Green photo-oxidation of styrene over W–Ti composite catalysts

Green photo-oxidation of styrene over W–Ti composite catalysts

Journal of Catalysis 309 (2014) 428–438 Contents lists available at ScienceDirect Journal of Catalysis journal homepage: www.elsevier.com/locate/jca...

2MB Sizes 0 Downloads 56 Views

Journal of Catalysis 309 (2014) 428–438

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Green photo-oxidation of styrene over W–Ti composite catalysts Mario J. Muñoz-Batista a, Anna Kubacka a,⇑, Rafal Rachwalik b, Belén Bachiller-Baeza a, Marcos Fernández-García a,⇑ a b

Instituto de Catálisis y Petroleoquímica, CSIC, C/Marie Curie 2, 28049-Madrid, Spain Institute of Organic Chemistry and Technology, Cracow University of Technology, Warszawska 24, 31-155 Kraków, Poland

a r t i c l e

i n f o

Article history: Received 10 September 2013 Revised 22 October 2013 Accepted 26 October 2013 Available online 27 November 2013 Keywords: Photo-catalysis Partial oxidation Styrene Titania Tungsten

a b s t r a c t A series of WO3–TiO2 composite catalysts with variable quantities of tungsten was prepared by a singlepot microemulsion procedure and evaluated in the gas-phase selective photo-oxidation of the aromatic hydrocarbon styrene. Samples improve the performance of TiO2 by a factor ca. 3 and increase significantly the selectivity to valuable products (as styrene oxide) upon UV and, more importantly, sunlighttype excitation. Optimum performance in terms of both activity and selectivity was achieved under a sunlight-type renewable energy source. A complete structural (bulk and surface) and electronic characterization using Transmission Microscopy, X-ray diffraction (XRD), X-ray Photoelectron (XPS), Raman, and UV–visible spectroscopies and calorimetry with the help of the ammonia probe molecule was carried out. The study was able to detect and quantify the presence of different W entities as a function of the tungsten content of the catalysts. Three species were found to be present throughout the series; oligomeric W species and WO3 nanoparticles in close contact with anatase-TiO2 and non-contacting WO3 platelets. From a spectro-kinetic analysis carried out with the help of electron paramagnetic resonance (EPR), we provide evidence that these W-containing species play different roles in reaction. Activity and selectivity improvements with respect to the bare titania are intimately related with presence of oligomeric W species at anatase surfaces. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction Heterogeneous photocatalysis by nanocrystalline semiconductors corresponds to a relatively novel technology applied to environmental abatement in both liquid- and gas-phase phases. This is essentially based on the excellent performance and stability of titania, the most prominent photocatalytic material, for the mineralization of typical pollutants, including refractory or nonbiodegradable molecules, under mild conditions, e.g., room temperature and atmospheric pressure and using oxygen (air) as oxidant agent [1–4]. Among titania single nanophases, anatase is clearly the most active phase largely because it shows a correct balance between its surface chemistry-related properties and the adequate physical properties for efficient handling of lighttriggered charge carriers, allowing them to be involved in chemical steps at the surface [4]. In contrast to studies devoted to photo-degradation of pollutants, relatively fewer studies were conducted on the application of photocatalysis for product synthesis using selective oxidation ⇑ Corresponding authors. Fax: +34 915854760. E-mail addresses: [email protected] (A. Kubacka), [email protected] (M. FernándezGarcía). 0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcat.2013.10.021

routes [5–7]). Within the field of selective photo-oxidations, a particular interesting and active topic concerns the partial oxidation of organic molecules, mostly hydrocarbons and alcohols, to give either aldehyde derivatives with wide use in the chemical industry (flavor, confectionary, beverage, other activities) or to epoxides with application in both chemical and plastic industries [6–14]. Such selective photo-catalytic processes capture a great deal of attention in recent years in the quest to switch to more benign, sustainable process that makes use of environmentally friendly reagents, free of harmful organic solvents and dangerous/costly oxidants, and using mild temperature/pressure operation conditions. Titania fulfils all such conditions for a green process; however, a main drawback is that titanium dioxide is an UV absorber and has to be modified in order to allow the use of renewable energy sources as the sun [1–7]. This modification seeks for the enhancement of the light to chemical energy conversion efficiency by increasing the number of useful photon from the 3–5% corresponding to the UV range to ca. 45–50%, corresponding to the UV–visible range of solar light. Profiting from both UV and visible lights in selective-photooxidation reactions would thus require the careful engineering of a significant number of physico-chemical variables of a TiO2-based

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

nanomaterial. First and as mentioned, the managing of light absorption of the anatase phase appears as a key goal. Several strategies have been pursued to address this point, being most popular ones related to the electronic modification of titania through cationic [1,2,4,15,16], anionic [1,2,4,17–20], doping or both [4,21–23], or the use of additional, visible-light-sensitize phases in intimate contact with TiO2 [4,24,25]. In all cases, the goal is to allow the absorption of visible light photons while making a non-negative or, less frequently, positive impact on charge recombination [16,17,24,21,22,26] To reach this objective, here we approach a rational design; by starting from the synthesis of a highly selective (e.g., no total oxidation products, CO2) TiO2 nanomaterial, we address the two-fold objective of increasing activity, e.g., improving light to chemical conversion, with the simultaneous (more specific) improvement of sunlight-based performance by surface-sensitization of the TiO2 phase with tungsten [16,27,28]. This is done here within a single-pot synthesis procedure, in order to ensure the minimum modification of the bare TiO2-anatase phase. Secondly, the light handling-related requirements as mentioned above have to be woven with our aim to adequately control selectivity to transform styrene into valuable chemicals. This unified goal corresponds to an essentially open yet very active field of research. In this context, the use of W at surface of titania nanomaterials has been shown not only to improve catalytic activity in photo-oxidation of organics, like butyl-acetate [29], 2-propanol [30,31], toluene [32,33], chlorophenols [34,35], and oxalic acid [36], but also to increase selectivity for partial vs. total oxidation for hydrocarbon transformation [21,22,37], or hydrogen production on photo-reforming of alcohols/acids ([38,39]). Thus, W can potentially fulfil all light- and chemical-related objectives upon adequate engineering of physico-chemical properties of the base TiO2-anatase material. Note that the situation is relatively different from the easy one of increasing activity either upon UV and/or visible light, with has been previously described in a number of cases [16,27–36]. To this end, we studied the transformation of styrene into valuable products upon both ultraviolet and sunlight excitation and analyze both the enhancement of activity and modification of selectivity as a function of the number and nature of the W species present at the surface of anatase. A gas-phase photo-catalytic process is presented as it has an inherent advantages vs. liquid-phase reactions; among them, some were mentioned before, like the absence of harmful organic solvents (not always but frequently used in selective photo-oxidation) [5–7], while others, like the easy recovering (and potentially facile regeneration if needed) of the catalyst, can be now added to the list [1–7]. In this study, we will provide evidence that starting from an anatase nanomaterial showing 100% selectivity for photo-oxidation of this aromatic hydrocarbon into aldehydes, we can adequately modify the selectivity to obtain valuable chemical products as a function of the existent surface W species nuclearity, this done without loss but instead a significant enhancement of photo-activity. A multitechnique approach (morphological analysis, XRD, XPS, Raman, Calorimetry of probe molecules, UV–visible and analysis of optical properties) was used to structurally/electronically characterize the specific W surface modifiers which can be used to maximize photo-oxidation activity/selectivity. The fate of carriers after illumination was investigated using EPR. Opposite trends of activity as selectivity as a function of W content; while photo-activity is maximized for low W content, selectivity to valuable chemicals is increasing with the W content. The rationale being that this result is interpreted on the basis the characterization study which provides strong evidence of the key role played by oligomeric W species present at the surface of or in close contact with TiO2-anatase.

429

2. Experimental section 2.1. Catalyst preparation Materials were prepared using a single-pot microemulsion preparation method using n-heptane (Scharlau) as organic media, Triton X-100 (Aldrich) as surfactant and hexanol (Aldrich) as cosurfactant. A TiO2 reference sample was obtained as a first step using a microemulsion with acetic acid (to fix the initial pH to a value of 5.0) into the aqueous phase and titanium tetraisopropoxide as precursor. In all composite samples and the WO3 reference, ammonium tungsten oxide (Alfa Aesar) was introduced in the aqueous phase of a microemulsión. After 30 min of agitation, a stoichoimetric (to obtain the corresponding W(VI) hydroxide) quantity of tetramethylammonium-hydroxide (TMAH) was introduced from the aqueous phase of a similar microemulsion. For nanocomposite samples, after 5 min of contact and pH adjustment with acetic acid (pH 5), titanium tetraisopropoxide was introduced into the previously resulting microemulsion drop by drop from a mixture with isopropanol (2:3). Water/Ti and water/surfactant molar ratios were, respectively, 110 and 18 for all samples [40,41]. The resulting mixture was stirred for 24 h, centrifuged, and the separated solid precursors rinsed with methanol and dried at 110 °C for 12 h. After drying, the solid precursors were subjected to a heating ramp (2 °C min1) up to 525 °C, maintaining this temperature for 2 h. Samples names are T for the titania reference, and xW for the composite ones where x is the molar content of WO3 (in relation to a fixed amount of titania corresponding to 1 mol).

2.2. Characterization details The BET surface areas and average pore volume and sizes were measured by nitrogen physisorption (Micromeritics ASAP 2010). XRD profiles were obtained with a Seifert D-500 diffractometer using Ni-filtered Cu Ka radiation with a 0.02° step and fitted using the Von Dreele approach to the Le Bail method [42]; particle sizes and microstrain were measured with XRD using the Willianson– Hall formalism [43]. Ti/M composition was analyzed by using inductively coupled plasma and atomic absorption (ICP-AAS; Perkin–Elmer, Optima 3300 DV). Transmission electron microscopy (HTEM) and X-ray energy dispersive spectra (XEDS) were recorded on a JEOL 2100F TEM/ STEM microscope. XEDS analysis was performed in STEM mode, with a probe size 1 nm, using the INCA x-sight (Oxford Instruments) detector. Specimens were prepared by depositing particles of the samples to be investigated onto a copper grid supporting a perforated carbon film. Deposition was achieved by dipping the grid directly into the powder of the samples to avoid contact with any solvent. XPS data were recorded on 4  4 mm2 pellets, 0.5 mm thick, prepared by slightly pressing the powered materials which were outgassed in the prechamber of the instrument at room temperature up to a pressure <2  108 to remove chemisorbed water from their surfaces. The SPECS spectrometer main chamber, working at a pressure <109 torr, was equipped with a PHOIBOS 150 multichannel hemispherical electron analyzer with a dual X-ray source working with Ag Ka (hm = 1486.2 eV) at 120 W, 20 mA using C 1s as energy reference (284.6 eV). Surface chemical compositions were estimated from XP-spectra, by calculating the integral of each peak after subtraction of the ‘S-shaped’ Shirley-type background using the appropriate experimental sensitivity factors and the CASAXPS (version 2.3.15) software. The microcalorimetric studies of ammonia adsorption were carried out in a differential heat-flow microcalorimeter (Tian-Calvet type C80; Setaram) connected to a conventional volumetric

430

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

apparatus. The adsorption temperature was maintained at 80 °C in order to limit physisorption. Each sample was evacuated overnight at 150 °C and cooled to adsorption temperature under vacuum. Equilibrium pressure was measured by means of a Baratron pressure transducer MKS Instrument. UV–visible transmission or diffuse reflectance spectroscopy experiments were performed with a Shimadzu UV2100 apparatus with a nominal resolution of ca. 1 nm using, for diffuse experiments, BaSO4 as reference. To interpret the optical properties of the solids, the local net radiation field (also called local rate of photon absorption; ea,s) at the sample position x,y,z was calculated as described in the Supporting information. The ea,s values describe the beam attenuation suffered by the light while traversing the samples. This thus corresponds to a measurement of the light– sample interaction at the specific geometry of the reactor system and provides evidence of the exact differences expected among samples concerning light absorption properties [44]. In this way, we can see how the different materials perform either under UV or sunlight. Band gap values of the samples were obtained using the equation ahm = A (hmEg)n/2 (where a, h, m, Eg, and A are absorption coefficient, Planck constant, radiation frequency, band gap and a constant, respectively), assuming an indirect band gap semiconductor behavior (n = 4). Band gap typical error is 0.03 eV [45]. The electron paramagnetic resonance (EPR) measurements were done with a Bruker ER200D spectrometer operating in the X-band and calibrated with a DPPH standard. For the 5,5-dimethyl-1-pyrroline N-oxide (DMPO) spin-trapping EPR experiments, the samples were suspended in water (at a concentration of 0.6 g L1) and were sonicated for 4 min. An aqueous solution (0.01 M) of DMPO spin trap (supplied by Sigma) was prepared and kept on ice during the whole set of experiments. Bidistilled water (Elix-10) was employed for these preparations. 100 ll of the solid suspension and 100 ll of the DMPO solution were mixed into an EPR flat quartz cell under atmospheric air and irradiated at different times, through a spectroscopic Pyrex glass filter with a cut-off at ca. 220 nm, with light excitation source identical to that employed for the photokilling tests (280/500 nm), being then immediately transferred to the spectrometer cavity for EPR analysis. A small radical concentration decay (of ca. 5% on average) was observed in the dark during the course of spectrum recording. The latter were obtained at 298 K at ca. 9.75 GHz microwave frequency, 19.5 mW microwave power, 100 kHz modulation frequency, 1 G modulation amplitude and 2  105 spectrometer gain. No significant signal saturation was observed in those conditions. Blank experiments were also performed over mixtures of 100 ll of the DMPO solution and 100 ll of water to check the absence of radical formation in the absence of solid under the employed conditions. 2.3. Photo-catalytic experimental details Gas-phase selective photo-oxidation tests were carried out with styrene and using a set-up described elsewhere [32,33]. In the Supporting information, we summarized information concerning the reaction set-up geometry and lamp characteristics (Figs. S1 and S2). Activity and selectivity for the gas-phase photooxidation were tested in a continuous flow annular photoreactor containing ca. 40 mg of photocatalyst as a thin layer coating on a pyrex tube. The corresponding amount of catalyst was suspended in 1 ml of ethanol, painted on a pyrex tube (cut-off at ca. 290 nm), and dried at RT. The reacting mixture (100 ml/min) was prepared by injecting styrene (P99%; Aldrich) into a wet (ca. 75% relative humidity, RH) 20 vol.% O2/N2 flow before entering to the photoreactor, yielding an organic inlet concentration of ca. 700 ppmv. Under such conditions, the reaction rate shows a zero-order kinetics with respect to the total flow and organic pollutant/oxygen concentrations. After flowing the mixture for 6 h (control test) in the dark,

the catalyst was irradiated by four fluorescent daylight lamps (6W, Sylvania F6W/D) with a radiation spectrum simulating sunlight (UV content of 3%; main emission lines at 410, 440, 540, and 580 nm; full description at Supporting information), symmetrically positioned outside the photoreactor. Similar tests were carried out in selected samples using UV lamps (Sylvania F6WBLT-65; 6W, maximum at ca. 350 nm; full description at Supporting information). Reaction rates were evaluated (vide supra) under steady-state conditions, typically achieved after ca. 6–10 h from the irradiation starting. No change in activity was detected for all samples within the next 24 h. The concentration of reactants and products was analyzed using an on-line gas chromatograph (Agilent GC 6890) equipped with HP-PLOT-Q/HP-Innowax columns (0.5/0.32 mm I.D.  30 m) and TCD (for CO2 measurement)/FID (organic measurement) detectors. Carbon balance was achieved in the 96–100% range in all experiments. The photonic efficiency for the reaction under UV and visible lights has been determined from the reaction rate and the flux of incoming photons accordingly to [46]:



Dm J  A  Dt

where Dm = number of styrene moles consumed; J = flux of photons; A = illuminated area; Dt = change in time. The flux of photons was calculated by means of the following equation:



Ik NA  h  c

where I = light intensity (W/m2); k = wavelength; NA = Avogadro constant; h = Planck’s constant; and c = speed of light. 3. Results and discussion 3.1. Structural and electronic characterization xW samples were obtained by a single-pot microemulsion procedure, which ensures the production of mesoporous, high surface area materials. A summary of morphological properties and chemical composition for all samples included in this work is presented in Table 1. Fig. 1A displays the XRD patterns of the samples. The presence of the TiO2 and WO3 single-phase references in Fig. 1 makes easy the assignment of peaks. All samples showed the anatase structure (JCPDS card 78-2486, corresponding to the I41/amd space group) with cell parameters summarized in Table 2. The presence of a monoclinic WO3 phase (JCPDS card 83-0951, corresponding to the P21/n space group) becomes evident as the W content of the materials grows. Some differences in relative intensity among peaks displayed by such phase are observed with respect to the WO3 reference, indicating certain influence of TiO2 in the preferential growth of the nanoparticles. The cell parameters of the monoclinic phase are also included in Table 2. In such table, most evident differences among samples are related to a decrease in the TiO2 c-parameter, indicating a rather modest change in tetragonality and cell volume by effect of WO3. The modest changes in both observables indicates that W is not present as a doping agent into the anatase structure [32,33]. Fig. 1B displays a representative example of an XRD profiles corresponding to the 0.1 W and 0.25 W samples after reaction. The plot shows the absence of noticeable modifications in both the anatase TiO2 and monoclinic WO3 phases. Primary particle size and strain parameters for the two detected phases are summarized in Table 3. Data for fresh and used samples are constant and thus a single datum is displayed for each sample. For TiO2, we observed a size of ca. 15.9 nm and a relatively high

431

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438 Table 1 Sample chemical composition and morphological properties.

a

Samples

W/Ti AAS-ICP

BET area (m2 g1)

Pore volume (cm3 g1)

Pore size (nm)

Porositya (%)

T 0.1 W 0.25 W 0.5 W

– 0.08 0.20 0.53

65.0 65.6 74.9 70.8

0.0608 0.0619 0.0674 0.0771

4.1 3.7 3.5 3.7

18.9 20.5 23.5 27.7

Estimated based on the density of samples and the pore volume determined using the adsorption branch of the N2 isotherm.

Intensity / a.u.

A

20

30

40

50

60

70

80

2θ / degrees

Intensity / a.u.

B

20

30

40

50

60

70

80

2θ / degrees Fig. 1. XRD patterns for the (A) fresh samples, and (B) used samples.

strain. According to results reported in the literature, a linear relationship between size and microstrain was observed for pure (and doped) nanosized anatase-type samples [47,48]. This suggests that our sample has roughly a 2 increase in strain with respect to other microemulsion anatase samples obtained using the same organic-phase/surfactant/cosurfactant recipe and having similar particle size (12–15 nm) [49]. Such fact seems a consequence of the preparation pH and ionic strength at the aqueous phase [47,49].

The size and strain characteristics are relatively maintained in the presence of WO3 although a small decrease in particle size is observed as a typical stabilization result, already described in the presence of W [50]. The monoclinic WO3 phase presents a primary particle size close to 25 nm in the two samples where it is detected. The structural characteristics of the samples were further analyzed with the help of Raman spectroscopy (Fig. 2). In all samples, mainly peaks due to the titania anatase phase at ca. 144, 195, 399, 517, and 639 cm1 [49,51] are observed in Fig. 2A. In the presence of W, e.g., for the xW samples, we can differentiate between two types of new signals. A first contribution comes from the XRDdetected monoclinic WO3, with main peaks at 125, 262, 322, 708, 798 cm1 [52]. An additional contribution, not present neither in the bare TiO2 and WO3 references, is additionally observed in Fig. 2A for composite samples. This corresponds to a line at ca. 970 cm1 related to the W@O stretching of surface wolframyl entities. The Raman frequency can be correlated with a WAO distance of ca. 1.73 Å [53–55]. Fig. 2B gives an example of the evolution of the samples under reaction conditions. In agreement with XRD, the anatase TiO2 and monoclinic WO3 main phases do not suffer any significant alteration; however, the dispersed, surface W species are modified under reaction. Fig. 2C presents an expanded view of the ca. 970–1000 cm1 region. As can be seen, the evolution of W@O entities concerns both the Raman frequency and intensity. Talking about the first, the frequency shift would correspond to a shortening of the WAO bond distance of ca. 0.03 Å [53]. The red shift indicates that the degree of polymerization of surface W species grows under reaction [56]. This evolution occurs in all samples, ending up in very similar species, although the extension of the polymerization seems larger for samples with the lower W content. There is also a loss of W@O stretch intensity for 0.1 W and 0.25 W samples which is probably related to a more heterogeneous structural situation after reaction. In any case, a semiquantitative measurement of the number of surface-W species can be carried out with the help of previous XPSRaman studies of W species present at surface and near-surface species at the anatase surface [32,33,57]. The amount of the dispersed, W species is included in Table 4 and amounts from ca. 60% in the 0.1 W to ca. 25% for the 0.5 W one. The structural characterization was completed with a TEM/ XEDS analysis of selected samples. TEM images are shown in Fig. 3 while XEDS results corresponding to W/Ti atomic ratio at different locations of the samples are displayed in Fig. 4. Two images of the 0.1 W sample (Fig. 3A and B) show the presence of particles of the two oxides, TiO2 and WO3, interwoven at nanometric level. The corresponding XEDS ratios provide evidence of a relatively homogeneous distribution of the both components (Fig. 4). For

Table 2 Cell parameters for TiO2 and WO3 phases detected by XRD. Samples

T 0.1 W 0.25 W 0.5 W

TiO2 Anatase Cell parameters (A)

WO3 (monoclinic) Cell parameters (A)

a

c

a

b

c

3.79 3.79 3.80 3.80

9.49 9.49 9.46 9.43

– – 7.36 7.34

– – 7.54 7.53

– – 7.71 7.70

432

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

Table 3 XRD-derived size and microstrain analysis for TiO2 and WO3 phases. TiO2 Size (nm)/microstrain hn2i1/2

Samples T 0.1 W 0.25 W 0.5 W

15.9/22.9 E 14.7/21.6 E4 15.1/21.0 E4 14.2/22.9 E4

– – 26.1/12.9 E4 25.6/13.5 E4

Table 4 Ti and W binding energies, W/Ti average molar ratio and the molar W fraction for Ti reference and xW samples.

A

Samples

Intensity / c.p.s.

T 0.1 W 0.25 W 0.5 W

XPS binding energy (eV) Ti (2p3/2)

W (4d5/2)

458.8 458.8 458.8 458.8

– 247.3 247.3 247.3

W/Ti XPSa

W/Ti XEDSa

W (%)b

– 0.20 0.47 0.65

– 0.27 (0.26) – 6.18 (2.46)

– 60/10 33/22 25/25

a W/Ti average molar ratio (we report the mean and the median – the latter in parenthesis) obtained by XPS or XEDS. b Molar W fraction (percentage) present at non-crystalline WOx entities located at TiO2 surfaces in fresh/used samples.

200

400

600

800

Wavenumber / cm

1000

1200

-1

Intensity / c.p.s

B

200

400

600

800

1000

Wavenumber / cm-1

C

972

Intensity / c.p.s.

1004

800

WO3 Size (nm)/microstrain hn2i1/2

4

850

900

950

1000

1050

Wavenumber / cm-1 Fig. 2. Raman spectra of the samples. (A) Fresh samples; (B) example for used sample; (C) Detail of the W@O region for fresh and used samples.

0.1 W, the average W/Ti ratio measured by XEDS is 0.27, which is moderately larger than the 0.08 obtained by chemical analysis. This is further analyzed with the help of XPS (Table 4), but a qualitative explanation relies in the presence of oligomeric W species at the anatase surface. For the 0.5 W sample, the W/Ti atomic ratio calculated by XEDS is less homogeneous through the sample. According to Fig. 4, this is a result of contributions from zones where W is well below 10 at.% (cation basis) and a characteristic TEM view (Fig. 3C) provide evidence of a different morphology with respect to ones having a more even contribution of both cations (Figs. 3A/B). According to the TEM/XEDS analysis, the presence of platelets seems characteristic of very large particles of WO3. Note the absence (or rather minor contribution) of such entities in 0.1 W according to XEDS. We can thus conclude that the XRD-Raman-TEM-XEDS study indicates the presence of three types of W-containing entities. The 0.1 W sample only shows two, the oligomeric W species at anatase surface and (as below detailed, scarce) monoclinic WO3 nanoparticles with relatively high interaction with anatase. The 0.25 W and 0.5 W samples have the additional presence of large WO3 platelet-type entities without significant contact with anatase. Comparison of W/Ti average atomic ratios obtained by XEDS and XPS (Table 4) further supports the presence of an additional W-containing phase present in W-dominated zones of the sample and exclusive of the 0.25 W/0.5 W materials. Only the oligomeric W species shows evolution under reaction conditions, with indications consistent with a larger degree of polymerization achieved under reaction. According to Table 4, the importance of the polymerization process increases (e.g., affects to a larger fraction) as the W content of the nanocomposite decreases. The electronic properties of the samples were analyzed by the joint use of XPS and UV–visible spectroscopies. The XPS analysis provides evidence of the presence of Ti(IV) and W(VI) in all composite samples, without detecting significant contribution of any reduced state of both cations (see Table 4) [58]. So, although W is present in several distinct entities, its oxidation state is constant through the sample series. The UV–visible spectra of the samples are dominated by the band gap of titania anatase, which is located at ca. 3.1 eV for the TiO2 reference here synthesized (Table 5) [59]. A progressive shift of the band gap to lower energy is observed in Fig. 5, with the numerical values reported in Table 5 accordingly to the indirect gap nature of both semiconductors [4]. Such a shift is

433

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

35 1.0

W/Ti at. ratio

30

W/Ti at. ratio

25

0

2

4

6

8

0

2

4

6

8

10

12

14

16

18

20

22

24

10

12

14

16

18

20

22

24

0.1W 0.5W

0.8 0.6 0.4 0.2

20

0.0

15

Scan Number

10 5 0 0

2

4

6

8

10

12

14

16

18

20

22

24

Scan Number Fig. 4. EDX analysis of the W/Ti atomic ratio at several positions of the samples. Inset is an expanded view of the low W/Ti atomic ratio region.

Table 5 Band gap energy of the samples assuming indirect band gap semiconductors. Samples

Band gap (eV)

T 0.1 W 0.25 W 0.5 W

3.08 3.02 2.95 2.88

observed in the presence of oligomeric W species dispersed in oxide surfaces [60], which according to Table 4 grows with W content (0.055, 0.064, 0.083 mol per mol of material in, respectively, 0.1 W, 0.25 W, and 0.5 W samples). Fig. 5 also displays an example of the UV–visible spectra obtained after reaction. As a general rule, we see the essentially unmodified (average) band gap energy with respect to fresh samples, but the presence of significant additional intensity in the region above 400–450 nm. The latter is mainly related to the surface accumulation of carbonaceous residues although the modifications suffered by oligomeric W species may likely be an additional factor [33]. 3.2. Surface and photocatalytic properties

Fig. 3. TEM images of different zones of the 0.1 W (upper, middle) and 0.5 W (down) samples. See text for details.

consistent with the corresponding average of the TiO2 and WO3 band gap energies which, for our references, take the values of 3.1 and 2.7 eV, respectively. Some gap localized density of states around 400–450 nm is also visible in Fig. 5. These are typically

Surface structural characteristics of the samples were analyzed in the previous section by a combined XPS-Raman study but here we focus on chemical properties of the surface and, more concretely, on surface acidity using calorimetry. Acidity of the materials was tested with the help of ammonia as a probe; calorimetry data reported in Fig. 6 indicate that WOx presence on bare TiO2 increases moderately acidity, particularly by enhancing the weight of the more acidic OH fraction, typically measured by the Q100 (e.g., Qdiff > 100 kJ mol1) observable [61]. Total acidity (measured for Qdiff above 70 kJ mol1) show, on the other hand, relatively modest differences among composite samples. The Q70/Q100 ratio per mol of sample is 0.67, 0.65, and 0.63 for, respectively, 0.1 W, 0.25 W and 0.5 W samples. The analysis suggests, as mentioned, modest differences in acidity among xW samples, mainly appearing for acid sites below 100 kJ mol1. Above such value we see only marginal differences among xW samples. The strong acidity of W oxides is well known [62] but presence of W at anatase surface also increases acidity with respect to the bare titania reference. Similarly to anatase-based solid solutions with acidic cations (W, Zr), this may be grounded in the formation of new hydroxyl entities (Brønsted-type acidity) [21,22]. Summarizing, from the minimum

434

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

A

160

-1

120

Qdiff / kJ·mol

Kubelka-Munk

140

2

100 80 60 40

300

350

400

450

500

550 0

Wavelength / nm

2

4

6 -2

µmol NH 3·m

Fig. 6. NH3 adsorption profiles for xW and T reference samples.

500

600

700

800

900

Wavelength / nm Fig. 5. UV–visible spectra for the samples. (A) fresh samples and (B) example for used samples.

amount of tungsten included in the composite samples we observed an increase (with respect to titania) of strong Brønsted acidity. Such increase is more or less constant through the series of xW samples. The different structural and surface properties of the nanocomposite samples render materials with quite different photo-catalytic properties. Such differences are illustrated in data presented in Fig. 7, which correspond to observables at pseudo-steady state, characteristic of the experimental conditions used (e.g., high relative humidity which minimize surface poisoning and allows long term stability). Absence of photolysis and significant polymerization of styrene was checked in blank tests. Note also that the WO3 reference is not included in the plot as activity is more than 10 times lower. This agrees with previous results using other aromatic hydrocarbons [32]. Photo-oxidation of styrene using titania-based nanomaterials (like P25) typically produces CO2 as total oxidation product, and benzaldehyde (plus acetaldehyde), styrene oxide and/or acetophenone as partial oxidation products. Fig. 8 presents the chemical structures of the partial oxidation products detected in this study within a simplified reaction scheme (see below). Note that carbon balance was achieved above 96% in all catalytic tests, and thus only rather minor products (like acetophenonene) are not displayed in Fig. 8. In this figure, a first point of discussion (even in the simplified form of the mechanism presented in the figure) is the production of 1-phenyl-ethanol (1PE) through a styrene oxide (SO) intermediate or by a direct hydration reaction of styrene. The first

3.0

1.0

2.5

0.8

2.0

0.6

1.5

0.4

1.0

0.2

PE / %

400

-2 -1

300

10

2

path seems favored in the literature [6,8,9,21,22,63]. In any case Fig. 8 only attempts to show the products detected as the mechanism is so complex that has not been previously studied and would require the extensive use of in situ, synchronous (e.g., time-resolved, light-triggered EPR/IR experiments which has not been previously published in a complex reaction like this – see for example Refs. [6,8,9,21,22,63]) techniques. Fig. 8 also includes information about the active centers involved in the generation of the products detected according to information below described. Presence of W in the material alters both activity and selectivity in adequate directions and upon both UV and sunlight excitations. Moreover, changes in these observables are roughly independent of the excitation wavelength (Fig. 7). Starting with the activity, we observed a maximum enhancement of ca. 2.7 (with respect to the nanoanatase and P25 samples) for the 0.1 W sample. This occurs under sunlight-type excitation. The enhancement factor decreases to 2.0 by the 0.5 W. Similar enhancement factors can be obtained while considering the photonic efficiency (calculated using light intensity data at the sample surface – Fig. 9). Although these calculations are only semiquantitative as they did not account for the number of electrons involved in the photo-chemical reaction, the real calculation of the photon efficiency is, as demon-

r reac. * 10 / mol m s

Kubelka-Munk

B

UV Sunlight

0.5

0.5W

0.25W

UV Sunlight 0.1W

0.0

T

Fig. 7. Experimental result for styrene photo-oxidation reaction rate (right axis) and photonic efficiency (left axis) under UV- and Sunlight-type irradiation.

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

435

Fig. 8. Simplified reaction scheme showing main products detected and the active sites involved in each case.

strated by Ohtani, rather complex [64]. In spite of this, useful information can be obtained if considering trends among our samples. So, the only significant difference from the behavior of the reaction rate observable is the moderately higher performance of the materials under UV than visible light. The modification of selectivity occurs gradually with the W presence. Incremental production of CO2 as well as of more interesting (than BA) partial oxidation products such as styrene oxide (SO) and 1-phenyl-ethanol (1PE) were detected. The corresponding decrease in BA production is more acute upon UV conditions. CO2 and SO products run in parallel with the increase in the W content while 1PE seems to grow significantly only for the 0.5 W sample. To interpret activity we must first consider the analysis of lightsolid interaction. Fig. 9 provides a quantitative answer. Considering first results upon UV excitation, Fig. 9 show that, although the chemical nature of the samples is quite different, the absorption of light show differences below 7.5%. This indicates that the activity boosting in the presence of W is mostly related to subsequent light handling events. An EPR study was carried out to analyze the influence of W in the fate of charge carriers and how this affects photoactivity. As noted in the literature, there is evidence that OH radical attack is the responsible for initial steps of aromatic hydrocarbons (including styrene) degradation, see Refs. [6,8,9,21,22,63]. A measure of OH radical species can be obtained through a DMPO-assisted EPR measurement. UV irradiation of DMPO-containing sample suspensions gives rise to a signal with 1:2:2:1 intensity pattern for all samples (Fig. 10). Its EPR parameters (g = 2.0056, aN = 14.9 G, aH = 14.9 G) are characteristic of DMPO-OH adducts [65–68]. Fig. 10 shows the temporal behavior of the signal for the 0.1 W sample. A comparison of the initial rate of OH for the first minutes (before multiple additions, within consecutive reactions, of OH radicals to DMPO molecules would drive to the formation of diamagnetic species) is customarily used to compare the sample power to generate this hole-related radical species [65–70]. More importantly, the correlation plot presented in Fig. 11 between the EPR analysis and the UV reaction rates (note that according to Fig. 7 such correlation is similar for visible light reaction rates – results not shown) provides conclusive evidence that the different W species present at the catalytic samples modify charge carrier separation and the number of hole-related surface species able to produce chemistry. Such hole-species are, as demonstrated by

the figure, involved in a kinetically-relevant step of the reaction. In this context, Fig. 12 sketched the structural and electronic contact of the two W-containing entities in contact with titania (see details at Supporting information), in an attempt to provide the rationale to interpret the light handling properties after absorption and, particularly, the fate of the chemically active hole-related species detected by EPR. Note that both W-containing nanostructures contain W6+ as main species although differ in their structural and electronic properties. Concerning UV excitation, we can see that the two W-containing phases in contact with TiO2 make opposite effects; while oligomeric W species will inject holes into the photocatalytically active anatase, the WO3 nanoparticles will withdraw them. The opposite trends for hole-related flow observed for the two W-containing species in contact with anatase generate a trade-off in photochemical properties. The behavior of the sample series can be explained by the relative quantity of the different W species. The amount of oligomeric W species in contact with anatase can be measured (as mentioned, 0.055/0.064/0.083 mol per mol of sample in 0.1 W/0.25 W/0.5 W), while the corresponding amount of WO3 nanoparticles is only known for 0.1 W (0.018 mol/mol material). As the interface contact of anatase with WO3 nanoparticles is significantly lower (per W atom) than that of oligomeric species, maximum hole-related production and photochemical activity is achieved with the 0.1 W. This is confirmed by the correlation plot presented in Fig. 11. For 0.25 W and particularly 0.5 W, an important increase in both WO3 nanoparticles in contact with anatase and ‘isolated’ WO3 platelets may exist. The first is detrimental for activity, while the second has a low (e.g., 1 order of magnitude lower) contribution to activity. The similar hole production and photo-activity of 0.1 W and 0.25 W would indicate that the chemically determinant structural difference among these two samples is restricted to the presence of WO3 platelets. For 0.5 W, the more drastic decrease in both observables would also be related to a contribution from WO3 platelets, but a more significant issue would be related to the partial (vs. complete for remaining samples) modification of the active oligomeric W species observed by Raman. For visible light, we first note a significant increase in light absorption by a factor of 1.7 (Fig. 9) from anatase to the 0.5 W sample. According to the previous discussion, this is a combined effect related to the (average) band gap decrease (Table 5) and the presence of W-related localized states (Fig. 4). For visible light excita-

436

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

0.1W 20 G

4 min

2 min

1 min

Fig. 10. EPR spectra under UV light irradiation of the DMPO-OH signal obtained after different time contact in the presence of the 0.1 W sample.

2.2

0.25W 0.1W

10

-2

rreac.* 10 / mol m s

-1

2.0 1.8

0.5W

1.6 1.4 1.2

T 1.0 0.8 1000

1500

2000

2500

3000

3500

4000

Initial Rate OH / a.u. Fig. 11. Correlation plot between the initial rate of OH radical formation and reaction rates displayed in Fig. 7 for UV excitation.

Fig. 9. Local rate of photon absorption (mW cm2) for selected samples. From top to bottom: T and 0.5 W samples under UV (first two panels) and Sunlight-type (last two panels) irradiation. See Supported information for calculations of this observable.

tion, the interface plays a role by which oligomeric W species inject holes in anatase while some minor withdraw of electrons (those generated from localized states of titania) from TiO2 may occur (such charge flow is not explicitly considered in Fig. 12). WO3 nanoparticles displayed in Fig. 12 absorb light but do not contribute positively to interface charge handling.

We may thus suggest that upon UV or visible light the role of oligomeric species is positive, allowing charge separation and injecting holes, the likely active species in photocatalytic reactions, in the anatase component of the samples. This explains the incremental activity (respect to the bare anatase reference) upon both UV and sunlight for xW samples and the maximum for the 0.1 W sample. The larger increase in activity upon sunlight vs. UV of xW samples with respect to the anatase reference (Fig. 7) may be directly associated with the increase in light absorption, as evidenced in Fig. 9. The selectivity results are more complex to interpret but general trends can be here discussed. Increasing production of CO2 is typically associated with an increment in acidity [71]. Concerning aromatic hydrocarbons, acidity helps in ring opening steps, which leads to CO2 production as a final product. Ring opening would

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438

437

4. Conclusions

Fig. 12. Scheme corresponding to the structural and electronic contact between Wand Ti-containing phases present in xW samples. Vertical black arrows correspond to energy differences between electronic states/bands, while blue/red arrows correspond to charge carrier events generated upon UV/visible light. WO3 represents nanoparticles in contact with the titania phase, while WxOy (W6+-species) represents oligomeric species at titania surfaces. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 6 Selectivity to styrene oxidation products for xW and TiO2 and P25 reference samples. BA: Benzaldehyde; 1PE: 1-phenyl-ethanol; SO: styrene oxide. Selectivity (%) Samples

P25 T 0.1 W 0.25 W 0.5 W

Sunlight-type

UV light

CO2

BA

SO

1PE

CO2

BA

SO

1PE

62.3 0 10.3 12.6 16.4

36.7 95.1 81.0 76.1 61.4

0 2.9 5.6 8.1 13.7

1.0 1.9 3.1 3.3 8.5

81.6 0 12.9 11.2 18.5

14.9 94.3 76.0 71.2 55.4

0 3.3 7.5 13.8 20.0

1.5 2.3 3.5 3.8 6.1

occur at an early stage of the reaction, with high enough kinetic constant to compete with partial oxidation routes in the case of pure titania [72]. According to Fig. 6, acidity scales in a similar way among the xW samples as the CO2 selectivity; e.g., a relatively smooth, continuous change trough the series, with obvious correlation with the content of W of the solid. As summarized in Fig. 8, BA seems primary differentiated from other partial oxidation products as, according to Table 6, is mainly related to isolated (e.g., non-interacting with tungsten) anatase sites. The surface of titania is thus modified by presence of W species with direct consequences in BA production. Consequently, the other partial oxidation products (SO/1PE) would be related to the presence of novel sites from titania and likely associated with oligomeric W species. Interesting to note is that combined selectivity of SO and 1PE is a factor of 1.5 larger than the maximum observed in previous W-doped TiO2 systems [21,22,37]. The trend in selectivity can be discussed in terms related to the structural situation of the oligomeric W species under reaction (Raman, Fig. 2). Most significant changes are observed between 0.25 W and 0.5 W samples; as said, the last sample is the only one which present two different oligomeric W species under reaction. This and the known relationship between polymerization degree and acidity in oligomeric W species [56] would suggest than more polymerized/acidic species would favor SO while less polymerized/acidic species are related to the sudden increase in 1PE detected for 0.5 W with respect to remaining xW samples. Unfortunately, a XAS study of such hypothesis is precluded by the heterogeneous distribution of W in the samples.

A series of three WO3–TiO2 composite catalysts, 0.1 W, 0.25 W and 0.5 W, were prepared by microemulsion using a single-pot procedure. A multitechnique XRD-XPS-Raman-TEM-XEDS study of the materials was able to detect and quantify the presence of different W entities as a function of the WO3 content of the catalysts. Three species were found to be present throughout the series; oligomeric W species and WO3 nanoparticles in close contact with anatase-TiO2 and non-contacting WO3 platelets. For 0.1 W, we observed exclusively the presence of the two W species in contact with anatase, with dominance of the oligomeric entities, while the remaining samples additionally contain WO3 platelets. Composite samples improve the performance of TiO2 by a factor of 2.7 and increase significantly the selectivity to valuable products (SO and PE) upon both UV and sunlight excitation. Optimum performance in terms of activity was achieved with the 0.1 W sample, while selectivity toward highly valuable chemicals as styrene oxide was maximized for the 0.5 W sample. Of note is the fact that activity is improved to a large degree upon a sunlight-type renewable light source than upon UV light. The study was able to provide evidence that activity and selectivity improvements are related with oligomeric W species present at reaction conditions. The interface contact between anatase of this size-limited W-containing species appears critical for the improvement of the photocatalytic properties though a wise handling of charge carriers. Such species suffer evolution under reaction conditions and a correlation between their degree of polymerization and inherent acidity and the specific, high added-value partial oxidation product originated from the radical attack to the vinyl moiety of styrene is proposed. Acknowledgments A. Kubacka and M.J. Muñoz-Batista thank MINECO for support through, respectively, the postdoctoral, Ramón y Cajal’’ and predoctoral FPI programs. R.R. wants to acknowledge EU (Project no.UDA-POKL.04.01.01-00-029/10-00) for financial support of a stay at Madrid (ICP-CSIC). Financial support by MINECO is also acknowledged (project CTQ2010-14872/BQU). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcat.2013.10.021. References [1] M.R. Hoffman, S.T. Martin, W. Choi, D.W. Bahneman, Chem. Rev. 95 (1995) 69– 121. [2] O. Carp, C.L. Huisan, A. Reller, Prog. Solid State Chem. 32 (2004) 33–145. [3] H. Thu, M. Karkmaz, E. Puzenat, C. Guillard, J.M. Herrmann, Res. Chem. Intermediat. 31 (2005) 449–543. [4] A. Kubacka, G. Colón, M. Fernández-García, Chem. Rev. 112 (2012) 1555–1614. [5] A. Maldotti, A. Molinari, R. Amadelli, Chem. Rev. 102 (2002) 3811–3839. [6] Y. Shiraishi, T. Hirai, J. Photochem. Photobiol. C 9 (2008) 157–189. [7] L. Palmisano, V. Augugliaro, M. Bellardita, A. Di Paola, E. García-López, V. Loddo, G. Marcí, G. Palmisano, S. Yurdakal, ChemSusChem 4 (2011) 1431– 1438. [8] M.A. González, S.G. Howell, S.A. Sikdar, J. Catal. 183 (1999) 159–162. [9] X. Li, C. Kutal, J. Mater. Sci. Lett. 21 (2002) 1525–1527. [10] P. Du, J.A. Moujlin, G. Mul, J. Catal. 238 (2006) 342–351. [11] V. Augugliaro, T. Caronna, V. Loddo, G. Marci, G. Palmisano, L. Palmisano, Y. Yurkadal, Chem. Eur. J. 14 (2008) 4640–4650. [12] S. Yurkadal, G. Palmisano, V. Loddo, V. Augugliaro, L. Palmisano, J. Am. Chem. Soc. 130 (2008) 1568–1569. [13] S. Higashimoto, N. Suetsugu, M. Azuma, H. Ohue, Y. Sakata, J. Catal. 274 (2010) 76–83. [14] D. Tsukamoto, Y. Shiraishi, Y. Sugano, S. Ichikawa, S. Tanaka, T. Hirai, J. Am. Chem. Soc. 134 (2012) 6309–6315. [15] M. Anpo, M. Takeuchi, J. Catal. 216 (2003) 505–516.

438 [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46]

M.J. Muñoz-Batista et al. / Journal of Catalysis 309 (2014) 428–438 A. Kubacka, M. Fernández-García, G. Colón, J. Catal. 254 (2008) 272–281. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269–277. N. Serpone, J. Phys. Chem. B 110 (2006) 24287–24331. C. Belver, R. Bellod, S.J. Stewart, F.G. Requejo, M. Fernández-García, Appl. Catal. B 65 (2006) 309–314. S. Livraghi, M.R. Chierotti, E. Giamello, G. Magnacca, M.C. Paganini, G. Cappeletti, C.L. Bianchi, J. Phys. Chem. C 112 (2008) 17244–17254. A. Kubacka, B. Bachiller-Baeza, G. Colón, M. Fernández-García, Appl. Catal. B 93 (2010) 274–281. A. Kubacka, G. Colón, M. Fernández-García, Appl. Catal. B 95 (2010) 238–244. H. Gao, B. Lu, F. Liu, X. Shao, Int. J. Photoenergy (2012). 453018-6. D. Robert, Catal. Today 122 (2007) 20–33. A.M. Balu, B. Baruwati, E. Serrano, J. Cot, J. García-Martínez, R.S. Varma, R. Luque, Green Chem. 13 (2011) 2750–2758. X. Zhang, K. Udawa, Z. Liu, S. Nishimoto, C. Xu, Y. Lu, H. Sakai, M. Ave, T. Marakoi, A. Kujishima, J. Photochem. Photobiol. A 202 (2009) 39–46. J.H. Pan, W.I. Lee, Chem. Mater. 18 (2006) 847. K.K. Akurati, A. Vital, J.-P. Dellemann, K.K. Michalor, T. Graule, D. Ferri, A. Baiker, Appl. Catal. B 79 (2008) 53–62. V. Keller, P. Benhanrd, F. Garin, J. Catal. 215 (2003) 129–138. Y.T. Kwon, K.Y. Song, I. Lee, G.J. Choi, Y.R. Do, J. Catal. 191 (2000) 192–199. S.Y. Chai, Y.J. Kim, W.I. Lee, J. Electroceram. 17 (2006) 909–912. A. Fuerte, M.D. Hernández-Alonso, A.J. Maira, A. Martínez-Arias, M. FernándezGarcía, J.C. Conesa, J. Soria, G. Munuera, Chem. Commun. (2001) 2718–2719. A. Fuerte, M.D. Hernández-Alonso, A.J. Maira, A. Martínez-Arias, M. FernándezGarcía, J.C. Conesa, J. Soria, G. Munuera, J. Catal. 212 (2002) 1–9. A.K.L. Sajjad, S. Shamaila, B. Tian, F. Chen, J. Zhang, Appl. Catal. B 11 (2009) 397–405. S.A.K. Legari, S. Sajjad, F. Chen, J. Zhang, Chem. Eng. J. 3 (2011) 906–915. J.M. Hermanm, M.N. Mozzanega, P. Pitchat, J. Photochem. 22 (1983) 333–343. A. Kubacka, B. Bachiller-Baeza, G. Colón, M. Fernández-García, J. Phys. Chem. C 113 (2009) 8553–8556. H. Yang, L. Guo, W. Yang, H. Liu, J. Power Sources 159 (2006). 1305.1306. É. Karácsonyi, L. Baia, A. Dombi, V. Danciu, K. Mogyorósi, L.C. Pop, G. Kovács, V. Cosßoveanu, A. Vulpoi, S. Simon, Zs. Pap, Catal. Today 208 (2013) 19–27. P.G. De Gennes, C. Taupin, J. Phys. Chem. 86 (1982) 2294–2303. M. Fernández-García, X. Wang, C. Belver, J.C. Hanson, J.A. Rodriguez, J. Phys. Chem. C 111 (2007) 674–682. A. Le Bail, H. Duroy, J.L. Forquet, Mater. Res. Bull. 23 (1988) 447–453. G.K. Williamson, W.H. Hall, Acta Metall. 1 (1953) 22–31. G.E. Imoberdof, H.A. Irazoqui, A.E. Cassano, O.M. Alfano, Ind. Eng. Chem. Res. 44 (2005) 6075–6082. J. Tauc, Mater. Res. Bull. 5 (1970) 721–729. C. Baumanis, D.W. Bahnemann, J. Phys. Chem. C 112 (2008) 19097.

[47] L.E. Depero, L. Sangetti, B. Allieri, E. Bontempi, A. Marino, M. Zocchi, J. Crys. Growth 198 (199) (1999) 516–521. [48] S. Vives, C. Meunier, Powder Diff. 24 (2009) 205–212. [49] M. Fernández-García, A. Martínez-Arias, A. Fuerte, J.C. Conesa, J. Phys. Chem. B 109 (2005) 6075–6083. [50] M. Fernández-García, A. Martínez-Arias, J.C. Hanson, J.A. Rodríguez, Chem. Rev. 104 (2004) 4063–4104. [51] M. Mikami, S. Nakamura, O. Kitao, H. Arakawa, Phys. Rev. B 66 (2002) 155213– 155216. [52] M. Sadakane, K. Sasaki, H. Kunioku, B. Ohtani, R. Abe, W. Ueda, J. Mater. Chem. 20 (2010) 1811–1818. [53] F.D. Hardscastle, I.E. Wachs, J. Raman Spectrosc. 26 (1995) 397–406. [54] F.Y. Chen, I.E. Wachs, J. Catal. 217 (2003) 468–477. [55] A. Baserga, V. Russo, F. Di Fondo, A. Bailini, C. Cattaneo, C.S. Casari, A. Li Bassi, C.E. Bottani, Thin Solid Films 515 (2007) 6465–6469. [56] T. Onfroy, V. Labarbier, G. Clet, M. Houalla, J. Mol. Catal. A 318 (2010) 1–7. [57] A.M. Márquez, J.J. Plata, Y. Ortega, J.Fdez.J.Fdez. Sanz, J. Phys. Chem. C 115 (2011) 16970–16976. [58] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, Handbook of X-Ray Photoelectron Spectroscopy, in: G.E. Muilenberg (Ed.), Perkin-Elmer, Minesota, 1978. [59] H. Irie, K. Kamiya, T. Shibanuma, S. Miura, D.A. Tryk, T. Yokohama, K. Hashimoto, J. Phys. Chem. C 113 (2009) 10761–10767. [60] D.G. Burton, M. Shtein, R.D. Wilson, S.L. Soled, E. Iglesia, J. Phys. Chem. B 103 (1999) 630–640. [61] A. Aroux, A. Gervasini, J. Phys. Chem. 94 (1990) 6371–6377. [62] P. Carniti, A. Gervasini, A. Aroux, J. Catal. 150 (1994) 274–283. [63] M. Lim, Y. Zhou, B. Wood, L. Zhou, V. Rudolph, G.Q. Lu, Environ. Sci. Technol. 43 (2009) 538–543. [64] B. Ohtani, Chem. Lett. 37 (2008) 167–189. [65] M.D. Hernández-Alonso, A.B. Hungría, A. Martínez-Arias, M. Fernández-García, J.M. Coronado, J.C. Conesa, J. Soria, Appl. Catal. B 50 (2004) 167. [66] M.A. Grela, M.E.J. Coronel, A.J. Colussi, J. Phys. Chem. 100 (1996) 16940. [67] E.G. Janzen, N. Sankuratry, Y. Kotake, J. Magn. Reson. 111 (1996) 254. [68] D. Dvoranova´, V. Brezova´, M. Mazur, M.A. Malati, Appl. Catal. B 37 (2002) 91. [69] Q. Chen, H. Shi, W.S. Li, Y. Xu, D.W. Wu, Catal. Sci. Technol. 2 (2012) 1213– 1220. [70] A. Kubacka, M.J. Muñoz-Batista, M. Ferrer, M. Fernández-Garcia, Appl. Catal. B 140–141 (2013) 680–690. [71] Q. Wang, M. Zhang, C. Chen, W. Mao, J. Zhao, Angew. Chem. Int. Ed. 49 (2010) 7976–7979. [72] T.D. Bui, A. Kimura, S. Ikeda, M. Matsuma, J. Am. Chem. Soc. 132 (2010) 8543– 8558.