Titanium dioxide (TiO₂) has been widely studied and applied in photocatalysis, solar energy conversion, and lithium-ion batteries owing to its ready availability, low cost, non-toxicity, and high stability [1-3]. In recent years, TiO₂ films have been widely studied, because of the difficulty of recovering powdered TiO₂ [4, 5]. Hence, TiO₂ films are typically deposited on a substrate in the form of nanoparticles, to form polycrystalline and amorphous films [6, 7]. Recently, films of TiO₂ nanotube arrays prepared by anodic oxidation technique have been widely used in photocatalysis, dye sensitized solar cells, and energy storage materials [8-10]. Single-crystalline anatase TiO₂ tetragonal nanosheet-array films have also been successfully prepared by hydrothermal methods [11]. The TiO₂ morphology and structure are among the most important aspects that affect their light absorption, charge separation and transfer, reactant adsorption, and photocatalytic activity [12-15].

TiO₂ mesocrystals have recently emerged and received attention as a new class of porous TiO₂ materials with oriented superstructures of arranged TiO₂ nanocrystal building blocks [16-22]. Recent articles have focused on the preparation of TiO₂ materials with tailored framework structures through the addition of different templates and additives [23-29]. However, the preparation of mesoporous single crystal TiO₂ films has not yet been reported.

Modification of TiO₂ mesocrystals with Au nanoparticles also enables visible-light-driven photocatalytic performance. Au nanoparticles have a localized surface plasmon resonance (SPR), which can absorb light in the visible region owing to a resonant oscillation of free electrons [30].

Herein, we used a facile, direct annealing approach to fabricate a TiO₂ mesocrystal film. The morphology and crystal phases of the film were controlled and the photocatalytic activity was studied for photocatalytic hydrogen generation. Moreover, we found that Au nanoparticles deposited on a TiO₂ mesocrystal film induced visible light-driven photocatalytic activity, which shows great potential for practical applications.

TiF₄, NH₄NO₃, chloroauric acid, Na₂SO₄ and K₂Cr₂O₇ were purchased from Aladdin (AR, Shanghai, China). P123 (amphiphilic triblock copolymer PEO-PPO-PEO ((EO)₂₀(PO)₇₀(EO)₂₀)) and NH₄F (AR) were purchased from Sigma-Aldrich. All the chemicals were used as received.

A titanium sheet (20 × 30 mm²) was cleaned by ultrasonic irradiation for 30 min in acetone, ethanol, and deionized water (volume ratio = 1:1:1) after being polished. The precursor solution was prepared from TiF₄, H₂O, NH₄NO₃, and P123 (molar ratio = 93:32000:444:1). The TiO₂ mesocrystal films were prepared by titanium sheet dip-coating in the precursor solution. The treated titanium sheets were calcined in air at 300, 400, 500, 600, and 800 ℃ for 1 h at a heating rate of 5 ℃/min, respectively.

Chloroauric acid solution was added to the above precursor solution at contents of 0.5%, 1%, 1.5%, and 2% (mass ratio in Au/TiO₂). According to the above method, the sheets were calcined in air at 400 ℃ for 1 h.

The morphology was characterized with a scanning electron microscope (SEM, Hitachi S4800) and a transmission electron microscope (TEM, JEOL JEM-2010). The crystal structure was characterized by X-ray diffraction (XRD, D/MAX-2000 with Cu K_α radiation). The pH value was measured with a pH meter (Mettler Toledo Delta 320). Photoelectrochemical measurements were performed in a conventional three-electrode configuration with an electrochemical station (CHI660). The TiO₂ mesocrystal film was used as the working electrode, and a platinum sheet and saturated calomel electrode (SCE) were used as the counter and reference electrodes, respectively. The transient photocurrent was measured with the use of a 10 s on-off cycle at a bias voltage of 0.5 V or zero in 50 mL Na₂SO₄ solution (0.5 mol/L). UV-Vis diffuse reflectance spectra were recorded with a spectrophotometer (Shimadzu, UV2600) with an integrating sphere attachment in the range 200 to 800 nm and with BaSO₄ as reflectance standard.

Photocatalytic hydrogen evolution: A double-cell reactor was used for photocatalytic hydrogen generation. The chambers were separated by ion membranes. Two pools were injected with 30 mL ethylene glycol solution (2 mol/L) and 30 mL sodium sulfate solution (0.5 mol/L), respectively. Three electrodes system were used for the hydrogen production tests. The TiO₂ mesocrystal film was used as the working electrode, and a platinum sheet and SCE were used as the counter and reference electrodes, respectively. The photocatalytic hydrogen production was initiated by a xenon lamp with a filter (λ < 400 nm). The amount of H₂ evolved within 1 h of irradiation was determined with a gas chromatograph (Ceaulight, GC-7900).

Photocatalytic reduction of Cr(Ⅵ): A Au/TiO₂ mesocrystal film was immersed in an aqueous solution containing Cr(Ⅵ) ions (10 ppm, pH ≈ 3 controlled by HCl), in a home-made reactor stirred for approximately 30 min to reach adsorption-desorption equilibrium in the dark [19, 31]. The photocatalytic reduction of Cr(Ⅵ) under UV or visible light irradiation was initiated by a xenon lamp with a filter (λ < 400 nm or > 420 nm). The concentration of Cr(Ⅵ) was analyzed by a UV spectrophotometer (UV 7502/PC) at the characteristic wavelength [32], from which the reduction rate could be calculated as: (1 -C/C₀) (C is the test concentration, C₀ is the initial concentration).

The structure of TiO₂ mesocrystal film annealed at 400 ℃ was characterized by SEM imaging. The TiO₂ mesocrystal film showed a stacked sheet structure with a size of several micrometers (Fig. 1(a)). A porous structure was clearly observed on the surface (Fig. 1(b)). As we have previously reported, the addition of NH₄F can control the thickness of the resulting TiO₂ plates [19]. Different thickness of the TiO₂ plates can be achieved through the use of different TiF₄/NH₄F contents at the same calcination temperature (400 ℃). As the NH₄F content is gradually increased (Fig. 2(a), (b)), the (001) facets of the TiO₂ plates are increased and the resulting crystals become thinner. Many gaps appear between neighboring sheets, which can increase the specific surface area of the films, and are beneficial for the photocatalytic performance. As the amount of NH₄F is increased further (Fig. 2(c)), the TiO₂ plates become very thin resulting in structural collapse. Therefore, the optimum conditions for synthesis of the TiO₂ mesocrystal film were a ratio of TiF₄/NH₄F volume ratio = 1:0.8 and calcination temperature of 400 ℃. A cross sectional SEM image (Fig. 2(d)) shows that the thickness of the TiO₂ crystal layer was approximately 1.33 μm. Furthermore, high-resolution TEM (HRTEM) images (Fig. 2(e), (f)) of a particle of the TiO₂ mesocrystals showed that the obtained TiO₂ was a single crystal and highly crystallized, with well-resolved (001) (d = 0.18 nm) crystalline lattices in selected area electron diffraction (SAED) patterns.

In the annealing process, the precursors of Ti⁴⁺, F^-, NH₄⁺, and H₂O undergo a series of reactions during evaporation of water at low annealing temperatures to form NH₄TiOF₃ (see Fig.S2 in Supporting Information (SI)). As the annealing temperature is increased further, NH₄TiOF₃ is easily transformed into TiO₂. When large amounts of nitrogen and fluorine are removed, the volume of the nanoparticles decreases, and gaps or pores form between the nanoparticles, resulting in porous TiO₂ materials that consist of anatase single-crystalline nanoparticles with dominant (001) facets [18, 19]. These TiO₂ nanosheets were assembled onto titanium substrates to form TiO₂ films. These TiO₂ nanosheets have a porous structure. The specific surface area was approximately 50 m²/g, and the pore size was mainly concentrated at 20 nm, showing a broad distribution (Fig.S3).

The XRD patterns of the films calcined at 400 ℃ (Fig. 3) indicated that the crystal structures and crystallinity were unchanged by the addition of NH₄F. All the patterns showed strong diffraction peaks from metallic Ti and well-defined characteristic peaks of anatase TiO₂.

The calcination temperature had a considerable influence on the crystal formation and crystallinity of the prepared TiO₂ mesocrystal film. As shown in Fig. 4, all of the TiO₂ films annealed at different temperatures from 300 to 500 ℃ had well-defined characteristic diffraction peaks of anatase TiO₂. As the temperature was increased from 300 to 400 ℃, the crystallinity of the TiO₂ film increased. However, when the calcination temperature reached 600 ℃, rutile TiO₂ appeared, and the percentage of the rutile phase increased with increasing temperature.

The morphology of the TiO₂ single crystals changed markedly with increasing calcination temperature. When the calcination temperature ranged from 300 to 500 ℃, the TiO₂ crystals became smaller and thicker (Fig. 2(b), Fig.S1(a), (b)). When the temperature reached 600 ℃, the crystal blocks were damaged, which might have been caused by rapid pyrolysis at high temperature (Fig.S1(c), (d)).

H₂ evolution was used to evaluate the photocatalytic performance of the TiO₂ mesocrystal film. The amount of H₂ evolved increased as the volume ratio of TiF₄ to NH₄F was increased from 1:0 to 1:0.8, and reached a maximum amount of 41.7 μmol/(h·cm²) at a volume ratio 1:0.8. The increased specific surface area and active reaction sites caused by gaps between neighboring sheets resulted in this enhancement of photocatalytic H₂ evolution performance (Fig. 5(a)). The H₂ generation performance of the TiO₂ mesocrystal films formed at different calcination temperatures were also investigated, and the results are shown in Fig. 5(b). Owing to the improved crystallinity compared with that of the film obtained at 300 ℃, the film calcinated at 400 ℃ exhibited increased H₂ generation ability. Owing to the optimum calcination conditions and crystalline morphology, the TiO₂ mesocrystal film calcinated at 400 ℃ demonstrated the best photocatalytic H₂ generation performance. However, the amount of H₂ evolved decreased when the calcination temperature was greater than 400 ℃. This decrease was attributed to the appearance of rutile TiO₂ through the phase transformation occurring at high temperature calcination (Fig. 4) and the crushing of crystal blocks (Fig.S1).

Photoelectric measurements were performed to characterize the photoelectric properties of the TiO₂ mesocrystal film. As shown in Fig. 5(c), the photocurrent density increased as the concentration of NH₄F was increased. The TiO₂ mesocrystal film prepared at volume ratio of TiF₄/NH₄F = 1:0.8 exhibited the highest photocurrent output of approximately 0.92 mA/cm², which was 3.2 times as high as that of the TiO₂ mesocrystal film with no NH₄F added. However, with more NH₄F added, the photocurrent output from the films decreased. This was mainly attributed to structural collapse leading to poor contact with the substance. This was also considered to be the main factor leading to the decrease of photocatalytic performance. The photocurrents of the TiO₂ mesocrystal film measured at different calcination temperatures were measured under UV irradiation (Fig. 5(d)). The photocurrent generated by the film calcined at 400 ℃ was 9 times as high as that calcined at 300 ℃. This result reflects the easier migration of photogenerated charge carriers in the film calcined at 400 ℃ compared with that calcined at 300 ℃, owing to the greater crystallinity of the former. As the annealing temperature was increased further, the photocurrent decreased considerably. This decrease was attributed to damage to the structure of the sample and a change of the crystal phase (Fig. 4 and Fig.S1). The results of the photoelectric performance also reflect a change of the photocatalytic performance.

To achieve the visible-light-driven photocatalysis, the TiO₂ mesocrystal film was modified with Au nanoparticles based on the optimized condition described above. The XRD patterns of Au/TiO₂ films showed that there was no obvious change of the crystalline phase and crystallinity as the Au loading was increased (Fig.S4). Characteristic diffraction peaks of Au were not found in the patterns, which could be attributed to the high dispersion of Au nanoparticles. Furthermore, the addition of Au had almost no effects on the morphology of the films (Fig.S5). The loaded Au nanoparticles induced absorption in the visible spectrum owing to SPR; The absorbance at 500–800 nm gradually increased with the introduction of the Au nanoparticles (Fig.S6). The results of photocatalytic Cr(Ⅵ) reduction performance under visible light irradiation are shown in Fig. 6. The reaction system was stirred for approximately 30 min to reach adsorption-desorption equilibrium in the dark. After light irradiation, the Cr(Ⅵ) ions were rapidly reduced. Modification of the TiO₂ with Au nanoparticles resulted in an increase of the photocatalytic activity owing to the SPR of the Au nanoparticles.

In summary, we prepared TiO₂ mesocrystal films by a direct annealing method. The morphology and crystal phase of the films were controlled by adjusting the ratio of NH₄F and calcination temperature. H₂ evolution was used to evaluate the photocatalytic performance of the TiO₂ mesocrystal film through the use of a three-electrode system. When the volume ratio of TiF₄ to NH₄F was 1:0.8 and the calcination temperature was approximately 400 ℃, the H₂ production was 41.7 μmol/(h·cm²). We found that Au nanoparticles loaded onto the TiO₂ mesocrystal film induced highly efficient visible light photocatalytic properties. Our work provides a new highly efficient photocatalytic film electrode, which could be widely applied in photocatalysis and photoelectric catalysis.