As a promising photocatalyst, TiO2 has been widely investigated for the photocatalytic degradation of hazardous pollutants from wastewater or air because of its high oxidative power, photostability, and nontoxicity [1, 2]. However, the widespread, practical use of TiO2 has been restricted because of its limited photocatalytic efficiency. To address this issue, various modification methods have been developed, such as controllable synthesis of various nanostructures [3, 4, 5, 6], doping with different ions [7, 8, 9, 10], sensitization by absorbed molecules or quantum dots [11], coupling with different band-gap semiconductors [12, 13], and modification by noble metals, transition metal ions, or oxides [14, 15, 16, 17, 18, 19, 20]. Amongst them, the development of TiO2 hollow nanostructures has been considered one of the most important strategies for improving the photocatalytic performance because TiO2 with well-defined hollow structures can present the integrated advantages of high specific surface area, low density, good surface permeability, and strong light-harvesting capacity, and thus lead to superior photocatalytic performance [21, 22, 23, 24, 25]. For example, it was reported that hollow TiO2 spheres with unique urchin-like morphology and tunable interior structure exhibited greatly enhanced photocatalytic activity, which was attributed to multiple reflections of ultraviolet light within the sphere interior voids [26]. Furthermore, it was also shown by Yu et al. [27] that the photocatalytic activity of anatase TiO2 hollow spheres was higher than that of TiO2 nanoparticles because of the high surface area (174 m2/g) and unusual hierarchical nanoporous structure, which allowed more effective transport of the objective molecules to the active sites on the framework shells. Recently, Yin et al. [28] presented mesoporous hollow TiO2 shells that show significantly enhanced photocatalytic activity toward the degradation of organic molecules because of small anatase grains of controllable crystallinity, high surface area, and excellent dispersity in water. Even so, it is still highly desirable to further improve the photocatalytic performance of excellent TiO2 photocatalysts with well-defined hollow structure from the viewpoint of practical applications.
In this context, modifying hollow TiO2 nanomaterials with noble metals such as Au and Ag is a meaningful and efficient strategy because they can act as electron sinks to effectively transfer the photogenerated electrons of TiO2, thus improving the corresponding photocatalytic performance. To this end, Yu et al. [29] prepared Ag-TiO2 composite hollow spheres using a microwave-hydrothermal strategy followed by a photochemical reduction process. Do et al. [30] synthesized a novel Au/TiO2 hollow nanostructured photocatalyst via surface coating using a layer-by-layer technology, then the adsorption of Au precursors, followed by high-temperature calcination and finally SiO2 etching. In addition, Ag/TiO2 nanotubes have been obtained in which TiO2 nanotubes were fabricated by an alkaline hydrothermal method followed by an annealing process; then, a catechol derivative, 3-(3,4-dihydroxyphenyl)propionic acid (diHPP), was employed as both linker and reducer of Ag+ to Ag nanoparticles [31]. These above-mentioned photocatalysts exhibited excellent photocatalytic activity. However, these approaches are usually complicated and employ multiple steps in the preparation of noble metal-modified TiO2 hollow nanomaterials. Therefore, it is crucial and a great challenge to develop facile and versatile strategies to synthesize them, while related reports are very rare.
In this paper, a facile template-induced synthesis method was first employed to prepare Ag/TiO2 hollow octahedra using Ag2O octahedra as templates and TiF4 as the precursor at 60 °C. The novel synthetic strategy is described as follows. First, the TiO2 nanoparticles are coated on the surface of Ag2O templates to form the shells of Ag/TiO2 hollow octahedra based on template-directed deposition. Simultaneously, the Ag2O templates can be in situ removed by dissolving the Ag2O octahedra with HF solution produced via the hydrolysis reaction of TiF4 in the reaction system. Furthermore, a rapid reaction can be expected because of the weak basicity of Ag2O, which can facilitate the hydrolysis rate of TiF4 in turn. More specifically, Ag nanoparticles can also be simultaneously deposited on the inside and outside surface of the TiO2 shell by effectively using the photosensitive properties of Ag2O and Ag+ ions under light irradiation, along with the formation of TiO2 hollow octahedra. This work may also provide new insights into preparing other noble metal-modified and hollow nanostructured photocatalysts.
All reagents were of analytical grade, supplied by Shanghai Chemical Reagent Ltd. (PR China), and used as received without further purification.
Octahedral Ag2O particles were synthesized by a wet chemical method according to a previous report [32]. In a typical procedure, 1.865 mL of H2O, 1.5 mL of NH4NO3 (0.1 mol/L), and 1.5 mL of AgNO3 (0.1 mol/L) were sequentially added to a glass vial, and then 135 µL of NaOH solution (2.0 mol/L) was added dropwise. The mixture was sonicated for 10 s and kept at 35 °C for 30 min. Subsequently, 1 mL of NaOH solution (1.5 mol/L) was added to the mixed solution with magnetic stirring, and its color immediately turned from colorless to brown in the vial. After the solution was stirred for 5 min, the resulting precipitation was centrifuged and washed with distilled water and absolute ethanol three times, respectively. Finally, the Ag2O sample was incubated in 10 mL of H2O overnight in the dark for further use.
The as-prepared Ag2O (0.075 mmol) was redispersed in 9 mL of water by sonication for 30 s. The suspension was heated to 60 °C with stirring; subsequently, 1 mL of TiF4 aqueous solution with various concentrations (0.01, 0.05, 0.1, or 0.3 mol/L) was immediately injected into the suspension. After the reaction continued for 30 min under the irradiation of a fluorescent lamp, the resultant suspension was filtered and washed with distilled water three times. The as-obtained samples were dried at 60 °C for 6 h. To improve the crystallinity, these samples were finally calcined at various temperatures (400, 600, or 800 °C) for 1 h.
Morphological analysis was performed using a Quanta x50 field-emission scanning electron microscope (FE-SEM, FEI, USA) and a JEM-2100F transmission electron microscope (TEM, JEOL, Japan). X-ray diffraction (XRD) data were collected on a Rigaku Ultima III X-ray diffractometer (Japan). Ultraviolet-visible (UV-Vis) analysis was performed on a UV-2450 UV-visible spectrophotometer (Shimadzu, Japan).
The evaluation of the activity of the prepared samples for the photocatalytic decolorization of methyl orange (MO) aqueous solution was performed at ambient temperature. First, 0.05 g of the prepared sample was dispersed into 10 mL of MO solution (15 mg/L) in a culture dish with a diameter of ca. 5 cm. For the evaluation of UV-light photocatalytic activity, a 4-W 365-nm UV lamp (Shenzhen LAMPLIC Science Co., China) was used as a light source. The average light intensity striking the surface of the reaction solution was approximately 20 mW/cm2, as measured by a UV meter (made in the photoelectric instrument factory of Beijing Normal University), with a peak intensity of 365 nm. The concentration of MO was determined by a UV-vis spectrophotometer (UV-1240, Shimadzu, Japan). After UV-light irradiation for some time, the reaction solution was centrifuged to measure the concentration of MO. As for the MO aqueous solution with a low concentration, its photocatalytic decomposition is a pseudo-first-order reaction and its kinetics may be expressed as ln(c0/c) = kt, where k is the apparent rate constant, and c0 and c are the MO concentrations at the initial state and after irradiation for t min, respectively [33].
Ag2O octahedron templates were obtained by a simple wet chemical method, and their corresponding morphologies were confirmed by FESEM and TEM. As shown in Fig. 1(a), the Ag2O templates present well-defined and uniform octahedral morphologies, with main edge lengths of 800-1000 nm. The inset is a TEM image of a representative Ag2O octahedron particle, which exhibits a hexagonal structure. It is well known that the hexagon is one of the typical projections of an octahedron. Furthermore, the strong contrast in the TEM image indicates that Ag2O particles are solid. The crystal structures of these Ag2O octahedra can be analyzed using XRD characterization. The XRD results (Fig. 2) clearly show that all of the diffraction peaks of these Ag2O samples could be indexed to the body-centered cubic structure of Ag2O (JCPDS 41-1104), and no diffraction peaks from impurities were detected. In addition, the intensity ratios of the (111) and (200) diffraction peaks for Ag2O octahedra were obviously larger than that of the standard pattern, which can be ascribed to the formation of Ag2O octahedra with predominately (111) planes.
The prepared Ag2O octahedra can be used as an effective template to prepare TiO2 hollow octahedra. The typical Ag/TiO2 hollow octahedra were obtained via the reaction of Ag2O and TiF4 aqueous solution (0.05 mol/L) at 60 °C for 30 min. Fig. 1(b) shows an SEM image of the as-obtained Ag/TiO2 sample, which reveals that the sample has nearly identical morphologies to the Ag2O templates. More specifically, the as-obtained octahedral samples have main edge lengths ranging from 800 to 1200 nm, which are slightly larger than that of the Ag2O templates. This result suggests that TiO2 particles are possibly deposited on the surface of Ag2O octahedra during the initial hydrolysis of TiF4. The inset shows the morphology of a broken Ag/TiO2 octahedron, which indicates that the as-prepared Ag/TiO2 samples are hollow and the thickness of the Ag/TiO2 shell is ca. 25 nm. The TEM image in Fig. 1(c) shows the hollow octahedral structures, where the electron beams can easily penetrate the thin shells of the Ag/TiO2 hollow octahedra. The TEM image in Fig. 1(d) further reveals that the shell thickness of Ag/TiO2 hollow octahedra is ca. 25 nm, which is consistent with that shown in Fig. 1(b). More importantly, a number of small nanoparticles with sizes of 5-10 nm are homogeneously dispersed on the inside and outside surface of the Ag/TiO2 hollow octahedron, as shown in the red circles. To determine the components of these small nanoparticles, a high-resolution TEM image is shown in Fig. 1(e). The spacing of adjacent lattice planes (shown in the corresponding red circle) is ca. 0.24 nm, which can be ascribed to the interplanar spacing of the (111) planes of cubic Ag. On the contrary, the clear lattice fringes of TiO2 are not observed in Fig. 1(e), which implies that the formed TiO2 shell is amorphous. The X-ray energy dispersion spectrum (EDS) in Fig. 1(f) shows that the hollow octahedra consist of Ag, Ti, and O, and their atomic percentages of Ag, Ti, and O are 5.23%, 17.74%, and 77.03%, respectively. The XRD pattern of as-prepared Ag/TiO2 hollow octahedra also reveals the amorphous phase of TiO2 because only the diffraction peak of metallic Ag is observed (Fig. 2(b)).
Obviously, because the concentration of TiF4 aqueous solution has an effect on the formation of Ag/TiO2 hollow octahedra, it is possible that the various Ag/TiO2 textures can be controllably obtained by adjusting the TiF4 concentration. According to simple calculation, 0.04 mol/L TiF4 is required to dissolve the Ag2O templates in this study. When the TiF4 concentration is 0.01 mol/L, it can be clearly shown in Fig. 3(a) that the yolk-shell octahedra of Ag2O/TiO2 are formed because Ag2O templates cannot be dissolved completely. With increasing TiF4 concentration to 0.1 mol/L, the hollow octahedra of Ag/TiO2 are synthesized again and their shell thickness rises to ca. 100 nm (Fig. 3(b)), which is obviously thicker than that obtained with 0.05 mol/L TiF4 (Fig. 1). This also indicates that the shells of Ag/TiO2 hollow octahedra can be controlled simply by adjusting the TiF4 concentration. However, the Ag2O templates are rapidly dissolved with a higher concentration of TiF4 (0.3 mol/L) and no precipitation is found in the reaction system.
To reveal the formation mechanism of Ag/TiO2 hollow octahedra, time-dependent experiments were carried out by intercepting the intermediate products after reaction for various times (0, 5, 10, and 30 min); their corresponding morphologies are shown in Fig. 4(a). Prior to reaction (0 min), only Ag2O octahedral particles exist (Fig. 4(a-1)). After reaction for 5 min, it can be clearly seen that the yolk-shell octahedron is formed (Fig. 4(a-2)). Furthermore, the thickness of the thin shell is ca. 10 nm, and the edge length of the inner core is ca. 950 nm. When increasing the reaction time to 10 min, similar yolk-shell octahedra are still found, as shown in Fig. 4(a-3). However, they are obviously different from that in Fig. 4(a-2) because the shell thickness becomes thicker and the size of the core becomes smaller. When the reaction time reaches 30 min, the inner core disappears and only hollow octahedra can be seen (Fig. 4(a-4)).
On the basis of the above TEM results, the formation of Ag-TiO2 hollow octahedra is illustrated in Fig. 4(b). In our reaction system, the hydrolysis of TiF4 and dissolution of Ag2O are controlled by the following reactions, respectively [27]:
It is easily understood that the hydrolysis rate of Ti precursors should be effectively controlled to obtain well-defined coating. In this respect, the hydrolysis rate of TiF4 is slower and easily adjusted as a result of the relatively stable Ti-F, compared with that of other Ti precursors [34]. Therefore, choosing TiF4 as the precursor is suitable for the construction of Ag/TiO2 hollow octahedra in our system. In addition, the weak basicity of Ag2O can slightly accelerate the hydrolysis rate by neutralizing HF, which is a hydrolysis product of TiF4. As a result, the heterogeneous nucleation and growth of TiO2 is also sped up on the surface of Ag2O templates. As shown in the schematic illustration (Fig. 4(b)), once the TiF4 aqueous solution is added to the Ag2O suspension, the thin shells of TiO2 hollow octahedra are rapidly formed in a very short time (3 min at 60 °C), attributed to the synergetic effect of TiF4 and Ag2O templates; at the same time, Ag2O templates are completely dissolved within 30 min at 60 °C, leading to hollow octahedral structures. Furthermore, to obtain metallic Ag modification on the surface of TiO2 hollow octahedra, the designed irradiation by a fluorescent lamp is considered in our system because Ag2O is a typical light-sensitive material and is decomposed into metallic Ag nanoparticles [35]. Therefore, under light illumination, Ag nanoparticles are formed and deposited on the shells of formed TiO2 hollow octahedra. Furthermore, Ag+ ions formed after the dissolution of Ag2O are also unstable and can be photochemically decomposed into Ag when exposed to light. This is a possible reason why Ag nanoparticles are modified on both the inside and outside surfaces of TiO2 hollow octahedra.
According to the above analysis, there are some specific advantages associated with the synthetic method we used to obtain Ag/TiO2 hollow octahedra. First, the synthetic condition is mild and fast in our case, while a long time of several hours or a high reaction temperature is usually required to finish the hydrolysis of TiF4 in previous reports [27, 34]. Second, the Ag2O templates can be easily removed during the reaction in our case, while removing hard templates usually requires additional treatment such as acid treatment, alkaline treatment, or calcination [36, 37]. Third, Ag nanoparticles can be deposited on the inside and outside surface of TiO2 hollow octahedra by a facile template-induced synthesis method in our case. In addition, it is also possible to control and adjust the amount of deposited Ag nanoparticles by tuning the intensity of the fluorescent lamp (corresponding work is being carried out in our group).
To achieve superior photocatalytic performance, the as-prepared Ag/TiO2 hollow octahedra were further treated by high-temperature calcination. Fig. 5 shows the corresponding SEM images of Ag/TiO2 samples calcined at 400, 600, and 800 °C, respectively. For simplicity, these samples are denoted Ag/TiO2-400, Ag/TiO2-600, and Ag/TiO2-800. From Fig. 5(a) and (b), it can be clearly seen that the morphologies of Ag/TiO2-400 and Ag/TiO2-600 are nearly the same as that of the uncalcined Ag/TiO2 sample, which indicates that calcination at 400 and 600 °C does not destroy the texture of Ag/TiO2 hollow octahedra. However, a morphology change for sample Ag/TiO2-800 can be clearly observed from Fig. 5(c). Specifically, a few flaws appear on the dense shells of Ag/TiO2 hollow octahedra, which indicates that the hollow octahedra are destroyed when calcined at 800 °C. The phase structures of Ag/TiO2-400, Ag/TiO2-600, and Ag/TiO2-800 were revealed by XRD. For Ag/TiO2-400, only the diffraction peaks of Ag were detected, as showed in Fig. 6. When Ag/TiO2 is calcined at 600 °C, obvious diffraction peaks of anatase TiO2 in Ag/TiO2-600 are present, in addition to that of Ag in the XRD pattern (Fig. 6). Generally, the temperature of phase transition from anatase to rutile is ca. 500 °C, which suggests that the photocatalytic activity of anatase is higher than that of rutile at the same reaction condition. As a consequence, there is another advantage of as-prepared Ag/TiO2 hollow octahedra, i.e., they not only can keep the stable anatase phase, but they also show good crystallinity. There are many factors that can greatly affect the transformation temperature of TiO2 from the anatase to rutile phase. For example, previous reports show that the adsorption of fluoride ions on the surface of TiO2 can result in a higher phase transformation temperature [38, 39]. In our case, some possible reasons are that the samples have small primary grains of TiO2, and Ag nanoparticles in Ag/TiO2 can block the growth of these small TiO2 primary grains. However, when the calcination temperature is raised to 800 °C, the crystal phase of TiO2 in Ag/TiO2-800 is converted to the rutile phase (Fig. 6), along with the morphology change as shown in Fig. 5(c). Based on these results, only Ag/TiO2-400 and Ag/TiO2-600 were chosen to conduct the following photocatalytic test.
Ag/TiO2 hollow octahedra can be further examined by UV-vis spectra, as shown in Fig. 7, because the UV-vis spectrum is also very sensitive to Ag nanoparticles because of their strong visible-light absorption caused by local surface plasmon resonance (LSPR). From Fig. 7, it is clear that a sharp absorption edge and an obvious absorption shoulder emerge in UV-vis spectra for all samples. More specifically, the absorption edge of TiO2 lies at ca. 450 nm, which is a red-shift compared with that of pure TiO2. This suggests that Ag atoms are possibly doped into the lattice of TiO2. In addition, the absorption shoulder is in the range of 450-800 nm, which is ascribed to the LSPR of Ag nanoparticles. This result indicates that the as-prepared Ag-TiO2 samples are actually composed of Ag and TiO2.
The photocatalytic performance of Ag/TiO2, Ag/TiO2-400, and Ag/TiO2-600 was evaluated by photocatalytic decolorization of MO aqueous solution under UV-light irradiation. In the dark, no change in the concentration of MO was observed in the presence of various photocatalysts. Furthermore, UV-light illumination in the absence of photocatalysts did not result in the photocatalytic decolorization of MO. Figure 8(a) shows the UV-light photocatalytic performance of various samples. For uncalcined Ag/TiO2, a relatively low photocatalytic activity was observed and the k value is calculated to be 0.001 min-1. Ag/TiO2-400 and Ag/TiO2-600 present obviously enhanced photocatalytic activity compared with uncalcined Ag/TiO2. The calculated k values of Ag/TiO2-400 and Ag/TiO2-600 are ca. 0.04 and 0.11 min-1, respectively. The mechanism of enhanced photocatalytic performance for Ag/TiO2 hollow octahedra under UV-light irradiation is illustrated in Fig. 8(b). Once Ag/TiO2 samples are irradiated by UV light, the photogenerated electrons at the conduction band of TiO2 can rapidly transfer to Ag nanoparticles to reduce oxygen molecules, while the photogenerated holes on the valance band of TiO2 diffuse to the surface to oxidize adsorbed MO molecules. Therefore, their enhanced photocatalytic performance should benefit from the rapid transfer of photogenerated electrons in our case, which can be ascribed to (1) short diffusion distances because of the thin shell of Ag/TiO2 hollow octahedra and Ag deposition on their inside and outside surfaces, (2) the formed Schottky barrier between Ag and TiO2 to facilitate interfacial charge transfer of photogenerated electrons of TiO2, and (3) the high crystallinity of TiO2, which is crucial because it can effectively inhibit the recombination of photogenerated electrons and holes.
Ag/TiO2 hollow octahedra were synthesized by a facile template-induced method using Ag2O octahedra as templates and TiF4 as the precursor at 60 °C for 30 min. The Ag2O templates can be in situ removed by dissolving the Ag2O octahedra with HF solution produced via the hydrolysis reaction of TiF4 in the reaction system. Ag nanoparticles were also deposited on the inside and outside surfaces of TiO2 shells by effectively using the photosensitive properties of Ag2O and Ag+ ions under light irradiation, along with the formation of TiO2 hollow octahedra. The shell thickness of Ag/TiO2 hollow octahedra and the yolk-shell octahedra of Ag2O/TiO2 was controlled by changing the concentration of TiF4 in the reaction system. The Ag/TiO2 hollow octahedra exhibited high UV-light photocatalytic activity because of (1) the short diffusion distances between photogenerated electrons and holes because of the thin shells of Ag/TiO2 hollow octahedral, (2) the deposition of Ag nanoparticles on the inside and outside surfaces of TiO2 shells, and (3) the rapid interfacial charge transfer between TiO2 shells and Ag nanoparticles. This work may also provide new insights into preparing other Ag-modified and hollow nanostructured photocatalysts.