Titanium(IV) fluoride (TiF₄; 98%, Shanghai Darui Chemicals Co. Ltd., Shanghai, People' Republic of China) was used in this study as the precursor to provide both titanium and fluorine sources. Tert-butanol (C₄H₁₀O; ≥ 98%, Sinopharm Chemical Reagent Co., Ltd., Shanghai, People' Republic of China) was used as the solvent in the solvent-thermal process. Thiourea (CH₄N₂S; ≥ 99%, Sinopharm Chemical Reagent Co., Ltd.) was used to provide nitrogen and sulfur sources in the calcination process. Commercially available Degussa P25 TiO₂ nanoparticles (Evonik Industries, Essen, Germany) were used for the comparison with nitrogen/sulfur co-doped TiO₂ nanocrystals with highly active {001} facets on their visible-light photocatalytic performance.

In a typical synthesis 12 , 1.6 g of TiF₄ was dissolved in 400 mL of tert-butanol under continuous stirring to obtain a transparent faint yellow solution. Then, the solution was transferred to the autoclave and allowed to alcoholize at 160°C for 3 days. The precipitated powders were filtered, washed with 0.1 M NaOH solution 10 to remove residual solvent for three times, and then dried at 50°C overnight to obtain the as-prepared TiO₂ powders, which was denoted as T0. For the introduction of nitrogen/sulfur co-doping, the as-prepared TiO₂ powders were mixed with thiourea 23 35 at a 2:1 weight ratio and then calcinated for 2 h at different temperatures of 300°C, 400°C, and 500°C, respectively, in a sealed tubular furnace. The obtained samples were denoted as T3, T4, and T5, correspondingly.

X-ray diffraction (XRD) experiments were conducted on a D/MAX-2004 X-ray powder diffractometer (Rigaku Corporation, Tokyo, Japan) with Ni-filtered Cu (λ =0.15418 nm) radiation at 56 kV and 182 mA to analyze the crystal structure and crystallite size of obtained powder samples. Their morphologies were examined by transmission electron microscopy (TEM) on a JEOL 2010 TEM (JEOL Ltd., Tokyo, Japan) operated at 200 kV, with a point-to-point resolution of 0.28 nm. TEM samples were prepared by dispersing a thin film of these powder samples on Cu grids. Their BET specific surface area values were measured by the N₂ adsorption/desorption isotherm with an Autosorb-1 Series Surface Area and Pore Size Analyzer (Quantachrome Instruments, Boynton Beach, FL, USA). X-ray photoelectron spectroscopy (XPS) measurements were made using an ESCALAB250 X-ray photoelectron spectrometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) with an Al K anode (1486.6 eV photon energy, 300 W). The UV–vis spectra of these powders were measured on a UV-2550 spectrophotometer (Shimadzu Corporation, Kyoto, Japan).

MB (Acros Organics, Morris Plains, NJ, USA) was used as a model organic pollutant for the static photocatalytic degradation experiment under visible-light illumination. A powder sample was placed at the bottom of a 50 × 10-mm petri dish, and 4 ppm of MB solution was added into the petri dish at a fixed concentration of 1 mg photocatalyst/mL solution. Samples T0 and T3 were used in the photocatalytic degradation of the MB experiment, and P25 TiO₂ powder was also used for comparison purposes under the same experimental conditions. The covered petri dishes were illuminated by a 300-W xenon lamp (PLS-SXE300, Beijing PerfectLight Technology Co., Ltd., Beijing, People's Republic of China), which has a glass filter to ensure a zero light intensity below 400 nm. The light intensity striking the MB solution was ca. 10 mW/cm², as measured by a Multi-Sense optical radiometer (Beijing Normal University Photoelectricity Instruments Plant, Beijing, China). The visible-light illumination time varied from 5 to 30 min. After recovering the photocatalyst by centrifugation, the light absorption of the clear solution was measured by a UV-2550 spectrophotometer (Shimadzu Corporation).

Wild-type E. coli AN 387 (ATCC 15597, the American Type Culture Collection, Manassas, VA, USA) were used for the photocatalytic inactivation experiment. After overnight culture, the cells were diluted to a cell suspension (ca.10⁷ cfu/mL) in buffer solution (0.05 M KH₂PO₄ and 0.05 M K₂HPO₄, pH 7.0) prior to the photocatalytic inactivation. All solid or liquid materials were autoclaved for 30 min at 121°C before use. For E. coli inactivation under visible-light illumination, the same xenon lamp was used, and the light intensity striking the cell suspensions was also at ca. 10 mW/cm². In the photocatalytic inactivation experiment, the aliquot of the 10-mL E. coli cell suspension was pipetted into a sterile 50 × 10-mm petri dish with the photocatalytic powder sample placed in the bottom at a fixed concentration of 1 mg photocatalyst/mL solution. Samples T0 and T3 were used in the photocatalytic inactivation experiments, and the P25 TiO₂ powder was also used for comparison purposes under the same experimental conditions. At regular time intervals, 100 μL of aliquots of the powder-treated cell suspensions was withdrawn in sequence. After appropriate dilutions in the buffer solution, aliquots of 100 μL were spread onto an agar medium plate and incubated at 37°C for 15 h. The number of viable cells in terms of colony-forming units was counted. Analyses were in duplicates, and control runs were carried out each time under the same experiment conditions but without any photocatalytic materials.

Figure 1 shows the crystal structure and morphology of the as-prepared pure TiO₂ nanocrystals (T0). Figure 1a displays the XRD pattern of T0 synthesized by the solvent-thermal process, which clearly demonstrates that all XRD peaks belonged to anatase TiO₂ (JCPDS No. 21–1272; space group: I41/amd(141)) with no rutile phase observed. Figure 1b shows the TEM observation of T0, which demonstrates that these nanocrystals were well faceted; square, hexagon, and rhombus shapes could be observed due to their different orientations. The high-resolution TEM (HRTEM) image of a single crystal with the rhombus shape (Figure 1c) demonstrates clear atomic planes with a lattice spacing of 0.352 nm, which correspond to the {101} lattice planes of anatase TiO₂. Figure 1d shows the HRTEM image of the side view of a single crystal. Atomic planes with a lattice spacing of 0.235 nm could be easily observed, which correspond to the {001} lattice planes of anatase TiO₂. From the HRTEM image observations, {001} lattice planes of the as-prepared pure TiO₂ nanocrystals were parallel to their top and bottom planes. Thus, exposed {001} facets were present in these as-prepared pure TiO₂ nanocrystals. From TEM/HRTEM observations and the symmetries of anatase TiO₂ nanocrystals 3 , the schematic geometry of the as-prepared pure TiO₂ nanocrystals could be described as the insert of Figure 1b. Their side length was approximately 20 to 40 nm, and their thickness was approximately 6 to 10 nm. From their size and geometry, the percentage of exposed highly reactive {001} facets in these as-prepared pure TiO₂ nanocrystals could be calculated by the method described by Zhu et al. 12 . The calculation was made on 20 TiO₂ nanocrystals, and the average percentage of {001} facets was determined at approximately 60%.

To introduce nitrogen/sulfur co-doping, these as-prepared pure TiO₂ nanocrystals were calcinated with thiourea at different temperatures of 300°C, 400°C, and 500°C, respectively, in a sealed tubular furnace. Figure 2 demonstrates the morphological evolution of T3, T4, and T5. After the heat treatment at 300°C, sample T3 kept the well-faceted geometry as the sample T0 (Figure 2a). {001} lattice planes of T3 were still parallel to their top and bottom planes, so the sample T3 also had exposed highly reactive {001} facets (Figure 2b). Compared with that of the sample T0, the side length of T3 decreased, while the thickness of T3 increased. With the increase of the calcination temperature to 400°C and 500°C, a clear geometry change could be observed for the samples T4 and T5 as demonstrated in Figure 2c,d, respectively. The well-faceted geometry of T0 gradually disappeared, and nanoparticles with non-uniform shapes appeared.

The XRD patterns of T3, T4, and T5 (not shown) demonstrate that all XRD peaks still belonged to anatase TiO₂ (JCPDS No. 21–1272; space group: I41/amd(141)) with no rutile phase observed. Thus, the calcinations of T0 with thiourea from 300°C to 500°C did not change the crystal structure of the obtained TiO₂ nanocrystals. However, the intensity of the (101) peak steadily increased with the increase of the calcination temperature, which suggests that the thermally stable {101} facets were enlarged. This observation is consistent with the TEM results. Thus, the calcination temperature should be carefully controlled to maintain the exposed highly reactive {001} facets. Under the current experimental conditions, the calcination temperature of 300°C is appropriate to maintain the exposed highly reactive {001} facets. BET specific surface area values of the samples T0 and T3 were measured. Because its size was in the nanometer range, the sample T0 had a large specific surface area of approximately 99 m²/g, which could largely enhance its contact efficiency with contaminants. After being calcinated with thiourea at 300°C, the specific surface area of T3 showed only a moderate decrease to approximately 70 m²/g. Thus, the sample of T3 also had a large specific surface area, which is desirable for its photocatalytic performance.

To investigate the chemical composition of TiO₂ nanocrystals with a high percentage of {001} facets after calcinations with thiourea, XPS investigations were conducted to obtain semi-quantitative composition data. Figure 3a shows a representative XPS survey spectrum of T3 (calcinations at 300°C), which clearly demonstrated the existence of Ti, O, N, S, F, and C. The existence of the C 1s peak could be attributed to the widespread presence of carbon in the environment. The relative element composition ratio was determined by multiplex high-resolution scans over Ti 2p, O 1s, N 1s, S 2p, and F 1s spectral regions. Multiplex high-resolution scans over N 1s, S 2p, and F 1s spectral regions are shown in Figure 3b,c,d, respectively. Figure 3b demonstrates that the N 1s peak could be best fitted by a combination of three N 1s peaks at 397.3, 398.5, and 399.8 eV, respectively. The N 1s peak at 397.3 eV is associated with Ti-N bonding 22 23 24 , which clearly demonstrated that nitrogen atoms were doped into the anatase lattice and replaced a small portion of oxygen atoms during the synthesis process, while the N 1s peaks at 398.5 and 399.8 eV represent some organic impurities or chemisorb N₂ on the sample surface. The doped N accounted for about 10.5% of the total N; thus, the doped N/Ti atomic ratio was determined at approximately 8.9%. This observation is different from some previous reports on nitrogen-doped TiO₂ with a high percentage of {001} facets wherein no doped N was found to form the Ti-N bonding 30 36 , which may be attributed to the different doping methods/conditions used. Figure 3c demonstrates that the S 2p peak could be best fitted by a combination of two S 2p peaks at 164.0 and 168.7 eV, respectively. The S 2p peak at 164.0 eV corresponded to the Ti-S bonding, while the S 2p peak at 168.7 eV corresponded to the surface S atoms adsorbed as SO₂ molecules 27 28 . The doped S accounted for about 34.8% of the total S; thus, the doped S/Ti atomic ratio was determined at approximately 1.0%. Figure 3d demonstrated that the F 1s peak position was at 684.2 eV, which corresponded to F⁻ ions physically adsorbed on the sample surface 3 33 34 . From the above XPS analysis results, it is clear that nitrogen/sulfur co-doping was introduced after calcinations with thiourea and that nitrogen/sulfur co-doped TiO₂ nanocrystals with a high percentage of {001} facets were successfully created by our approach. XPS analysis was also conducted on T4 and T5 samples with a higher calcination temperature, which showed a similar result of nitrogen/sulfur co-doping. For the sulfur dopant, the doping concentration did not change much. However, an obvious loss of nitrogen dopant was observed with the increase of the calcination temperature. The doped N/Ti atomic ratio sharply dropped from 8.9% for T3 to 0.67% for T5, which could largely deteriorate its visible-light absorption capability.

The optical properties of nitrogen/sulfur co-doped TiO₂ nanocrystals with a high percentage of {001} facets were investigated by measuring their diffuse reflectance spectra. The optical absorbance was then approximated from the reflectance data by the Kubelka-Munk function, as given by Equation 1:

F R = 1 − R 2 2 R ,

where R is the diffuse reflectance 37 . Figure 4a shows the light absorbance curves (in terms of Kubelka-Munk equivalent absorbance units) of nitrogen/sulfur co-doped TiO₂ nanocrystals with {001} facets (T3, T4, and T5), which were compared with that of pure TiO₂ nanocrystals without doping (T0) and commercially available P25 TiO₂ nanoparticles. P25 TiO₂ nanoparticles are widely used in the photocatalyst studies as a model TiO₂ with high photocatalytic performance. As expected, the light absorbance of P25 TiO₂ had the characteristic spectrum with the fundamental absorbance stopping edge at approximately 400 nm, so its light absorbance was limited mainly in the UV light range. Pure TiO₂ nanocrystals with a high percentage of {001} facets, however, demonstrated a moderate light adsorption capability from 400 to 500 nm, which had not been reported before. This observation may be attributed to the nanosize of the sample T0, which largely increased its {001} surface/bulk ratio. Ariga et al. 38 demonstrated using their scanning tunneling microscopy study that a surface state-mediated visible-light activity could occur on nanostructured TiO₂ {001} facets. Thus, the largely enhanced {001} surface/bulk ratio of the sample T0, compared with that of micro-sized TiO₂ single crystals with a high percentage of {001} facets, may lead to this observed visible-light absorbance. With nitrogen/sulfur co-doping, the visible-light absorbance of TiO₂ nanocrystals with a high percentage of {001} facets was largely enhanced. T3 demonstrated a much higher light absorbance from approximately 380 to 800 nm than T0, which could be attributed to the anion-doping effect from the co-doping of nitrogen and sulfur. Thus, it should possess a largely enhanced photocatalytic performance than pure TiO₂ nanocrystals with a high percentage of {001} facets. With the increase of the calcination temperature, the visible-light absorbance of T4 and T5 decreased gradually. T4 still possessed a good visible-light absorbance, while the light absorbance of T5 was similar with that of T0. This observation suggests that the calcination temperature for anion doping in our synthesis process is critical to the light absorbance of these samples. With the increase of the calcination temperature (over 300°C), the decomposition of the anion-doping source of thiourea accelerated, and more and more nitrogen and sulfur got lost without being doped into the anatase lattice. This observation is in agreement with the XPS measurement results that the anion-doping concentration decreased with the increase of the calcination temperature.

From the above studies of morphology, anion-doping concentration, and light absorption performance, it is clear that the sample T3 had the highest percentage of {001} facets, anion-doping concentration, and visible-light absorption. Thus, the visible-light photocatalytic activities of the nitrogen-doped TiO₂ nanocrystal photocatalyst with a high percentage of {001} facets were evaluated with the sample T3 by its photocatalytic degradation of MB and photocatalytic inactivation of E. coli under visible-light illumination (λ > 400 nm) and compared with that of T0 and commercially available P25 TiO₂ nanoparticle photocatalyst. Photocatalytic degradation of MB was conducted by exposing the MB solution with various photocatalysts under visible light for varying time intervals (from 5 to 30 min). After the centrifugation to recover photocatalysts, the light absorption of the clear solution was measured, and the remaining percentage of MB in the solution was calculated by the ratio between the light absorptions of photocatalyst-treated and photocatalyst-untreated MB solutions. Figure 4b shows the MB residue ratio changes in treated MB solutions with different treatment times. P25 TiO₂ demonstrated a fair degradation performance on MB under visible-light illumination, which could be attributed to its mixed nature of both anatase and rutile phases of TiO₂. Its superior dispersity in an aqueous environment also contributes to its efficient contact with organic pollutants. After a 30-min treatment, its degradation effect slowed down, and the residual MB percentage was approximately 30%. Without nitrogen doping, the sample T0 showed a slightly better degradation effect on MB compared with the P25 TiO₂ nanoparticles under visible-light illumination, which could be attributed to its moderate light adsorption capability in the wavelength range from 400 to 500 nm as demonstrated in Figure 4a. The sample T3, however, demonstrated a much faster degradation effect on MB under visible-light illumination than both P25 TiO₂ and T0. Within just 5 min, most of the MB in the solution was degraded, and the residual MB percentage dropped to just approximately 14%. After a 30-min treatment, the residual MB percentage was near zero, suggesting a complete degradation of MB.

The photocatalytic activity of sample T3 was further demonstrated by its bactericidal effect on the viability of E. coli cells and compared with that of sample T0 and P25 TiO₂. The photocatalytic inactivation experiments on E. coli were conducted by exposing the cells suspended in the buffer solution with various photocatalysts under visible-light illumination (λ > 400 nm) for varying time intervals. The survival ratio of E. coli was determined by the ratio of N _t/N ₀, where N ₀ and N _t are the numbers of colony-forming units at the initial time and each following time interval, respectively. Figure 4c shows the E. coli survival ratio under various treatments. P25 TiO₂ showed no obvious bactericidal effect under visible-light illumination. After a 2-h treatment, the E. coli survival ratio was still around 95%. The sample T0 showed a slightly better photocatalytic inactivation effect on E. coli under visible-light illumination compared with P25 TiO₂ nanoparticles. After a 2-h treatment, the E. coli survival ratio was around 85%. The sample T3 demonstrated a much better photocatalytic inactivation effect on E. coli under visible-light illumination. The survival ratio of E. coli showed a continuous decreasing trend with the increase of visible-light illumination time. After a 2-h treatment, the survival ratio of E. coli dropped to around 50%. This observation is in agreement with the photocatalytic degradation experiment results demonstrated in Figure 4b. Thus, the combination of the large specific area due to their nanosize, high percentage of active {001} facets, and enhanced visible-light adsorption due to nitrogen/sulfur co-doping optimized the photocatalytic performance of TiO₂ under visible-light illumination.