Enhanced UV-Visible Light Photocatalytic Activity by Constructing Appropriate Heterostructures between Mesopore TiO2 Nanospheres and Sn3O4 Nanoparticles

Novel TiO2/Sn3O4 heterostructure photocatalysts were ingeniously synthesized via a scalable two-step method. The impressive photocatalytic abilities of the TiO2/Sn3O4 sphere nanocomposites were validated by the degradation test of methyl orange and •OH trapping photoluminescence experiments under ultraviolet (UV) and visible light irradiation, respectively. Especially under the visible light, the TiO2/Sn3O4 nanocomposites demonstrated a superb photocatalytic activity, with 81.2% of methyl orange (MO) decomposed at 30 min after irradiation, which greatly exceeded that of the P25 (13.4%), TiO2 (0.5%) and pure Sn3O4 (59.1%) nanostructures. This enhanced photocatalytic performance could be attributed to the mesopore induced by the monodispersed TiO2 cores that supply sufficient surface areas and accessibility to reactant molecules. This exquisite hetero-architecture facilitates extended UV-visible absorption and efficient photoexcited charge carrier separation.


Introduction
As a stable, low-cost and environmentally benign material, nanoscaled titanium dioxide (TiO 2 ) with unique structural and functional properties has become a widely used semiconductor photocatalyst for various solar-driven clean energy technologies [1]. Tailoring the morphology of TiO 2 photoanode is a preferred route to achieving high performance in solar cells due to its enhanced properties, such as high surface area, faster electron transport, lower electron-hole recombination rate and good light-harvesting features [2,3]. Nevertheless, the wide optical bandgap of TiO 2 , which seriously limits its light harvesting capability, leaving about 96% of the solar light energy wasted [4]. Compared with the solution of generating donor or acceptor states in the band gap by adding impurities, rationally designing and constructing the surface heterostructures would be a more efficient strategy for achieving an excellent photocatalyst [5].
Recently, Sn 3 O 4 , a novel non-stoichiometric oxide, has raised particular interest in the field of photocatalysis-especially in terms of its catalytic behavior under visible light irradiation-due to a suitable band-gap inside visible light (2.2-2.9 eV) and a distinct surface structure composed of both valences of tin [6,7]. Several studies have shown its great potential as an auspicious photocatalyst under visible light, both for generating hydrogen and degrading dyes [8]. However, some drawbacks have hindered its performance in practice. As a semiconductor with a relatively narrow band gap, pure Sn 3 O 4 generally leads to a fast recombination of photoexcited electron-hole pairs, which ultimately decreases its degradation rate [9].

Synthesis of Samples
In this research, we propose a two-step synthesis method to obtain TiO 2 /Sn 3 O 4 nanocomposites by preparing TiO 2 core via sol-gel route first and then synthesizing Sn 3 O 4 on the surface of TiO 2 .
(1) Synthesis of core TiO 2 nanospheres [19]: 3.5 mL tetrabutyl titanate was dissolved in the 50 mL ethylene glycol while stirring vigorously for 10 h. Then, the mixture was immediately poured into a solution containing 170 mL acetone and 2.7 mL deionized water under constant stirring, until white precipitation appeared. The acquired precipitate was calcined in air at 500 • C for 1 h to produce the TiO 2 powders; (2) Coating TiO 2 with Sn 3 O 4 : 0.2 g of the TiO 2 product described above and 5.0 mmol SnCl 2 ·2H 2 O were mixed with 25 mL deionized water, followed by the addition of 12.5 mmol Na 3 C 6 H 5 O 7 ·2H 2 O and 2.5 mmol NaOH under magnetic stirring. During this process, Sn(II) ions were attached to the surface of hydroxyl-rich TiO 2 spherical colloids through inorganic grafting. The resulting precursor was then transferred to a 50 mL Teflon-lined stainless autoclave and maintained at 180 • C for 12 h. Finally, the collected powder was washed several times with deionized water and ethanol, and dried at 60 • C for 12 h.

Characterization of Samples
In order to obtain the physical and chemical properties of as-prepared samples, several characterizations were conducted. X-ray diffraction (XRD) patterns were recorded by a Rigaku D/MAX-2500 diffractometer with Cu Kα radiation (Rigaku, Tokyo, Japan). Raman spectra were obtained using a HORIBA HR800 spectrometer with an Nd:YAG laser at a wavelength of 532 nm (Horiba Yvon, Paris, France). Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images were acquired from the ZEISS SUPRA 55 (Zeiss, Oberkochen, Germany)and JEOL JEM-2010 (JEOL, Tokyo, Japan), respectively. X-ray photoelectron spectra (XPS) were recorded on a scanning X-ray microprobe PHI Quantera II (Ulvac-PHI, Chigasaki, Japan). The nitrogen adsorption-desorption isotherm was measured at 77 K on an Autosorb-iQ2-MP analyzer (Quantachrome Instruments, Boynton Beach, FL, USA). The absorption spectra were carried out by UV-visible spectrophotometer (Lambda 950, Perkin-Elmer, Shelton, WA, USA) and the hydroxyl radicals (•OH) trapping photoluminescence spectra were examined by a fluorescence spectrophotometer (Hitachi F-4500, Hitachi, Tokyo, Japan) using excitation a wavelength of 315 nm.

Photocatalytic Experiments
The photocatalytic activity of TiO 2 /Sn 3 O 4 nanocomposites was evaluated via methyl orange (MO) degradation rate. 80 mL aqueous suspension of MO (20 mg/L) and 80 mg of photocatalyst powder were placed in a 100 mL beaker. Prior to irradiating, the suspensions were magnetically stirred in the dark for 40 min to establish adsorption-desorption equilibrium. A 250 W mercury lamp and a 500 W Halogen lamp with a 420 nm cut-off filter were used as the UV and visible light sources, respectively. After given irradiation time intervals, aliquots of the mixed solution were collected and centrifuged to remove the catalyst particulates for analysis. Four consecutive cycles were tested. The samples were washed thoroughly with water and dried after each cycle.
Using terephthalic acid as a probe molecule, the hydroxyl radicals (•OH) at the photo-illuminated sample/water interface were examined by a special photoluminescence (PL) technique. Terephthalic acid reacts readily with •OH, producing 2-hydroxyterephthalic acid, a great fluorescent material with a unique photoluminescence peak at 426 nm [20], which makes it easy to detect by fluorescence spectrum (excitation wavelength: 315 nm, fluorescence peak: 426 nm). In a typical experiment, 80 mL 0.5 mM terephthalic acid and 2 mM NaOH aqueous solution were completely mixed and then transferred into a 100 mL beaker. The rest of steps are the same as for the degradation of MO. Figure 1a illustrates the XRD pattern of the TiO 2 /Sn 3 O 4 nanocomposites. All the diffraction peaks can be indexed to the anatase TiO 2 (JCPDS 21-1272, marked with black •) and triclinic Sn 3 O 4 (JCPDS 16-0737, marked with red $), validating the high purity of the synthesized TiO 2 /Sn 3 O 4 composite phase. Compared with the single-phase TiO 2 and Sn 3 O 4 , all the diffraction peaks are broad and weak, which indicates that the crystallinities are slightly reduced [21]. This result may be attributed to lattice distortion induced by interfacial strain because of different lattice parameters between Sn 3 O 4 and TiO 2 [22]. In addition, Raman spectroscopy results (Figure 1b) further confirm the purity of the synthesized TiO 2 /Sn 3 O 4 composite phase. Specifically, Raman activities of 144, 196, 396, 520 and 638 cm −1 were assigned to the anatase TiO 2 [23], and the 133, 143, 170 and 238 cm −1 Raman peaks could be attributed to the Sn 3 O 4 , in accordance with previous reports [7,24]. The textures of the as-synthesized TiO 2 /Sn 3 O 4 , TiO 2 , Sn 3 O 4 and P25 were characterized by N 2 physisorption experiments. The Brunauer-Emmett-Teller (BET) surface area data of samples are provided in Table 1. The N 2 adsorption-desorption isotherm and pore-size distribution of TiO 2 /Sn 3 O 4 nanocomposites are shown in Figure 1c. The results display that the TiO 2 /Sn 3 O 4 nanocomposites possess an average pore diameter of 2.733 nm and a larger surface area of 68.1 m 2 /g than the as-prepared TiO 2 (0.04 m 2 /g), Sn 3 O 4 (35.2 m 2 /g), and the reported TiO 2 /Sn 3 O 4 nanobelt heterostructure (51.5 m 2 /g) [9]. Such a high surface-to-volume ratio for the TiO 2 /Sn 3 O 4 nanocomposites might be of extreme good value in photocatalytic processes, as they would provide more active sites for the adsorption of reactant molecules, and their optical absorbance would increase at visible wavelengths [25]. Figure 1d shows the UV-visible diffusion reflectance spectra (DRS) and plots of (F(R)hv) 1/2 versus photo energy (hv) of the TiO 2 /Sn 3 O 4 along with spectra of the pristine TiO 2 and Sn 3 O 4 for comparison. The absorption spectra of the TiO 2 /Sn 3 O 4 nanocomposites exhibit the mixed absorption properties of both the components. In particular, the absorption edge for TiO 2 /Sn 3 O 4 nanocomposites is clearly shifted towards visible region (near 505 nm). The optical band gap determined from the plot of the Kubelka-Munk function was found to be 2.46 eV, compared to the observed values of 3.22 eV and 2.61 eV for TiO 2 and Sn 3 O 4 , respectively. These data reveal the Sn 3 O 4 /TiO 2 nanocomposites have a lower band gap than the pure Sn 3 O 4 and TiO 2 nanoparticles, which is consistent with the published literature [9], and can be explained by the reduced crystallinity of both materials [1], as shown by XRD analysis.  The chemical composition and valence state were characterized by X-ray photoelectron spectroscopy (XPS). The full range of XPS spectra, ranging from 0 to 1000 eV, of TiO 2 /Sn 3 O 4 nanocomposites are shown in Figure 2a. No impurities were observed in the spectra, which is consistent with the results of XRD and Raman. Figure 2b shows the curve fitting data of the Sn 3d core-level spectra. Moreover, the Sn 3d doublet characterized by Sn 3d 3/2 -Sn 3d 5/2 splitting peak can be clearly observed. The prominent peak of Sn 3d 5/2 level is dissolved into two peaks centered at 486.77 and 486.15 eV, which can be attributed to Sn(IV) and Sn(II) configurations [26], respectively. The Sn 3d 3/2 spectra exhibit two peaks at 495.14 and 494.51 eV, which are assigned to Sn 2+ and Sn 4+ [26]. As shown in Figure 2c, the binding energies (BE) of Ti 2p 3/2 and Ti 2p 1/2 are 458.5 and 464.2 eV respectively, which are ascribed to the Ti 4+ oxidation states [27]. On the basis of the above discussion, it can be concluded that the TiO 2 /Sn 3 O 4 sample is composed of Ti(IV), Sn(II and IV), and O, which is in good agreement with the XRD and Raman results. In addition, the calculated Ti/Sn ratio is 0.20, indicating that most of the surface of the TiO 2 nanocrystals is covered by Sn 3 O 4 nanocrystals (SEM and TEM experiments further confirm this result, and will be discussed later). spectra (DRS) and plots of (F(R)hv) 1/2 versus photo energy (hv) of the TiO2/Sn3O4 along with spectra of the pristine TiO2 and Sn3O4 for comparison. The absorption spectra of the TiO2/Sn3O4 nanocomposites exhibit the mixed absorption properties of both the components. In particular, the absorption edge for TiO2/Sn3O4 nanocomposites is clearly shifted towards visible region (near 505 nm). The optical band gap determined from the plot of the Kubelka-Munk function was found to be 2.46 eV, compared to the observed values of 3.22 eV and 2.61 eV for TiO2 and Sn3O4, respectively. These data reveal the Sn3O4/TiO2 nanocomposites have a lower band gap than the pure Sn3O4 and TiO2 nanoparticles, which is consistent with the published literature [9], and can be explained by the reduced crystallinity of both materials [1], as shown by XRD analysis. The morphology and microstructure of the as-prepared TiO 2 /Sn 3 O 4 nanocomposites were carefully analyzed by microscopy. Generally, lots of interleaved Sn 3 O 4 nanoplates are able to self-assemble into an ordinary flower-like nanostructure, as shown in Figure 3a. Similarly, by introducing highly monodispersed TiO 2 nanospheres of~130 nm diameter (Figure 3b) into the growth environment, the homogeneous Sn 3 O 4 nanoparticles started to grow on the surface of each individual TiO 2 core with intimate contact, thus forming an interface of two different semiconductors that would facilitate photo-excited electron transfer and photon-generated carrier separation (Figure 3c,d). Noticeably, the composites inherit a favorable dispersion in the solution, and fully contact with the absorbate, which is positive for the outstanding photocatalytic performance. However, it should be noted that the TiO 2 /Sn 3 O 4 nanocomposites are not completely covered by Sn 3 O 4 nanocrystals. Furthermore, these advantageous heterojunctions with 10-20 nm sizes wrapping uniformly onto the surface of TiO 2 nanospheres were verified by TEM observation, as shown in Figure 1e    The photocatalytic activities of P25, TiO 2 , Sn 3 O 4 and TiO 2 /Sn 3 O 4 heterostructures were evaluated by the degradation of MO in water under UV-and visible-light irradiation (Figure 4a,b).The degradation of the MO solution under identical experimental conditions, but with no photocatalyst, is provided for comparison. The degradation efficiency of the as-synthesized TiO 2 /Sn 3 O 4 heterostructures was defined as C/C 0 , where C 0 is the initial concentration of MO after equilibrium adsorption, and C is the concentration during the reaction. Both blank experiment results showed that MO could not be decomposed without photocatalyst under UV-or visible-light irradiation. In contrast, the photodegradation efficiency of TiO 2 /Sn 3 O 4 nanocomposites was 95% within 30 min under UV-light irradiation, which is superior to the as-prepared TiO 2 nanospheres (62%) and Sn 3 O 4 nanoplates (70%). Furthermore, the MO decomposition efficiency found for the TiO 2 /Sn 3 O 4 photocatalyst was comparable to that determined under the same experimental conditions for the reference P25 catalyst; that is, 99% after 10 min. Additionally, in the visible-light irradiation experiment (Figure 4b   For a better understanding, the photocatalytic kinetics of the samples was analyzed using the Langmuir-Hinshelwood model, as shown in Figure 4c,d. All of the data follow a first-order reaction model, and the calculated apparent kinetic rate constants (κ) are summarized in Table 1. We found that, under UV irradiation, the TiO2/Sn3O4 exhibited a much faster photo-decomposition activity (κ = 0.24 min −1 ) than the pure TiO2 (0.028 min −1 ) and Sn3O4 (0.064 min −1 ), and was as fast as the P25 (0.24 min −1 ). Furthermore, under visible-light irradiation, the calculated value of κ for the TiO2/Sn3O4 sample (κ = 0.052 min −1 ) was twice as high as that for the neat Sn3O4 (0.024 min −1 ), and more than a For a better understanding, the photocatalytic kinetics of the samples was analyzed using the Langmuir-Hinshelwood model, as shown in Figure 4c,d. All of the data follow a first-order reaction model, and the calculated apparent kinetic rate constants (κ) are summarized in Table 1. We found that, under UV irradiation, the TiO 2 /Sn 3 O 4 exhibited a much faster photo-decomposition activity (κ = 0.24 min −1 ) than the pure TiO 2 (0.028 min −1 ) and Sn 3 O 4 (0.064 min −1 ), and was as fast as the P25 (0.24 min −1 ). Furthermore, under visible-light irradiation, the calculated value of κ for the TiO 2 /Sn 3 O 4 sample (κ = 0.052 min −1 ) was twice as high as that for the neat Sn 3 O 4 (0.024 min −1 ), and more than a dozen times higher than that for the single TiO 2 (0.0010 min −1 ) and the P25 (κ = 0.0023 min −1 ). In addition, the TiO 2 /Sn 3 O 4 heterostructures could be recycled and reused at least four times without significant loss of efficiency (Figure 4e,f), which demonstrates its great potential as an efficient and Nanomaterials 2017, 7, 336 8 of 11 stable photocatalytic material. These remarkably good performances can be attributed to the improved UV-and visible-light absorption efficiency, and the high photo-excited carrier-separation rate resulting from the novel TiO 2 /Sn 3 O 4 heterostructures.

Results and Discussion
Based on all of the results above, a possible mechanism for charge transfer and photocatalytic process can be proposed (Scheme 1). As illustrated in Figure 1d, the diffusion reflectance spectra (DRS) and plots of (F(R)hv) 1/2 versus photo energy (hv) indicate that the bandgap of Sn 3 O 4 (2.61 eV) is smaller than that of TiO 2 (3.22 eV). Additionally, the potentials of the valence band (VB) and conduction band (CB) of Sn 3 O 4 are higher than those of TiO 2 , so the heterostructure of TiO 2 /Sn 3 O 4 belongs to typical type-II heterojunction [9]. When Sn 3 O 4 contacts TiO 2 cores to form a heterojunction, the difference in chemical potential causes band bending at the interface of the junction [28], which drives photoexcited electrons to transfer from Sn 3 O 4 to TiO 2 , and photoexcited holes to migrate in the opposite direction, until the Fermi levels of TiO 2 and Sn 3 O 4 reach equilibrium. The possible mechanisms for charge transfer and hydroxyl radical (•OH) generation under UV-and visible-light irradiation will be discussed separately. (1) Upon UV illumination, electrons in the VB could be excited to the CB of both oxides, simultaneously forming the same number of holes in the VB. This is due to the fact that the Sn 3 O 4 nanoparticles were not fully coated as a shell onto the TiO 2 nanospheres (Figure 3c,d) and that the suitable bandgap of TiO 2 and Sn 3 O 4 is lower than the energy of ultraviolet photons. Next, the photo-generated electrons were collected by the TiO 2 particles and the holes by the Sn 3 O 4 particles; that is, electrons transferred from Sn 3 O 4 to TiO 2 , and holes migrated from TiO 2 to Sn 3 O 4 (compare Scheme 1a with Figure 1d). The unique behavior that electrons and holes preferentially accumulate on different materials would result in a great separation of photo-generated carriers, and thus reduce the charge recombination rate, ultimately increasing carrier lifetime. As a consequence, the formation efficiency of hydroxyl radicals (•OH)-a strong oxidant for most pollutants [9,20,29]-by the reaction of holes with surface hydroxyl groups or physisorbed water molecules at the Sn 3 O 4 surface and the production rates of •OH and superoxide radicals (O 2 − ) radicals resulting from the reactions of electrons with dissolved oxygen molecules and water molecules will be massively enhanced; this will increase the volume of oxidant inside the system. (2)  The photocatalytic oxidation of dyes occurs through the reactive species, which came into being after the light absorption and electron-hole formation by the photocatalyst [30]. Terephthalic acid photoluminescence probing technique (TAPL) was employed to examine the generation of active •OH radicals [31]. Figure 5a,b gives the •OH-trapping photoluminescent spectra of TiO 2 /Sn 3 O 4 nanocomposites in TA solution with UV-and visible-light irradiation, respectively. The increased photoluminescence intensity confirms that the •OH radicals are mainly responsible for the photodegradation process, and it also verifies the photocatalytic activity of the TiO 2 /Sn 3 O 4 nanocomposites.

Conclusions
In summary, this study demonstrates a facile route to synthesizing TiO 2 /Sn 3 O 4 nanocomposites that not only display enhanced photocatalytic performance in UV irradiation, but also allow a significant level of visible light photocatalytic activity. The large surface area derived from the monodispersed mesopore TiO 2 /Sn 3 O 4 nanospheres and unique TiO 2 /Sn 3 O 4 heterojunctions are considered to be major contributions to supplying abundant active sites and separating photogenerated carriers, respectively. The strengthened photocatalytic performances will greatly promote the practical application of the TiO 2 /Sn 3 O 4 nanocomposites in eliminating organic pollutants from wastewater, and producing hydrogen by splitting.