Article pubs.acs.org/Langmuir
Photochemical Grafting of Organic Alkenes to Single-Crystal TiO2 Surfaces: A Mechanistic Study Ryan Franking,† Heesuk Kim,†,§ Scott A. Chambers,‡ Andrew N. Mangham,‡,⊥ and Robert J. Hamers*,† †
Department of Chemistry, University of WisconsinMadison, 1101 University Avenue, Madison, Wisconsin 53706, United States Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352, United States
‡
S Supporting Information *
ABSTRACT: The UV-induced photochemical grafting of terminal alkenes has emerged as a versatile way to form molecular layers on semiconductor surfaces. Recent studies have shown that grafting reactions can be initiated by photoelectron emission into the reactant liquid as well as by excitation across the semiconductor band gap, but the relative importance of these two processes is expected to depend on the nature of the semiconductors, the reactant alkene and the excitation wavelength. Here we report a study of the wavelength-dependent photochemical grafting of alkenes onto single-crystal TiO2 samples. Trifluoroacetamide-protected 10-aminododec-1-ene (TFAAD), 10-N-BOC-aminodec-1ene (t-BOC), and 1-dodecene were used as model alkenes. On rutile (110), photons with energy above the band gap but below the expected work function are not effective at inducing grafting, while photons with energy sufficient to induce electronic transitions from the TiO2 Fermi level to electronic acceptor states of the reactant molecules induce grafting. A comparison of rutile (110), rutile (001), anatase (001), and anatase (101) samples shows slightly enhanced grafting for rutile but no difference between crystal faces for a given crystal phase. Hydroxylation of the surface increases the reaction rate by lowering the work function and thereby facilitating photoelectron ejection into the adjacent alkene. These results demonstrate that photoelectron emission is the dominant mechanism responsible for grafting when using short-wavelength (∼254 nm) light and suggest that photoemission events beginning on mid-gap states may play a crucial role.
■
INTRODUCTION The grafting of molecular layers to semiconductors has emerged as an important tool to control interfacial properties and reactivity by providing different types of chemical/ biochemical functional groups1,2 and the ability to tune the optical and electronic properties of the surface.2−4 The UVinduced photochemical grafting of terminal alkenes to semiconductors is of particular interest because it provides a facile, nondestructive method to covalently bond organic molecules to a wide range of semiconductors including Si,1,5−8 diamond,2,9,10 SiC,11 and more recently metal oxides such as SiO2,12 WO3,13 SnO2,14 ZnO2,15 and TiO2.15−17 TiO2 is of interest for its widespread use in solar energy,3 selective molecular catalysts,18 and biological materials.17,19 Photochemical grafting is an attractive way to link molecules to surfaces of TiO2 and other materials because it does not require high temperatures or ultrahigh vacuum, offers simple photopatterning procedures, and because the resulting layers can be very stable even in aqueous environments.2,16 Previous studies have shown the alkenes will graft to nanocrystalline TiO2 samples when illuminated with ultraviolet light at 254 nm,16,17,20 but a mechanistic understanding of how the grafting reactions are initiated has been lacking. Recent mechanistic studies8,10,21−23 have shown that grafting of alkenes onto the group IV semiconductors diamond and silicon can be initiated by two processes: (1) photoexcitation of bulk electron−hole pairs6−8 and (2) photoejection of electrons.21−23 © 2012 American Chemical Society
Both processes create valence-band holes that make the surface subject to nucleophilic attack to the alkenes, grafting the molecules to the surface.6,22,24,25 Photoejection grafting processes on silicon and diamond are facilitated by using bifunctional molecules that combine one functional group that can act as an electron acceptor with the alkene group that initiates nucleophilic attack on the positively charged surface sites.8,21−23,26 Because these two funcional groups impact differnt steps in the overall kinetic pathway, they do not necessarily need to be on the same molecule.8,26 The relative importance of exciton formation and of photoelectron emission may be expected to depend on the specific substrate and reactant molecules, as well as the wavelength used during illumination.7,8,21 In contrast to diamond and silicon, TiO2 is well-known as a photo-oxidation catalyst because of its deeper-lying valence band and overall wide band gap.27 However, the potential importance of exciton versus photoemission as an initiation mechanism has not been established on TiO2 or other metal oxides. Here, we report an investigation of the photochemical grafting of alkenes to singlecrystal TiO2 samples of both anatase and rutile. Our results show dramatically increased grafting yield at short wavelengths where photoejection of electrons from the TiO2 to the alkene Received: May 29, 2012 Revised: June 29, 2012 Published: July 2, 2012 12085
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
acceptor states is expected, and and there is very little or no significant grafting using light of longer wavelengths even at photon energies above the bulk band gap. These results demonstrate that photoelectron emission into acceptor states of the adjacent reactant liquids is the dominant process that initiates grafting of alkenes to TiO2 surfaces.
■
EXPERIMENTAL METHODS
Single-Crystal Sample Preparation. Rutile (001) and (110) single crystals were obtained from CrysTec GmbH. Before use, they were cleaned in fresh piranha solution consisting of a mixture of 98% concentrated H2SO4 and 30% hydrogen peroxide in a 3:1 volumetric ratio for at least 30 min and rinsed with 18 MΩcm deionized (DI) water (prepared using a Barnstead Nanopure system) to remove organic layers. Caution! Piranha solution is extremely dangerous, and proper safety precautions must be followed. The bare crystals were then etched in 48% HF acid for 15 min, rinsed with DI water, and dried. Caution! HF solution is extremely dangerous, and proper safety precautions must be followed. Finally, the crystals were annealed at 900 °C for 1 h inside a TiO2 box made from a 99.99% pure TiO2 sputter target (Kurt J. Lesker Company). In order to test the effects of surface hydroxyl groups, where noted, rutile (101) surfaces were soaked in 1 M NaOH after annealing for 10 min, rinsed with DI water, and dried. Single-crystal anatase (001) films were epitaxially grown by oxygen plasma assisted molecular beam epitaxy on SrTiO3(001) substrates at Pacific Northwest National Laboratory. Before use, the crystals were cleaned by immersing in deionized water while illuminated with ultraviolet light at 254 nm (∼10 mW/cm2) for at least 30 min to remove organic contaminates. Anatase (101) surfaces were obtained cut and polished from SurfaceNet GmbH, derived from natural samples. Before use, these surfaces were hand polished with 0.02 μm silica colloid for 3 min, etched in 48% HF for 1 min, and annealed at 300 °C for 3 h. For each of the crystal faces, these preparation methods led to atomically flat surfaces as observed by atomic force microscopy (AFM); representative AFM images are shown in the Supporting Information. Alkene Photochemical Grafting. Figure 1a shows the three alkenes investigated here. These alkenes, trifluoroacetamide-protected 10-aminododec-1-ene (TFAAD), 10-N-BOC-aminodec-1-ene (tBOC), and 1-dodecene, were stored under argon. Inside an argon glovebox, a few drops of the relevant alkene were placed on a freshly prepared sample. The sample was pressed against a UV-grade fusedsilica window and sealed in an airtight stainless steel reaction cell. The sample was then illuminated though the window for the reaction time specified using the relevant light source as described below. After UV reaction, the sample was sonicated in acetone and electronic grade methanol for 5 min each. Finally, the sample was blown dry under nitrogen. Typical reactions were performed using a low-pressure mercury quartz grid lamp with a primary wavelength of 254 nm and a total power output of 10 mW/cm2 measured at the sample location. In order to test the linearity of the reaction with illumination intensity, wire mesh screens were used to reduce the lamp intensity to 3.3 and 1.1 mW/cm2. The screens were placed sufficiently far from the sample to maintain uniform illumination across the sample. When possible, only samples reacted simultaneously are quantitatively compared. Wavelength-specific measurements at 254 and 350 nm were made using a combination of a 450 W high-pressure HgXe lamp and separate band-pass filters. The filters had a full width at half-maximum of 10 nm. The 254 nm filter passed 2.5 mW/cm2, and the 350 nm filter passed 10.5 mW/cm2. A water cell was used to eliminate IR radiation from the lamp. In order to test the effect of 185 nm illumination on the grafting of dodecene, a low-pressure mercury lamp with enhanced output at 185 nm (Fuller Ultraviolet, model 1020 high-ozone UV lamp) was used with a total power output of 7 mW/cm2. Characterization. X-ray photoelectron spectroscopy (XPS) was performed in an ultrahigh vacuum XPS system using an Al Kα source (nominal 1486.6 eV photon energy) with a quartz-crystal monochromator and a hemispherical analyzer, typically using an analyzer
Figure 1. (a) Structures of the three molecules used in studies reported here. (b) High-resolution XPS data from rutile (110) surfaces grafted with TFAAD and t-BOC using 254 and 350 nm light. A spectrum of TFAAD reacted in the dark is also shown for comparison. The spectra were normalized to the O(1s) peak height for ease of comparison. (c) Surface N coverage calculated from the N(1s) data.
resolution of 0.1−0.2 eV. The electron takeoff angle was 45°, and peak positions and areas were calculated by fitting raw data to Voigt functions after a Shirley background correction.28 A low-energy electron gun was used to neutralize sample charging as needed, and spectra were shifted to place the lowest C(1s) peak at 284.8 eV.29 Calculated areas were normalized by the corresponding atomic sensitivity factors for the X-rays incident at 90° from the analyzer (C(1s) = 0.296; O(1s) = 0.711; Cl(2p) = 0.77; Ti(2p) = 1.798).30 Molecular coverage was calculated from the peak areas taking into account electron attenuation through the organic film, and error bars are determined from estimated measurement uncertainties and errors in peak fitting as described previously.16 12086
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
■
Article
RESULTS Wavelength Dependence of TFAAD and t-BOC Grafting. We first compared grafting of two molecules, TFAAD and t-BOC, on rutile (110) surfaces using light with wavelengths of 254 nm (4.88 eV) and 350 nm (3.54 eV). Both wavelengths correspond to photon energies above the 3.0 eV band gap of rutile.31 Previous studies have shown that the trifluoroacetamide group and the tert-butoxycarbonyl groups can act as electron acceptors. Figure 1b shows high-resolution XPS of the F(1s), O(1s), Ti(2p), N(1s), and C(1s) peaks for each of the samples after reaction as well as a dark control sample that was exposed to TFAAD but not illuminated with UV light. Each spectrum was normalized to give the same O(1s) peak height since the O(1s) signal arises almost entirely from the bulk TiO2. Figure 1b shows XPS spectra of a rutile (110) surface after reaction with TFAAD under 254 nm light (2.5 mW/cm2) for 18 h. The C(1s) spectra show a large peak shifted to 284.8 eV as well as a shoulder at 286.5 eV and two small satellite peaks at 289 and 293.1 eV. These peaks are consistent with previous reports for TFAAD grafted onto silicon, diamond, and amorphous carbon.1,5,10,23 We attribute the 284.8 eV peak to alkyl carbon from the alkyl chain of TFAAD, the 286.5 eV peak to the amide carbon bonded solely to nitrogen, the 289 eV peak to the amide carbon bound to nitrogen and oxygen, and the 293.1 eV peak to the CF3 group.9 The N(1s) and F(1s) regions show peaks at 400.8 and 689.7 eV, respectively, consistent with the presence N and F atoms from TFAAD and the C(1s) peak assignments. We conclude that TFAAD successfully grafts to the surface under 254 nm illumination. Figure 1b also shows XPS spectra of a rutile (110) sample after grafting with TFAAD using 350 nm light (10.5 mW/cm2) for 18 h. The C(1s) region shows the main peak shifted to 284.8 eV in addition to a shoulder at 265.5 eV and a satellite peak at 289.1 eV. In contrast to the sample illuminated with 254 nm light, the sample illuminated with 350 nm light shows little to no peak signal at 293.1 eV. The nitrogen region shows a small peak at 400 eV, and the fluorine region shows a peak at 689.2 eV. A small F(1s) peak is present on the dark control sample that may arise from a small amount of F remaining on the surface from the HF treatment or physisorption of a small amount of TFAAD. XPS data obtained of the surface before any exposure to TFAAD are shown in the Supporting Information. The C(1s), N(1s), and F(1s) peaks are significantly stronger on the sample exposed to 254 nm light compared with those grafted using 350 nm light. Since both TFAAD and t-BOC have one nitrogen atom per molecule and there is no detectable background contamination from N, we use the nitrogen coverage to compare the molecular coverage for each of the molecules. Figure 1c shows the coverage calculated from the N(1s) peak area. The total coverage for the 254 nm sample is 6.7 ± 1.5 N atoms/nm2. The 350 nm sample yields 1.5 ± 0.3 N atoms/nm2. For comparison, a dark control that was exposed to TFAAD in the dark under otherwise identical conditions shows 0.6 ± 0.2 N atoms/nm2. Since the 350 nm light is approximately four times more intense (∼5.8 times as many photons/s) than the 254 nm light, we conclude that 254 nm illumination leads to grafting that is ∼39 times more efficient per photon than that obtained with 350 nm illumination. In order to increase the generality of our observations with TFAAD, we repeated the 254 and 350 nm reactions with a
second molecule, t-BOC. XPS spectra for t-BOC are also shown in Figure 1b. With 254 nm illumination, the C(1s) region shows the main peak shifted to 284.8 eV with a shoulder at 286.8 eV and a satellite at 289.4 eV. The results are consistent with those observed previously for photochemical grafting of t-BOC on other surfaces.23 We attribute the 284.8 eV peak to alkyl carbon from the alkyl chain of t-BOC, the 286.8 eV peak to the amide carbon bound solely to the nitrogen, and the 289.4 eV to the amide carbon bound to nitrogen and oxygen. The N(1s) region shows a peak at 400.2 eV consistent with the nitrogen from the t-BOC molecule. We conclude that, similar to TFAAD, t-BOC successfully grafts to the rutile surface at 254 nm. The XPS spectra of the sample grafted with t-BOC again show significantly more efficient grafting when using light at 254 nm compared with that at 350 nm. Figure 1b compares the coverage calculated for both samples from the N(1s) peak area. Illumination with 254 nm light yields 7.7 ± 1.5 N atoms/nm2, while 350 nm illumination yields only 1.9 ± 0.4 N atoms/nm2. We conclude that, as with TFAAD, 254 nm illumination leads to a much more efficient grafting reaction than 350 nm illumination. Wavelength Dependence of 1-Dodecene Grafting. To further investigate the effect of wavelength and surface preparation on grafting efficiency, we investigated the grafting of 1-dodecene to the rutile (110) surface using a 254 nm lowpressure Hg lamp (10 mW/cm2) and a 185 nm (6.7 eV) enhanced Hg lamp (7 mW/cm2). We also examined altering the reactivity of the rutile surface at 254 nm by first hydroxylating it with a 1 M NaOH solution for 10 min. Figure 2a shows XPS spectra of the resulting samples along with a dodecene dark control. Each of the spectra for the grafted and control samples shows a peak shifted to 284.8 eV in the C (1s) region, which we attribute to alkyl carbon. The coverage calculated for each of the samples is compared in Figure 2b. The sample exposed using 254 nm light yielded 18 ± 3 C atoms/nm2 (1.5 molecules/nm2), while the dark control showed 19 ± 3 atoms/nm2 (1.6 molecules/nm2). Since there is no statistical difference between the dark control and 254 nm sample, we conclude that 1-dodecene does not graft in significant yield to the acid-treated sample using 254 nm light. However, 1-dodecene can be induced by graft using shorter-wavelength light, as the sample illuminated at 185 nm shows a C coverage of 29 ± 5 C atoms/nm2 (2.4 molecules/ nm2). In addition, 254 nm light-induced grafting of 1-dodecene to the base-treated sample, which yielded 34 ± 6 C atoms/nm2 (2.8 molecules/nm2). We conclude that while there is little to no reaction of 1-dodecene at 254 nm with an as-prepared rutile surface, reaction can be initiated at 185 nm or by first hydroxylating the rutile surface with a basic solution. Intensity Dependence of Grafting. In order to help compare reactions performed at different illumination intensities and rule out nonlinear effects, we grafted TFAAD to the rutile (110) surface at three different illumination intensities, adjusting the time in order to achieve the same fluence (total integrated number of incident photons). The grafting was preformed with the 254 nm lamp for 2 h at 10 mW/cm2, 6 h at 3.3 mW/cm2, and 18 h at 1.1 mW/cm2. Using the nitrogen peak to determine TFAAD coverage, Figure 3 shows the coverage calculated from XPS for each of the samples: 10 mW/cm2 yields 5.6 ± 0.8 C atoms/nm2, 3.3 mW/ cm2 yields 4.7 ± 0.8 C atoms/nm2, and 1.1 mW/cm2 yields 6.7 ± 1.1 C atoms/nm2 (4.7, 3.9, and 5.6 molecules/nm2, 12087
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
scales (data provided in the Supporting Information), we conclude that under the conditions of our experiments the molecular coverage depends only on the total fluence. Surface and Crystal Structure Dependence of the Grafting. In order to examine how the grafting reaction depends on the crystallographic structure of the underlying substrate, we grafted TFAAD to rutile (110), rutile (001), anatase (101), and anatase (001) under a fixed set of conditions (254 nm Hg lamp, 4 h). For bulk samples, the rutile structure is the lower-energy polymorph and is generally less reactive than anatase.32,33 For anatase, the (101) surface is more stable, but the (001) surface is generally accepted to be the most active catalytically and photochemically,34 with the (101) surface containing a greater concentration of surface hydroxyl groups.35 The (110) surface is the most stable surface of rutile.34 Figure 4a shows XPS spectra of the four samples after grafting. Each of the four crystal faces show a main carbon peak shifted to 284.8 eV, with a shoulder at 286.5 eV and satellite peaks at ∼289 and ∼293 eV. The nitrogen and fluorine regions show peaks at 401 and 689.9 eV, respectively. All of these peaks are consistent with TFAAD present on the surface, and we conclude that TFAAD successfully grafts to all four crystal faces. Quantitatively, we observe differences between the rutile and anatase samples. Figure 4b compares the calculated coverage from the N(1s) peak for each of the grafted crystal faces: rutile (110) yields 6.7 ± 1.2 N atoms/nm2; rutile (001) yields 6.5 ± 1.2 N atoms/nm2; anatase (101) yields 4.4 ± 0.8 N atoms/ nm2; and anatase (001) yields 3.6 ± 0.7 N atoms/nm2. We consistently find that rutile shows slightly higher reactivity toward photochemical grafting than anatase; however, there are no significant differences in reactivity between different crystal faces of a given bulk crystal phase.
■
Figure 2. (a) High-resolution XPS data for 1-dodecene grafting to single-crystal rutile (110) surfaces at 254 and 185 nm and to a hydroxylated rutile (110) surface at 254 nm. A dark control sample is also shown. The spectra were normalized to the O(1s) peak height for ease of comparison. (b) Carbon coverage calculated from the C(1s) peak areas of each of the spectra.
DISCUSSION Our data show several key results. First, while both TFAAD and t-BOC graft to the rutile surface at 254 nm (4.89 eV), both alkenes are unreactive when using longer-wavelength light at 350 nm (3.55 eV) that is still well above the band gap of both anatase (3.2 eV) and rutile (3.0 eV).31 On the basis of the fact that the 350 nm light yielded a significantly lower coverage while having a higher flux (intensity), we conclude that the grafting efficiency, defined as the number of molecules grafted per incident photon, is approximately 39 times higher when using 254 nm light compared with 350 nm light. Second, in contrast to TFAAD and t-BOC, 1-dodecene shows little to no grafting at 254 nm but can be induced to graft using shorter wavelengths of 185 nm (6.71 eV) (or, as discussed below, by intentional hydroxylation of the surface). Finally, while little difference in grafting efficiency is observed between specific crystal faces, both crystal faces of the rutile phase show a higher grafting efficiency than the anatase faces. Excitons versus Electron Ejection. Previous work on the photochemical grafting of alkenes onto surfaces such as amorphous carbon,23,26 diamond,21,22 and silicon21 has shown the importance of a photoelectron ejection process during the grafting. On amorphous carbon, the alkene reactivity was shown to correlate to the electron acceptor levels of the alkenes.23 On diamond, this observation was expanded to correlate the grafting rate to the photoelectron ejection yield.10,22 These results paint a picture where the electron must be ejected from the semiconductor surface for efficient grafting to occur. Though the actual grafting reaction occurs by reaction of the alkene bond with surface holes, the ejection of
Figure 3. Nitrogen coverages calculated from the N(1s) peak of XPS spectra from a single-crystal rutile (110) surface reacted with TFAAD at different illumination intensities and constant fluence.
respectively). Since these coverage values are similar and do not show a trend of increasing grafting with increasing intensity and since TFAAD grafting does not self-terminate within these time 12088
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
where t is the length of time of illumination and P is the probability that an incident photon induces a photoemission event. Note that in this case the concentration of holes increases linearly with time in the absence of a surface oxidation process (such as reaction with an alkene) that can restore overall neutrality to the bulk. A previous study found that illumination of TiO2 thin films on metallic substrates with UV light from a deuterium lamp yielded photoemission with an overall probability of 10−3 to 10−2.37 Assuming a typical hole lifetime of ∼10 ns36 and even using a conservative estimate of P ∼ 10−6 shows that within ∼10 ms the concentration of excess surface holes produced by irreversible photoemission exceeds the steady-state concentration created by band gap excitation. Thus, provided that the wavelength of the excitation source exceeds the energy required to remove an electron into the adjacent medium (in our case, the reactant alkenes), photoemission becomes a more favorable process for creating surface holes. Our results are qualitatively similar to those reported previously on grafting of molecules to p-type diamond.22 However, an important distinction between TiO2 and prior results on group IV semiconductors is that the valence band of TiO2 is much lower in energy, such that electron cannot be ejected directly from the valence band. The absolute energy of diamond’s Fermi level is ∼4.6 eV below vacuum,22 close to that of n-type TiO2. In a previous study on wavelength-dependent photoemission from diamond surfaces into organic liquids, photoemission was observed into TFAAD with photon energies shorter than ∼4.1 eV and into tBOC with energies above ∼4.6 V, while no photoemission (and no grafting) was observed into dodecene.22 On the basis of these observations, the absolute acceptor levels of TFAAD, t-BOC, and dodecene was estimated,22 leading to an overall energy level picture illustrated in Figure 5. The ordering of the electron acceptor levels depicted here is also in qualitative agreement with earlier density functional calculations.23 To establish the absolute energies of the TiO2 samples, the Fermi level position of the TiO2 samples must be known. Work function measurements for rutile (110) surfaces vary from 5.2 to 4.0 eV38−40 depending on measurement method and sample preparation. Higher work functions are typically found on vacuum-annealed surfaces, while those that have surface −OH groups have significantly lower values.41,42 Gutmann and co-workers41 showed that UV illumination can permanently lower the work function of rutile from ∼4.8 to ∼4.0 eV and that of anatase from ∼5.2 to 4.4 eV. In both cases, the lowering of the work function was attributed to UV-induced dissociation of water on the surfaces. Measurements on single crystals have typically shown that rutile (110) and (100) surfaces have similar work function,40 while under identical conditions, work functions on anatase surfaces are typically ∼0.2−0.5 eV larger than those of rutile.41,43,44 Both anatase and rutile are n-type materials, with the Fermi typically about 0.2−0.3 eV away from the conduction band minimum due to the presence of surface defects.44−46 Using the above values for the alkene acceptor levels and the approximate values for TiO2, an energy level diagram can be established as shown in Figure 5. Here we have assumed a TiO2 work function of 4.5 eV, the approximate mid-range of values reported previously. Our data suggest that, in the case of TiO2, the electron emission originates from surface states lying within the band gap below the Fermi energy. Previous studies of photoemission from both anatase and rutile TiO2 have noted substantial photoemission from mid-gap surface states,47−49 and mid-gap
Figure 4. (a) High-resolution XPS spectra of TFAAD grafted at 254 nm to single-crystal rutile (110), rutile (001), anatase (101), and anatase (001) surfaces. The data were normalized to the O(1s) peak height for ease of comparison. (b) Nitrogen coverage calculated from the N(1s) peak area of each of the spectra.
electrons from the diamond surface plays a key role by efficiently produce surface holes. The importance of photoelectron ejection to the alkene can be understood on the basis of rate theory. Under steady-state conditions and in the absence of surface recombination, the rate at which electron−hole pairs are created per unit area of surface is simply given by I/Ephoton, where I is the irradiance (W/cm2) and Ephoton is the energy per photon. For a carrier lifetime τ (previously determined to be ∼10 ns),36 this yields a surface density of [h]surface = Iτ/Ephoton. In this equation, [h]surface represents the surface density (number per unit area) of holes that would be present under steady-state illumination if all of the holes formed through the depth of the sample were concentrated at the surface. For a photoemission process, the number of holes per unit area created at the surface is given by [h]surface = IPt/Ephoton, 12089
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
likely a result of the reduced work function of hydroxylated surfaces rather than an increase in the density of reactive surface sites. Crystal Face Dependence. The different crystal phases and surface orientations of TiO2 are well-known to have different reactivity toward water and organic adsorption.34,60−62 Since previous work indicated that the grafting reaction occurs between the alkene and surface hydroxyl groups,16 the reactivity of the alkenes may have been expected to correlate to the ability of each surface to form surface hydroxyls from water. For example, less water dissociation has been experimentally observed at rutile (110) surfaces than rutile (001) surfaces,34 and therefore, we might expect greater grafting to occur at the rutile (001) surface. Our results, however, show little difference in alkene grafting between rutile (110) and rutile (001) and instead indicate a greater difference between the two different bulk crystal phases. Studies of the work function of rutile surfaces have shown little dependence on the crystal face; a change of only 0.09 eV was observed between rutile (110) and rutile (100).39,40 In contrast, differences between the anatase and rutile phases are much larger, with experiments showing that rutile has a work function 0.2−0.5 eV smaller than anatase.41,43,44 Rutile also has a smaller band gap of 3.0 eV compared with 3.2 eV for anatase.31,63,64 The smaller work function of rutile is also accompanied by a higher density of states near the valence-band edge,41 which should further enhance the photoemission yield from rutile compared with anatase. On the basis of the data above, we conclude that the higher yield of photochemical grafting on rutile compared with anatase is a direct result of the lower work function of rutile, which places its Fermi level ∼0.2−0.5 eV higher than that of anatase. This facilitates the photoexcitation of electrons from surface states to acceptor levels present on the molecules in solution, creating the holes that are necessary for subsequent reaction. The smaller band gap of rutile may enhance bulk absorption, but as discussed earlier, the photoemission process is ultimately more effective at inducing grafting, such that the bulk absorption is expected to have little effect on the grafting. This is further supported by the fact that the anatase samples are derived from natural samples with large amounts of impurities (giving the samples a deep purple color). These impurities are expected to shorten the lifetime of photoexcited electron−hole pairs. Thus, if bulk absorption was important, then a much larger difference would be expected between the natural anatase samples and the highly pure, synthetic rutile samples.21
Figure 5. Energy level diagram versus vacuum, illustrating the relationship between the TiO2 work function, incident photon energies, and the liquid alkene acceptor levels. Values shown here are for rutile, assuming a work function of 4.5 eV. Band gap excitation (not shown) can also occur for all of the incident photon energies.
states are known to play a crucial role in charge transport in TiO2 nanoparticle films.49−53 The diagram in Figure 5 is consistent with our results showing that while 350 nm (3.55 eV) illumination is above the band gap of both rutile (3.0 eV) and anatase (3.2 eV) rutile and therefore sufficient for creating excitons, it cannot easily eject electrons from states near the Fermi level into the alkene acceptor levels, resulting in only low concentrations of holes and a low grafting rate. In contrast, illumination at 254 nm (4.88 eV) is above the work function of the rutile and likely capable of ejecting electrons directly onto the TFAAD and t-BOC acceptor levels but is yet ineffective in grafting of 1-dodecene. However, even shorter wavelengths of 185 nm (6.70 eV) can also induce grafting of 1-dodecene. Effect of Surface Hydroxylation. Our experiments demonstrate that while 1-dodecene exhibited little or no grafting on a surface treated with piranha solution, grafting of 1dodecene could be initiated by light at 254 nm if the surface was hydroxylated with a basic solution. Previous studies of TiO2 have shown that surface oxygen and −OH groups act as traps for holes.54−56 Our previous studies involving hydrolysis experiments and detailed FTIR of the grafted molecules proposed that grafting occurs via nucleophilic attack by the organic alkenes to surface −OH groups, linking the alkenes to the surface via Ti−O−C bonds.16 Thus, the density of surface −OH groups might be expected to influence the reactivity. In addition, previous studies have shown that the work function of rutile TiO2 decreases from ∼4.9 to 4.0 eV as the concentration of surface hydroxyl groups increases.57−59 As depicted in Figure 5, a decrease in work function to 4.0 V (from the 4.5 assumed in Figure 5) would be sufficient to allow 254 nm photons to emit electrons into 1-dodecene, in agreement with our experimental observations. Therefore, we conclude that the increased grafting of 1-dodecene on hydroxylated surfaces is
■
CONCLUSIONS With this work, we have examined the photon wavelength and surface dependence of UV photochemical grating of 1-alkenes onto TiO2. The results show the grafting reaction rate is highly dependent on the relationship between the incident photon energy and both the work function of the material and the alkene acceptor level. When the photon energy is sufficient to allow excitation from states at or below the TiO2 Fermi level to an acceptor level of the adjacent alkene, the rate of the grafting is greatly enhanced. For alkenes with high acceptor levels such as 1-dodecene, grafting efficiency can be improved by either decreasing the photon wavelength or decreasing the substrate work function. These results demonstrate a new pathway for controlling and understanding alkene UV grafting efficiency by tuning the substrate work function in relation to the incident 12090
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
(8) Huck, L. A.; Buriak, J. M. Toward a Mechanistic Understanding of Exciton-Mediated Hydrosilylation on Nanocrystalline Silicon. J. Am. Chem. Soc. 2012, 134, 489−497. (9) Strother, T.; Knickerbocker, T.; Russell, J. N.; Butler, J. E.; Smith, L. M.; Hamers, R. J. Photochemical Functionalization of Diamond Films. Langmuir 2002, 18, 968−971. (10) Nichols, B. M.; Butler, J. E.; Russell, J. N.; Hamers, R. J. Photochemical Functionalization of Hydrogen-Terminated Diamond Surfaces: A Structural and Mechanistic Study. J. Phys. Chem. B 2005, 109, 20938−20947. (11) Rosso, M.; Giesbers, M.; Arafat, A.; Schroen, K.; Zuilhof, H. Covalently Attached Organic Monolayers on SiC and SixN4 Surfaces: Formation Using UV Light at Room Temperature. Langmuir 2009, 25, 2172−2180. (12) ter Maat, J.; Regeling, R.; Yang, M. L.; Mullings, M. N.; Bent, S. F.; Zuilhof, H. Photochemical Covalent Attachment of Alkene-Derived Monolayers onto Hydroxyl-Terminated Silica. Langmuir 2009, 25, 11592−11597. (13) Cardiel, A. C.; Benson, M. C.; Bishop, L. M.; Louis, K. M.; Yeager, J. C.; Tan, Y. Z.; Hamers, R. J. Chemically Directed Assembly of Photoactive Metal Oxide Nanoparticle Heterojunctions via the Copper-Catalyzed Azide-Alkyne Cycloaddition “Click” Reaction. ACS Nano 2012, 6, 310−318. (14) Shah, S.; Benson, M. C.; Bishop, L. M.; Huhn, A. M.; Ruther, R. E.; Yeager, J. C.; Louis, K. M.; Hamers, R. J. Chemically Assembled Heterojunctions of SnO2 Nanorods with TiO2 Nanoparticles via “Click” Chemistry. J. Mater. Chem. 2012, 22, 11561−11567. (15) Hamers, R. J.; Chambers, S. A.; Evans, P. E.; Franking, R.; Gerbec, Z.; Gopalan, P.; Kim, H.; Landis, E. C.; Li, B.; McCoy, M. W.; Ohsawa, T.; Ruther, R. E. Molecular and Biomolecular Interfaces to Metal Oxide Semiconductors. Phys. Status Solidi C 2010, 7, 200−205. (16) Franking, R. A.; Landis, E. C.; Hamers, R. J. Highly Stable Molecular Layers on Nanocrystalline Anatase TiO2 through Photochemical Grafting. Langmuir 2009, 25, 10676−10684. (17) Li, B.; Franking, R.; Landis, E. C.; Kim, H.; Hamers, R. J. Photochemical Grafting and Patterning of Biomolecular Layers onto TiO2 Thin Films. ACS Appl. Mater. Interfaces 2009, 1, 1013−1022. (18) Ebitani, K.; Fujie, Y.; Kaneda, K. Immobilization of a LigandPreserved Giant Palladium Cluster on a Metal Oxide Surface and Its Nobel Heterogeneous Catalysis for Oxidation of Allylic Alcohols in the Presence of Molecular Oxygen. Langmuir 1999, 15, 3557−3562. (19) Topoglidis, E.; Cass, A. E. G.; Gilardi, G.; Sadeghi, S.; Beaumont, N.; Durrant, J. R. Protein Adsorption on Nanocrystalline TiO2 Films: An Immobilization Strategy for Bioanalytical Devices. Anal. Chem. 1998, 70, 5111−5113. (20) Franking, R.; Hamers, R. J. Ultraviolet-Induced Grafting of Alkenes to TiO2 Surfaces: Controlling Multilayer Formation. J. Phys. Chem. C 2011, 115, 17102−17110. (21) Wang, X. Y.; Ruther, R. E.; Streifer, J. A.; Hamers, R. J. UVInduced Grafting of Alkenes to Silicon Surfaces: Photoemission versus Excitons. J. Am. Chem. Soc. 2010, 132, 4048−4050. (22) Wang, X. Y.; Colavita, P. E.; Streifer, J. A.; Butler, J. E.; Hamers, R. J. Photochemical Grafting of Alkenes onto Carbon Surfaces: Identifying the Roles of Electrons and Holes. J. Phys. Chem. C 2010, 114, 4067−4074. (23) Colavita, P. E.; Sun, B.; Tse, K. Y.; Hamers, R. J. Photochemical Grafting of N-Alkenes onto Carbon Surfaces: The Role of Photoelectron Ejection. J. Am. Chem. Soc. 2007, 129, 13554−13565. (24) Wang, X.; Colavita, P. E.; Metz, K. M.; Butler, J. E.; Hamers, R. J. Direct Photopatterning and SEM Imaging of Molecular Monolayers on Diamond Surfaces: Mechanistic Insights into UV-initiated Molecular Grafting. Langmuir 2007, 23, 11623−11630. (25) Wang, X. Y.; Landis, E. C.; Franking, R.; Hamers, R. J. Surface Chemistry for Stable and Smart Molecular and Biomolecular Interfaces via Photochemical Grafting of Alkenes. Acc. Chem. Res. 2010, 43, 1205−1215. (26) Colavita, P. E.; Streifer, J. A.; Sun, B.; Wang, X. Y.; Warf, P.; Hamers, R. J. Enhancement of Photochemical Grafting of Terminal
photon wavelength. Even though TiO2 is a well-known photooxidizer when illuminated with above-band-gap light,27,65 our results show that the grafting of organic alkenes requires shortwavelength light able to photoexcite electrons from mid-gap surface states or possibly from electrons in the conduction band into acceptor levels of the adjacent reactant molecules. The irreversible electron emission leaves the sample positively charged and enhances the subsequent reaction with the electron-rich vinyl group of the organic alkenes.
■
ASSOCIATED CONTENT
S Supporting Information *
AFM images of samples after initial treatment with piranha solution or HF solution. XPS spectra of the clean rutile (101) surface immediately after HF treatment and annealing. TFAAD kinetic curve at 254 nm on rutile (001). This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Present Addresses §
Korea Institute of Science and Technology, Seoul, Korea. Cree, Inc., Raleigh, NC.
⊥
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation Grant CHE0911543. The growth of anatase (001) epitaxial films at PNNL was supported by the U.S. Department of Energy, Office of Science, Division of Chemical Sciences under Award #399998 and #48526, and was performed in the Environmental Molecular Sciences Laboratory, a national science user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory.
■
REFERENCES
(1) Strother, T.; Cai, W.; Zhao, X. S.; Hamers, R. J.; Smith, L. M. Synthesis and Characterization of DNA-Modified Silicon (111) Surfaces. J. Am. Chem. Soc. 2000, 122, 1205−1209. (2) Yang, W. S.; Auciello, O.; Butler, J. E.; Cai, W.; Carlisle, J. A.; Gerbi, J.; Gruen, D. M.; Knickerbocker, T.; Lasseter, T. L.; Russell, J. N.; Smith, L. M.; Hamers, R. J. DNA-Modified Nanocrystalline Diamond Thin-Films as Stable, Biologically Active Substrates. Nat. Mater. 2002, 1, 253−257. (3) O’Regan, B.; Grätzel, M. A Low-Cost, High-Efficiency Solar-Cell Based on Dye-Sensitized Colloidal TiO2 Films. Nature 1991, 353, 737−740. (4) Linford, M. R.; Fenter, P.; Eisenberger, P. M.; Chidsey, C. E. D. Alkyl Monolayers on Silicon Prepared from 1-Alkenes and HydrogenTerminated Silicon. J. Am. Chem. Soc. 1995, 117, 3145−3155. (5) Cicero, R. L.; Linford, M. R.; Chidsey, C. E. D. Photoreactivity of Unsaturated Compounds with Hydrogen-Terminated Silicon(111). Langmuir 2000, 16, 5688−5695. (6) Stewart, M. P.; Buriak, J. M. Exciton-Mediated Hydrosilylation on Photoluminescent Nanocrystalline Silicon. J. Am. Chem. Soc. 2001, 123, 7821−7830. (7) Sun, Q. Y.; de Smet, L.; van Lagen, B.; Giesbers, M.; Thune, P. C.; van Engelenburg, J.; de Wolf, F. A.; Zuilhof, H.; Sudholter, E. J. R. Covalently Attached Monolayers on Crystalline Hydrogen-Terminated Silicon: Extremely Mild Attachment by Visible Light. J. Am. Chem. Soc. 2005, 127, 2514−2523. 12091
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
TiO2, and SrO and the Associated Valence Band Offsets with Si(001). J. Vac. Sci. Technol., B 2004, 22, 2205−2215. (47) Borodin, A.; Reichling, M. Characterizing TiO2(110) Surface States by Their Work Function. Phys. Chem. Chem. Phys. 2011, 13, 15442−15447. (48) Sanjines, R.; Tang, H.; Berger, H.; Gozzo, F.; Margaritondo, G.; Levy, F. Electronic-Structure of Anatase TiO2 Oxide. J. Appl. Phys. 1994, 75, 2945−2951. (49) Wendt, S.; Sprunger, P. T.; Lira, E.; Madsen, G. K. H.; Li, Z. S.; Hansen, J. O.; Matthiesen, J.; Blekinge-Rasmussen, A.; Laegsgaard, E.; Hammer, B.; Besenbacher, F. The Role of Interstitial Sites in the Ti3d Defect State in the Band Gap of Titania. Science 2008, 320, 1755− 1759. (50) Bahnemann, D. W.; Hilgendorff, M.; Memming, R. Charge Carrier Dynamics at TiO2 Particles: Reactivity of Free and Trapped Holes. J. Phys. Chem. B 1997, 101, 4265−4275. (51) Huber, R.; Sporlein, S.; Moser, J. E.; Gratzel, M.; Wachtveitl, J. The Role of Surface States in the Ultrafast Photoinduced Electron Transfer from Sensitizing Dye Molecules to Semiconductor Colloids. J. Phys. Chem. B 2000, 104, 8995−9003. (52) Zhao, H.; Zhang, Q. L.; Weng, Y. X. Deep Surface Trap Filling by Photoinduced Carriers and Interparticle Electron Transport Observed in TiO2 Nanocrystalline Film with Time-Resolved Visible and Mid-IR Transient Spectroscopies. J. Phys. Chem. C 2007, 111, 3762−3769. (53) Fabregat-Santiago, F.; Mora-Sero, I.; Garcia-Belmonte, G.; Bisquert, J. Cyclic Voltammetry Studies of Nanoporous Semiconductors. Capacitive and Reactive Properties of Nanocrystalline TiO2 Electrodes in Aqueous Electrolyte. J. Phys. Chem. B 2003, 107, 758−768. (54) Lawless, D.; Serpone, N.; Meisel, D. Role of OH Radicals and Trapped Holes in PhotocatalysisA Pulse-Radiolysis Study. J. Phys. Chem. 1991, 95, 5166−5170. (55) Micic, O. I.; Zhang, Y. N.; Cromack, K. R.; Trifunac, A. D.; Thurnauer, M. C. Trapped Holes on TiO2 Colloids Studied by Electron-Paramagnetic-Resonance. J. Phys. Chem. 1993, 97, 7277− 7283. (56) Serpone, N.; Lawless, D.; Khairutdinov, R.; Pelizzetti, E. Subnanosecond Relaxation Dynamics in TiO2 Colloidal Sols (Particle Sizes Rp=1.0−13.4 nm)Relevance to Heterogeneous Photocatalysis. J. Phys. Chem. 1995, 99, 16655−16661. (57) Hugenschmidt, M. B.; Gamble, L.; Campbell, C. T. The Interaction of H2O with a TiO2(110) Surface. Surf. Sci. 1994, 302, 329−340. (58) Silva, A. G.; Bundaleski, N.; Moutinho, A. M. C.; Teodoro, O. Dynamics of Water Adsorption on TiO2 Monitored by Work Function Spectroscopy. Appl. Surf. Sci. 2012, 258, 2006−2009. (59) Marques, H. P.; Canario, A. R.; Moutinho, A. M. C.; Teodoro, O. Tracking Hydroxyl Adsorption on TiO2 (110) through Secondary Emission Changes. Appl. Surf. Sci. 2009, 255, 7389−7393. (60) Bullard, J. W.; Cima, M. J. Orientation Dependence of the Isoelectric Point of TiO2 (Rutile) Surfaces. Langmuir 2006, 22, 10264−10271. (61) Henderson, M. A. The Influence of Oxide Surface-Structure on Adsorbate ChemistryDesorption of Water from the Smooth, the Microfaceted and the Ion-Sputtered Surfaces of TiO2(100). Surf. Sci. 1994, 319, 315−328. (62) Ohno, T.; Sarukawa, K.; Matsumura, M. Crystal Faces of Rutile and Anatase TiO2 Particles and Their Roles in Photocatalytic Reactions. New J. Chem. 2002, 26, 1167−1170. (63) Serpone, N.; Lawless, D.; Khairutdinov, R. Size Effects on the Photophysical Properties of Colloidal Anatase TiO2 ParticlesSize Quantization or Direct Transitions in This Indirect Semiconductor. J. Phys. Chem. 1995, 99, 16646−16654. (64) Reddy, K. M.; Reddy, C. V. G.; Manorama, S. V. Preparation, Characterization, and Spectral Studies on Nanocrystalline Anatase TiO2. J. Solid State Chem. 2001, 158, 180−186.
Alkenes at Surfaces via Molecular Mediators: The Role of SurfaceBound Electron Acceptors. J. Phys. Chem. C 2008, 112, 5102−5112. (27) Linsebigler, A. L.; Lu, G. Q.; Yates, J. T. Photocatalysis on TiO2 SurfacesPrinciples, Mechanisms, and Selected Results. Chem. Rev. 1995, 95, 735−758. (28) Shirley, D. A. High-Resolution X-ray Photoemission Spectra of Valence Bands of Gold. Phys. Rev. B 1972, 5, 4709. (29) Swift, P. Adventitious CarbonThe Panacea for Energy Referencing. Surf. Interface Anal. 1982, 4, 47−51. (30) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook of X-ray Photoelectron Spectroscopy; Perkin-Elmer Corporation: Eden Prairie, MN, 1992. (31) Tang, H.; Prasad, K.; Sanjines, R.; Schmid, P. E.; Levy, F. Electrical and Optical-Properties of TiO2 Anatase Thin-Films. J. Appl. Phys. 1994, 75, 2042−2047. (32) Ding, Z.; Lu, G. Q.; Greenfield, P. F. Role of the Crystallite Phase of TiO2 in Heterogeneous Photocatalysis for Phenol Oxidation in Water. J. Phys. Chem. B 2000, 104, 4815−4820. (33) Sclafani, A.; Herrmann, J. M. Comparison of the Photoelectronic and Photocatalytic Activities of Various Anatase and Rutile Forms of Titania in Pure Liquid Organic Phases and in Aqueous Solutions. J. Phys. Chem. 1996, 100, 13655−13661. (34) Diebold, U. The Surface Science of Titanium Dioxide. Surf. Sci. Rep. 2003, 48, 53−229. (35) Arrouvel, C.; Digne, M.; Breysse, M.; Toulhoat, H.; Raybaud, P. Effects of Morphology on Surface Hydroxyl Concentration: A DFT Comparison of Anatase-TiO2 and Gamma-Alumina Catalytic Supports. J. Catal. 2004, 222, 152−166. (36) Colbeau-Justin, C.; Kunst, M.; Huguenin, D. Structural Influence on Charge-Carrier Lifetimes in TiO2 Powders Studied by Microwave Absorption. J. Mater. Sci. 2003, 38, 2429−2437. (37) Laan, M.; Aarik, J.; Josephson, R.; Repan, V. Low Current Mode of Negative Coronas: Mechanism of Electron Emission. J. Phys. D.: Appl. Phys. 2003, 36, 2667−2672. (38) Batzill, M.; Katsiev, K.; Gaspar, D. J.; Diebold, U. Variations of the Local Electronic Surface Properties of TiO2(110) Induced by Intrinsic and Extrinsic Defects. Phys. Rev. B 2002, 66, 235401. (39) Nakamura, R.; Okamura, T.; Ohashi, N.; Imanishi, A.; Nakato, Y. Molecular Mechanisms of Photoinduced Oxygen Evolution, PL Emission, and Surface Roughening at Atomically Smooth (110) and (100) N-TiO2 (Rutile) Surfaces in Aqueous Acidic Solutions. J. Am. Chem. Soc. 2005, 127, 12975−12983. (40) Imanishi, A.; Tsuji, E.; Nakato, Y. Dependence of the Work Function of TiO2 (Rutile) on Crystal Faces, Studied by a Scanning Auger Microprobe. J. Phys. Chem. C 2007, 111, 2128−2132. (41) Gutmann, S.; Wolak, M. A.; Conrad, M.; Beerbom, M. M.; Schlaf, R. Effect of Ultraviolet and X-ray Radiation on the Work Function of TiO2 Surfaces. J. Appl. Phys. 2010, 107, 103705. (42) Di Valentin, C.; Tilocca, A.; Selloni, A.; Beck, T. J.; Klust, A.; Batzill, M.; Losovyj, Y.; Diebold, U. Adsorption of Water on Reconstructed Rutile TiO2(011)-(2 × 1):TiO Double Bonds and Surface Reactivity. J. Am. Chem. Soc. 2005, 127, 9895−9903. (43) Xiong, G.; Shao, R.; Droubay, T. C.; Joly, A. G.; Beck, K. M.; Chambers, S. A.; Hess, W. P. Photoemission Electron Microscopy of TiO2 Anatase Films Embedded with Rutile Nanocrystals. Adv. Funct. Mater. 2007, 17, 2133−2138. (44) Zhang, X. R.; Lin, Y. H.; He, D. Q.; Zhang, J. F.; Fan, Z. Y.; Xie, T. F. Interface Junction at Anatase/Rutile In Mixed-Phase TiO2: Formation and Photo-Generated Charge Carriers Properties. Chem. Phys. Lett. 2011, 504, 71−75. (45) Kumarasinghe, A. R.; Flavell, W. R.; Thomas, A. G.; Mallick, A. K.; Tsoutsou, D.; Chatwin, C.; Rayner, S.; Kirkham, P.; Warren, S.; Patel, S.; Christian, P.; O’Brien, P.; Grazel, M.; Hengerer, R. Electronic Properties of the Interface between p-CuI and Anatase-Phase N-TiO2 Single Crystal and Nanoparticulate Surfaces: A Photoemission Study. J. Chem. Phys. 2007, 127, 114703. (46) Chambers, S. A.; Droubay, T.; Kaspar, T. C.; Gutowski, M. Experimental Determination of Valence Band Maxima for SrTiO3, 12092
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
Langmuir
Article
(65) Thompson, T. L.; Yates, J. T. Surface Science Studies of the Photoactivation of TiO2-New Photochemical Processes. Chem. Rev. 2006, 106, 4428−4453.
12093
dx.doi.org/10.1021/la302169k | Langmuir 2012, 28, 12085−12093
