Strengthened Removal of Tetracycline by a Bi/Ni Co-Doped SrTiO₃/TiO₂ Composite under Visible Light

Abstract

A two-step hydrothermal method was used to first obtain a SrTiO₃/TiO₂ composite then to dope the composite with Bi, Ni and Bi/Ni. Morphology, crystalline structures, surface valances and optical features of SrTiO₃/TiO₂ and Bi-, Ni-, Bi/Ni-doped SrTiO₃/TiO₂ were assessed. XRD and XPS analysis showed that Bi and Ni were successfully doped and existed in Bi(3+) and Ni(2+) oxidation state. UV–vis analysis further revealed that the bandgap energies of TiO₂ and SrTiO₃/TiO₂ were calculated to be 3.14 eV and 3.04 eV. By comparison, Bi, Ni and Bi/Ni doping resulted in the narrowing of bandgaps to 2.82 eV, 2.96 eV and 2.69 eV, respectively. The removal ability of SrTiO₃/TiO₂ and doped SrTiO₃/TiO₂ were investigated with tetracycline as the representative pollutant. After 40 min of exposure to visible light, Bi/Ni co-doped SrTiO₃/TiO₂ photocatalyst was able to remove 90% of the tetracycline with a mineralization rate of about 70%. In addition, first-order removal rate constant was 0.0074 min⁻¹ for SrTiO₃/TiO₂ and increased to 0.0278 min⁻¹ after co-doping. The strengthened removal by co-doped photocatalyst was attributed mainly to the enhanced absorption of visible light as co-doping resulted in the decreases of bandgap energies. At the same time, the co-doped material was robust against changes in pH. Removal of tetracycline was stable as pH changed from 5 to 9. Tetracycline removal was inhibited to a certain degree by the presence of nitrate, phosphate and high concentration of humic acid. Moreover, the co-doped material exhibited strong structural stability and reusability. In addition, a photocatalysis mechanism with photogenerated holes and ·O₂⁻ radicals as main oxidative species was proposed based on entrapping experiments and EPR results.

1. Introduction

Antibiotics are widely recognized as effective in fighting against pathogenic infection in humans and animals. Studies have shown that the majority (more than 70%) of antibiotics are used for animals raised for food [1]. After consumption, antibiotics were usually discharged in their original forms or as metabolites through feces and urine [2,3]. These antibiotic residues could then directly or indirectly enter the aquatic and terrestrial environment. The omnipresence of antibiotics raised concerns about their adverse effects on ecological systems [4,5].

Tetracyclines are widely used as human and veterinary medicines for their broad-spectrum antibacterial ability [6]. Disposal of unused or expired tetracycline medications and excretion by humans and animals resulted in tetracycline pollution of environment. Concentrations of tetracycline in aqueous systems were usually in the ng/L or μg/L level and could reach the mg/L level in wastewater from pharmaceutical factories [7]. Technologies for tetracycline removal included advanced oxidation [8], adsorption [9], membrane separation [10], biodegradation [11] and photocatalysis [12]. Among them, photocatalysis is regarded as one of the most promising approaches because of its high removal efficiency and mineralization ability. Tetracylines’ removal via photocatalysis involved generation of reactive radicals such as hydroxyl radical (·OH), superoxide radical (·O₂⁻) and hole (h⁺) under irradiation [13]. A variety of photocatalysts such as metal oxides [14], layered double hydroxide [15], metal sulfides [16] and g-C₃N₄ [17] have been developed to removal tetracycline.

Perovskite-type oxides with a general formula of ABO₃ were extensively applied as catalysts in photocatalytic processes, energy storage and conversion [18,19]. A in the structure represents an alkaline earth metal or a rare element, while B is a transition metal cation or p-block metal. Among them, titanate perovskites (ATiO₃), where A is usually Ca, Sr, Ba, etc., are renowned for their broad bandgap and exceptional electronic, optical, magnetic and photocatalytic characteristics [20]. Owing to their robustness against photo corrosion and physicochemical stability, titanate perovskites were widely used in photocatalytic processes. Moreover, changing the coordination of the active atoms could effectively change the electronic structure of perovskite catalysts and improve catalytic performance.

Strontium titanate (SrTiO₃) is notable for its cubic-like structure and superior thermal stability. It is a semiconductor with a wide bandgap energy of 3.2 eV and has been extensively investigated in the degradation of both organic and inorganic contaminants [21,22]. Recent studies on strontium titanate focused on the enhancement of its photocatalytic ability under visible light. This can be achieved through the creation of heterojunctions or introduction of impurities into the lattice structure [23]. Doping was found to be an effective method for the introduction of lattice defects and modification of electronic structure. Doping with lanthanum, niobium or iron was found to lead to obvious improvement of electrochemical properties [24].

In this research, a SrTiO₃/TiO₂ composite was obtained via hydrothermal treatment of TiO₂ and Sr(NO₃)₂. Bi, Ni and Bi/Ni were doped into the SrTiO₃/TiO₂ to investigate the effects of single or dual doping. Changes in microstructures and photocatalytic properties before and after doping were compared. Tetracycline was used as a representative pollutant to evaluate their photocatalytic performance under visible light. In addition, the effects of application conditions (dosage, pH and presence of other substance) on the tetracycline removal efficiency were also studied in detail. Mechanism of removal was explored through active substance capture. The ultimate goal was to provide insight into doping as a method to enhance photocatalytic removal of antibiotics.

2. Results and Discussion

2.1. Characteristics of Photocatalysts

2.1.1. Morphology

Figure 1 shows the SEM images of TiO₂, S-TO, Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO. TiO₂ showed a well-defined structure with a relatively uniform particle size. The image of S-TO revealed the presence of a cubic structure, indicating the generation of SiTiO₃ on the TiO₂ surface [25]. Morphologies of Bi- or Ni-doped S-TO (Figure 1c,d) were quite similar to that of S-TO. The co-doping of Bi and Ni together led to a highly compact and dense structure with blurred surface contours, as shown in Figure 1e. This could be due to the deformation of crystal defects by doping. N₂ adsorption–desorption isotherms were also obtained and are shown in Figure S1 of the Supplementary Materials. All photocatalysts showed type IV isotherms with H3 hysteresis loop, indicating irregular pore structures.

Table 1. BET surface areas and pore volumes of the photocatalysts.

Sample	BET Surface Area (m2/g)	Pore Volume (cm3/g)	Average Pore Size (nm)
TiO2	50	0.18
S-TO	63.3	0.33	20.8
Bi/S-TO	59.8	0.30	20.4
Ni/S-TO	65.6	0.32	20.1
Bi/Ni/S-TO	63.3	0.29	18.8

lists the BET surface areas, average pore sizes and pore volumes. TiO₂ data were obtained from a supplier. S-TO composite manifested a significantly higher surface area and pore volume compared with the original TiO₂. The pore structures of S-TO were maintained after Bi, Ni doping, as proven by only slight changes in surface area, pore volume and pore size. This is in accordance with SEM observation that doping did not change significantly the morphology of the SrTiO₃/TiO₂ structure.

2.1.2. Crystal Structures

XRD patterns are shown in Figure 2. Distinct diffraction peaks of TiO₂ found at 2θ of 25.3°, 37.8°, 48.1°, 54.0°, 55.1°, 62.7° and 75.0° corresponded to the (101), (004), (200), (105), (211), (204) and (215) crystal planes of anatase TiO₂, respectively (PDF No.21-1272) [26]. S-TO, on the other hand, revealed characteristic diffraction peaks at 2θ of 22.8°, 32.2°, 39.8°, 46.5°, 57.7° and 67.8°. These matched the crystal planes of SrTiO₃ at (100), (110), (111), (200), (211) and (220), respectively (PDF No. 35-0743) [27]. The positions of the diffraction peaks remain basically unchanged after the introduction of Bi, Ni or Bi/Ni into S-TO. No Bi-, Ni-related peaks were found. Contents of Bi and Ni were relatively low (4.0 wt% and 8.0 wt%, respectively) and were most likely dispersed in the S-TO structure, thus not detected by XRD. However, after doping, the intensity of the (101) diffraction peaks of TiO₂ weakened. This suggested that the Bi, Ni doping increased the lattice defects and decreases the crystallinity of S-TO. In addition, the diffraction peak at 25.3° shifted slightly to a higher angle (Figure 2b), proving decreases in lattice constant when dopant was incorporated into the original structure [28]. Furthermore, crystal sizes and inter-planar spacing between atoms (d-spacing) were estimated by Debye–Scherer formula and Bragg’s Law using (101) plane. Results are shown in Table S1 of the Supplementary Materials. Crystal size decreased slightly after doping. The average crystal sizes of TiO₂, S-TO, Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO are 27.1 nm, 26.2 nm, 25.4 nm, 23.2 nm and 20.3 nm, respectively. It seems that the doping of Ni and Bi in the S-TO framework prevented crystal growth. As shown in Table S1, d-spacing values are all 3.51 nm, indicating that Bi, Ni doping did not change the crystallinity, and Bi, Ni were most likely inserted into the S-TO lattice.

2.1.3. Elemental Composition and Chemical State Analysis

XPS spectra of Bi/Ni/S-TO were obtained to illustrate the elemental composition and chemical state. Binding energies were corrected using carbon at 284.8 eV as standard. The survey spectrum in Figure 3a shows the presence of Ti, O, Sr, Bi and Ni, suggesting that Bi and Ni were successfully incorporated. Figure 3b–f provide the exhaustive XPS results of O 1s, Ti 2p, Sr 3d, Bi 4f and Ni 2p. For O 1s, the peak at the 531.5 eV energy band position is close to the lattice oxygen, while the one at 528.5 eV corresponded to the adsorbed oxygen [29]. The XPS profile of Ti 2p revealed two characteristic peaks (Figure 3c), consisting of Ti 2p_3/2 (457.98 eV) and Ti 2p_1/2 (463.95 eV), with a difference in binding energy ∆E of 5.96 eV. In addition, a satellite peak at 471.44 eV was also found. This indicated that Ti is present in the material in the tetravalent oxidation state [30]. The Sr spectrum (Figure 3d) included Sr 3d_5/2 (132.74 eV) and Sr 3d_3/2 (134.47 eV), indicating that Sr is present in the form of Sr²⁺. Bi consists also of two characteristic peaks at binding energies of 158.58 eV and 163.88 eV, as shown in Figure 3e. These were assigned to Bi 4f7/2 and Bi 4f5/2 orbitals, revealing that Bi is of trivalent state [31]. In Figure 3f, The XPS spectra of Ni 2p show peaks at binding energies of 855.51 eV and 879.25 eV. These are the characteristic peaks of Ni 2p_3/2 and Ni 2p_1/2 spin-orbit doublets. The other two peaks at 861.43 eV and 871.17 eV were the satellite peaks of Ni 2p_3/2 and Ni 2p_1/2 orbitals. It could infer that Ni was in a +2 oxidation state [32]. The ionic radii of Ti⁴⁺, Sr²⁺, Bi³⁺ and Ni²⁺ are 0.06 nm, 0.18 nm, 0.069 nm and 0.103 nm, respectively. The similarity of radius makes it possible for Bi or Ni to substitute Ti or Sr in the SrTiO₃/TiO₂ framework.

2.1.4. Optical Properties

The effects of Bi, Ni doping on light absorbance were revealed by UV–vis absorbance spectra, as shown in Figure 4a. The visible light absorption of Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO was significantly enhanced compared to TiO₂ and S-TO. The absorbance edges of S-TO after Bi, Ni or Bi/Ni doping showed systematic redshifts, and a widest absorbance edge of approximately 480 nm for Bi/Ni/S-TO was observed. This showed that incorporation of Bi and Ni could enhance the photo-absorption of visible light. With the doping of Ni and Bi, the difference in electronegativity may lead to the redistribution of electron density and thus reconstruction of the internal electric field. The may result in the inhibition of the complexation of e⁻-h⁺ pairs.

Bandgap energies (Eg) were next calculated from Tauc Equation (1) [33].

$$\alpha h\upsilon =A(h\upsilon -E_g{)}^{n/2}$$

(1)

where α, h, υ and A were absorbance parameter, Plank’s parameter, light frequency and absorbance, respectively. The value of n could be 1 and 4 for direct and indirect transition semiconductors, respectively. Results are shown in Figure 4b. The bandgap energies of TiO₂ and S-To were 3.14 eV and 3.04 eV, respectively. Single doping of Bi or Ni narrowed the bandgap energy to 2.82 eV and 2.96 eV, while co-doping further narrowed the bandgap to 2.69 eV. The narrowing of bandgap energy could result in the enhancement of photocatalytic performance [34]. Nematollahi et al. [35] also observed that, with Bi doping, the bandgap energy was narrower. They argued that this is because the energy state of Bi³⁺ 6s is located above the TiO₂ valance layer (O 2p state).

2.2. Photocatalytic Tetracycline Removal

Results from characterization showed that Bi, Ni doping broadened the response to visible light and narrowed the bandgap energy of SrTiO₃-TiO₂ without causing obvious changes in SrTiO₃-TiO₂ morphology. Next, tetracycline was selected as the representative pollutant to compare the photochemical performance of TiO₂, SrTiO₃-TiO₂ as well as Bi-, Ni-doped SrTiO₃-TiO₂. Figure 5a shows the changes in tetracycline removal at different reaction time under visible light by different photocatalyst. The dark reaction was carried out for 30 min before light exposure; 5–10% of the initial tetracycline was removed during the dark stage by TiO₂, S-TO, Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO. This could be attributed to adsorption. With illumination, removal of tetracycline by TiO₂ was still limited. Even after 120 min of reaction, removal efficiency did not go beyond 14%. By comparison, removal by S-TO reached 58% at 120 min. Doping could greatly enhance tetracycline removal. Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO were able to remove 77%, 80% and 90%, respectively, of the initial tetracycline after only 40 min. Additionally, the doping of Bi and Ni improved not only efficiency but also removal rate. Kinetics calculation manifested that tetracycline removal fitted first-order kinetics. Reaction rate constants of different photocatalysts are shown in Figure S2 of the Supplementary Materials. Bi/Ni/S-TO achieved the highest reaction rate constant of 0.0278 min⁻¹, which is 26.2 times higher than that of the original TiO₂ and 3.7 times higher than that of S-TO. Comparison of the five photocatalysts showed that Bi/Ni/S-TO achieved the best tetracycline removal performance. Therefore, factors that affect Bi/Ni/S-TO removal, degree of mineralization of removed tetracycline and the material’s reusability and stability were next investigated in detail. The removal efficiency and kinetics by Bi/Ni/S-TO were comparable to similar studies of photocatalysis. Peng et al. [36] achieved a tetracycline removal of 98% after 150 min via a ZnLa_xBi_2−xO₄ structure. In a study of Maggu et al. [37], 91% of tetracycline was removed by Bi₂O₃/Sb₂S3 with a rate constant of 0.01749 min⁻¹.

Figure 5b is the effect of Bi/Ni/S-TO dosage on removal of tetracycline. Tetracycline removal increased significantly from 42% to 91% at 40 min with the increase in dosage from 0.4 g/L to 1.2 g/L, then levelled off between 1.2 g/L and 1.6 g/L. Removal efficiency actually dropped slightly as dosage increased further to 2.0 g/L. Higher dosage means more active sites, thus an initial increase in removal. However, as dosage increased to higher than 2.0 g/L, the trend reversed. This could be explained by the fact that an excessive amount of the material would lead to an increase in turbidity which, in turn, reduced light penetration and production of active substances in the system [38]. The optimal dosage was selected to be 1.2 g/L.

Removal of tetracycline by Bi/Ni/S-TO under different pH was also explored. Tetracycline is an amphoteric compound. It exists mainly in the cationic form at pH < 3.3, while at pH around 7.68, zwitterions predominated [39]. At pH > 7.68, it is in the anionic form [40]. This change in electrochemical properties may affect the interaction between tetracycline and Bi-Ni-S/TO. As shown in Figure 5c, Bi/Ni/S-TO was robust against pH changes. Removal efficiency remained stable at 90% at pH 5–9 and only declined in extreme acidic (pH < 3) and alkaline (pH > 11) conditions. The most favorable pH condition for Bi/Ni/S-TO was near neutral. This may be because under neutral conditions, the h⁺ and free radicals in the system are in a co-activated state. When pH is too low, the H⁺ in the system increased, which could eliminate ·OH by forming water. At high pH, both tetracycline and Bi/Ni/S-TO were negatively charged, which resulted in charge repulsion and reduced contact between pollutant and photocatalyst.

Besides fluctuation in pH, wastewater polluted by tetracycline often also contained many co-existing substances. In this study, anions Cl⁻, SO₄²⁻, CO₃²⁻/HCO₃⁻, NO₂⁻, PO₄³⁻ and humic acid were selected as representatives to investigate their effects. The concentration of anion was set at 20 mmol/L. As shown in Figure 5d, the effects of SO₄²⁻ and CO₃²⁻/HCO₃⁻ were negligible, while the addition of Cl⁻ resulted in about 5% decline. In contrast, the presence of NO₂⁻ and PO₄³⁻ was highly inhibitory, and removal of tetracycline dropped from 90% at 40 min to 23% and 40%, respectively. Cao et al. [41] found that NO₂⁻ and PO₄³⁻ would compete with the tetracycline for the active sites on the catalyst surface. Additionally, NO₂⁻ and PO₄³⁻ increased the polarity of the catalyst surface, which also prevented tetracycline from reaching the active sites. As a matter of fact, H₂PO₄⁻ was commonly used as radical entrapper as it can react with h⁺ and ·OH to generate less reactive radicals ·H₂PO₄ [42]. Overall, the effect of anions followed the order of PO₄³⁻ > NO₃⁻ > Cl⁻ > SO₄²⁻ ≈ CO₃²⁻/HCO₃⁻.

Humus is omnipresent in natural water bodies. It is believed that 45% to 85% of the dissolved organic carbon in water bodies was humus. Most of the humus in nature exists as humic acid, humin and fulvic acid with concentration around 10 mg/L. Therefore, the concentrations of humic acid were set at 1 mg/L, 5 mg/L, 10 mg/L and 20 mg/L (Figure 5e). The removal rate of tetracycline increased slightly (about 5%), indicating that the presence of low concentration of humic acid may be conducive to photocatalytic reaction. As a matter of fact, humic acid is a natural photosensitive substance and could adsorb light energy and be conducive to energy conversion. However, the inhibitory effect on tetracycline removal increased gradually with the increase in humic acid concentration. At 20 mg/L, tetracycline removal efficiency was 85.9% at 40 min as compared with more than 90% when no humic acid was present. The decline was caused probably by competition for active radicals between humic acid and tetracycline. In summary, both anions and organics had a significant impact of photocatalytic removal of tetracycline. It is important that when used for real wastewater, pre-treatment may be necessary to remove, for instance, N and P first.

Although Bi/Ni/S-TO showed high removal of tetracycline, it is also important to identify the percentage of mineralization, that is, the percentage that is degraded to CO₂ and H₂O. TOC removal efficiency was commonly used to determine the extent of mineralization. As shown above, more than 90% of the tetracycline was removal after 40 min of reaction. By comparison, TOC removal reached only about 40% at 40 min and was 70% as time was extended to 100 min. The lagging of TOC removal shows that tetracycline degradation by photocatalytic process was most likely a step-by-step process, and tetracycline was degraded to intermediate organics and then completely mineralized. That is, mineralization took a longer time to finish, and complete mineralization was not achieved even after 120 min of reaction.

The reusability and stability of Ni/Bi/S-TO was studied by multiple run of tetracycline removal, and the results are shown in Figure 5g. Ni/Bi/S-TO showed exceptional stability. There is no obvious decrease in removal ability after five runs. XRD pattern (Figure 5h) remained basically unchanged after five uses, showing that crystalline structures were maintained.

2.3. Mechanism of Tetracycline Removal

2.3.1. Free Radical Trapping and EPR Analysis

In order to determine the main reactive species involved during photocatalytic removal by Bi/Ni/S-TO of tetracycline, triethanolamine (TEA), tert-butanol (TBA) and p-benzoquinone (BQ) were used, respectively, as h⁺, ·OH and ·O₂⁻radical trapping reagents. Results of tetracycline removal with and without these reagents are shown in Figure 6a. Tetracycline removal decreased slightly with the addition of TBA. However, only 58.9% and 25.1% of tetracycline was removed with the introduction of TEA and BQ, respectively, and 38.5% and 72.3% lower than that when no reagent was added. The inhibitory effect of TEA and BQ indicated that h⁺ and ·O₂⁻ were the two main active substances in the photocatalytic removal of tetracycline, while ·OH only played a secondary role.

EPR spectroscopy was next used to verify the generation of h⁺ and ·O₂⁻ by Bi/Ni/S-TO in Figure 6b,c. In Figure 6b, four specific peaks of ·O₂⁻ were observed under visible light, while no ·O₂⁻ signal was observed in darkness. The intensity of ·O₂⁻ signal weakened as time prolonged. Figure 6c shows peaks of h⁺ under irradiation, which was absent in darkness. The EPR spectra verified that Bi/Ni/S-TO composites can produce h⁺ and ·O₂⁻ under visible light excitation.

2.3.2. Mechanism of Tetracycline Removal by Bi/Ni/S-TO

The forbidden band width of the Bi/Ni/S-TO was calculated from the bandgap energy to be 2.69 eV, as shown in Figure 4b. According to the VB-XPS spectrum in Figure 7a, E_VB.NHE is 1.51 eV. Therefore, the valence band (E_VB) is calculated to be around 1.27 eV based on Equation (2). And the conduction band (E_CB) is −1.42 eV.

$${\mathrm{E}}_{\mathrm{V}\mathrm{B}.\mathrm{N}\mathrm{H}\mathrm{E}}=\mathsf{\phi }+{\mathrm{E}}_{\mathrm{V}\mathrm{B}.\mathrm{X}\mathrm{P}\mathrm{S}}-4.44\mathrm{e}\mathrm{V}$$

(2)

where E_VB.NHE is the E_VB of the corresponding standard hydrogen electrode, φ is the work function of the instrument (4.2 eV).

Based on the experimental and characterization results, a possible mechanism for removal of tetracycline by Bi/Ni/S-TO was deduced (Figure 7b). Photogenerated electrons excited by visible light were trapped by Bi³⁺ and Ni²⁺ on Bi/Ni/S-TO which inhibit the carrier complexation and improve the photocatalytic efficiency [43]. In addition, based on the free radical trapping experiments, it is understood that ·O₂⁻ and h⁺ are the main active species involved in the photocatalytic removal of tetracycline by Bi/Ni/S-TO. And the standard redox potential (−0.046 eV vs. normal hydrogen electrode, NHE) of O₂/·O₂⁻ is between VB and CB of Bi/Ni/S-TO. The dissolved oxygen in water can react with photogenerated electrons to generate ·O₂⁻ which, in turn, oxidized and decomposed tetracycline [44]. This explained the high mineralization rate, as shown in Figure 5g. On the other hand, VB potential of the Bi/Ni/S-TO material under light is negative compared to the standard redox potential of H₂O/·OH (2.26 eV vs. NHE), and therefore did not directly generate ·OH. However, a small amount of radicals ·OH can be produced by the conversion of ·O₂⁻ to H₂O₂ and further decomposition. This is in accordance with results from entrapping experiment that ·OH contributed only to a small degree to tetracycline removal [45]. Meanwhile, the h⁺ left by Bi/Ni/S-TO in VB could have participated to remove tetracycline. With ·O₂⁻ and h⁺, tetracycline was oxidized to smaller molecules, and some are mineralized to H₂O and CO₂. In summary, the photocatalytic mechanism of Bi/Ni/S-TO composites follows the charge migration pathway, as shown in Figure 7b.

3. Materials and Methods

3.1. Materials

TiO₂ (≥98.0%), strontium nitrate (Sr(NO₃)₂) (≥99.5%), bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O)(≥99.0%), nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O) (≥99.0%) were all of analytical grade and purchased from Sinopharm Chemical Reagent Co., Shanghai, China.

3.2. Material Preparation

3.2.1. Preparation of SrTiO₃/TiO₂

SrTiO₃/TiO₂ was prepared by a two-step hydrothermal method [46,47]. First, 0.16 g of strontium nitrate and 0.1 g of titanium dioxide were added to 20 mL of distilled water. The mixture was stirred for 30 min then transferred to a Teflon-lined autoclave. The mixture was allowed to react under 160 °C for 8 h then cooled naturally. The resultant powder was washed with distilled water and ethanol and dried at 60 °C to obtain SrTiO₃/TiO₂ (labelled as S-TO).

3.2.2. Preparation of Bi/SrTiO₃/TiO₂ and Ni/SrTiO₃/TiO₂

An amount of 0.1 g of S-TO was added to 20 mL of bismuth nitrate (3.71 mmol/L) or nickel nitrate (1.89 mmol/L) solution, respectively. That is, the theoretic percentages of Bi and Ni doping were 4.0 wt% and 8.0 wt%, respectively. After thorough mixing, the pH of the mixture was adjusted to 12 with 1 mol/L NaOH then transferred into an autoclave and subjected to react at 150 °C for 24 h, the resultant materials were Bi-doped or Ni-doped SrTiO₃/TiO₂. Thus, obtained materials were washed, dried at 60 °C and labelled as Bi/S-TO and Ni/S-TO.

3.2.3. Preparation of Bi/Ni/SrTiO₃/TiO₂

An amount of 0.1 g of S-TO was added to 40 mL of solution containing bismuth nitrate and nickel nitrate. Concentrations of bismuth nitrate and nickel nitrate were 1.85 mmol/L and 0.95 mmol/L, respectively. Thus, the percentages of Bi and Ni doping were also kept at 4.0 wt% and 8.0 wt%, respectively. Solution pH was then adjusted to 12 and the mixture allowed to react in an autoclave at 150 °C for 24 h. The resulting material was washed and dried to obtain the Bi/Ni/SrTiO₃/TiO₂ (labelled as Bi/Ni/S-TO).

3.3. Characterizations

The crystalline structure and morphology of the materials were analyzed via X-ray diffraction (XRD, Rigaku Ultima IV, Tokyo, Japan) and scanning electron microscopy (SEM, ZEISS Gemini 300, Oberkochen, Germany). Surface compositions, electronic states and defects and optical properties were obtained by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, Tokyo, Japan), FLS1000 photoluminescence spectrometer (Edinburgh Instrument Company, Livingston, UK) and UV–vis diffuse reflection spectrometer (Lamada950, Platinum Elmer Co., Ltd., Tokyo, Japan), respectively.

3.4. Evaluation of Photocatalytic Performance

The photocatalyst tested included TiO₂, S-TO, Bi/S-TO, Ni/S-TO and Bi/Ni/S-TO. The evaluations of photocatalytic performance of these materials were carried out in photochemical equipment (Model BXU034, Guangzhou Xingchuang Electronics Co., Ltd., Guangzhou, China). For each photocatalyst, 30 mg of the material was added to 10 quartz tubes with 25 mL of tetracycline solution (20 mg/L). One tube without the addition of photocatalyst was used as control. The mixture was stirred in the dark for 30 min before it was exposed to visible light. A 300 W xenon lamp equipped with a UV cut-off glass filter was used as the visible light source. One quartz tube was taken out every 20 min and centrifuged. The supernatant was analyzed for concentrations of tetracycline and TOC with an ultraviolet–visible spectrophotometer (UV-5100, Metash, Shanghai, China) and a multi N/C 3100 analyzer (Analytikjena, Jena, Germany). To test the reusability, the photocatalyst underwent 5 successive runs. Specifically, after each test of tetracycline removal, the material was collected by centrifugation then washed thoroughly and dried for another use.

3.5. Investigation on Factors Affecting Photocatalytic Performance

The effects of dosage, pH and the presence of other substances on the tetracycline removal were explored. Dosage ranged from 0.4 to 2.0 g/L. The initial pH was set at 3, 5, 7, 9 and 11. The presence of Cl⁻, SO₄²⁻, CO₃²⁻/HCO₃⁻, NO₃⁻, PO₄³⁻ and humic acid on tetracycline removal were also investigated. The concentrations of co-existent anion were 20 mmol/L, while the concentration of humic acid varied from 5 to 20 mg/L. The photocatalytic reaction procedure was the same as in Section 3.4.

All tests above were conducted in triplet, and averages and standard deviations were calculated.

3.6. Free Radical Trapping and Electron Paramagnetic Resonance (EPR) Analysis

Tert-butanol (TBA), triethanolamine (TEA) and p-benzoquinone (BQ) were selected as free radical trapping reagents. An amount of 30 mg of photocatalyst was added to quartz tubes containing 25 mL of tetracycline solution of 20 mg/L. Then, 1 mL of TBA, TEA or BQ with a concentration of 1 mmol/L was added. The procedure on photochemical reaction was the same as in Section 3.4. In addition, EPR analysis was carried out via Bruker (Billerica, MA, USA) EMXplus-6/1 with 5,5-dietyl-1-pyrroline-N-oxide (DMPO) and 2,2,6,6-tetramethylpiperidinyl-1-oxide (TEMPO) as trapping reagents.

4. Conclusions

Photocatalyst Bi/Ni/S-TO was synthesized by a two-step hydrothermal method. Characterization results showed that although doping of Bi/Ni did not significantly change the morphology and pore structure of SrTiO₃/TiO₂, it increased the adsorption of visible light by narrowing the bandgap energy. Meanwhile, the formation of photogenerated electron–hole pairs was promoted, and the separation efficiency of photogenerated carriers was improved. Tetracycline removal studies revealed that Bi/Ni/S-TO showed the highest removal efficiency and rate. After 40 min, 90% of the initial tetracycline was removed. In addition, the removal was dependent on dosage, initial concentration, pH and presence of other materials. High dosage of Bi/Ni/S-TO may result in high turbidity, thus low light penetration. There is also an optimal pH range of 5–9. Attention should be paid to the presence of anions the like of NO²⁻ and PO₄³⁻, which could severely inhibit photocatalytic reaction. On the other hand, a small amount of HA (<5 mg/L) may actually enhance tetracycline removal, while high concentration was inhibitory. Furthermore, Bi/Ni/S-TO exhibited remarkable stability and reusability, indicating their potential applications in real situation. Free radical entrapping experiment reaction and EPR analysis identified ·O₂⁻ and h⁺ as the main active species. The weight percentages of Bi and Ni doping were set at 4.0% and 8.0%, respectively, in this research. In light of the obvious improvement in photocatalytic performance, it is advisable to expand the dosage to a wider range to optimize the doping process.