Enhanced photocatalytic activity of single-phase, nanocomposite and physically mixed TiO2 polymorphs

Applied Catalysis A: General 489 (2015) 51–60

Contents lists available at ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Enhanced photocatalytic activity of single-phase, nanocomposite and physically mixed TiO2 polymorphs Renata Kaplan, Boˇstjan Erjavec, Albin Pintar ∗ Laboratory for Environmental Sciences and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia

a r t i c l e

i n f o

Article history: Received 18 July 2014 Received in revised form 29 September 2014 Accepted 9 October 2014 Available online 18 October 2014 Keywords: TiO2 polymorphs Photocatalysis Water treatment Bisphenol A Synergistic effect

a b s t r a c t In this study, testing of TiO2 polymorphs (anatase, rutile, brookite) and their mixtures (anatase/rutile, anatase/TiO2 -B) in heterogeneous photocatalytic oxidation process was conducted at ambient conditions in a batch slurry reactor. The efficiency of bare TiO2 catalysts was evaluated based on the degree of bisphenol A (BPA) removal, which is a well-known endocrine disrupting compound (EDC). The obtained results indisputably show that BPA removal is strongly affected by catalyst morphology, crystallite size, structure and specific surface area. Detailed interpretation of catalyst properties combined with BPA removal rates leads to the conclusion that photocatalytic oxidation is the most prominent either by using pure anatase particles or high surface area anatase/TiO2 -B nanocomposite. However, the highest extent of mineralization was observed in the presence of high specific surface area nanotubular anatase/TiO2 -B nanocomposite. Interestingly, when anatase and rutile particles were physically mixed, an additional beneficial effect on BPA degradation was observed. Interpretation of the obtained results shows that a synergistic effect between the respective phases takes place, and consequently enhances the overall activity. This phenomenon was explained by the proposed mechanism of overall hydroxyl radicals concentration increment due to transfer of OH• formed on the surface of anatase particles (via H2 O oxidation with photogenerated holes in the valence band) to rutile particles. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Titanium dioxide (TiO2 ) has been used as a white pigment from ancient times, and consequently, it is safety to humans and the environment is guaranteed by history [1]. It can mainly be found in three different crystalline forms: anatase, rutile and brookite. Anatase adopts a tetragonal structure in which four edges are shared per octahedron, but there is no corner sharing. Similarly, rutile also has a tetragonal structure in which two opposite edges of octahedron are shared in order to form a linear chain along [0 0 1] direction. Further, chains are joined through corner connections. On the other hand, brookite has an orthorhombic structure where octahedron share three edges and also corners [2]. Rutile is the stable form, while anatase and brookite are metastable and can be transformed to rutile when heated. In recent years TiO2 -B has also attracted attention, due to its exclusive relatively opened structure. It is formed when layered hydrogen titanates are subjected to heating [3]. All crystallographic forms of nanocrystalline TiO2 are of great importance in the view point of applications, since they

∗ Corresponding author. Tel.: +386 1 47 60 237; fax: +386 1 47 60 460. E-mail address: [email protected] (A. Pintar). http://dx.doi.org/10.1016/j.apcata.2014.10.018 0926-860X/© 2014 Elsevier B.V. All rights reserved.

are applicable in a wide variety of fields, such as optics, cosmetics and solar cells [4]. Anatase can be synthesized by various methods such as sol–gel and hydrothermal synthesis [5,6], while rutile can be prepared via phase transition from anatase using either calcination at temperatures over 600 ◦ C [7] or hydrothermal synthesis [8]. There have been many studies on their performance in photocatalytic processes [9,10]. On the other hand, there have only been a few reports on the synthesis of brookite-type TiO2 . The difficulty in preparing pure brookite TiO2 samples is probably the reason for the small number of studies on its photocatalytic performance [11]. It is well established that anatase is the most photocatalytically active TiO2 polymorph, compared to rutile and brookite. The high activity can be contributed to prolonged lifetimes of charge carriers and spatial charge separation [12]. Since efficiency of photocatalysis is largely limited by detrimental electron–hole recombination, coupling of two polymorphs of TiO2 has proved to be an efficient approach for enhancing electron–hole separation and inhibiting electron–hole recombination. Especially anatase/rutile nanostructure proved to be very efficient, since coupling of anatase and rutile allows the transport of excited electrons from anatase to rutile with lower band energy, and therefore charge recombination can be suppressed [2]. Beside this, there have been some reports on the synergistic effect between physically mixed anatase and rutile

52

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

powders, where higher activity was observed compared to singular phases [13]. Activity of TiO2 in catalytic applications is, besides crystal form, also influenced by various parameters like crystallinity, impurities, surface area and density of surface hydroxyl groups [14]. It has been widely applied in advanced oxidation processes (AOPs), since they have the capability of exploiting the high reactivity of in situ produced hydroxyl radicals in order to achieve complete abatement and thorough mineralization of different organic pollutants [15–18]. There are many AOPs, such as wet air oxidation (WAO), catalytic wet air oxidation (CWAO), photocatalytic oxidation, foto-Fenton process and ozonation [19]. Among them, heterogeneous photocatalytic oxidation process can be used for oxidation of a wide variety of organics [20]. It makes use of a semiconductor metal oxide as a catalyst and of oxygen as oxidizing agent. TiO2 is the most widely utilized photocatalyst due to its highly distinctive properties; water insolubility, cost effectiveness, durability and resistance to abrasion [21,22]. TiO2 as a semiconductor has the ability to activate organic pollutants by hydrogen radical abstraction by valence-band holes, as well as oxygen by conduction-band electrons, because it can easily derogate from the stoichiometric composition, which results in an oxygen deficient surface [23]. High surface area of a catalyst is advantageous, since the number of surface vacant sites for oxygen activation is directly connected to the specific surface area of materials [23]. In the following contribution, the preparation of TiO2 polymorphs and their influence on the degradation process of water dissolved toxic aromatic compound bisphenol A (BPA), an endocrine disrupting compound [24], is systematically investigated. BPA is known to interfere with normal hormone functions and therefore poses a threat to aquatic life and humans as well [25]. According to Yamamoto et al. [26], it was detected in seven out of ten investigated landfill leachates in concentrations ranging from non-detectable to 17.2 mg/l, thus making it an appropriate candidate as a testing molecule in AOPs. The influence of physical mixtures of anatase and rutile on BPA removal was addressed and explained via a proposed mechanism. To the best of our knowledge, there have yet been no reports made on the influence of TiO2 polymorphs and their mixtures (coexisting nanocomposites and physical mixtures) on the removal of EDCs. Especially, physical mixtures of TiO2 polymorphs seem to be a grey area in understanding of the enhanced activity of mixed TiO2 phases. Photocatalytic experiments were conducted in a batch slurry reactor. Catalyst efficiency was determined by withdrawing liquid-phase samples at pre-selected time intervals, and determining the remaining BPA and total organic carbon (TOC) content by using high-performance liquid chromatography (HPLC) and TOC analyses, respectively. Morphological and textural properties of tested catalysts were examined using SEM, XRD, TPD, UV–vis-DR and BET analytical techniques. CHNS analyses of fresh and spent catalysts were also applied in order to evaluate the amount of carbonaceous deposits on catalyst surface during the reaction, which enabled us to estimate the suitability of prepared catalysts for a long-term purification of BPA in the process of photocatalytic oxidation.

2. Experimental 2.1. Catalyst preparation Pure anatase (A) phase was prepared by using 0.5 g of titanate nanotubes with BET surface area of 400 m2 /g as a precursor (the preparation procedure is available elsewhere [27]). TiO2 nanotubes with a layered structure were dispersed in 75 ml of 0.001 M H3 PO4 solution by means of ultrasonic homogenizer. Afterwards, the mixture was hydrothermally treated in a Teflon-lined autoclave for 24 h at 180 ◦ C. The white pasty material was thoroughly washed with

distilled water and dried in vacuum under cryogenic conditions for 24 h. Pure rutile (R) phase was prepared using room temperature synthesis route, similar to that proposed by Zhang et al. [28], where a certain amount of TTIP (Sigma–Aldrich, ≥97.0%) was dropwise added to 37 ml of 2-propanol (Sigma–Aldrich, p.a., ≥99.8%) under vigorous stirring. Afterwards the obtained solution was subjected to mixing in the presence of an ice bath, while dropwise adding 100 ml of 1 M HCl to form a misty mixture. Stirring in an ice bath was carried on for the next 4 h. The employed procedure has led to the formation of a transparent liquid which was left standing for 21 days at room temperature. White precipitates, which formed during this time period, were collected by means of centrifugation and thoroughly washed with distilled water and dried under cryogenic conditions for 24 h. After drying the material was annealed at 450 ◦ C. Pure brookite (B) phase was prepared using synthesis route similar to that reported by Nguyen-Phan et al. [29]. TTIP (Sigma–Aldrich, ≥97.0%) was dropwise added to 50 ml of 2propanol (Sigma–Aldrich, p.a., ≥99.8%) while mixing. Afterwards, 200 ml of 0.1 M NaOH was added and left under vigorous stirring for 2 h. The mixture was later hydrothermally treated in a Teflon-lined autoclave for 72 h at 200 ◦ C. The white pasty material obtained was thoroughly washed with distilled water and dried at 60 ◦ C in an oven for 18 h. Anatase/rutile (AR) nanocomposite was synthesized using ultrasound assisted sol–gel technique, similar to that reported by Prasad et al. [30]. First, a certain amount of titanium (IV) isopropoxide (Sigma–Aldrich, ≥97.0%) (TTIP) was dropwise added to 50 ml of 2-propanol (Sigma–Aldrich, p.a., ≥99.8%) under vigorous stirring. The formed mixture has been subjected to sonication using ultrasonic bath in order to ensure optimal distribution of TTIP in 2-propanol. Subsequently, urea (Kemika) was added in molar ratio of TiO2 :urea = 1.1:1 in order to induce the formation of crystal lattice defects [31]. After the addition of urea, the mixture was again subjected to sonication in order to achieve optimal mixing of all the present compounds. The obtained sol was converted into gel by dropwise addition of distilled water under forceful mixing. After the hydrolysis, stepwise preparation procedure was continued with sonication in order to achieve formation of localized hot spots by collapse of bubbles due to cavitation [32]. The prepared sample was dried at 60 ◦ C in an oven for 18 h. After drying the obtained material was annealed at 700 ◦ C for 3 h. For the preparation of anatase/TiO2 -B (ATB) nanocomposite, the material obtained in anatase/rutile preparation process (without the annealing step) was used as a precursor for hydrothermal synthesis of high surface area material with homogeneous morphology. The powder was dispersed in 10 M NaOH using ultrasonic homogenizer and later sealed in a Teflon-lined autoclave, which was filled up to 70% of its volume. The synthesis took place at 130 ◦ C for 12 h. The resulting white material was filtered and thoroughly washed (i) with distilled water and (ii) 0.1 M HCl in order to promote proton exchange mechanism. To preserve the structure, the material was dried in vacuum under cryogenic conditions for 48 h. After drying the material was annealed at 400 ◦ C with the aim of inducing changes in the crystal structure, morphology, surface properties and crystallite size, which could further influence catalytic properties during photocatalytic oxidation of the model pollutant BPA. 2.2. Catalyst characterization The prepared materials were characterized by means of scanning electron microscopy (Carl Zeiss, model FE-SEM SUPRA 35VP). X-ray powder diffraction patterns were acquired using PANalytical ˚ X’pert PRO MPD diffractometer with Cu K␣1 radiation (1.54056 A)

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

53

in reflection geometry. Data collection was conducted in the range between 10 and 90◦ (steps of 0.034◦ ). Specific surface area of materials, total pore volume and average pore size were determined by measuring nitrogen adsorption and desorption isotherms at −196 ◦ C (Micromeritics, model TriStar II 3020). BET (Brunauer, Emmett, Teller) theory was applied in order to calculate the specific surface area of materials, while pore size distribution was calculated from the desorption branch of the corresponding nitrogen isotherms, employing the BJH method. Diffuse reflectance UV–vis spectra of the prepared materials were recorded using a UV-VIS spectrophotometer (Perkin Elmer, model Lambda 35) equipped with the RSA-PE-19M Praying Mantis accessory. The materials were scanned with the speed of 120 nm/min and slit set to 4 nm. Background correction was performed using a white reflectance standard (Spectralon® ). Photoluminescence (PL) spectra were recorded using LS55 Fluorescence Spectrometer (Perkin Elmer, model LS55) with an excitation wavelength light set to 300 nm. The dehydration process of the prepared materials was monitored using TPO/TGA analysis (Perkin Elmer, model Pyris 1), where samples were heated up to 750 ◦ C in air atmosphere with heating rate of 10 ◦ C/min. Acidity measurements of the prepared polymorphs were performed using Pyris 1 by Perkin Elmer. Each material was heated at 180 ◦ C for 120 min under N2 flow and then cooled down to room temperature. The surface of the solids was saturated with n-propylamine for 10 min. The excess of n-propylamine was purged with N2 flow for 60 min. TPD of n-propylamine was measured by heating to 750 ◦ C at 10 ◦ C/min. The amount of carbon accumulated on the catalyst surface during photocatalytic experiments was determined by means of CHNS elemental analysis (Perkin Elmer, model 2400 Series II).

were suspended in water by means of ultrasonification. Prior to the illumination period, the suspension was kept in dark to allow establishing of sorption equilibrium. The reactor content was illuminated by UVA high pressure mercury lamp (150 W, with a maximum at  = 365 nm). The lamp was placed in a water cooling jacket positioned vertically in the centre of the slurry. Representative 2-ml aqueous-phase samples for subsequent analyses were periodically collected from the reaction suspension in predetermined time intervals and filtered through a membrane filter (Sartorius, 0.45 ␮m) in order to remove catalyst particles.

The effectiveness of photocatalytic oxidation was evaluated by determining temporal BPA conversions as a model pollutant, and concentrations of potential intermediates formed during the oxidative destruction using various analytical techniques (HPLC, TOC, IC). Measurements performed with HPLC apparatus (Spectra SystemTM ) were conducted in the isocratic analytical mode using 100 mm × 4.6 mm BSD Hypersil C12 2.4 ␮m column with the flow rate of 0.5 ml/min and methanol:ultrapure water ratio of 70:30 (UV detection at  = 210 nm) for measuring the concentration of BPA. The level of mineralization was determined by measuring the total organic carbon (TOC) content with an advanced TOC analyzer (Teledyne Tekmar, model Torch), which is equipped with a high-pressure NDIR detector, and applying a high-temperature catalytic oxidation (HTCO) method carried out at 750 ◦ C. Ion chromatography (IC) was employed to determine the composition of liquid-phase samples using Dionex ICS-3000 apparatus and an IONPAC AG11-HC analytical column.

2.3. Photocatalytic oxidation experiments

3. Results and discussion

Photolytic and photocatalytic degradation of BPA (C0 = 10 mg/l in ultrapure water with 18.2 M cm resistance) was studied in a 250 ml batch slurry reactor at atmospheric pressure. The reactor unit was thermostated at T = 20 ◦ C (Julabo, model F25), magnetically stirred (600 rpm) and continuously sparged with purified air (45 l/h). The catalyst concentration was in the range of 0–125 mg/l. Before admixing the catalyst into the model solution, the powders

3.1. Catalyst characterization

2.4. Analysis of end-product solutions

SEM micrographs of materials synthesized in the present study are shown in Fig. 1. Sample A with large (200 nm) and uniform elongated nanoparticles (Fig. 1a) exhibited BET surface area of 40 m2 /g. Low temperature synthesis of R phase (Fig. 1b) led to the formation of nanoparticles approximately 100 nm in length and 30 nm in

Fig. 1. SEM micrographs of (a) anatase (A), (b) rutile (R), (c) brookite (B), (d) anatase/TiO2 -B (ATB) mixture and (e) anatase/rutile (AR) mixture.

54

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

Fig. 2. XRD diffractograms of anatase (A), rutile (R), brookite (B), anatase/TiO2 -B (ATB) mixture and anatase/rutile (AR) mixture.

width, with BET surface area of 99 m2 /g. After annealing at 450 ◦ C in air atmosphere for 3 h, BET surface area decreased to 32 m2 /g. On the other hand, particles of sample B (Fig. 1c) showed very low BET surface area of 8 m2 /g and interesting humming-top like geometry. ATB material shown in Fig. 1d is in the form of well-developed and randomly orientated titanate nanotubes with BET surface area of 208 m2 /g (after annealing at 400 ◦ C). AR nanocomposite shown in Fig. 1e is a highly crystalline material with BET surface area of 9 m2 /g (after annealing at 700 ◦ C). Table 1 summarizes the values of measured BET surface area, total pore volume, average pore width and crystallite size of the as-prepared materials (A and B) and annealed solids (R, AR, ATB), which were used in subsequent BPA oxidation experiments. The average crystallite size of all prepared materials, including TiO2 P25 (Degussa) which was used for the comparison, were calculated by means of Scherrer equation for the main diffraction peak of anatase (1 0 1), rutile (1 1 0) and brookite (1 2 1). The calculated values are presented in Table 1, together with N2 physisorption data. Fig. 2 shows XRD diffractograms of the synthesized samples. AR and ATB materials, accordingly to XRD measurements, consist of anatase and rutile or anatase and TiO2 -B phases, respectively. TiO2 -B is a less known TiO2 polymorph, perceiving a broad band at 2 = 14.02◦ [33], which is reflecting its poor crystallinity. This phase is interesting due to its unique crystallographic structure [34]. It is reported that its presence can enhance the photoactivity of TiO2 surface, since it behaves as a site where conduction-band electrons accumulate [35]. N2 adsorption–desorption isotherms of A, R, B, ATB, AR samples and reference material TiO2 P25 are presented in Fig. S1a. ATB can be defined as a mesoporous material, corresponding to the range of pore size from 2 to 50 nm [36]. According to the IUPAC

classification, ATB and R solids correspond to type IV isotherms (adsorption and desorption in porous materials), which are typical for mesoporous materials. Characteristic feature of type IV isotherms is the hysteresis loop (type H3, typical for slit-shaped pores or plate-like particles) [36]. ATB material has a hysteresis loop which is approaching the relative pressure P/P0 = 1, what suggests the presence of macropores (>50 nm) [37]. On the other hand, A and R samples, which were prepared using a synthesis route that involves a calcination step at high reaction temperature (450 ◦ C), exhibit significantly smaller area of the hysteresis loop. However, the latter with a stepwise adsorption and desorption branch can hardly be observed in the case of TiO2 P25 sample, solid B prepared at 200 ◦ C, and AR mixture annealed at 700 ◦ C. These materials are approaching to the type II isotherms, which are typical for multilayer physical adsorption of gases by nonporous or macroporous adsorbent, where monolayer–multilayer adsorption can occur. The corresponding BJH pore size distributions are presented in Fig. S1b. It is clear that pore size distribution is significantly affected by temperature used for the preparation of individual materials. The ATB catalyst exhibits a wide pore size distribution with a peak in the range from 10 to 20 nm, which corresponds to mesopores. With higher annealing temperature or severe hydrothermal treatment (i.e., in the case of A, R, B and AR materials) we can observe increased crystal growth of TiO2 nanoparticles and aggregation of smaller pores, therefore, the surface area decreases and the pore size increases. The ATB solid contains inner pore channels, due to its nanotubular structure and therefore shows higher pore volume comparing to other prepared materials. As regards the commercial TiO2 P25 sample, the pore size distributions could not be observed, which is in accordance with the negligible hysteresis loop presented in Fig. S1a. The thermal stability of the prepared titanium dioxide polymorphs and their mixtures was evaluated using thermogravimetric analysis (TGA). The obtained results are presented in Fig. 3. Samples A, B, R and AR show minimal weight loss of 2.3, 1.9, 1.1 and 0.1 wt.%, respectively. Insignificant mass decrease in the case of AR sample can be attributed to high annealing temperature, which was used during the preparation procedure (700 ◦ C). If we compare it to the starting material used for the preparation of AR, it is possible to see that there is no further weight loss after annealing at temperatures higher than 400 ◦ C. The first and most substantial decrease in mass occurs at temperatures up to 200 ◦ C, which is attributed to the evaporation of physically adsorbed water [38]. Overall weight loss amounts to 44.1 wt.%. Preparation procedure of sample R involves annealing at 450 ◦ C, which has proved as a sufficiently high temperature for the dehydration of the as-prepared wheat-like nanocrystals. It is possible to see in Fig. 3 that dehydration of the starting material (R starting material) is completed at approximately 400 ◦ C. In the case of ATB sample, low weight loss of 5.7 wt.% was found, as a consequence of previously incorporated annealing step in the material preparation (i.e., 450 ◦ C), where most of the adsorbed water and intralayered OH groups were already removed. The ATB starting material shows typical thermogravimetric

Table 1 Specific surface area (SBET ), total pore volume (Vpore ), average pore width (dpore ) and average crystallite size of the prepared materials and commercial TiO2 P25 sample. Material

SBET (m2 /g)

Vpore (cm3 /g)

dpore (nm)

A R B ATB AR P25

40 32 8 208 9 52

0.31 0.32 0.04 0.70 0.04 0.15

26.3 34.4 21.3 10.1 16.0 10.2

a

Crystallite size could not be determined due to low intensity of the corresponding peaks.

Crystallite size (nm) Anatase

Rutile

Brookite

33.6 – – 6.7 44.8 23.0

– 17.0 –

– – 139.2 – – –

a a

35.0

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

Fig. 3. Thermogravimetric diagram of anatase (A), rutile (R), brookite (B), anatase/TiO2 -B (ATB) mixture, anatase/rutile (AR) mixture, and their corresponding starting materials (in the case of R, AR and ATB samples).

diagram of the as-prepared protonated titanate nanotubes with a total weight loss of 15.3 wt.%. Typically, weight loss below 300 ◦ C is usually ascribed to the removal of adsorbed water and intralayered OH groups [39]. Above 300 ◦ C, the initial protonated structure begins its transformation from titanate nanotubes to anatase. This phenomenon is accompanied with the titanate nanotubes to nanorods morphological transformation. Based on the above presented results, it can be concluded that the heat treatment regime significantly affects the physicochemical properties of particulated solids. The optical behavior of prepared catalysts was studied by measuring their UV–vis-DR spectra. Characteristic absorption band of titanium dioxide-based materials can be assigned to the transition of an electron from the valence band (VB) to the conduction band (CB), leaving a hole behind (Eq. (1)) [40]: TiO2 + h → h+ VB + e− CB

(1)

Table 2 summarizes band gaps of various catalysts (A, R, B, AR, ATB), which were prepared in this study. The spectral data show strong cut-offs where the absorption of light is minimal (Fig. S2a). Accordingly, using Eq. (2), band gap energies were calculated from the obtained data (Fig. S2b). In this equation, E stands for band gap energy (J), h is the Planck constant (6.626 × 10−34 J s), C is the speed of light (3.0 × 108 m/s) and  is the cut-off wavelength (nm): Eg =

h∗C 

(2)

Eq. (2) results to the following values of the band gap energy: 3.13 eV for A, 2.99 eV for R and 3.15 eV for B, respectively. The obtained band gap energies are in good agreement with values reported in the literature: 3.15 [41], 3.0 and 3.13 eV [42] for anatase, rutile and brookite, respectively. Among examined materials, ATB mixture shows the band gap absorption edge positioned deeper into the UV region (390 nm), what results in a slightly wider band gap energy (Eg = 3.19 eV) compared to other solids. It is common for nanotubular structures to absorb photons from 350 to 400 nm, corresponding to the band gap energies of 3.1–3.5 eV [43]. In the case of R and AR samples (Eg = 2.99 and 2.94 eV, respectively), slight redshift is observed, comparing to the prepared materials. The red-shift can be ascribed to high annealing temperatures used in the preparation procedures (450 and 700 ◦ C, respectively), which usually reflect in a decrease of band gap [44]. Additionally, all the obtained band gap energies were evaluated also using the Kubelka–Munk analysis (Fig. S2c). The results show good agreement with the above reported data.

55

Fig. 4. Photolytic and photocatalytic degradation of BPA in the presence of anatase (A), rutile (R), brookite (B), anatase/TiO2 -B (ATB) mixture and anatase/rutile (AR) mixture. Reaction conditions: atmospheric pressure, T: 20 ◦ C, C0 (BPA): 10 mg/l, V0 (BPA): 250 ml, pH0 : 5.2, C (catalyst): 125 mg/l, t(irradiation under UV light) : 60 min.

TPD characterization by means of n-propylamine was conducted in order to evaluate the acidic properties of each catalyst. Fig. S3 shows the obtained TPD curves of the prepared oxides. The peaks in high and low temperature regions can be attributed to desorption of n-propylamine from strong and weak acid sites, respectively. The TPD curve of ATB mixture shows the presence of two distinguished peaks, while brookite shows three. Other materials indicate merely one peak. Temperatures, at which these peaks appear, together with the corresponding amounts of acid sites for each material, are presented in Table 2. Values are 150, 90, 60, 40 and 850 ␮mol/g for anatase (A), rutile (R), brookite (B), anatase/rutile (AR) and anatase/TiO2 -B (ATB) samples, respectively. The influence of BET surface area on the amount of acid sites is clearly evident. The highest number of acid sites is present in the ATB solid, which exhibits the highest surface area (208 m2 /g). A and R solids follow the same trend, while sample B and AR mixture (with BET surface area of 8 and 9 m2 /g, respectively), have 60 and 40 ␮mol/g of acid sites. All these data combined show merely a linear trend. This is in good agreement with the results reported by Papp et al. [45], where it was shown that surface acidity of TiO2 is strongly connected to its surface area. Density of acid sites on sample B was the highest (7.5 ␮mol/m2 ) among the presented TiO2 polymorphs and their mixtures, while sample R showed the most scarce distribution (2.8 ␮mol/m2 ). 3.2. Photocatalytic oxidation over single-type catalysts The photocatalytic activities of TiO2 polymorphs were assessed by means of degradation of water dissolved model pollutant bisphenol A (BPA) under UV light irradiation for 60 min of operation. Fig. 4 demonstrates BPA conversion obtained in the presence of TiO2 polymorphs in comparison to photolytic BPA oxidation (blank sample). In any of heterogeneously photocatalyzed runs performed in this study, no dissolution of titania in the liquid phase was observed. In order to determine adsorption equilibrium, all experiments underwent a dark period (30 min). The obtained results show that in the case of B and R solids no decrease of BPA concentration by adsorption on the surface is observed. On the other hand, A, AR and ATB samples show slight BPA uptake of 5.5, 7.3 and 3.7%, respectively. When comparing the results of photocatalytic oxidation of aqueous BPA (C0 = 10 mg/l), it was confirmed that BPA has a photo-resistant character in the presence of UV light. No evident degradation was further observed

56

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

Table 2 Wavelength cut-offs obtained from UV–vis diffuse reflectance spectra, corresponding calculated band gaps and amount and density of acid sites determined with Npropylamine adsorption for anatase (A), rutile (R), brookite (B), anatase/TiO2 -B (ATB) mixture and anatase/rutile (AR) mixture. Material

Wavelength cut off (nm)

Band gap (eV)

Density of acid sites (␮mol/m2 )

Amount of acid sites (␮mol/g)

Peaks of n-propylamine desorption (◦ C)

A R B ATB AR

396 415 395 390 423

3.13 2.99 3.15 3.19 2.94

3.8 2.8 7.5 3.0 4.4

150 90 60 850 40

300 354 104 (highest), 249, 481 117, 390 (highest)

a

a

Peak could not be determined due to very low amount of acid sites.

with the use of highly crystalline and low BET surface area B phase. The highest degradation rate was achieved using the AR mixture (94% conversion at t = 60 min) and ATB mixture (94% conversion at t = 60 min). These results are very interesting, because the same performance in BPA removal was achieved, despite the fact that the ATB mixture has BET surface area which is 23-fold higher than that of AR mixture. Pure A solid exhibits significantly higher activity for photocatalytic oxidation of BPA (90% conversion at t = 60 min), comparing to R (47% conversion at t = 60 min) and B (8% conversion at t = 60 min) samples. These findings indisputably indicate that BPA conversion is affected by morphology, crystallite size and surface area of the individual catalysts. ATB sample is composed of high specific surface area particles with low crystallinity, while AR solid is defined as a highly crystalline powder with low BET surface area. Pure A and R phases have similar BET surface areas of 40 and 32 m2 /g, respectively. These statements stress the importance of anatase phase in the photocatalytic reaction, because it has been proved that less developed BET surface area enables necessary degradation rate of organic pollutant, if the demand of appropriate crystallinity is fulfilled. AR material is mainly composed of anatase (97%), accompanied with a small amount of rutile (3%), where mass fractions of anatase and rutile were determined from relative XRD diffraction intensities of [1 0 1] and [1 1 0] reflections for anatase and rutile, respectively, using Eqs. (3) and (4) [46]: Anatase (%) = Rutile (%) =



 0.79A  R + 0.79A

∗ 100

(3)



1 ∗ 100 (R + 0.79A)/R

(4)

The presence of small amounts of rutile is beneficial, if comparing to pure anatase phase with higher surface area, because better charge separation was attained over the solid containing both polymorphs. The B phase showed negligible activity for the removal of BPA (Fig. 4), probably due to the low BET surface area and presence of large crystallites. When dealing with organic compounds, photocatalytic oxidation kinetics is often modeled using Langmuir–Hinshelwood (L–H) equation with the assumption that the reaction takes place on the surface of photocatalyst particles. In the present work, L–H kinetic model was employed in order to describe the mineralization process of BPA. The initial expression (Eq. (5)) where r is reaction rate (mg/(l min)), k is the reaction rate constant (1/min), K is the adsorption constant (1/g) and C (mg/l) is concentration at time t (min): r=

−dC =k∗ dt

 KC  1 + KC

(5)

can be simplified to a pseudo-first order equation (Eq. (6)), due to low initial concentration (KC is much lower than 1, so the reaction is of first order): r=

−dC = k ∗ KC = k ∗ C dt

(6)

where k is the apparent first-order rate constant (1/min). The integrated form of Eq. (6) is: 

C(t) = C0 ∗ e−k ∗t

(7)

where C0 stands for initial concentration (mg/l). By using logarithmic function, Eq. (7) can be linearly expressed as: ln

C  0

C

= k ∗ t

(8)

A plot of ln C0 /C vs. time (Eq. (8)) represents a straight line, the slope of which upon linear regression equals the apparent firstorder rate constant k . At the time t = t1/2 , the concentration of the model pollutant is equal to exactly one half of the one at t = 0, so half-life for each experiment can be determined using the following expression (Eq. (9)): t1/2 =

ln 2 k

(9)

The reaction half-lives values demonstrate that the presence of different phases on the catalyst surface has a positive effect on the removal of BPA, because short half-lives of 15.1 min (AR mixture) and 14.9 min (ATB) were obtained. The solid materials composed of more than one phase are benifitial, due to better charge separation over different phases. Interestingly, the AR mixture shows high activity towards BPA degradation, despite its very low specific surface area (9 m2 /g). When BPA is degraded in the presence of a commercial catalyst TiO2 P25 (not shown in the graphs) with BET surface area of 50 m2 /g (phase composition in ratio of 80/20 for anatase/rutile), we obtain a half-life of 2.9 min. The comparison of obtained results demonstrates that almost similar catalytic performance was observed, despite 5.6-fold lower BET surface area of the synthesized nanocomposite. This is probably due to higher proportion of anatase phase and/or different defect structure, which is regarded as beneficial when TiO2 is used in photocatalytic applications [47]. On the contrary, considerably longer reaction half-life of pure A phase was registered (t1/2 = 18.1 min). The high photocatalytic activity of TiO2 catalysts consisted of mixed phases is in large part due to the synergistic activation of the rutile phase by anatase. The rutile phase has the ability to extend the photoactive range further into the visible part, therefore, harvesting more light and electron transfer from rutile to anatase trapping sites obstructs charge recombination [48]. Unlike that, the reason for poor activity of pure R (half-life for BPA degradation was 65.6 min) originates from the prompt rates of recombination. Similary, pure B particles suffer from the low BET surface area combined with unsuficient charge separation, resulting in a lenghty reaction half-life (t1/2 = 693.1 min). As stated before, good performance of ATB mixture can be attributed to its high BET surface area and abundance of surface acid sites. High activity of AR material can be also contributed to its narrow band gap (Eg = 2.94 eV) comparing to other prepared materials, which minimizes the required energy to excite electrons to the conduction band. Similar band gap width was observed for R powder (Eg = 2.99 eV), however, much lower activity towards BPA removal was determined. This data further emphasizes the benefits of applying nanocomposites as photocatalysts,

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

57

Table 3 Carbon content (measured by means of CHNS elemental analysis) on the surface of fresh and spent catalysts used in the photocatalytic oxidation process, and TOC conversions combined with TOC percentages deposited on catalyst surface. Material

TOC removal (%)

TCfresh (mg/g)

TCspent (mg/g)

TOCaccumulated (%)

A R B AR ATB

14.4 12.9 – – 79.9

1.43 0.18 0.16 0.17 0.38

1.34 0.74 0.51 0.63 1.95

– 9 – – 25.3

Reaction conditions: atmospheric pressure, T: 20 ◦ C, C0 (BPA): 10 mg/l, V0 (BPA): 250 ml, C (catalyst): 125 mg/l, pH0 : 5.2, t(irradiation under UV light) : 60 min.

giving rise to suppressed charge recombination and, consequently, higher oxidation rates compared to pure phases. Table 3 shows the TOC removals obtained after 60 min irradiation of BPA solution under UV light in the presence of various photocatalysts. TOC conversions were monitored in order to determine the degree of mineralization of the model pollutant. The obtained results show that up to 80% of the initial BPA could be mineralized to CO2 and H2 O using high surface area ATB mixture. On the contrary, the AR mixture was unable to mineralize BPA, though similar BPA conversion was attained in the presence of this material. In view of this, AR is capable to degrade BPA, but lacks the ability to completely mineralize the obtained intermediates. When the photocatalytic oxidation over AR was prolonged from 60 to 180 min, it was possible to achieve 94.1% and 13% conversion of BPA and TOC, respectively (not shown in the paper). This implies that considerably longer residence time is needed in order to assure mineralization of the formed intermediates. In this regard, insufficient TOC removals were obtained in the presence of pure A and R phases, and even negligible TOC conversion in the presence of brookite. In addition, Table 3 shows the carbon content on the surface of fresh and spent catalysts determined by means of CHNS elemental analysis. Initial carbon content (TCfresh ) was rather low, being in the range from 0.16 to 0.18 mg/g for R, B and AR samples, while A and ATB samples have slightly higher amount of carbonaceous deposits on the surface. This is strongly connected to the BET surface area of the prepared materials. Atmospheric impurities have higher tendency to adsorb on the materials with higher specific surface area. On the other hand, amount of carbon on fresh sample A is even higher. This is attributed to the use of high surface area titanate nanotubes as a precursor for its preparation. It can be seen in Table 3 that during the photocatalytic experiments small amounts of BPA or its derivatives were adsorbed on the catalyst surface. In the case of sample B and AR mixture, carbonaceous deposits on the surface can be attributed to BPA adsorption, since no TOC conversion was attained. Carbon content on sample A is even lower after its use in the process of heterogeneous photocatalysis; therefore, it is possible to conclude that anatase surface is capable to oxidize the surplus of accumulated carbon. In the case of R catalyst, 9% of removed TOC was actually deposited on the surface as carbonaceous deposits; therefore, the real TOC removal equals to 3.9%. The amount of carbonaceous deposits on ATB sample is 25.3%, while real TOC conversion is still the highest among the prepared materials (54.6%). Carbon-based elemental analysis indicates the potential for photocatalytic oxidation of BPA over crystalline A, which has less expressed deposition of carbonaceous deposits and exhibits high rate of BPA removal, but unfortunately results in rather low TOC conversion. Fig. 5 shows results of IC analyses of treated liquid-phase samples. It can be seen that residual TOC can be to some extent attributed to the presence of non-oxidized and refractory organic acids (acetic, formic and oxalic acid). The highest amount of carboxylic acids (sum of concentrations of acetic, formic and oxalic acid) was detected in the liquid samples obtained after the

Fig. 5. Concentration of carboxylic acids (acetic, formic and oxalic acid) in endproduct solutions after 60 min irradiation under UV light in the presence of anatase (A), rutile (R), brookite (B), anatase/TiO2 -B (ATB) mixture and anatase/rutile (AR) mixture.

photocatalytic oxidation of BPA using A and R solids (2.16 and 1.23 mg/l), respectively, followed by ATB mixture (0.87 mg/l), AR mixture (0.84 mg/l) and B (0.65 mg/l). One can see that similar amount of acids was produced in the presence of ATB and AR catalysts, despite their different performance concerning the TOC removal. High surface area is essential for achieving high mineralization rates, since the number of surface vacant sites for the production of hydroxyl radicals and oxygen activation is directly connected to the specific surface area of materials. ATB solid, which is rich in surface vacant sites, ensures high TOC removal, leaving out solely a small portion of refractory organic acids. On the other hand, AR sample suffers from low surface area and, therefore, only a small portion of BPA is transformed to organic acids, but not completely mineralized. We believe that BPA is transformed to intermediates such as 4-isopropylphenol and 4-ethylphenol [24]. Regarding the TOC removal over ATB and AR samples, a remarkable difference was observed, which can be attributed to the significant difference in specific surface area and, consequently, amount of surface acid sites (refer to Table 2). Due to the refractory character of acetic and formic acid, they are unlikely to be completely oxidized in the heterogeneous photocatalytic process in the absence of noble metals. However, results presented in Fig. 5 show that oxalic acid is more efficiently degraded in the presence of high specific surface area ATB sample compared to AR sample. Both solids are subjected to acidic conditions during BPA removal (pH of solution, being in the range of 4.8–4.9, is below pHPZC ), giving rise to positively charged surface. At these conditions, oxalic acid (pKa = 1.23) is predominantly in the form of oxalate ion (C2 O4 2− ), which is likely attracted by the positively charged catalyst surface. Since the amount of surface acid sites on ATB sample surpasses these on AR sample, the adsorption of C2 O4 2− on the surface of the former is greatly enhanced, resulting in higher TOC conversion. 3.3. Photocatalytic oxidation over physical mixtures of catalysts In addition to runs carried out in the presence of single-type catalysts, photocatalytic tests were performed in this study using a physical mixture consisting of either two (manatase :mrutile = 1:1, manatase :mbrookite = 1:1, mrutile :mbrookite = 1:1) or three polymorphs (manatase :mbrookite :mrutile = 1:1:1). Total concentration of catalysts was kept equal to 125 mg/l. The obtained BPA and TOC conversions as a function of time are presented in Fig. 6a. Results show that over 90% BPA conversion was achieved using the anatase/brookite

58

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

physical mixture, while physical mixtures of anatase/rutile and rutile/brookite resulted in 80 and 44% BPA conversion, respectively. It is known that composites of brookite and anatase can also suppress the recombination of electron and hole pairs [49]. The photocatalytic examination of anatase/brookite physical mixture clearly pointed out that better charge separation over both solids was attained, due to synergy between particles after sonication in water [13]. Simultaneously, TOC conversion increased from 14.4 (anatase) to 33.8% (anatase/brookite). On the other hand, the TOC conversion of rutile/brookite mixture (2.6%) and anatase/rutile (4.5%) was worsened compared to the TOC removal over pure rutile (12.9%). The physical mixture of all three singular polymorphs resulted in BPA and TOC conversions of 70 and 7.2%, respectively. Anatase is commonly considered as the most active TiO2 polymorph, since its charge carriers have more prolonged lifetimes [12], but anatase/rutile nanocomposites have even higher potential due to the spatial separation of photogenerated charge carriers. Li and Gray [12] reported that the photoactivity of anatase powders cannot be improved by physically mixing with rutile powders, but we argue against this statement. Further, we decided to mimic the behavior of our best performing mixed-phase nanocomposite AR (refer to Fig. 6b). Like previously mentioned, the anatase/rutile ratio in the AR sample is 97/3. The physical mixture, using A and R singular phases, was

prepared according to the anatase/rutile ratio in AR. In this way, after 60 min of irradiation under UV light, we were able to completely convert BPA and achieve 27.1% TOC removal. This even further supports the earlier statement that anatase activity can be enhanced when it is physically mixed with rutile. Also, the amount of rutile present in the physical mixture plays a decisive role in the photocatalytic activity, since A (97%):R (3%) ratio (Fig. 6b) proved to be more effective than A (50%):R (50%) ratio (Fig. 6a), due to higher anatase activity. Since BET surface areas of A and R singular phases are higher than that of AR (40, 32 and 9 m2 /g, respectively), their concentrations in the photoreactor were correspondingly reduced in order to compensate the difference in surface areas (normalized sample in Fig. 6b). The normalized sample enabled 50% BPA removal, but no TOC conversion was obtained. Based on these results we can indisputably confirm that the anatase/rutile physical mixture has a beneficial effect on the removal of BPA, but is still outperformed by the mixed-phase nanocomposite of anatase and rutile (AR sample). Similar results were presented by Ohno et al. [13], where a synergistic effect between anatase and rutile for the photocatalytic oxidation of naphtalene was reported. Synergy can exist when different phases are separated but form agglomerates (like in TiO2 P25 powder). Agglomeration of two or three

Fig. 6. (a) BPA and TOC conversions obtained over physical mixtures of different polymorphs (manatase :mrutile = 1:1, manatase :mbrookite = 1:1, mrutile :mbrookite = 1:1) or all three polymorphs (manatase :mbrookite :mrutile = 1:1:1); (b) BPA and TOC conversions obtained over AR, physical mixture of A (97 wt.%) and R (3 wt.%), and physical mixture of A and R, normalized to BET surface area of AR. Reaction conditions: atmospheric pressure, T: 20 ◦ C, C0 (BPA): 10 mg/l, V0 (BPA): 250 ml, pH0 : 5.2, C (catalyst): 125 mg/l, t(irradiation under UV light) : 60 min.

Fig. 7. (a) BPA conversions obtained over anatase (A), rutile (R), physical mixture of A (50%) and R (50%) determined experimentally, and physical mixture of A (50%) and R (50%) calculated; (b) BPA conversions obtained over AR nanocomposite, physical mixture of A (97%) and R (3%) determined experimentally, and calculated BPA conversion over physical mixture of A (97%) and R (3%). Reaction conditions: atmospheric pressure, T: 20 ◦ C, C0 (BPA): 10 mg/l, V0 (BPA): 250 ml, pH0 : 5.2, C (catalyst): 125 mg/l, t(irradiation under UV light) : 60 min.

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

59

Fig. 8. Proposed mechanism for enhanced formation of hydroxyl radicals over anatase/rutile physical mixture.

respective phases and consecutively synergy between them can be achieved by sonication in water [13]. The synergistic effect between the anatase and rutile phases was further confirmed by calculations presented in Fig. 7. Apparent first-order rate constants k for A and R singular phases were used to calculate apparent first-order rate constants for physical mixtures; consequently, time dependent BPA conversions were determined using Eq. (7). Results presented in Fig. 7a confirm the assumption that the activity of singular A and R phases can indeed be enhanced with physical mixing (or when the phases are immobilized in the nanocomposite – see Fig. 7b), since experimentally determined conversions are higher than the calculated ones. To further address and explain this phenomenon, we propose the mechanism presented in Fig. 8. The contact between anatase and rutile particles (e.g., by collisions or formation of temporary agglomerates in aqueous medium due to intense mixing) induces the transfer of hydroxyl radicals formed on the surface of anatase particles (via H2 O oxidation with photogenerated holes in the valence band) to rutile particles. The latter consequently act as a sink of anatase-generated OH• radicals. In this way, the stripped anatase surface can uptake newly generated OH• radicals leading to higher concentration of these species at the surface of connected dissimilar-phase particles. Both, anatase and rutile particles, are believed to enable the formation of OH• radicals on their surfaces, though the structure of anatase generates sufficiently higher concentrations of surface OH radicals, due to longer lifetimes of charge carriers generated in anatase particles compared to those generated in rutile particles [12]. The prolonged charge carrier lifetime ensures higher concentration of valence band holes at any specific time unit, which, as stated above, are responsible for water oxidation and, consequently, formation of OH• radicals. We can say that the concentration gradient of OH• radicals is responsible for the transport of reactive species to less occupied surface (i.e., rutile particles). For example, if an anatase particle possesses 6 OH• radicals per specific surface area and time, and a rutile particle possesses 3, which could be reasonably supposed on the basis of a higher concentration of anatase photogenerated holes, then an OH• radical is transferred to the surface of adjacent rutile particle. The transferred anatase OH• radical is instantly replaced by a newly formed specie; thus, the overall (considering both particles) specific concentration (2 OH• ) is higher than in separated dissimilar-phase particles (␳1 OH• ). In this case, this would amount to ten OH• radicals, instead of nine for separated particles. The proposed mechanism presented in Fig. 8 explains the higher activity of anatase/rutile physical mixture for the oxidation of water-dissolved BPA, and is more plausible than the transfer of charge carriers over different-phase solvated particles. However, the transfer of anatase generated

electrons to adjacent rutile particles with a lower conduction band was verified by characterizing a dry physical mixture using both UV–vis-DR and PL characterization techniques (Figs. S4 and S5). Yet again, this demonstrates how the discharged anatase surface and enhanced charge separation are essential for efficient formation of OH• radicals and increased oxidation rate of BPA. On the other hand, the synergistic effect could also originate from collided anatase and rutile particles possessing diverse acid-base surface properties. On such occasion, the concentration of surface OH groups can be locally altered, and consequently the formation of OH• radicals. However, according to Table 2, this scenario is less likely to happen, since the surface properties of these two catalysts are too similar. The above presented findings are of great importance regarding the enhancement of catalytic activity. Although anatase/rutile nanocomposites show high activity for the removal of water dissolved organic matter, they often lack of high BET surface area, which is essential for achieving a complete mineralization of pollutants in different environmental applications. Despite of several reports for the preparation of anatase/rutile nanocomposites, little to none exhibit such high activity in the wide variety of applications as TiO2 P25 [50–52]. Since pure high surface area anatase [53,54] and rutile [55] can be easily prepared, the overall catalytic activity can be upgraded by using physical mixtures of singular anatase and rutile nanoparticles. 4. Conclusions To summarize, various titanium dioxide-based catalysts were prepared, as singular and mixed-phase nanoparticles (i.e., anatase, rutile, brookite, anatase/rutile and anatase/TiO2 -B) in order to evaluate the influence of polymorphism and their physicochemical properties (BET surface area, crystallinity, morphology, acidity, etc.) on the removal of bisphenol A in heterogeneous photocatalytic process. It was determined that high crystallinity of catalysts is beneficial, but in order to achieve total mineralization, high surface area is of greater importance. The anatase phase exhibits significantly higher activity for photocatalytic oxidation of BPA, comparing to rutile and brookite, respectively. Activity of anatase can be further enhanced, if it is physically mixed with rutile. When anatase and rutile particles are in contact, the latter acts as a sink of OH• radicals formed on the surface of anatase particle. In this way, the stripped anatase surface can uptake newly generated OH• radicals, leading to higher concentration of these species at the overall surface of connected dissimilar-phase particles, like in the newly proposed mechanism. Physical mixtures of anatase and rutile show that a synergistic effect between the respective phases takes place,

60

R. Kaplan et al. / Applied Catalysis A: General 489 (2015) 51–60

and consequently enhances the overall activity for BPA removal, but they are still surpassed by the mixed-phase nanocomposite of anatase and rutile (when normalized to the BET surface area of individual materials). CHNS elemental analysis of fresh and spent catalysts confirmed that the anatase phase outperformed other polymorphs due to significantly less expressed deposition of carbonaceous species during the photodegradation runs. Similarly, high surface area anatase/TiO2 -B mixture is also prominent, since it shows high BPA and TOC conversions.

[18] [19] [20] [21]

Acknowledgement

[27]

The authors acknowledge the financial support of the Ministry of Education, Science and Sport of the Republic of Slovenia through research program P2-0150. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apcata. 2014.10.018.

[22] [23] [24] [25] [26]

[28] [29] [30] [31] [32] [33] [34] [35] [36]

References [1] K. Hashimoto, H. Irie, A. Fujishima, Jpn. J. Appl. Phys. 44 (2005) 8269–8285. [2] Y. Liao, W. Que, Q. Jia, Y. He, J. Zhang, P. Zhong, J. Mater. Chem. 22 (2012) 7937–7944. [3] A.R. Armstrong, G. Armstrong, J. Canales, P.G. Bruce, Angew. Chem. Int. Ed. 43 (2004) 2286–2288. [4] S. Bakardjieva, V. Stengl, L. Szatmary, J. Subrt, J. Lukac, N. Murafa, D. Niznansky, K. Cizek, J. Jirkovsky, N. Petrova, J. Mater. Chem. 16 (2006) 1709–1716. [5] X. Ding, X. Liu, Mater. Sci. Eng. A 224 (1997) 210–215. [6] R.F. de Farias, J. Colloid Interface Sci. 239 (2001) 584–586. [7] D.A. Hanaor, C.C. Sorrell, J. Mater. Sci. 46 (2011) 855–874. [8] W.W. So, S.B. Park, S.J. Moon, J. Mater. Sci. Lett. 17 (1998) 1219–1222. [9] D. Zhang, Russ. J. Phys. Chem. A 87 (2013) 129–136. [10] R.I. Bickley, T. Gonzales-Carreno, J.S. Lees, L. Palmisano, R.J.D. Tilley, J. Solid State Chem. 92 (1991) 178–190. [11] H. Kominami, Y. Ishii, M. Kohno, S. Konishi, Y. Kera, B. Ohtani, Catal. Lett. 91 (2003) 41–47. [12] G. Li, K.A. Gray, Chem. Phys. 339 (2007) 173–187. [13] T. Ohno, K. Sarukawa, K. Tokieda, M. Matsumura, J. Catal. 203 (2001) 82–86. [14] T. Kawahara, Y. Konishi, H. Tada, N. Tohge, J. Nishii, S. Ito, Angew. Chem. Int. Ed. 41 (2002) 2811–2813. [15] R. Andreozzi, V. Caprio, A. Insola, R. Marotta, Catal. Today 53 (1999) 51–69. ˜ [16] M. Pera-Titus, V. García-Molina, M.A. Banos, J. Giménez, S. Esplugas, Appl. Catal. B: Environ. 147 (2004) 219–256. [17] S. Esplugas, J. Giménez, S. Contreras, E. Pascual, M. Rodríguez, Water Res. 36 (2002) 1034–1042.

[37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55]

M. Klavarioti, D. Mantzavinos, D. Kassinos, Environ. Int. 35 (2009) 402–417. F. Luck, Catal. Today 53 (1999) 81–91. I.A. Balcioglu, Y. Inel, J. Environ. Sci. Health A31 (1996) 123–138. A. Mahyar, M.A. Behnajady, N. Modirshahla, Indian J. Chem. 49A (2010) 1593–1600. S. Min, F. Wang, Y. Han, J. Mater. Sci. 42 (2007) 9966–9972. A. Pintar, M. Besson, P. Gallezot, Appl. Catal. B: Environ. 30 (2001) 123–139. N. Watanabe, S. Horikoshi, H. Kawabe, Y. Sugie, J. Zhao, H. Hidaka, Chemosphere 52 (2003) 851–859. P. Chen, K.G. Linden, D.E. Hinton, S. Kashiwada, E.J. Rosenfeldt, S.W. Kullman, Chemosphere 65 (2006) 1094–1102. T. Yamamoto, A. Yasuhara, H. Shiraishi, O. Nakasugi, Chemosphere 42 (2001) 415–418. ´ A. Pintar, Appl. Catal. B: Environ. 132–133 B. Erjavec, R. Kaplan, P. Djinovic, (2013) 342–352. S. Zhang, C.Y. Liu, Y. Liu, Z.Y. Zhang, L.Y. Mao, Mater. Lett. 63 (2009) 127–129. T.D. Nyguyen-Phan, E.J. Kim, S.H. Hahn, W.J. Kim, E.W. Shin, J. Colloid Interface Sci. 356 (2011) 138–144. K. Prasad, D.V. Pinjari, A.B. Pandit, S.T. Mhaske, Ultrason. Sonochem. 17 (2010) 409–415. V.N. Kuznetsov, N. Serpone, J. Phys. Chem. C 111 (2007) 15277–15288. K.S. Suslick, Ann. Rev. Mater. Sci. 29 (1999) 295–326. T. Hongo, A. Yamazaki, Micropor. Mesopor. Mater. 142 (2011) 316–323. M.V. Koudriachova, J. Nano Res. 11 (2010) 159–164. M.M. Mohamed, B.H.M. Asghar, H.A. Muahten, Catal. Commun. 28 (2012) 58–63. S. Lowell, J.E. Shields, M.A. Thomas, M. Thommes, Characterization of Porous Solids and Powders: Surface Area, Pore Size and Density, 1st ed. 2004, Corr. 2nd printing, 2004, XIV, ed., Kluwer Academic Publishers, 2004. J. Yu, H. Yu, B. Cheng, C. Trapalis, J. Mol. Catal. A: Chem. 249 (2006) 135–142. J. Wang, W. Sun, Z. Zhang, Z. Jiang, X. Wang, R. Xu, R. Li, X. Zhang, J. Colloid Interface Sci. 320 (2008) 202–209. H. Zhang, G.R. Li, L.P. An, T.Y. Yan, X.P. Gao, H.Y. Zhu, J. Phys. Chem. C 111 (2007) 6143. K. Nagaveni, M.S. Hegde, N. Ravishankar, G.N. Subbanna, G. Madras, Langmuir 20 (2004) 2900–2907. J. Zhou, G. Tian, Y. Chen, J.-Q. Wang, X. Cao, Y. Shi, K. Pan, H. Fu, Dalton Trans. 42 (2013) 11242–11251. D. Reyes-Coronado, G. Rodríguez-Gattorno, M.E. Espinosa-Pesquiera, C. Cab, R. de Cross, G. Oskam, Nanotechnology 19 (2008) 145605–145614. J.J. Sene, W.A. Zeltner, M.A. Anderson, J. Phys. Chem. B 107 (2003) 1597–1603. Y.P. Varshni, Physica 34 (1967) 149–154. J. Papp, S. Soled, K. Dwight, A. Wold, Chem. Mater. 6 (1994) 496–500. K.J.A. Raj, B. Viswanathan, Indian J. Chem. 48A (2009) 1378–1382. S. Yang, L. Gao, J. Am. Ceram. Soc. 88 (2005) 968–970. D.C. Hurum, A.G. Agrios, K.A. Gray, J. Phys. Chem. 107 (2003) 4545–4549. J.C. Yu, L. Zhang, J. Yu, Chem. Mater. 14 (2002) 4647–4653. M. Qamar, M. Saquib, M. Munner, Desalination 186 (2005) 255–271. M. Qamar, M. Saquib, M. Munner, Dyes Pigments 65 (2005) 1–9. D.M.A. Alrousan, P.S.M. Dunlop, T.A. McMurray, J.A. Byrne, Water Res. 43 (2009) 47–54. D. Chen, F. Huang, Y-B. Cheng, R.A. Caruso, Adv. Mater. 21 (2009) 2206–2210. S. Patra, C. Davoisne, S. Bruyère, H. Bouyanfif, S. Cassaignon, P.-L. Taberna, F. Sauvage, Part. Part. Syst. Charact. 30 (2013) 1093–1104. X. Jiang, C.J. Brinker, Chem. Commun. 46 (2010) 6123–6125.