Enhanced Photocatalysis from Truncated Octahedral Bipyramids of Anatase TiO2 with Exposed {001}/{101} Facets.

In this study, we develop a new synthetic method to grow anatase TiO2 crystals composed of truncated octahedral bipyramids (TOBs) with exposed {001} and {101} facets by a vapor-solid reaction growth (VSRG) method. The VSRG method employs TiCl4(g) to react with CaO(s)/Ca(OH)2(s) at 823-1043 K under atmospheric pressure. The O-deficient pale-blue TOB TiO2 crystals display high amount of both {001} and {101} facets. Together, they decompose methylene blue photocatalytically under UV-visible (UV-vis) light irradiation. The most-efficient TOB catalyst VT923 (grown at 923 K, average edge length 400 nm, average thickness 200 nm, and surface area 4.20 m2/g) shows a degradation rate constant k, 0.0527 min-1. This is close to that of the P25 standard 0.0577 min-1. However, the surface area of P25 (46.8 m2/g) is about 12 times that of VT923. The extraordinary performance of VT923 is attributed to the presence of high amount of coexisting {001} and {101} facets to form effective surface heterojunctions. They would separate photogenerated electrons and holes effectively on {101} and {001} surfaces, respectively. For VT923, the {001}/{101} ratio is 0.764, which is close to 1, the highest value observed for all TOB samples grown in this study. The surface heterojunctions prolong the electron-hole separation so that VT923 demonstrates the excellent photocatalytic capability. In addition, residual Cl atoms on the exposed faces are easily removed to show clean TiO surface layers with sufficient amount of O-deficient sites in the current samples.

Introduction

Titanium dioxide (TiO2) has long been investigated for its potential use as a photocatalyst to degrade pollutants because of its abundance, nontoxicity, low cost, high photostability, and excellent activity.[1−9] Among three major TiO2 crystal phases, namely, rutile, anatase, and brookite, anatase has demonstrated the most-active photocatalytic performance.[10−12] The activity may be improved further by controlling the following factors. These are the following: (1) crystallinity, (2) doping, (3) generating surface defects, and (4) increasing active surface area. Another important factor is controlling the exposed TiO2 faces. It has been suggested that at atomic level the surface structure of a crystal facet affects the photocatalytic ability greatly.[13,14] In general, a crystal facet with high percentage of unsaturated atomic sites is usually highly reactive. It was demonstrated that due to the presence of high density of unsaturated Ti (Ti5c) and oxygen (O2c) sites on {001} facets, anatase {001} facets show much higher photocatalytic activity than {101} facets do.[15−17] However, under normal circumstances, much less {001} facets are exposed on crystals. This is due to their low thermodynamic stability, as shown by the high average surface energy 0.90 J/m2. The value is significantly higher than that of {101} facets, 0.44 J/m2.[18,19] In addition to the number of active sites of Ti5c and O2c, the {001}/{101} area ratio is important too. It is proposed that the “surface heterojunction” between {101} and {001} facets would assist the transfer and separation of electrons and holes, as shown for the truncated octahedral bipyramidal (TOB) crystal model in Figure .[20−22] It is suggested that the valance band energy of {001} facets is higher than that of {101} facets. Consequently, the photogenerated electrons and holes migrate, respectively, to {001} and {101} facets to assist the oxidation and the reduction reactions during the photocatalytic process. Clearly, by assuming that the oxidation and the reduction rates on exposed {001} and {101} facets, respectively, are comparable, crystals with equal amounts of both facets would show the best catalytic performance. Thus, searching for a route to fabricate TiO2 with equal amounts of exposed {001} and {101} facets has been under intensive investigation. In 2008, microcrystals of anatase TiO2 with 47% {001} and 53% {101} exposed facets were synthesized in HF(aq) under a hydrothermal condition.[23] However, the F atoms introduced into the reaction not only changed the surface energies and the areas of the facets but also adsorbed strongly on the crystal surfaces to decrease the reactivity.[24]

Figure 1

{001} and {101} facets surface heterojunction.

In this experiment, we report the preparation of TOB anatase TiO2 crystals with different amount of {001} and {101} facets via a new vapor–solid reaction growth (VSRG) route. We find that the synthesis conditions and the ratio of {001}/{101} facets affect their photocatalytic capability significantly.

Results and Disscusion

Through a new VSRG route, TiCl4(g) was reacted with solid mixtures of CaO/Ca(OH)2 (60:40 wt %) in a hot-wall rector at 823–1043 K to generate two products, a pale-blue powder and a blue solid. After X-ray diffraction (XRD) analyses (Figure S1 in the Supporting Information), we confirm that they are the anatase- and the rutile-phase TiO2 (Joint Committee of Powder Diffraction Standard, JCPDS No. 21-2172 and 21-1272), respectively.[16] The rutile-phase TiO2 is still under investigation and will be discussed elsewhere. The pale-blue products, T823, T873, T923, and T1043 which grew at 823, 873, 923, and 1024 K, respectively, were characterized by scanning electron microscopy (SEM) and XRD (Figures S2 and S3 in the Supporting Information). The SEM images show that the crystals are 200–800 nm in size. From the XRD analyses, we confirm that all of the products are pure anatase TiO2. With increasing temperature of growth, not only the crystal size enlarges but also the crystallinity increases as well. A summary of the data of the products is listed in Table .

Table 1

Summary of the Samples

sample	% {001} facets	grain size (nm)a	BET (m2/g)	degradation rate k (min–1)
T823	0.15	19 ± 1	b	0.0144
T873	0.28	31 ± 2	5.4 ± 0.1	0.0319
T923	0.43	35 ± 4	4.2 ± 0.1	0.0483
T1043	0.39	39 ± 4	3.6 ± 0.1	0.0370
VT923	0.43	35 ± 3	4.2 ± 0.1	0.0527
AT923	0.43	35 ± 3	4.3 ± 0.1	0.0412
OT923	0.08	29 ± 2	b	0.0370

Calculated by the full width at half-maximum value of the (200) signal in the XRD pattern.

The yield of the reaction is too low to collect enough sample for Brunauer–Emmett–Teller (BET) studies.

Among these, T923 shows a more uniformly distributed crystal size (Figure S2). Thus, T923 is selected for detailed investigations. A representative set of characterization data of T923 is presented in Figure . The SEM images of T923, Figure a,b, shows numerous uniform TOB-shaped crystals. Their widths and thicknesses are estimated to be in the range 400–600 and 200–250 nm, respectively. The energy-dispersive X-ray (EDX) spectrum (Figure c) shows the signals of Ti (31 atom %) and O (68 atom %). The result agrees well with the composition of TiO2. In Figure d, the XRD pattern of T923 displays the reflections corresponding to anatase-phase TiO2. A transmission electron microscopic (TEM) image in Figure e shows a TOB with a width 600 nm. From the high-resolution TEM (HRTEM) image shown in Figure f, the interplanar distance 0.19 nm is observed for the single crystal. This is indexed to the distance between (200) crystal planes of anatase TiO2.[25,26] The selected-area electron diffraction (SAED) shows a dot pattern in Figure g for T923. This confirms that the single crystal sample is indexed to the [001] zone axis of anatase TiO2.

Figure 2

(a) Low- and (b) high-magnification SEM images, (c) EDX, (d) XRD, (e) low- and (f) high-magnification TEM images, and (g) ED of T923.

The pale-blue color of T923 suggests that the sample may be O-deficient.[29] The thermogravimetric analysis (TGA) result presented in Figure S4 in the Supporting Information shows 1.38 wt % increase after the sample was heated in O2(g). From this, the original T923 composition is determined to a formula TiO1.96. T923 was further heated at 773 K in vacuum to remove traces of Ca and Cl atoms incorporated into the product during the growth. Because the oxide did not decompose at this temperature, we assume that the composition of VT923 is still TiO1.96. On the other hand, after T923 was heated at 773 K in air, all of the Ti atoms in the sample AT923 were fully oxidized. Thus, most of the O deficiency is removed in the sample.

The chemical states and binding energies of the surface elements on T923, VT923, and AT923 were investigated by X-ray photoelectron spectra (XPS). As shown in the survey spectra in Figure S5, the sample surfaces display signals from Ti and O atoms majorly. In addition, signals from traces of C, Si, Ca, and Cl elements are also observed. In Figure , their high-resolution XPS are shown. The Ti 2p spectrum of T923 in Figure a shows two major peaks at 458.6 and 464.4 eV from 2p3/2 and 2p1/2 electrons, respectively. They are resolved to major signals for Ti atoms in the 4+ state and minor shoulders at 457.9 and 463.6 eV for the ones in the 3+ state.[28,29] For VT923, the 4+ state 2p3/2 and 2p1/2 electrons are observed at 458.2 and 463.9 eV, while the 3+ electrons are found at 457.1 and 463.3 eV, respectively. These are very similar to the values of the T923 electrons. The integrated intensity ratios of the 3+/4+ electrons of T923 and VT923 are comparable. This suggests that both have similar amounts of Ti3+ atoms. For AT923, the 4+ state 2p3/2 and 2p1/2 electrons shift to low energies of 458.3 and 464.1 eV, respectively. In addition, the intensities of Ti3+ 2p3/2 and 2p1/2 electrons of the sample decrease as well. This indicates that the heat treatment in air removed a significant amount of O deficiency. Figure b shows the XPS signals of O 1s electrons. For T923, two signals are found at 529.5 and 531.2 eV. They are assigned to the O atoms of TiO2 and those of the surface Ti–OH groups.[30,31] After the heat treatment at 773 K, the intensities of the 529.5 eV signals did not change much. Figure c shows the presence of Ca atoms in T923, as suggested by the 2p1/2 and 3p3/2 peaks found at 350.9 and 347.2 eV, respectively.[8,14,32] Also, as displayed in Figure d, Cl 2p1/2 and 3p3/2 peaks are observed at 200.2 and 198.1 eV for T923, respectively. Both Ca and Cl signals diminish significantly in the spectra of VT923 and AT923. The data suggest that both signals were from CaCl2, a reaction byproduct that was evaporated after T923 was heated at 773 K. Other low-level elements, such as C and Si, found in the survey spectra are assigned to adventitious carbon atoms from air and evaporated silicon containing species from the eroded quartz tube reactor.

Figure 3

High-resolution XPS spectra of (a) Ti 2p, (b) O 1s, (c) Ca 2p, and (d) Cl 2p of T923, VT923, and AT923.

The electron paramagnetic resonance (EPR) results of T923 and AT923 are shown in Figure . For T923, a signal at a g value 1.96 is ascribed to the presence of Ti3+–O vacancies in the anatase crystals.[37] Additionally, g values 1.94 and 1.90 are assigned to the O vacancies on the surface and in the subsurface layers of the crystals, respectively.[33,34,35] For AT923, which was annealed in air, all of the signals decrease significantly. Only a weak one from the defects in the bulk still exists.

Figure 4

EPR spectra of T923 and AT923.

In this VSRG process, pale-blue TOB-shaped anatase TiO2 crystals are obtained. We have found that as the reaction temperature increases, the size and crystallinity of the crystals also increase. The growth of TOB TiO2 crystals is affected by several important factors. First of all, we discover that the solid reactant must contain some Ca(OH)2(s). Without this ingredient, no anatase TiO2 would grow in the reaction. We speculate that during the ramping stage, Ca(OH)2(s) dehydrolyzes to generate H2O(g).[27] The vapor reacts with TiCl4(g) to form suitable Ti- and O-containing precursors, which then decompose and grow into TOB TiO2. On the other hand, when pure Ca(OH)2(s) was employed, more H2O(g) would be generated. In this case, only octahedron-shaped TiO2 (OT923, Figure S6) were grown. The second important factor is the amount of TiCl4(g) in the gas phase. The SEM images and the XRD patterns of the products grown at 923 K under TiCl4(g) supply rates of 1.7–5.0 bubbles/s are displayed in Figures S7 and S8 in the Supporting Information. Clearly, T923, grown at 3.2 bubbles/s, shows more homogeneous TOB-shaped crystals.

The TiO2 crystals were investigated for their photocatalytic degradation capabilities of aqueous methylene blue (MB) solutions. The experimental observations of T923 are shown in Figure S9 in the Supporting Information as an example. Before the UV–visible light irradiation, T923 was mixed with MB (25 ppm) and placed in a dark environment for 60 min. Then, a spectrum was recorded every 5 min for 30 min. Combined with the results from other experiments (blank, T823, T873, T1043, and P25), C/C0 and ln(C/C0) are plotted against time and shown in Figure . In the blank experiment, no photocatalyst was added. Also, the result form the commercially supplied P25 is used for comparison. From the results in Figure a, it is clear that the photocatalytic activities of the samples are in the order P25 > T923 > T1043 > T873 > T823. The degradation rate constants k are derived from the data shown in Figure b and listed in Table . Apparently, P25 shows the best catalytic ability, while T923 exhibits the second-best performance. However, the superior performance of P25 is only superficial. When nitrogen adsorption/desorption isotherms displayed in Figure are carefully analyzed by Brunauer–Emmett–Teller (BET) method, T923 presents a surface area 4.2 m2/g, which is only 1/12 of that of P25, 46.8 m2/g. Nonetheless, the degradation rate constant k (0.0483 min–1) of T923 is close to that of P25 (0.0577 min–1). This suggests that the photocatalytic ability of the P25 surface is inferior to that of T923 even though P25 possesses the high surface area and the so-called “phase heterojunction” structure, which could enhance the photocatalytic performance of P25.[36,37] Thus, we conclude that the TOB-shaped TiO2 crystals demonstrated excellent photocatalytic ability as predicted by literature.[13] Below, we will discuss the major factors affecting the performance of the TOB-shaped TiO2. These include areas of the exposed crystal facets and concentration of the defects.

Figure 5

Photodegradation performance of MB by T823, T873, T1043, T923, and P25 (a) C/C0 vs time and (b) ln(C/C0) vs time.

Figure 6

Nitrogen adsorption/desorption isotherms of T923 and P25.

Raman spectroscopy was employed to investigate the exposed facets of TOB-shaped TiO2.[38] In Figure , the spectra show characteristic anatase TiO2 peaks at 144, 394, 514, and 636 cm–1. These peaks are assigned to Eg (symmetric O–Ti–O stretching), B1g (antisymmetric O–Ti–O bending), A1g (symmetric O–Ti–O bending), and Eg (symmetric O–Ti–O stretching) vibrations, respectively. As shown in Figure , all Eg peaks (144 cm–1) of the samples are normalized to the same intensity. With increasing the growth temperature from 823 to 1043 K, the B1g and A1g intensities increase simultaneously.[38] T923 shows the maximum intensities. From the intensity ratios of A1g/Eg (144 cm–1), the exposed {001}/{101} facet ratios (in percentage) of the TOB-shaped TiO2 crystals are determined. The percent {001} facets are estimated and listed in Table .[38]

Figure 7

Raman spectra of TiO2 samples T823, T873, T923, and T1043.

As summarized in Table , Figure displays that the MB degradation rate constants k increase as the percentage of {001} facets on the TOB-shaped crystals increases. As a result, T923, one of the TOB samples not treated further, displays the highest percentage of {001} facets, 42.5%. It also exhibits the highest degradation rate constant k, 4.84 × 10–2 min–1. The exceptional reactivity of the samples correlates well with the high percentage of {001} facets. We propose that the excellent performance originates from the phenomenon of surface heterojunction.[13,14] According to the theory, surface heterojunctions form between adjacent {001} and {101} facets. Thermodynamically, the surface heterojunctions are beneficial to the transfer and separation of photogenerated electrons and holes. As shown in Figure , while the electrons migrate from {001} to {101} facets, the holes tend to migrate the other way, from {101} to {001} facets. Consequently, TOB crystals with high amounts of coexposed {001} and {101} facets show much-enhanced photocatalytic activities. T923 demonstrates the best photocatalytic activity among the untreated samples. Its {001}/{101} ratio is 0.74. The value is the closest to 1 among the untreated samples.

Figure 8

Correlation between degradation rate constant k (black) and {001/101} facet ratio (red) of T823, T873, T923, and T1043.

The second factor that affects the reactivity is the defect concentrations in the TOB crystals. As described above, all untreated samples are pale-blue, indicating that they are O deficient. For example, T923 shows a composition TiO1.96. After T923 was oxidized at 723 K in air for 5 h, a fully oxidized white sample, AT923, was obtained. After analyzing the Raman result of AT923, we conclude that the sample contains 42.6% {001} facets. While this is comparable to that of T923, AT923 has much lower O deficiency. As a result, the rate constant k of AT923 decreases significantly to 4.12 × 10–2 min–1 (Figure S10). In contrast, after T923 was annealed at 723 K under vacuum without oxidation for 5 h, a pale-blue sample, VT923, was obtained. Because of the oxidization-free environment, the amount of O deficiency in VT923 was maintained. The process not only removed the impurities (see Figure ) but also increased the amount of {001} facets to 43.3%. This corresponds to a {101}/{001} ratio 0.764. Consequently, VT923 demonstrates the highest degradation rate constant, 5.27 × 10–2 min–1, among all TOB samples investigated in this study.

In addition to the exposed crystal facets and the concentration of the defects on the catalyst surface, surface area is another important factor that may affect the photocatalyst performance. However, surface area is not the one that determines the performance of the TOB crystals. As mentioned above, T923, which has a much smaller surface area, performs nearly as good as P25 does. As displayed in Figure , the performance of VT923 (specific surface area 4.32 m2, k = 5.27 × 10–2 min–1) is even closer to that of P25 (specific surface area 46.84 m2, k = 5.71 × 10–2 min–1). While our samples have much smaller specific surface areas, the photocatalysis efficiencies are much higher those that of P25. To further confirm this observation, we increase the amount of VT923 in the photocatalysis experiment to 0.1 g. This way, the catalysts would provide a total surface area comparable to that of P25. Results of the photocatalytic degradation of MB are shown in Figure S11. Clearly, with the same surface area, the performance of VT923 exceeds that of P25 greatly.

Last, the crystallinity of the TOB TiO2 may show some effects on their catalytic performance.[39] Using the XRD patterns shown in Figure S4 in the Supporting Information, the averaged grain sizes of the samples are estimated from the Schrrer equation using the full width at half-maxima values of the (200) peaks summarized in Table . As the growth temperature was increased, the grain size increased as well. This indicates that the crystallinity of the TOB particles was raised at high growth temperatures. Nonetheless, the degradation rate constant of T1043 (3.7 × 10–2 min–1) is not the highest one. It is slightly smaller than that of T923 (4.84 × 10–2 min–1). This corresponds well with the BET surface areas of the samples, 4.2 and 3.8 m2, for T923 and T1043, respectively (Figure S12).

Photoluminescence (PL) spectroscopy can provide information related to the electron–hole recombination efficiency. With electron–hole pair recombination after a photocatalyst is irradiated, photons are emitted, resulting in photoluminescence; as the PL signal decreases, the photocatalytic ability will increase.[40,41] As depicted in Figure , holes oxidize H2O molecules to •OH radicals, while electrons reduce O2 molecules to •O2– radicals. Therefore, weaker PL signals suggest that more radicals are generated by holes and electrons. Consequently, the photocatalysis is enhanced. On the other hand, for materials with easy recombination pathways, the PL intensity is strong. In the present case, the surface heterojunctions between {001}/{101} facets can effectively separate the electrons and holes into different faces. In turn, this reduces the rate of their recombinations.[20] Thus, we may use the PL intensities to rationalize the observed photocatalytic performances of the TiO2 samples. In Figure S13, the PL spectra of AT923, VT923, and commercial anatase TiO2 are displayed. Clearly, the PL intensities of the samples are in the following order of decreasing strength: commercial anatase TiO2 > AT923 > VT923. VT923. In general, the sample with more coexposed {001}/{101} facets shows the weakest intensity. In this situation, electrons and holes in VT923 are more effectively separated. Consequently, the probability of their recombination is reduced. In addition, VT923 has more O vacancies that could capture more photoinduced electrons to reduce their recombination with the holes. This additional nonradiative relaxation channel prolongs the life time of the excited states and leads to the high photocatalytic activity of VT923.[42] AT923, with a slightly lower {001}/{101} ratio and less O vacancies, displays a stronger PL intensity. The commercial sample, which does not provide clearly exposed crystal facets, demonstrates the strongest PL signal.

As summarized in Table S1, the rate constants of TOB-shaped TiO2 samples with coexposed {001}/{101} facets reported in literature are compared with those of VT923.[43−50] Clearly, the rate constant of VT923, which indicates its photocatalytic performance, surpasses those of the others. One clear difference between VT923 and the literature samples is their sizes. VT923 is much larger than the other ones reported previously. As a result, VT923 has a much smaller surface area. As discussed previously, the exposed surfaces of VT923 perform photodegradation of MB much more effectively. This much-enhanced performance of VT923 may originate from its synthetic route. All of the literature samples were prepared by hydrothermal methods in the presence of F(aq)– ions. The F(aq)– ions might cap the (001) faces so that the further growth of TiO2 layers was hampered. This shaped the crystals into the observed TOBs.[48] However, due to the strong bonding interaction, the F(aq)– ions capped on the TOB surfaces were difficult to remove. The blockage reduced the number of active surface sites, and consequently, the photocatalytic performances of these samples were degraded. In contrast, the VSRG of VT923 did not involve F-containing reactants. Instead, TiCl4(g) was employed. Thus, only Cl atoms covered the TOB crystal surfaces of VT923 weakly. We speculate that these Cl atoms were easily eliminated so that more exposed active sites were available for the photocatalytic decomposition of MB.

Conclusions

In this study, we successfully grow micrometer-sized TOB-shaped anatase TiO2 crystals using a VSRG route employing TiCl4(g) and Ca(OH)2(s)/CaO(s) as the reactants. The TOB crystals display highly exposed amount of both {001} and {101} facets. Residual Cl atoms on the exposed faces are easily removed to show clean TiO surface layers with sufficient amount of O-deficient sites. Although the surface areas of the samples are low, the TOB crystals show excellent performance toward photocatalytic degradation of MB. The surface area of the best-performed TOB sample, VT923, is only 1/12 of that of P25. Nonetheless, the degradation rate constant k of VT923, 0.0527 min–1, is close to that of P25, 0.0577 min–1. This is attributed to the presence of the extremely efficient surface heterojunctions generated between the exposed {001} and {101} facets. The heterojunctions would separate photogenerated electrons and holes effectively on the {101} and {001} surfaces, respectively. The {001}/{101} ratio of VT923 is 0.764, close to 1. This prolongs the electron–hole separation so that more reactive radicals could be generated. Consequently, the photocatalytic capability of the TOB crystals is enhanced significantly. We anticipate that TOB TiO2 crystals with smaller particle sizes and higher surface areas than those of VT923 might perform photocatalytic degradation of MB better than P25 does. The topic is under investigation.

Experimental Section

Preparation of TOB TiO2 via VSRG

Before the reaction, a hot-wall reactor system composed of a quartz tube (length: 85 cm, diameter: 27 mm) inside a Lindberg tubular furnace was heated at 1273 K and 10–5 Pa overnight. After the apparatus was loaded with a mixture of CaO(s)/Ca(OH)2(s) (0.5 g, 60:40 wt %, determined by TGA shown in Figure S14 in the Supporting Information) at room temperature, it was heated to 823–1043 K. Then, TiCl4(l) (Fluka 98%, 1.84 mL) was evaporated under 1.03 × 105 Pa with the assistance of a stream of Ar(g) at 1.7–5.0 bubbles/s for 20 h to react with the solid. The original solid reactants turned blue. The blue solids were identified preliminarily as rutile-phase crystals and are still under investigation. A full report on this will be discussed elsewhere. In addition, pale-blue powders T823, T873, T923, and T1043 (ca. 0.1 g) were collected from the reactor wall. They were generated from the reactions carried out at 823, 873, 923, and 1043 K, respectively. T923 was further processed at 773 K in vacuum for 5 h to remove traces of impurities in the pale-blue product. The as-processed sample is named VT923. In addition, T923 was annealed in air at 773 K for 5 h to remove the O deficiency. After the treatment, the white product AT923 was collected. Photocatalytic activities of the samples were examined.

Characterizations

All samples were verified using a powder XRD (Bruker AXS D8 Advance) with Cu Kα1 radiation (40 kV, 20 mA). SEM images and EDX spectra were acquired with a JEOL JSM-7401F operated at 15 keV. TEM, SAED, HRTEM images, and EDX data were acquired on a JEOL JEM-200 FTM at 200 kV. XPS data were measured by a ULVAC-PHI (PHI 5000 Versaprobe) Quantera SXM/Auger spectrometer. A Renishaw Raman spectrometer using a 632.8 nm laser was used to characterize the vibrational information of the samples. UV–vis diffuse reflectance spectra of the samples were recorded using a Hitachi UV–vis 3010 spectrophotometer. BET surface area measurements were performed on a Micrometrics ASAP 2020. EPR spectra were recorded using a JES-FA200 spectrometer. The PL spectra were recorded with a Jobin-Yvon Spex Fluoro-3 spectrofluorometer equipped with a 450 W Xe lamp. The data were analyzed by a Jobin-Yvon spectrometer HR460 with multichannel charge-coupled device detector. TGA studies were performed with a NETZSCH STA 409PC.

Photocatalytic Degradation of MB under UV–Vis Light Irradiation

The performance of TiO2 as a photocatalyst to degrade MB was evaluated under the illumination of a Xe lamp (XBO R 101 W, OSRAM). It was placed 10 cm above the surface of the sample composed of a MB solution (7.81 × 10–5 M, 80 mL). Before the photodegradation, 10 mg of TiO2 (10.0 mg, 1.25 × 10–1 mmol) was mixed with the solution. After the mixture was ultrasonicated in the dark at 293 K for 25 min, it was settled for 40 min to ensure that the complete adsorption/desorption equilibrium was reached before the illumination. After the photodegradation started, a portion of the MB solution (1.0 mL) was extracted to mix with deionised water (1.5 mL) every 5 min to make a sample solution for further UV–vis spectroscopic analysis. The maximum UV–vis absorbance of each sample was measured.