2.1. Characterization
XRD was used to analyse the crystalline phases of the prepared samples. As shown in
Figure 1, the diffraction angles of the anatase phases corresponding to (101), (004), (200), and (211) of anatase crystals are 25.3°, 37.9°, 48.1°, and 55.1°, respectively, which proves that anatase structures are present in all TiO
2-based materials. The diffraction peaks of the samples with dimethylformamide (DMF) as the solvent (diffraction peaks at 2 θ = 25.3°) are broader and weaker than those of the samples with ethanol (Et) as the solvent, which may indicate that the samples dissolved in ethanol have a better crystallisation effect compared to DMF. In addition, as shown in
Table 1, we calculated the crystallinity of the samples and the results also indicate that the samples dissolved in ethanol have higher crystallinity. This revealed a preference for ethanol as a solvent in the crystallization of tungsten-doped TiO
2 powders, over DMF. It is well known that crystallinity is crucial for photocatalytic activity. Materials with elevated crystallinity generally exhibit improved properties in photocatalysis attributed to the well-ordered and uniform structures, which facilitate charge transfer from the center to the surface [
38]. The analysis indicates that tungsten-doped TiO
2 synthesized utilizing WCl
6 as the tungsten precursor potentially exhibits enhanced photocatalytic performance, attributed to its superior crystallinity, compared with Na
2WO
4 and AMT. For all samples, no phases other than the anatase phase were detected. The formation of WO
3 was not evidenced by the characteristic peaks typically observed in XRD spectra, as noted by other authors. This may be due to the low concentration of WO
3, which was insufficient for detection by XRD [
29,
39]. The absence of a characteristic WO
3 peak in the XRD pattern of tungsten-doped TiO
2 suggests that tungsten ions either formed W–O–Ti bonds within the lattice or occupied interstitial sites [
40]. Several studies have shown that tungsten ions successfully substitute titanium ions within the TiO
2 crystal lattice due to the similar ionic radius (W
6+ at 0.060 nm and Ti
4+ at 0.0605 nm) [
40].
The grain size of the prepared material was calculated using the Scherrer formula. The results were presented in
Table 1. As illustrated in
Table 1, the average grain size of the tungsten-doped TiO
2 samples is smaller than that of the un-doped TiO
2 samples, indicating that doping with various tungsten precursors affected the crystallinity of materials and thus hinders the crystal growth.
The FTIR spectra of the pristine TiO
2 and the tungsten-doped TiO
2 materials between 4000 cm
−1 and 1000 cm
−1 are shown in
Figure 2. There are only two distinct absorption bands around 3450 cm
−1 and 1630 cm
−1 in
Figure 2, which represent the stretching vibrations of the water and hydroxyl groups, respectively [
41]. As illustrated in
Figure 2, the introduction of tungsten results in the slight enhancement of the peaks at 3450 cm
−1 and 1630 cm
−1, suggesting an increase in the presence of water and hydroxyl groups, respectively. This observation implies that the addition of tungsten may lead to a higher concentration of hydroxyl groups within the tungsten-doped TiO
2. The Lewis surface acidity of tungsten-doped TiO
2 increases with the addition of tungsten [
42], making it easier to adsorb water and form surface hydroxyl groups, while the –OH groups can capture photogenerated holes (h
+) and convert them into active •OH radicals [
43], thus improving the photocatalytic performance of the material.
The SEM and EDS mapping images of the pristine TiO
2 and tungsten-doped TiO
2 are shown in
Figure 3. It can be found that pristine TiO
2 exhibits textural characteristics with a dominant presence of irregularly shaped aggregates accompanied by interparticle voids [
44]. This observation highlights the agglomeration tendency of pristine TiO
2 [
20]. Tungsten-doped TiO
2 with different tungsten precursors of WCl
6, Na
2WO
4●2H
2O, and AMT have been displayed in
Figure 3b–d. Compared with
Figure 3a, the tungsten-doped TiO
2 samples were looser, which suppressed the tendency of agglomeration to some extent. Some studies suggest that this could prove advantageous for both adsorption and photocatalysis processes [
45,
46]. Compared with
Figure 3b–d, W3-TiO
2-Et exhibits a loosely structured morphology characterized by enhanced porosity, which potentially contributes to a significantly larger specific surface area compared to other tungsten-doped TiO
2. The microscopic chemical ingredient analysis of the tungsten-doped TiO
2 materials has been shown in
Figure 3e. The patterns show the presence of titanium (Ti), oxygen (O), and tungsten (W) without any other element. The elements of Ti, O, and W are identified, confirming the presence of measured atomic percentage of 26.99, 71.61, and 1.40%, respectively. It can be clearly seen that the atomic percentage of oxygen is 2.7 times that of titanium, which may be due to the presence of oxygen functionalities remaining on the surface of the tungsten-doped TiO
2 [
35].
Figure 3f–h shows the elemental mapping of oxygen (O), titanium (Ti), and tungsten (W) in different colors on W1-TiO
2-Et. The homogeneous distribution of three elements within the W1-TiO
2-Et inferred the successful synthesis of tungsten-doped TiO
2. This homogeneity suggests a well-integrated structure on materials which is crucial for the photocatalysis [
29].
TEM and HRTEM techniques were used to analyse the morphology and microstructure of TiO
2-ET and W1-TiO
2-Et nanoparticles. In the TEM analysis (
Figure 4), it was observed that the morphology of tungsten-doped TiO
2 and pristine TiO
2 nanoparticles is similar. Both materials are characterized by the presence of spherical nanoparticles, which aggregate to form larger clusters. The particle size distribution is between 6~10 nm, which is consistent with the calculation results of Scherrer’s equation (
Table 1). The HRTEM image in
Figure 4c revealed a fringe spacing of approximately 0.352 nm, which is consistent with the crystal growth direction of the anatase TiO
2 (1 0 1) plane, as evidenced by XRD measurements of the sample.
As shown in
Figures S2 and S3, the N
2 adsorption and desorption isotherm curves were recorded to study the specific surface area and corresponding pore size distribution of the tungsten doped samples [
47]. The IUPAC classification of the nitrogen adsorption and desorption isotherms for the examined samples revealed a type V pattern, indicating the mesoporous properties of the prepared material. Meanwhile, the pore size distribution of all the samples was mainly distributed in 0~50 nm, which also indicated the mesoporous properties of the prepared materials. As shown in
Table 1, except for the tungsten source with Na
2WO
4●2H
2O as the precursor, doping with other tungsten precursors increased the specific surface area, pore size, and pore volume of the materials and provided more active sites. The material with Na
2WO
4●2H
2O as the precursor had little effect on the specific surface area, but the pore size of the material with Na
2WO
4●2H
2O as the precursor was significantly increased compared to the pristine TiO
2, which may improve the adsorption and photocatalytic ability of the materials. In
Table 1, it was found that W3-TiO
2-DMF had the largest specific surface area of 350.39 m
2/g, while the pore volume and pore size of the samples with ethanol as a solvent were larger than those of the samples with DMF as a solvent.
The crystalline phase of as-prepared samples was analyzed by X-ray photoelectron spectroscopy (XPS). In the XPS spectrum of Ti 2p, two XPS signals appear with binding energies of 458.8 and 464.5 eV, which are contributed by Ti 2p
1/2 and Ti 2p
3/2 and originate from the Ti
4+ (
Figure S4) [
48,
49,
50].
Figure S4 shows that the binding energy at 529.88 eV corresponds to the crystal lattice oxygen of Ti–O or W–O, suggesting that W–O and Ti–O share O 1s orbitals in the W–O–Ti bond. The binding energy at 531.2 eV originates from hydroxyl groups bonded to Ti or W at the surface, and the binding energy near 532.0 eV originates from adsorbed water (H
2O
ads) or oxygen bonded to carbon (C–O) [
29,
51]. The O 1s binding energies of pristine TiO
2 and tungsten-doped TiO
2 materials were investigated in
Figure S5, and it was found that the tungsten-doped TiO
2 materials were shifted towards a higher binding energy in the O 1s orbital, which may be because the electronegativity of W (2.36) is greater than the electronegativity of Ti (1.54). The doping of TiO
2 with tungsten can alter the electron cloud density of oxygen, leading to a slight shift to the binding energy towards to higher place.
As illustrated in
Figure 5, the visual representation revealed that the samples synthesized using a range of tungsten precursors exhibited a co-doping pattern, characterized by the presence of both W
5+ and W
6+ ions. However, a different ratio of W
5+ to W
6+ was discerned among these samples. The disparity in the W 4f region between W1-TiO
2-Et (WCl
6), W2-TiO
2-Et (Na
2WO
4●2H
2O), and W3-TiO
2-Et (AMT) was employed as an illustrative example. The W 4f peak of W1-TiO
2-Et (WCl
6) was deconvoluted into four peaks, W
6+ 4f
5/2 at 37.55 eV, W
6+ 4f
7/2 at 35.58 eV, W
5+ 4f
5/2 at 36.70 eV, and W
5+ 4f
7/2 at 34.62 eV [
52,
53], while the XPS spectrum of pristine TiO
2 is flat in this region. This is consistent with the expected characteristic signature of pristine TiO
2. The analysis reveals that the signals of W2-TiO
2-Et (Na
2WO
4●2H
2O) and W3-TiO
2-Et (AMT) can be explicitly decomposed into four distinct peaks, each exhibiting comparable positions within the examined region. The W
5+/W
6+ ratios of W1-TiO
2-Et, W2-TiO
2-Et, W3-TiO
2-Et, W1-TiO
2-DMF, W2-TiO
2-DMF, and W3-TiO
2-DMF were 1:4.6, 1:5.2, 1:7.0, 1:7.0, 1:7.8, and 1:8.9, respectively. The study’s outcomes demonstrate the substantial influence of diverse tungsten precursors on the co-doping process involving W
5+ and W
6+. The investigation of tungsten precursors is crucial for understanding the physicochemical properties and photocatalytic performance of tungsten-doped TiO
2. However, the intricacies of the mechanism remain unexplored, necessitating comprehensive investigation for a clearer understanding.
2.2. Photocatalytic Degradation of TC and CIP
The photocatalytic activity of the tungsten-doped TiO
2 was evaluated by photodegradation of TC (50 mg/L) and CIP (50 mg/L). In contrast, the degradation of the TC and CIP solutions showed consistent results for the prepared samples. The activities of all W
5+/W
6+ co-doped samples were increased compared to the undoped TiO
2. As shown in
Figure 6a,b, the highest removal of TC and CIP was achieved by the prepared W1-TiO
2-Et composites, and the removal rate of W1-TiO
2-Et (77.24%, 80%) was more than 2.5 times that of pure TiO
2 (31%, 32%). Meanwhile, in the dark, it was found that different tungsten precursors had a great influence on the adsorption effect, and high adsorption of TC and CIP by W3-TiO
2-DMF was observed.
The kinetic analyses of TC and CIP degradation were based on the fitting of pseudo-first order equations with kinetic constants as shown in
Figure 7. During the photocatalytic degradation of TC and CIP, the highest K values of 0.0039 min
−1 (
Table 2) and 0.0044 min
−1 (
Table 3) were obtained for W1-TiO
2-Et, respectively. The increase in K value after the addition of tungsten precursors indicates that the co-doping of W
6+ and W
5+ accelerated the degradation of TC and CIP in the composites. The co-doping promotes carrier migration and reduces the photogenerated electron-hole complexation efficiency [
31].
In addition, we made comparisons with previous work. As shown in
Table 4 and
Table 5, a variety of photocatalysts for TC and CIP degradation are listed. By comparison, it was found that the larger concentration of TC and CIP was selected as the degradants in this work, and the prepared W
5+/W
6+ co-doped TiO
2 composites showed high removal rates for both degradants.
The effect of different W contents on the photocatalytic properties of tungsten-doped TiO
2 composites was investigated as shown in
Figure 8.
Figure 8a shows the removal curves of CIP, and
Figure 8b shows the apparent first-order rate constant k (min
−1) for CIP.
Figure 8a shows that the highest CIP removal rate (80%) of tungsten-doped TiO
2 composites was achieved when Ti/W = 3:0.1, but when the W content was further increased, the removal rate decreased instead.
Figure 8b also shows that the composites have the maximum K value for Ti/W = 3:0.1, indicating that Ti/W = 3:0.1 is the optimum W loading. The photocatalytic activity of tungsten-doped TiO₂ composites was observed to increase when the optimum tungsten loading was reached, beyond which certain surface reaction sites of photocatalytic activity may be hindered, thus limiting the reaction rate [
29].
In summary, the photocatalyst with ammonium metatungstate (AMT) as the tungsten precursor and DMF as the solvent exhibited the strongest adsorption capacity for ciprofloxacin (CIP) and tetracycline (TC) in the dark (52% and 48% in 2 h). However, the photocatalyst with WCl6 as the tungsten precursor and ethanol as the solvent (W1-TiO2-Et) showed the highest photodegradation performance for ciprofloxacin (CIP) and tetracycline (TC). The photodegradation performance to tetracycline (TC) on W1-TiO2-Et was 2.3, 2.8, and 7.8 times that of AMT, Na2WO4●2H2O as the tungsten precursors and pristine TiO2, respectively. The optimum tungsten loading was also investigated in this work, and it was found that Ti/W = 3:0.1 is the optimum tungsten loading.
2.3. Mechanism
It is worth highlighting some facts about tungsten-doped TiO2 which exhibited higher photocatalytic efficiency for the degraded TC and CIP, though the mechanism is far from understood.
As illustrated in
Figure 7, the photocatalytic efficiencies of the different tungsten-doped TiO
2 samples are significantly different. XPS analysis (
Figure 5) revealed that these materials exhibited different ratios of W
5+ and W
6+. The photocatalytic efficiency of the prepared samples may be related to the ratio of W
5+ and W
6+, and co-doping promotes carrier migration and reduces the photogenerated electron-hole complex efficiency. To confirm this hypothesis, electrochemical impedance spectroscopy (EIS) and photoluminescence (PL) analyses have been performed, and the results are shown in
Figure 9a,b.
The findings indicated that tungsten-doped TiO
2 exhibited a diminished photogenerated carrier radius and an extended photogenerated carrier lifetime in comparison to pristine TiO
2. This suggests that the separation rate and migration of photogenerated carriers in the W
5+/W
6+ co-doped composites are more rapid than in pristine TiO
2. W1-TiO
2-Et has the smallest photogenerated carrier radius and the longest photogenerated carrier lifetime, which confirms that the W1-TiO
2-Et has the highest photocatalytic activity. Meanwhile, it can be seen from
Figure 5 that the W
5+/W
6+ ratio of W1-TiO
2-Et is the largest; the higher percentage of W
5+ indicates the higher oxygen vacancy concentration [
53], which accelerates the migration rate of photogenerated carriers. The higher concentration of oxygen vacancies also makes it easier to adsorb water to form surface hydroxyls, which promotes the photocatalytic efficiency [
53]. It is believed that the phenomenon should be influenced by the molecular configuration of the tungsten precursors. Tungsten in AMT and Na
2WO
4●2H
2O is surrounded by oxygen and hydroxyl groups/H
2O, respectively. In contrast, tungsten in tungsten (VI) chloride (WCl
6) is also in the +6 oxidation state but is coordinated by chlorine atoms. The difference in coordination environment between these tungsten compounds results in the diverse ways in which tungsten can be bonded in the synthesis of tungsten-doped TiO
2, influenced by the steric hindrance effect and surface hydroxyl group. The combined effects underscore the efficiency of photocatalytic degradation in the removal of TC and CIP.
The solvent also impacted the photocatalytic efficiency of the synthesized samples obviously.
Figure 7 shows that tungsten-doped TiO
2 synthesized with ethanol exhibited enhanced photocatalytic efficiency compared to those synthesized in DMF. This phenomenon may be attributed to the varying conformations exhibited by ethanol and DMF molecules. As mentioned earlier, compared to DMF, using ethanol as solvent can increase the crystallinity, pore size, and pore volume of the tungsten-doped TiO
2 materials. This facilitates charge transfer from the center to the surface and increases the active sites, thus improving the photocatalytic activity. In addition, the gel formation process was faster in the DMF solvent than in the ethanol solvent during the preparation process, which may be due to the larger polarity and dielectric constant of DMF compared to ethanol. The reaction was slower when ethanol was used as a solvent and the nanoparticles nucleated uniformly in ethanol solvent [
60]. Ethanol can be used as a very dispersive solvent by the reaction and ethanol as a solvent plays an important role in controlling the crystal growth [
60].
Figure 10 shows the UV-vis absorption spectra of tungsten-doped TiO
2 with different tungsten precursors and solvents. In general, the absorption of tungsten-doped TiO
2 was stronger than that of undoped TiO
2 in visible light region. The red shift was observed with the incorporation of tungsten into TiO
2 materials. The band gaps of the prepared tungsten-doped TiO
2 are estimated by the Tauc plot as shown in
Figure S7. Pristine TiO
2 exhibited the E
g values about 3.3 eV, which is consistent with the value reported in literature [
61]. It is obvious that doping TiO
2 with tungsten slightly modifies the TiO
2 absorption edge in the visible region. This observed phenomenon has been elucidated by the quantum confinement effect [
34,
62]. It is true that tungsten doping in TiO
2 has been observed to reduce the band gap of photocatalysts, which is attributed to create new energy levels within the band gap [
49]. However, smaller particle size often leads to higher E
g values [
13]. As the size of the particles decreases, the degree of quantization of energy levels is stronger, leading to an increased band gap [
45]. That is, the narrow band gap and red-shifted adsorption edges of the samples analysed by UV-visible diffuse reflectance spectroscopy may indicate a better photocatalytic activity of the tungsten-doped TiO
2 samples [
29].
For the purpose of analysing the adsorption mechanism of the tungsten-doped TiO
2 samples in the dark, the zeta potential of the tungsten-doped TiO
2 samples have been measured (
Figure S8), focusing on their surface charge characteristics.
The pka1, pka2, and pka3 of TC are 3.3, 7.7, and 9.7, respectively [
63]. Positive charges dominate when the pH is below 3.3, both positive and negative charges are present when the pH is below 7.7 but above 3.3, and negative charges dominate when the pH is above 7.7 [
60]. CIP has two pKa of 5.9 and 8.9 at pH values between 5.9 and 8.9 [
64], at pH values below 5.9 the cation of CIP is dominant, and at pH values above 8.9, the anion of CIP is dominant [
65,
66]. This work is based on 50 ppm TC and 50 ppm CIP solutions at pH 3.2 and 5.3, so positive charges dominate in both 50 ppm TC and 50 ppm CIP solutions.
As shown in
Table 6, the samples with AMT as the tungsten precursors had less positive surface charge in TC and CIP solutions (pH = 3.2, 5.3). According to the principle of anisotropic attraction and anisotropic repulsion, the sample with AMT as the tungsten precursor has a better adsorption effect on ciprofloxacin solution in the dark. From
Table 1, it can be seen that the specific surface area of W3-TiO
2-DMF is larger than the other samples; therefore, combined with the BET potential and zeta potential, it can be concluded that W3-TiO
2-DMF has a stronger adsorption effect on the 50 ppm TC and 50 ppm CIP solutions, which is also consistent with the results in
Figure 6.
Free radical trapping tests were carried out on W1-TiO
2-Et as shown in
Figure 11a,b to determine which are the active species in the photocatalytic process of W1-TiO
2-Et [
67]. The photodegradation efficiency decreased when IPA, EDTA-2Na, and BQ were added to the TC and CIP solutions. IPA (71.3%, 76.24%), EDTA-2Na (56.6%, 63.49), and BQ (43%, 51.86). This result indicates that ·O
2− and h
+ are the main active species of W1-TiO
2-Et photocatalysts during the photodegradation of CIP and TC [
68].
To assess the reusability of the photocatalysts, the W1-TiO
2-Et photocatalysts were subjected to four cycles of reusability tests. The results are shown in
Figure 12: the photocatalyst exhibits excellent photocatalytic efficacy for TC and CIP degradation in all reusability tests. This proves that the sample is an excellent photocatalyst for reusability.