is a desire to shift the production of chemicals and fuels to renewable resources to reduce dependence on petrochemical resources; additionally, removal of noxious molecules from the atmosphere is important for improving air quality. Solutions for both of these problems require efficient transformations of
944 www.tcr.wiley-vch.de
organic molecules, and heterogeneous photochemistry provides one avenue for driving these key processes. The design and implementation of photochemistry to address these pressing problems related to energy and the environment requires a detailed understanding of the interplay between materials properties and chemical bonding and reactivity. Herein, selected examples are used to illustrate how the
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
powerful tools of surface chemistry are used to establish a molecular approach to understanding heterogeneous photochemical processes. Fundamental studies of titania are an illustrative case because the chemical and structural properties of titania have been widely studied, as discussed in several excellent reviews.[1–3] Fundamental studies of surface chemistry under ultrahigh vacuum and associated theoretical models establish guiding principles for photocatalysis as has been discussed in several reviews.[2,4–6] For example, the mechanism for photo-oxidation of methanol to methyl formate was determined using surface science tools,[7] providing insight into the analogous vapour phase process over a supported catalyst at higher pressures.[8] The molecular level insight into surface photochemistry provided by surface science studies provides the basis for understanding photocatalysis despite the vastly different reaction conditions, especially since most catalytic processes operate at low steady-state concentrations of reactants on the surface. In this account, the roles of defects and dopants in charge carrier recombination, bonding of molecular intermediates to the surface, treatment of excited states of the semiconductormolecule system, and coupling to thermal processes are discussed. Although titania is used to illustrate these effects, similar factors may play a role in other metal oxide photocatalysts in general.
Overview of Heterogeneous Photochemistry on Semiconductors Principles Excitation of electron-hole pairs in a semiconductor material is the first step in photocatalysis. The energy of the band gap in the semiconductor is critical because it determines the photon energy required for excitation of charge carriers. The alignment of the band edges with electronic states in molecules and
Cynthia M. Friend is the T. W. Richards Professor of Chemistry and Director of the Rowland Institute at Harvard University. She joined the Harvard faculty in 1982 after one year of postdoctoral research at Stanford University. Friend earned a doctorate in Physical Chemistry from the University of California (UC) Berkeley (1981) and a B.S. degree from UC Davis (1977). She has been widely honoured for her research in surface chemistry and catalysis. Friend is a Fellow of the American Association for the Advancement of Science and the American Chemical Society.
Chem. Rec. 2014, 14, 944–951
Fig. 1. Qualitative illustration of principles of photocatalysis applied to water splitting from reference [10]. This schematic representation does not take into account the interaction of the electron-hole pair following excitation which alters the energies of these charge carriers.
intermediates on the surface also determines the types of reactions that the charge carriers (electrons and holes) can promote because charge carriers must have sufficient energy to drive reactions and also be able to interact with specific states of molecules that lead to oxidation or reduction.[9–12] For example, water splitting to O2 and H2 is an important reaction that has the potential to produce clean fuels. The minimum energy required for this reaction is 1.23 eV (Figure 1); thus, a minimum requirement for a photocatalyst for this process is a material with a band gap of 1.23 eV or greater. The optical band gap of semiconductor materials provides a guide to which materials might be viable photocatalysts; however, the actual energies of the electrons in the conduction band and holes in the valence band will be different due to Coulombic interaction (Figure 2).[13,14] Hence, the additional energy required to create an excitonic pair of electrons and holes must be considered. In addition, the conduction and valence bands of a material must also have a proper alignment with specific states in surface-bound intermediates; in other words, there must be electronic states in a molecule or molecular intermediate that can accept or donate an electron for the desired reaction to occur. For example, there must be states of methoxy that overlap with the top of the valence band of TiO2 in order for photo-oxidation to occur (Figure 3). Hence, the electronic band gap and band alignment are key factors in determining photochemical efficacy of a material for a specific reaction. The same principles apply to other photochemical reactions; however, the requirements for band alignment and the overall energetics will be different for every specific reaction. To fully
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 945
THE CHEMICAL RECORD
Energy
Conduction band
Eexc
Egap
Eop Valence band Density of states
Fig. 2. Qualitative illustration of the effect of the difference between the fundamental and optical gap due to exciton binding. The difference is related to the electron and hole effective masses and is typically less than 0.1 eV.
TiO2
bridging site in the photo-dissociation of water using 400 nm light.[15,16] In this case, water photo-dissociation occurs; however, the overall production of O2 and H2 requires subsequent steps that could be either thermal or photochemical. Furthermore, the efficiency for the water photo-dissociation depends on the alignment of the energy of the excited state in the molecule and the top of the valence band, where the holes are made to drive the O–H bond scission. More generally, the photo-active species may be a new intermediate formed from an earlier thermal step. Thus, the electronic structure of both the photo-active molecular species on the surface and the material to which it is bound are important because the excitation of electron hole pairs in the material must be aligned with excited states that lead to photoreaction in the molecule. Thus, designing efficient and selective photochemical processes requires not only an understanding of overall energetics, but also understanding of the nature of key intermediates and of the dynamics of excited states. Below, the role of defects in determining the nature of surface bound species is discussed.
Energy
conduction band
Methoxy occupied orbitals
Overlap of states
Materials Design
TiO2 valence band
Density of states Fig. 3. Qualitative illustration of the hybridization of the highest occupied orbitals of methoxy with the top of the valence band of titania which allows for transfer of a hole in the valence band to the molecule so as to induce photo-oxidation.
understand the requirements for promoting a reaction with light, a detailed molecular scale mechanism is required because the electronic states of different molecular intermediates will be aligned differently with the bands of the semiconductor.[12] Understanding the elementary steps that lead to an overall reaction is a critical aspect of photochemistry on surfaces that is not widely considered in designing new materials for photochemistry. The example above for water only considers the overall reaction energetics and not the elementary steps in the overall process. Understanding elementary steps is critical because the energetics for both photochemical and thermal steps depend on the intermediates for these steps and their bonding to the surface. Recent studies using scanning tunnelling microscopy (STM) were used to demonstrate that a proton is transferred to a bridging oxygen on TiO2 to yield one OH group bound on the Ti five-coordinate rows and one at a
946 www.tcr.wiley-vch.de
There has been a major recent effort to design new materials for photochemical reactions. To date, transition metal oxides are among the most common materials used for photocatalysis and within this class, titania is most commonly used. Titania has the advantage that it is easy to prepare, is earth abundant and chemically versatile; however, it has an optical band gap in the ultraviolet (3.1 eV for rutile), making it rather inefficient for solar energy conversion. Titania also has some activity for water splitting which has made it a popular material to investigate.[17,18] Because there is great interest in using solar energy for photochemical transformations, there is a drive to develop materials with excitations that overlap the solar spectrum, which has the highest photon irradiance between ∼1–3 eV. Therefore, there is a major effort to develop photocatalysts with appropriate optical and chemical properties to promote photochemical processes using visible light predominant in the solar spectrum.[19,20] To this end, other transition metal oxides and a variety of semiconductors with smaller band gaps (Figure 4) as well as hybrid thereof are being designed[20,21] and tested[10,22–24] for a variety of reactions. For example, hybrids of TiO2 and CdSe have been demonstrated to absorb visible light via the CdSe leading to charge injection into the titania.[25] Another approach is to sensitize semiconductor materials, such as titania, with metallic structures. Recent work has demonstrated the potential value of using SPR excitation of Ag or Au nanoparticles as photocatalysts.[26–29] Recent studies of ethylene oxidation using Ag nanoparticles as photocatalysts[30,31]
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
Fig. 4. Schematic of the variation of the optical band gap and band alignment using water splitting as a reference. Figure from reference [10] by Maeda and Domen.
are an excellent example of the value of fundamental studies that provide insight into mechanism. There are a range of studies of SPR-based photocatalysts for various reactions, including water splitting,[28,29] the decomposition of organic dyes and other VOCs,[26,32] and methanol oxidation,[33] suggesting wide applicability to this approach; however, the effect of SPR excitation on specific elementary steps has not been determined for most of these cases. A full description of these effects must describe the time-dependent behaviour of these collective excitations. Most of the endeavours to design new photocatalysts have all focused on engineering the optical properties of these materials; however, a molecular scale understanding of chemical reactivity is required to determine how materials can be modified to achieve more efficient photocatalysis. As motivated above in the discussion of water photo-dissociation, the bonding of photo-active species and alignment of electronic states that lead to bond scission with the valence and conduction bands for hole- and electron-mediated processes, respectively, are an important part of designing photocatalysts for specific reactions. Therefore, the bonding and thermal reactivity of specific molecules on a particular photocatalyst must be considered. Using the case of reduced rutile TiO2(110) as a case study, the value of fundamental studies of reactivity in understanding key factors that control photochemical reaction are revealed.
Defects and Photochemistry Titania is one of the most widely investigated materials for photocatalysis even though it has a large band gap of ∼3 eV in the ultraviolet. There has been a concerted effort to modify the electronic and optical properties of TiO2 through doping with the goal of creating charge carriers and inducing photochemical events using visible light. Doping with other transition metals, e.g. Ni, Fe, W,[20] and with other anions, e.g.
Chem. Rec. 2014, 14, 944–951
nitride,[34,35] is one approach used. Most commonly, creation of defects in the titania through reduction is employed.[1] Defects in titania are created through thermal, electron- or photoinduced reduction. Materials with a high defect density can also be created using a variety of methods, including laserassisted oxidation.[36] Doping of titania with defects significantly alters the electronic and optical properties by introducing states into the gap.[1] In titania, the change in optical properties manifests itself with a dark blue colour with increasing degree of reduction. Oxygen deficiency in titania and other transition metal oxides creates electronic states in the band gap that lower the energy for electron-hole pair excitation. In the case of titania, this leads to absorption of visible light. Typical defects that occur in TiO2 are oxygen vacancies on the surface, e.g. bridging oxygen vacancies on the rutile TiO2(110) surface, and subsurface defects such as Ti interstitials. The deficiency of O both on the surface and in the bulk lead to excess electron density near these defects,[9,11] creating charge differences that can lead to site-specific reactivity. Exposure of reduced surfaces to gaseous O2 also creates oxygen adatoms on the surfaces of reduced titania, filling in most of the surface oxygen vacancies and compensating for the charge due to the excess Ti interstitials.[15] All of these vacancies can play a role in both thermal and photochemical processes on the surface. While the potential role of surface O vacancies is clearly established, only recently has the effect of subsurface, interstitial defects been demonstrated to be important in thermal reactions on titania. In photochemical processes, defects can affect the lifetimes of charge carriers. For example, O vacancies are thought to trap charge carriers leading to an increase in the rate of electron-hole pair recombination, thus decreasing the efficiency of heterogeneous photochemistry. Because photons penetrate into the bulk of a photocatalyst, electron-hole pairs are created even in the bulk of the solid. The absorption depth is estimated to be ∼100 nm for light with an energy of 3.4.eV, for example.[17] In order to promote photochemical processes, these charge carriers must remain separated and must migrate to the surface where reactants are bound. Photocatalyst architectures can be designed to optimize photon absorption and survival of charge carriers as they migrate to the surface—for example, by using nanowires. However, control of materials composition and structure at an atomic level is important for the best efficiency because of the need to control chemical reactivity. Understanding the elementary steps that lead to an overall photochemical reaction on a semiconductor surface is an important and often overlooked factor in determining the overall efficiency. Specifically, photochemical transformations depend on the binding of reactants on the surface because bonding determines their ground and electronically excited states, including how they evolve in time. Defects play an
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 947
THE CHEMICAL RECORD
important role in determining the binding of molecular species to the surface and, therefore, changing the nature of ensuing photochemical events. Many studies have demonstrated the importance of bridging oxygen vacancies on the surface in determining the rates of thermal reactions. For example, water dissociates to two OH groups on a bridging oxygen vacancy.[19] Similarly, methanol reacts to form CH3O bound to bridging oxygen vacancies transferring the H to a neighbouring bridging O.[21] Here, the roles of O adatoms and subsurface defects are discussed in more detail. Oxygen Adatoms Oxygen adatoms on the surface affect chemical reactivity on titania by leading to the formation of new intermediates. O adatoms are formed from exposure to TiO2 that is partially reduced. O2 reacts to fill surface vacancies and O adatoms form on the Ti five-coordinate rows in concert.[4,37] Thus, the net effect is to neutralize excess charge on the surface and to change the nature of potential reaction sites.[15,16] Oxygen adatoms promote the formation of specific intermediates that are formed thermally prior to photochemical activation. The new intermediates have very different photochemical efficiencies for reactions. These new intermediates have a different electronic structure, leading to changes in their alignment relative to the states in titania. As a result, the photochemical behaviour will almost certainly be different than the original reactant molecule. There are several well-documented examples of changes in photochemical efficiency due to the presence of O adatoms on titania. In the case of methanol, O adatoms on titania promote O–H bond breaking to form methoxy on TiO2(110). The O adatoms are all removed as water below 350 K.[7,18,38] Methoxy formed in this manner is photo-oxidized to formaldehyde irradiation with ultraviolet light (Figure 5). Notably, the rate of methoxy photo-oxidation is ∼50 times faster than methanol on a reduced surface.[7,20,38,39] This difference has been rationalized in terms of better level alignment of methoxy with the titania valence band than methanol.[12,20] Similar reactions are observed for ethanol on titania[10,22–24,40] and Ru-modified TiO2[23,25]—acetaldehyde is formed. For dense layers of methoxy, a second photochemical pathway develops form the photo-oxidation of formaldehyde.
Fig. 5. Schematic of initial photochemical event in the photo-oxidation of methoxy formed from methanol on TiO2(110). Adapted from reference [7].
948 www.tcr.wiley-vch.de
The ultimate product is methyl formate that is formed from reaction with a nearby methoxy (Figure 6). The fact that methyl formate is produced shows that methoxy is photooxidized more slowly than formaldehyde. If methoxy reacted faster, it would not be available for the coupling process. The mechanism for both the primary and secondary oxidation of methoxy was established based on fundamental studies using scanning tunnelling microscopy (STM) and isotopic labelling in conjunction with both photon-stimulated and thermal desorption measurements.[7] The efficiency of the photo-oxidation and the sequential nature of formaldehyde is illustrated by the photon-stimulated desorption of the products during irradiation (Figure 7). The example of methoxy photo-oxidation demonstrates how O adatoms affect the efficiency of photooxidation by promoting the formation of active surface intermediates with electronic structure that aligns better with the semiconductor states of titania. This case also demonstrates the value of fundamental studies in determining the pathways for photo-oxidation at a molecular level. A second example of the effect of O adatoms on photochemistry via formation of a new photo-active species is the case of ketone photo-oxidation on titania. Ketones are not photochemically active on reduced titania and they are relatively weakly bound to the surface.[2,41] Butyrophenone, for example, is highly mobile on the surface even at low temperature,
Fig. 6. Schematic of secondary photo-oxidation of formaldehyde and coupling with methoxy on titania to form methyl formate. Adapted from reference [4].
Fig. 7. Production of formaldehyde (red) and methyl formate (blue) as a function of irraditation time from photo-oxidation of methoxy on TiO2(110). The sharp rise in formaldehdye decays as the reactions deplete the surface of methoxy. Figure from reference [4].
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
Fig. 8. Titanium interstitials are required for formation of a benzaldehyde dimer (diolate) as illustrated in the schematic on the top right. The diolate is observed using scanning tunnelling microscopy (upper left). Density functional theory calculations show that the formation of the diolate is not favoured on a stoichiometric surface— interstitials are required. Corresponding simulations of the STM images are shown below the experimental image on the left.
forming a 1-D gas along the Ti five-coordinate rows.[41,43] Oxygen adatoms present on the surface react with ketones to form a surface complex that is photo-active.[2,42,43] This specific case clearly demonstrates the importance of O adatoms in determining the photochemical activity for organic oxidation reactions by formation of a distinct surface intermediate. Clearly, O adatoms play a major role in determining the types of intermediates that form on the titania surface which, in turn, affects photochemical efficiency. These results indicate that the presence of O2 during photo-oxidation plays a direct role in the formation of photo-active species. Indeed, the photo-oxidation pathways for trimethyl acetate on TiO2(110) strongly depend on the pressure of O2 in the gas phase during photo-reaction.[2,44] The initial photo-reaction is hole-mediated C–C bond breaking to yield CO2 and a t-butyl radical for all O2 pressures investigated. The ensuing reactions are altered by variation in the pressure of O2 because the distribution of the acetate and the oxygen on the surface is affected. Subsurface Defects While the roles of defects such as vacancies and adatoms in determining the nature of surface intermediates are clear and intuitive, less obvious is the fact that subsurface defects, e.g. Ti interstitials, are also important in determining reactive intermediates on the surface as illustrated by the case of the thermal reaction of benzaldehyde to yield stilbene and surface oxygen
Chem. Rec. 2014, 14, 944–951
on reduced TiO2(110). Interstitial Ti promotes the dimerization of benzaldehyde on reduced TiO2(110) as shown using scanning tunnelling microscopy (STM) and density functional theory (DFT) (Figure 8).[45] More recently, we used infrared spectroscopy to identify the diolate intermediate and to demonstrate a direct correlation of the amount formed and the degree of bulk reduction of the titania, providing further evidence for the key role of subsurface defects in promoting the formation of this intermediate.[46] The origin of this effect is that Ti interstitials present in the subsurface region donate charge to Ti4+ on the surface, leading to partial reduction. This increased electron density on the surface titanium ions weakens the C=O bonds and promotes C–C bond formation. Oxidation of the surface to neutralize this charge prevents the formation of the diolate.[47] While the photochemical activity of this species has not been investigated, it is clear that subsurface defects have a role on the bonding of species adsorbed on the surface. Hence, the bonding and possibly the very nature of photo-active intermediates present on a surface will be altered by subsurface defects.
Molecular Structure and Photo-oxidation Due to the importance of alignment of the electronic states of molecular species and the bands of the titania, the molecular structure of homologous sets of molecules also affects
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 949
THE CHEMICAL RECORD
Fig. 9. The differences in the rate of photo-decompostion of benzoate and trimethylacetate on reduced TiO2 is evident from STM images taken after illuminaton. Benzoate (top) does not undergo any detectable photo-reaction whereas trimethyl acetate (bottom) decomposes after exposure to light for only a few minutes.
photochemical efficiency. For example, there is a vast difference in the rate of photo-decomposition of benzoate and trimethyl acetate on reduced TiO2(110) (Figure 9).[2,48] While the origin of this effect is not understood in detail, it is interesting to note that the efficiency for photo-decomposition of these carboxylates correlates with the ease of C–C bond breaking. Specifically, the resonance stabilization of the carboxylate group via interaction with the phenyl ring most likely affects photochemical efficiency. Theoretical treatment of this effect would provide more insight into the large differences in photo-reaction efficiency.
Theoretical Treatment of Surface Photochemistry Theoretical treatment of photochemical events has the potential to provide deep insight into the design of more efficient processes. For example, the vast difference in the photochemical efficiency for decomposition of trimethyl acetate and benzoate could be probed by theory. In order to properly treat photochemical processes on surfaces, time dependent methods that include a description of electronically excited states and their evolution are required. Unfortunately, most theoretical studies invoked to describe photochemical processes have focused entirely on the bonding in the ground state.[42,49] There are few examples of the application of timedependent theoretical methods to surface photochemistry. For example, the photo-dissociation of methanol and phenol on anatase TiO2.[50] This study revealed a charge transfer from
950 www.tcr.wiley-vch.de
phenol to titania so as to create a molecular radical cation that subsequently reacts with oxygen and/or water to lead to the overall photo-decomposition reaction. The degree of charge transfer was greater for phenol than for methanol, which is consistent with experimental results cited in the paper.[50] Nonadiabatic molecular dynamics simulations have also been used to model electrons created through photon absorption on TiO2.[51] These hydrated (so-called “wet”) electrons are highly delocalized and require the presence of both water and titania to exist in this state. This example illustrates the importance of electron-phonon coupling and the need to consider strong interaction of the adsorbate and surface in modelling photo-induced processes. Nevertheless, ground state calculations can provide insight into photochemical mechanisms and efficiency. For example, comparison of the valence structure of methanol and methoxy on titania probed the level alignment with titania and concluded that methoxy will react more efficiently than methanol,[12,34,35] consistent with experimental observations.[1,38] These studies were coupled with time-dependent experimental measurements (two-photon photo-emission) that provide added insight about the temporal evolution of electronic states.[12] Further development of accurate time-dependent methods for treating excited states coupled with an understanding of the key intermediates that lead to photochemical reactions remains a challenge in the development of more efficient photochemistry on semiconductors. Theoretical studies have the potential to establish guiding principles in designing materials for specific photochemical reactions.
Outlook and Summary Development of effective strategies for efficient photochemical processes requires an understanding of the elementary steps that lead to the overall reaction. Specifically, the identification of photo-active intermediates on the surface and a description of excited state dynamics are an important and generally overlooked factor in heterogeneous photochemistry. While design of materials with appropriate optical properties is important, alignment of the electronic states of key reactive intermediates with states in the material is also critical. Thus, fundamental surface science studies that map out the reaction mechanism for photochemical processes on materials with controlled composition and structure play a critical role in understanding heterogeneous photochemistry. A more detailed understanding of charge carrier dynamics in photocatalytic materials, while not discussed in detail here, is also required. To this end, the development of new theoretical and experimental tools that probe the time evolution of excited states in conjunction with
Chem. Rec. 2014, 14, 944–951
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
Heterogeneous Photochemistry
existing methods that elucidate molecular bonding, structure and reactivity are needed as a complement to band-gap engineering of materials.
[24] [25]
Acknowledgements
[27]
The support of the U.S. National Science Foundation under grants DMR-0934480 and CHE-0956653 is gratefully acknowledged. I also thankfully acknowledge the assistance of Mr. Dmitry Vinichenko in preparing several figures.
[28]
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
[17] [18] [19] [20] [21] [22] [23]
[29] [30] [31] [32]
J. Tao, T. Luttrell, M. Batzill, Nat. Chem. 2011, 3, 296–300. M. A. Henderson, I. Lyubinetsky, Chem. Rev. 2013, 113, 4428–4455. C. L. Pang, R. Lindsay, G. Thornton, Chem. Soc. Rev. 2008, 37, 2328–2353. A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515–582. C. Sousa, S. Tosoni, F. Illas, Chem. Rev. 2013, 113, 4456– 4495. M. A. Henderson, Surf. Sci. Rep. 2011, 185–297. K. R. Phillips, S. C. Jensen, M. Baron, S.-C. Li, C. M. Friend, J. Am. Chem. Soc. 2013, 135, 574–577. H. Kominami, H. Sugahara, K. Hashimoto, Catal. Commun. 2010, 11, 426–429. C. Di Valentin, G. Pacchioni, A. Selloni, J. Phys. Chem. C 2009, 113, 20543–20552. K. Maeda, K. Domen, J. Phys. Chem. C 2007, 111, 7851– 7861. E. Finazzi, C. Di Valentin, G. Pacchioni, A. Selloni, J. Chem. Phys. 2008, 129, 154113. A. Migani, D. J. Mowbray, A. Iacomino, J. Zhao, H. Petek, A. Rubio, J. Am. Chem. Soc. 2013, 135, 11429–11432. L. Chiodo, J. M. Garcia-Lastra, A. Iacomino, S. Ossicini, J. Zhao, H. Petek, A. Rubio, Phys. Rev. B 2010, 82, 045207. W. Kang, M. S. Hybertsen, Phys. Rev. B 2010, 82, 085203. Z.-T. Wang, N. Aaron Deskins, I. Lyubinetsky, J. Phys. Chem. Lett. 2012, 3, 102–106. S. Tan, H. Feng, Y. Ji, Y. Wang, J. Zhao, A. Zhao, B. Wang, Y. Luo, J. Yang, J. G. Hou, J. Am. Chem. Soc. 2012, 134, 9978– 9985. L. Li, P. A. Salvador, G. S. Rohrer, Nanoscale 2014, 6, 24–42. A. Fujishima, K. Honda, Nature 1972, 238, 37–38. M. A. Henderson, Surf. Sci. 1996, 355, 151–166. P. Liao, E. A. Carter, Chem. Soc. Rev. 2013, 42, 2401–2422. Z. Zhang, O. Bondarchuk, J. M. White, B. D. Kay, Z. Dohnálek, J. Am. Chem. Soc. 2006, 128, 4198–4199. W.-T. Chen, V. Jovic, D. Sun-Waterhouse, H. Idriss, G. Waterhouse, Int. J. Hydrogen Energy 2013, 38, 15036–15048. S. Kundu, A. B. Vidal, M. A. Nadeem, S. D. Senanayake, H. Idriss, P. Liu, J. A. Rodriguez, D. Stacchiola, J. Phys. Chem. C 2013, 117, 11149–11158.
Chem. Rec. 2014, 14, 944–951
[26]
[33]
[34] [35] [36] [37]
[38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51]
P. V. Kamat, Acc. Chem. Res. 2012, 45, 1906–1915. I. Robel, V. Subramanian, M. Kuno, P. V. Kamat, J. Am. Chem. Soc. 2006, 128, 2385–2393. X. Zhang, Y. L. Chen, R.-S. Liu, D. P. Tsai, Rep. Prog. Phys. 2013, 76, 046401–046442. S. Linic, P. Christopher, H. Xin, A. Marimuthu, Acc. Chem. Res. 2013, 46, 1890–1899. S. Linic, P. Christopher, D. B. Ingram, Nat. Mater. 2011, 10, 911–921. W. Hou, S. B. Cronin, Adv. Funct. Mater. 2012, 23, 1612– 1619. P. P. Christopher, H. H. Xin, A. A. Marimuthu, S. S. Linic, Nat. Mater. 2012, 11, 1044–1050. S. Linic, P. Christopher, US Patent 20,130,122,396 2013. X. Chen, H.-Y. Zhu, J.-C. Zhao, Z.-F. Zheng, X.-P. Gao, Angew. Chem. Int. Ed. 2008, 47, 5353–5356. R. Sellappan, M. G. Nielsen, F. González-Posada, P. C. K. Vesborg, I. Chorkendorff, D. Chakarov, J. Catal. 2013, 307, 214–221. J. Tao, M. Yang, J. Chai, J. Pan, Y. P. Feng, J. Phys. Chem. C 2014, 118, 994–1000. S. S. Soni, M. J. Henderson, J.-F. Bardeau, A. Gibaud, Adv. Mater. 2008, 20, 1493–1498. E. C. Landis, K. C. Phillips, E. Mazur, C. M. Friend, J. Appl. Phys. 2012, 112, 063108–063108-5. A. C. Papageorgiou, N. S. Beglitis, C. L. Pang, G. Teobaldi, G. Cabailh, Q. Chen, A. J. Fisher, W. A. Hofer, G. Thornton, Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 2391–2396. M. Shen, M. A. Henderson, J. Phys. Chem. Lett. 2011, 2, 2707–2710. W. Yang, Z. Ren, Z. Ma, D. Dai, T. K. Minton, J. Phys. Chem. C 2013, 117, 5293–5300. P. M. Jayaweera, E. L. Quah, H. Idriss, J. Phys. Chem. C Nanomater Interfaces 2007, 111, 1764–1769. S. C. Jensen, A. Shank, R. J. Madix, C. M. Friend, ACS Nano 2012, 6, 2925–2930. M. A. Henderson, N. A. Deskins, R. T. Zehr, M. Dupuis, J. Catal. 2011, 279, 205–212. S. C. Jensen, K. R. Phillips, M. Baron, E. C. Landis, C. M. Friend, Phys. Chem. Chem. Phys. 2013, 15, 5193–5201. M. Henderson, J. M. White, H. Uetsuka, H. Onishi, J. Catal. 2006, 238, 153–164. L. Benz, J. Haubrich, S. C. Jensen, C. M. Friend, ACS Nano 2011, 5, 834–843. P. Clawin, C. M. Friend, K. Al-Shamery, Chem. Eur. J. 2014, DOI 10.1002/chem.201402102. S. C. Jensen, J. Haubrich, A. Shank, C. M. Friend, Chem. Eur. J. 2011, 17, 8309–8312. E. C. Landis, S. C. Jensen, K. R. Phillips, C. M. Friend, J. Phys. Chem. C 2012, 116, 21508–21513. Q. Guo, C. Xu, Z. Ren, W. Yang, Z. Ma, D. Dai, H. Fan, T. K. Minton, X. Yang, J. Am. Chem. Soc. 2012, 134, 13366–13373. S. Suzuki, T. Tsuneda, K. Hirao, J. Chem. Phys. 2012, 136, 024706. S. A. Fischer, W. R. Duncan, O. V. Prezhdo, J. Am. Chem. Soc. 2009, 131, 15483–15491.
© 2014 The Chemical Society of Japan and Wiley-VCH, Weinheim
www.tcr.wiley-vch.de 951
