In the last 30 years, semiconductor photocatalysis has become an important research topic due to its potential application in solving the problems of environmental pollution and energy crises, such as degradation of the organic pollutants [1-5], air purification [6-10], reduction of CO₂ to hydrocarbon fuel [11-13], and splitting water to produce hydrogen [14-17]. Among all the photocatalysts, TiO₂ is the most studied because of its nontoxicity, higher chemical stability, and oxidizability features [18, 19]. However, the wide band gap of anatase TiO₂ (3.2 eV) hampers its wide practical application due to the fact that it absorbs only a very small ultraviolet part (3%–4%) of solar light [20, 21]. To make use of the abundant amount of visible light (approximately 45%) from solar light, the development of visible-light-driven TiO₂ is of great importance, but remains a great challenge [3, 22, 23]. Up to now, many strategies have been used to extend the light-responsive region of TiO₂ from the ultraviolet (UV) to visible region, including dye sensitization [24, 25], semiconductor coupling [17, 26, 27], doping TiO₂ with metals or non-metals [22, 28, 29], noble-metal deposition [30], and surface-plasma effects [8, 31, 32]. Recent study also shows that modification of anatase TiO₂ with graphene can also result in a visible-light response [33-35].

In recent years, hollow nanostructures have attracted much attention due to their unique properties and potential practical applications [36-39]. According to the literature, many hollow nanostructures have been prepared, such as nanotubes [40] and hollow micro/nanospheres [17, 41-44]. Large fractions of void space in hollow structures have been successfully used to encapsulate and control release of sensitive materials such as drugs, cosmetics, and DNA [45]. Our group has previously reported the fabrication of TiO₂ hollow nanoboxes (TiO₂-HNBs) assembled from TiO₂ nanosheets (TiO₂-NSs) via a fluoride-induced self-transformation strategy by solvothermal treatment of cubic TiOF₂ in alcohol solution [45]. It was found that the photoreactivity of the obtained TiO₂-HNBs under UV irradiation was improved when compared with that of the discrete building block TiO₂-NSs. Upon considering that TiO₂-HNBs can still only be excited by UV light, herein we prepared visible-light-responsive TiO₂-HNBs by in situ modification of TiO₂-HNBs with methionine (C₅H₁₁NO₂S), a N-and S-containing biomacromolecule. We systematically studied the effect of calcination temperature on the structure and visible photocatalytic activity of methionine-modified TiO₂-HNBs.

All reagents and solvents were of reagent grade (Wuhan Guoyao Chemical Co., China) and used without further purification.

Precursor TiOF₂ nanocubes were synthesized by hydrothermal reaction. Briefly, 30 mL of acetate, 5 mL of hydrofluoric solution (40 wt%), and 15 mL of tetrabutyl titanate were mixed in a 100-mL PTFE beaker under magnetic stirring. The mixture was heated at 200 ℃ for 12 h in an autoclave. After cooling to room temperature, the resulting white powders were washed with water and ethanol several times and dried at 60 ℃ overnight.

To prepare methionine-modified TiO₂-HNBs, the mixture of 1.0 g of TiOF₂ and 0.05 g of DL-methionine in a crucible was calcined at 400 ℃ for 2 h with a heating rate of 1 ℃ min⁻¹. Similarly, the mixture of TiOF₂ and methionine was also calcined at different temperatures, and the prepared sample denoted TMx, where x represents the calcination temperature (300–500 ℃). For example, TM400 means that the sample was prepared by calcining the mixture at 400 ℃ for 2 h.

In order to study the effect of methionine on the structure and performance of TiO₂-HNBs, direct calcination of TiOF₂ in the absence of methionine was also performed under identical conditions, and the resulting sample denoted Tx, where x represents the calcination temperature (300–500 ℃).

The crystalline structure of the catalyst was characterized by powder X-ray diffraction (XRD) employing a scanning rate of 0.05° s⁻¹ in a 2θ range from 10° to 60°, in a Bruker D8 Advance diffractometer using monochromatized Cu K_α radiation. The morphology and microstructure of the as-prepared sample were analyzed by field-emission scanning electron microscopy (SEM) (Hitachi, Japan) and transmission electron microscopy (TEM) (G20, Tecnai, USA). Nitrogen adsorption-desorption isotherms were obtained from a nitrogen-adsorption instrument (ASAP 2020, USA), from which all the photocatalysts were degassed firstly at 150 ℃, followed by investigating the surface areas and pore-size distribution of the photocatalysts. The UV-Vis diffused reflectance spectrum (DRS) was collected using a spectrophotometer (Shimadzu UV-2550, Japan) from 200 to 800 nm using BaSO₄ as background. Fourier-transform infrared spectroscopy (FTIR) was recorded on a spectrometer (NeXUS 470) using the KBr pellet technique. X-ray photoelectron spectroscopy (XPS) spectra were recorded on a photoelectron spectrometer (VG Multilab 2000, VG, Inc., USA) using monochromatic Al K_α radiation under vacuum at 2 × 10⁻⁶ Pa. All the binding energies were referenced to the C 1s peak at 284.8 eV of the surface adventitious carbon.

The photocatalytic activity of the photocatalyst was evaluated both by photocatalytic decomposition of the organic dye rhodamine B (RhB) and NO oxidation under visible-light irradiation.

For RhB degradation, one cylindrical Pyrex^® flask with a capacity of approximately 50 mL was used as the photoreactor vessel in the reaction system. Approximately 50 mg of the photocatalyst was put into 50 mL of RhB solution with a concentration of 1.0 g L⁻¹. After being ultrasonicated and stirred in the dark overnight to establish adsorption-desorption equilibrium, the solution was illuminated by an Xe lamp (300 W) with a cut-off filter (λ > 420 nm). At given time intervals, 3.0 mL of suspension sample was taken and centrifuged to remove the photocatalyst particles. The concentration of the RhB in solution was monitored by UV-Vis spectroscopy at 554 nm.

NO oxidation was performed in a consecutive flow reactor at ambient temperature. The volume of the rectangular reactor constituted of stainless steel with a quartz glass cover was 4.5 L (L × W × H, 30.0 cm × 15.0 cm × 10.0 cm). A simulated visible-light source was acquired from an LED lamp (λ > 400 nm). First, 30 mL of deionized water was poured into 0.2 g of the as-obtained samples, and the mixed solution was subjected to ultrasonication for 30 min. The resulting suspension was then sedimentated onto a culture dish with a diameter of 11.5 cm. The dish containing the sample was heated at 60 ℃ for 3 h to thoroughly evaporate any water, and then cooled to normal temperature. NO gas was supplied from a compressed gas cylinder having 50 ppm of NO (N₂ balance) in line with traceable standards recommended by the U.S. National Institute of Standards and Technology. An air stream produced by a zero-air generator (Advanced Pollution Instrumentation, A Teledyne Technologies Co., Model 701) can be used for diluting the initial NO concentration to approximately 600 ppb. In a gas blender, the gas streams were totally premixed, and the gas-flow rate was set as 1.0 L min⁻¹ by employing a mass-flow controller. When the gas system reached adsorption- desorption balance, the lamp was turned on. A chemiluminescence NO_x analyzer (Advanced Pollution Instrumentation, A Teledyne Technologies Co., Model T200) was used for monitoring the concentrations of NO continuously, at a gas-flow rate of 1.0 L min⁻¹.

The photocurrent of the film electrode was measured on an electrochemical system (CHI-760e, Shanghai, China) using a standard three-electrode cell. The ITO/sample, measuring 2.0 cm × 1.5 cm, acted as the working electrode, a Ag/AgCl electrode was used as the reference electrode, and a platinum wire was used as the counter electrode. The ITO/sample working electrode was prepared by a dip-coating method: 50 mg of the photocatalyst was suspended in 1.0 mL of ethanol, which was then dip-coated onto the ITO glass electrode. 0.1-mol L⁻¹ Na₂SO₄ was used as an electrolyte. The photoelectric responses of the sample at light-on and light-off were measured using an LED lamp (3 W) emitting mainly at 420 ± 10 nm as the light source.

The surface photovoltage spectroscopy (SPS) measurements of the sample were carried out with a home-built apparatus [4, 46]. The SPS signals are the potential barrier change of the testing electrode surface between that in the presence of light and that in the dark.

XRD was used to investigate the changes of the phase structure of the photocatalyst with calcination temperature. From Fig. 1(a), it can be seen that a sharp peak at 2θ = 23.4° corresponding to TiOF₂ was observed, indicating the successful synthesis of TiOF₂ precursor [23, 47, 48]. Upon calcination of the precursor TiOF₂ at 300 ℃ for 2 h (T300 sample), the peak intensity for TiOF₂ decreases, while a broad peak exists at 2θ = 25.3°, corresponding to anatase TiO₂ (A-TiO₂), appears (Fig. 1(b)) [49]. This infers that anatase TiO₂ is a more stable phase than TiOF₂. When the calcination temperature increases to 350 ℃ (Fig. 1(c)), the precursor TiOF₂ almost completely transforms into anatase TiO₂. However, upon calcination of the mixture of TiOF₂ and methionine at 300 ℃, no indication of the formed anatase TiO₂ was observed for the TM300 sample. Phase transformation of TiOF₂ to anatase TiO₂ begins at 350 ℃ (TM350 sample), and the phase transformation completes at 400 ℃ upon calcination of the mixture of precursor TiOF₂ and methionine. This indicates that methionine can prevent TiOF₂-to-anatase TiO₂ phase transformation, which is possibly due to the fact that the presence of methionine retards the diffusion of fluoride ions from the core to the shell.

Consistent with the reported literature [23, 50-53], the cubic morphology of the as-prepared TiOF₂ precursor is confirmed by the corresponding TEM image (Fig. 2(a)). It can be seen that the synthesized TiOF₂ cubes have sidelengths of approximately 200–300 nm. When calcining the mixture of precursor TiOF₂ and methionine at 300 ℃ for 2 h, the morphology of the photocatalyst seems unchanged (Fig. 2(b)). However, careful examination shows the formation of core-shell structured nanocubes. When compared with the TEM image of the precursor TiOF₂ sample (Fig. 2(a)), the dark core of the cubes become much lighter, and the nanosheets around the cores can be observed, indicating that the phase transformation begins. This is in accordance with the XRD characterization results ((a) and (b) of Fig. 1B). Upon increasing the calcination temperature to 350 ℃, the hollow interiors of the nanocubes become obvious (Fig. 2(c)). Almost totally hollow structures are obtained upon further increase in the calcination temperature to 400 ℃, indicating the completion of TiOF₂ to anatase TiO₂ ((c) of Fig. 1B).

We proposed that TiOF₂ cubes will transform into TiO₂-HNBs during a solvothermal reaction when using alcohol as solvent by in situ transformation of TiOF₂ to anatase TiO₂ nanocrystals [45]:

(1)

(2)

However, in the present study, TiOF₂ cubes transforms into more stable anatase TiO₂-HNBs by direct calcination. The possible reaction is

(3)

It has been reported that fluoride ions facilitate the formation of high-energy anatase TiO₂-NSs due to the fact that the adsorption of F⁻ on the surface of TiO₂ nanocrystals can sharply reduce the surface energy of (001) facets [54]. Therefore, it is understandable that precursor TiOF₂ can in situ transform into anatase TiO₂-HNBs assembled from TiO₂-NSs with exposed high-energy (001) facets at high temperature. The inset of Fig. 2(d) clearly shows the sideview high-resolution TEM (HRTEM) image of an erected TiO₂ nanosheet for the TM400 sample, from which we can see a lattice spacing of approximately 0.235 nm, confirming the exposure of the (001) facets of anatase TiO₂ nanosheets.

The obtained TiO₂-HNBs are not stable, and decompose into discrete TiO₂-NSs at high temperatures, such as 450 and 500 ℃ (Fig. 2(e) and 2(f)).

Fig. 3 shows the SEM images of the TM400 sample at different magnifications, which further confirms the formation of TiO₂-HNBs assembled from TiO₂-NSs.

To obtain more detailed structural information of the prepared TiO₂-HNBs, nitrogen adsorption was used to measure the Brunauer-Emmett-Teller (BET) surface area and pore structure of the photocatalyst. It was found that both the BET surface area and pore volume (PV) of TMx nearly increase with increasing calcination temperature (Table 1). The BET surface area increases from 7.0 to 12.9 m² g⁻¹, and the PV increases from 0.022 to 0.038 cm³ g⁻¹ when the calcination temperature increases from 300 to 400 ℃, further confirming the heat-induced hollowing process, namely from a solid nanocube (TiOF₂) to hollow boxes (TiO₂-HNBs). The TM400 (12.9 m² g⁻¹) and T400 (11.6 m² g⁻¹⁾ samples have similar BET surface areas, which are smaller than that of hollow TiO₂-HNBs (approximately 40 m² g⁻¹) reported by Jia et al. [55] and that of TiO₂ hollow microspheres (20–40 m² g⁻¹) reported by Zhang et al. [56]. This is possibly due to the different preparation methods.

Fig. 4 shows the nitrogen-adsorption isotherms and corresponding pore-size distribution curves of the TM400 and T400 samples. It can be seen that the shape of the isotherms for both samples is of type IV with H3 hysteresis loop shapes, suggesting narrow, slit-shaped pores that are generally associated with plate-like particles, which agrees well with their sheet-like morphology (Figs. 2 and 3).

Light-harvesting ability is of great importance to the photoreactivity of photocatalysts. Here, UV-vis DRS were used to evaluate the effect of methionine modification on the light-responsive property of the photocatalyst. From Fig. 5, it can be seen that all the photocatalysts after methionine modification show obvious absorption in all the visible region, although this absorption steadily decreases with increasing calcination temperature. The long tail in the visible-light region of the TiO₂ sample usually comes from the doping or surface oxygen vacancy [29].

When compared with the T400 and TM400 samples, we can clearly see that the absorption of the TM400 sample in the visible-light region is much stronger than that of the T400 sample, indicating that the TM400 sample is a potential visible-light- responsive photocatalyst.

Fig. 6 compares the FTIR spectra of methionine-modified photocatalysts calcined at different temperatures. The peak for the Ti-O and O-Ti-O flexion vibrations of the TM300 sample appears at 563 cm⁻¹ (Fig. 6(a)) [29]. The intensity of the peak becomes much stronger for samples calcined at high temperatures, which is due to the phase transformation from TiOF₂ to anatase TiO₂ (Eq. (3)). The peak at 980 cm⁻¹ for the TM300 sample is attributed to the vibration of the Ti−F bond, which cannot be observed for TiO₂ samples calcined at high temperature. However, a newly formed peak at 1090 cm⁻¹ is observed for all samples calcined at temperatures higher than 350 ℃, which is attributed to the Ti−O−Ti vibration [29], further confirming the phase transformation from TiOF₂ to anatase TiO₂ during calcination. The peaks of 1630 and 3440 cm⁻¹ are from the surface −OH groups/H₂O due to the adsorption of moisture from the air [29, 33].

Both of the XPS survey spectra of the unmodified (T400) and methionine-modified TiO₂-HNBs (TM400) calcined at 400 ℃ for 2 h are compared in Fig. 7. It can be clearly seen that the T400 sample contains only Ti, O, and C, with sharp photoelectron peaks appearing at binding energies of 458 eV (Ti 2p), 531 eV (O 1s), and 285 eV (C 1s). The carbon peak is attributed to the residual carbon from the precursor solution and adventitious hydrocarbon from the XPS instrument itself [33]. However, two obvious connected peaks with binding energies of 159 and 164 eV, corresponding to S 2p, are found in the spectrum of the methionine-modified TiO₂-HNBs (TM400). Careful examination also shows a small peak centered at 400 eV in the TM400 sample, which is ascribed to the N 1s signal [29].

Fig. 8(A) and 8(B) compare the high-resolution XPS spectra of the T400 and TM400 samples in the Ti 2p and O 1s regions, respectively. From Fig. 8(A), two peaks of pure TiO₂-HNBs (T400) at 458.3 and 464.0 eV are assigned to the Ti 2p_3/2 and Ti 2p_1/2 spin-orbital-splitting photoelectrons in the Ti⁴⁺ chemical state, respectively [33]. However, the Ti 2p_3/2 binding energy of the methionine-modified TiO₂-HNBs sample (TM400) shifts from 458.3 to 458.8 eV when compared with the pure TiO₂-HNBs (T400). This approximate 0.5-eV shift in the XPS measurement indicates the presence of a strong interaction at the interfaces between titania and methionine. This intense interaction results in the formation of the electron-transfer channel, which is beneficial to the improvement of the photo-induced charge-separation rate during the photocatalytic process [57]. The strong interaction between TiO₂-HNBs and methionine also confirms from the increased O 1s XPS binding energy when comparing the spectrum of the TM400 sample with that of T400 sample (Fig. 8(B)).

Fig. 8(C) shows the corresponding high-resolution XPS spectrum of the N 1s region taken from the TM400 sample. The curve of the N 1s region can be deconvoluted into two peaks. The small peak (399.0 eV) is attributed to Ti−N, and the larger peak at 400.1 eV is assigned to the NH₃/NH₄⁺ adsorbed on the surface of the TiO₂-HNBs, respectively [22, 58].

Fig. 8(D) is the high-resolution XPS spectrum of the S 2p region of theTM400 sample. The S 2p curve contains two isolated peaks at binding energies of 158.7 and 164.1 eV, which can be attributed to organic S(−Ⅱ) and Ti-S, respectively [58]. The organic S(−Ⅱ) comes from the methionine (C₅H₁₁NO₂S), and the formation of the Ti-S bond reflects the fact that S atoms replace O atoms in the TiO₂ lattice.

To confirm the successful doping of N and S elements into the lattice of TiO₂-HNBs, EDX mapping was used to observe the distribution of N and S elements in the TM400 sample. As shown in Fig. 9, it can be seen that not only Ti, but also N and S elements, are homogeneously distributed on the surface of TiO₂-HNBs, which is consistent with the XPS characterization results (Fig. 8).

RhB is used as a probe to evaluate the visible-light photocatalytic activity of the photocatalyst under visible-light irradiation (λ > 420 nm). Fig. 10(a) and 10(b) are the degradation profiles of RhB in the presence of Tx and TMx photocatalysts. In the absence of any photocatalyst, RhB shows little degradation even under visible-light irradiation, reflecting the relative stability of RhB dye. Obvious degradation of RhB is observed in the presence of an illuminated photocatalyst, and the degradation obeys pseudo-zero-order kinetics. The apparent degradation rate constants are compared in Fig. 11. It can be seen that when calcination temperature is below 400 ℃, the modification of methionine plays a negative role in the photoreactivity of TMx. This is because the presence of methionine prevents the phase transformation of TiOF₂ to anatase TiO₂. It has been reported that TiOF₂ shows poor photocatalytic activity [49]. When the calcination temperature is between 400 and 450 ℃, the positive effect of methionine on the photoreactivity of TiO₂-HNBs becomes obvious. For example, the degradation rate constant (K_app) of RhB in the TM400 sample is calculated to be 0.15 mol L⁻¹ h⁻¹, which is 1.53 times higher than that of the unmodified T400 sample (0.098 mol L⁻¹ h⁻¹). When the calcination temperature is higher than 500 ℃, the positive effect of methionine modification can be neglected. This is due to the fact that it is difficult to dope N and S into the lattice of TiO₂ lattice at high temperature.

The efficient separation of photo-generated electrons and holes is very important to improve the photoreactivity of the photocatalyst [4]. To account for the high visible-light photocatalytic activity of methionine-modified TiO₂-HNBs, we compared the photocurrent response and surface photovoltage spectroscopy between the T400 and TM400 samples. From Fig. 12A, prompt generation of photocurrents with good reproducibility when the ITO/TiO₂-HNBs electrodes are illuminated can be seen. The photocurrent value of the TM400 based film is as high as 1.5 μA cm⁻², which is 3 times higher than that of T400 based film (0.5 μA cm⁻²). The photocatalytic activity of TiO₂ is highly related to the number of the separated photo-generated charge carriers. Thus, it can be deduced that the photocatalytic activity of T400 is higher than that of the T400 sample.

The SPS was also measured to confirm the separation efficiency of photo-generated carriers of the photocatalyst. From Fig. 12B, it can be seen that the SPS signal of T400 can only be observed in the UV region due to the large band gap of anatase TiO₂. However, TM400 not only shows a much higher SPS signal in the UV region, but also extends the signal to the visible-light region, indicating its visible photocatalytic activity.

The stability of the photocatalyst is of great importance from the viewpoint of practical applications. Therefore, we studied the reusability of the TM400 photocatalyst in the degradation of RhB for five consecutive cycles under visible-light irradiation (Fig. 13). We observe that the degradation rate of RhB in the presence of the TM400 photocatalyst almost remains unchanged even after being recycled five times, reflecting its good stability. The high stability of the methionine-modified TiO₂-HNBs photocatalyst makes it promising for use in the treatment of wastewater.

Upon considering that RhB can also be degraded via the photosensitization pathway, we measured the visible photocatalytic activity of the photocatalyst by photocatalytic oxidation of NO under visible-light irradiation, and the results are shown in Fig. 14. It can be seen that, similar to the results of RhB degradation, the TM400 sample shows the highest visible-light photoreactivity among all the photocatalysts with a NO decomposition rate of 20%, which is 2 times higher than that of the T400 sample (a NO decomposition rate of less than 10%). This result also confirms the enhanced visible photocatalytic activity of methionine-modified TiO₂-HNBs.

Visible-light-responsive TiO₂-HNBs assembled from TiO₂ nanosheets with exposed high-energy (001) facets were fabricated by simple calcination of a mixture of TiOF₂ cubes and methionine. During calcination, TiOF₂ cubes transform into more stable anatase TiO₂-HNBs via an in situ dissolution-recrystallization process, and both N and S elements are simultaneously doped into the TiO₂ lattice. Modification of TiO₂-HNBs with methionine not only greatly improves their light-harvesting ability, but also facilitates the efficient separation of photo-generated electron-hole pairs, therefore greatly enhancing the visible photocatalytic activity in the degradation of RhB and oxidation of NO. The present study provides a simple approach to designing highly efficient visible-light-responsive photocatalysts.