Oxygen Vacancy Mediated Band-Gap Engineering via B-Doping for Enhancing Z-Scheme A-TiO2/R-TiO2 Heterojunction Photocatalytic Performance

Fabrication of Z-scheme heterojunction photocatalysts is an ideal strategy for solving environmental problems by providing inexhaustible solar energy. A direct Z-scheme anatase TiO2/rutile TiO2 heterojunction photocatalyst was prepared using a facile B-doping strategy. The band structure and oxygen-vacancy content can be successfully tailored by controlling the amount of B-dopant. The photocatalytic performance was enhanced via the Z-scheme transfer path formed between the B doped anatase-TiO2 and rutile-TiO2, optimized band structure with markedly positively shifted band potentials, and the synergistically-mediated oxygen vacancy contents. Moreover, the optimization study indicated that 10% B-doping with the R-TiO2 to A-TiO2 weight ratio of 0.04 could achieve the highest photocatalytic performance. This work may provide an effective approach to synthesize nonmetal-doped semiconductor photocatalysts with tunable-energy structures and promote the efficiency of charge separation.


Introduction
The technology of semiconductor advanced oxidation processes (AOPs) has shown great potential in the degradation of organic pollutants in water since it was developed in 1976 [1]. As a typical n-type semiconductor material, titanium dioxide (TiO 2 ), has been widely used in the field of photocatalytic degradation due to its unique physical and chemical properties [2][3][4]. However, TiO 2 only responds to the ultraviolet part (4%) of sunlight due to its wide band gap (3.2 eV), and shows low carrier-separation efficiency (≤10%) and weak redox capability, severely restricting its further application [5][6][7].
Among various strategies to solve these issues, the design and fabrication of heterojunctions are promising due to their merits in separating charge carriers and in coupling the advantages of each component [8,9]. Because they are always constructed with two components using a staggered band structure, the common Type-II heterojunctions are usually formed. In this mode, the photogenerated electrons in the component having a higher CB position migrate to the component with a lower CB position, while the holes migrate in the reverse direction. Thus, the recombination of charge carriers is suppressed, and the spatial separation of charge carriers is promoted [10][11][12]. Compared with the traditional type II heterojunction, the typical direct Z-scheme heterojunction has the same band-structure configuration but a distinctly different charge-carrier transfer mode. In this mode, the electrons in the CB of the component with the higher CB position (stronger reduction abilities) and the holes in the VB of the component with the lower VB position (stronger oxidation abilities) are preserved, while the electrons and holes with inferior redox powder recombine [13,14]. Therefore, the Z-scheme heterojunction can not only effectively separate charge carriers, but also retain the highest redox potential of each individual component, which helps to obtain more active free radicals (•O 2 − and •OH) to participate in subsequent surface redox reactions [15,16].
To date, various TiO 2 based Z-scheme heterojunctions have been reported to have improved photoelectrochemical properties, including NiO/TiO 2 [17], MoS 2 /TiO 2 [18], Co 3 O 4 /TiO 2 [19], WO 3 /TiO 2 [20], Cu 2 O/TiO 2 [21], etc. However, preparation of the direct Z-scheme heterojunction remains a great challenge, for which the identification of suitable semiconductors and the construction of high-quality heterointerfaces still need to be addressed [22]. For most of the work that simply combines two or more materials, the interfacial resistance between the two phases will restrict charge transfer and affect their stability [23]. Recently, the fabrication of efficient phase junctions has been shown to be an effective way to promote charge separation, and to lead to enhanced photocatalytic activity [24][25][26]. Furthermore, for a TiO 2 based phase junction, anatase TiO 2 (A-TiO 2 ) and rutile TiO 2 (R-TiO 2 ) can be generated in-situ during the formation of TiO 2 in a sol-gel process by B-doping [27], implying the potential to form an A-TiO 2 /R-TiO 2 contacted intact heterointerface.
In addition, recombination at the heterojunction that resulted from the mismatched band alignment is highly detrimental to the efficiency of the photocatalytic process. Also, to form a high-efficiency Z-scheme photocatalyst, in addition to the staggered band structure of the two contacted semiconductors, the oxidation component should possess a lower VB position to display strong oxidation ability, while the other reduction component should have a higher CB position to exhibit strong reduction ability [13,28]. Therefore, engineering the interface and adjusting the band-alignment of the two components are critical for enhanced photocatalytic performance. Furthermore, the introduction of oxygen vacancies (Vo) has been reported to work as an efficient polarization strategy to improve the unsatisfied charge carriers separation during the photocatalytic process, which is critical to restrict the photocatalytic performance of semiconductors [29][30][31]. However, because they can also work as recombination centers, it is a double-edged sword to introduce Vo into the photocatalyst for charge carriers separation [32]. Thus, reasonably designed and constructed Vo concentrations are essential to enhance photocatalytic performance.
Today, it is well known that doping can produce oxygen vacancies in nanostructured TiO 2 , and the content of oxygen-vacancy defects can also be adjusted by the amount of B-doping [33][34][35]. However, the challenge lies in the choice of the synthesis route to control or tune the formation of oxygen-vacancy defects. In addition, boron doping has also been proved to be a feasible approach to modify the phase content of TiO 2 to form an A-TiO 2 /R-TiO 2 heterojunction [36]. Niu et al. demonstrated that the amount of B-doping could be optimized to prepare B-doped TiO 2 as an efficient visible-light-driven photocatalyst [27]. Wang et al. synthesized the B-doped TiO 2 with a tunable anatase/rutile ratio and proposed that the formed type-II phase junction facilitated the separation of charge carriers [37]. However, there are no studies on Z-scheme R-TiO 2 /A-TiO 2 photocatalysts that operate via synergistic regulation of the anatase/rutile ratio and the formation of oxygen vacancy defects. Moreover, the underlying regulation mechanism, including the essential relation between Vo tuning, the A-TiO 2 /R-TiO 2 junction, and photocatalytic activity, still needs to be explored [38].
Hence, in this study, we developed an effective approach to in situ formation of a Z-scheme B-doped A-TiO 2 /R-TiO 2 phase junction with adjustable band structures and oxygen vacancies. The photodegradation activity was enhanced by the Z-scheme charge transfer pathway formed among the phase junction, the markedly positively shifted band potentials, and the synergistically optimized oxygen vacancy defect contents. Moreover, the formation mechanism of the Z-scheme transfer path and the regulating mechanisms of the band structure and oxygen vacancy are studied in detail.

Preparation of the Catalysts
All samples were prepared by calcinating the prepared precursors. The precursor was obtained by fully mixing Ti, boron, and carbon-precursor solution, dried at 80 • C for 48 h, and then the precursor was heat-treated at 800 • C for 2 h at 5 • C/min in a muffle oven. In our work, the low-content carbon precursor was applied to form an oxygen-deficient environment by reacting with oxygen in the air during the heat-treatment, to incorporate oxygen vacancies into the prepared titanium oxide-based catalysts.

Evaluation of Photodegradation Activity
The RhB photodegradation activity of the catalysts under simulated sunlight irradiation is described as follows. Firstly, 30 mL of RhB solution (10 mg/L) was made, followed by the addition of 30 mg of the catalyst. After stirring for 15 min in the dark, the prepared solution was illuminated for 90 min using a 300 W Xe lamp with a wavelength in the range of 190-1100 nm as the light source. At regular intervals of 30 min, 5 mL of the solution was collected and tested by a UV-Vis spectrophotometer (UV2800-A, UNICO, Shanghai, China) with wavelength in the range of 200-800 nm, to analyze the residue RhB concentration according to the corresponding peak for the RhB centered at 554 nm under illumination. Figure 1a shows the phase composition of the prepared samples. For all samples, only peaks ascribed to rutile TiO 2 (R-TiO 2 ) (PDF 76-1938) and anatase TiO 2 (A-TiO 2 ) (PDF 83-2243) can be detected without the appearance of other phases [39,40]. Furthermore, from the magnified XRD patterns (Figure 1b), the (101) diffraction peak of A-TiO 2 and (110) diffraction peak of R-TiO 2 show an obvious shift to the lower 2θ value with increasing introduced boron content. Because the B 3+ cations (27 pm) are smaller than the Ti 4+ cations (60.5 pm), the resulting lattice shrinkage would cause the gradual shift of the diffraction peaks [41]. Therefore, these results indicate the successful introduction of boron into TiO 2 and the substitution of B 3+ in the crystal lattices of TiO 2 . In addition, the weight ratios of the R-TiO 2 to A-TiO 2 are calculated to be 0.1, 0.09, 0.05, 0.04, 0.06, and 0.07, respectively.  Figure 1a shows the phase composition of the prepared samples. For all samples, only peaks ascribed to rutile TiO2 (R-TiO2) (PDF 76-1938) and anatase TiO2 (A-TiO2) (PDF 83-2243) can be detected without the appearance of other phases [39,40]. Furthermore, from the magnified XRD patterns (Figure 1b), the (101) diffraction peak of A-TiO2 and (110) diffraction peak of R-TiO2 show an obvious shift to the lower 2θ value with increasing introduced boron content. Because the B 3+ cations (27 pm) are smaller than the Ti 4+ cations (60.5 pm), the resulting lattice shrinkage would cause the gradual shift of the diffraction peaks [41]. Therefore, these results indicate the successful introduction of boron into TiO2 and the substitution of B 3+ in the crystal lattices of TiO2. In addition, the weight ratios of the R-TiO2 to A-TiO2 are calculated to be 0.1, 0.09, 0.05, 0.04, 0.06, and 0.07, respectively.  Figure 2 shows the morphologies of the as-prepared catalysts, and the corresponding particle size statistics are shown in Figure S1. All samples show a uniform and fine particle distribution with an average particle size of about 75 nm. For the pure-TiO2 sample (Figure 2a), the particles are in spherical morphologies. Whereas, for the 14% B-TiO2 sample (Figure 2f), particles in spherical, ellipsoidal, and rhombic morphologies appeared. Moreover, microstructures of the 10% B-TiO2 sample investigated using TEM and HRTEM were shown in Figure 3. The 10% B-TiO2 sample shows a similar morphology to that of the 14% B-TiO2 sample, including spherical, ellipsoidal and rhombic morphologies. In the HRTEM images (Figure 3b), lattice fringes with the spacing of 0.249, 0.351, and 0.170 nm correspond to the (101) plane of R-TiO2, the (101) and the (105) plane of A-TiO2 [42,43], respectively, from which the fine-contacted interface between these two phases can be observed. Furthermore, the EDS result in Figure 3c-g reveal the uniform distribution of elements B, Ti, and O, confirming the successful introduction of element B and the formation of the A-TiO2/R-TiO2 heterostructure.  Figure 2 shows the morphologies of the as-prepared catalysts, and the corresponding particle size statistics are shown in Figure S1. All samples show a uniform and fine particle distribution with an average particle size of about 75 nm. For the pure-TiO 2 sample (Figure 2a), the particles are in spherical morphologies. Whereas, for the 14% B-TiO 2 sample (Figure 2f), particles in spherical, ellipsoidal, and rhombic morphologies appeared. Moreover, microstructures of the 10% B-TiO 2 sample investigated using TEM and HRTEM were shown in Figure 3. The 10% B-TiO 2 sample shows a similar morphology to that of the 14% B-TiO 2 sample, including spherical, ellipsoidal and rhombic morphologies. In the HRTEM images (Figure 3b), lattice fringes with the spacing of 0.249, 0.351, and 0.170 nm correspond to the (101) plane of R-TiO 2 , the (101) and the (105) plane of A-TiO 2 [42,43], respectively, from which the fine-contacted interface between these two phases can be observed. Furthermore, the EDS result in Figure 3c  XPS analysis was applied to find out the concentrations of defect states (oxygen vacancy) of TiO2 after the introduction of B (Figure 4). In the survey spectra (Figure 4a), besides the C calibration element, B, O, and Ti elements are detected for all the samples. Figure 4b shows the XPS spectra of B1s, in which a broad peak, centered at around 191 eV ascribed to a substitutional B that occupies O sites, can be observed for B-doped samples [36]. This result proves the successful incorporation of B into the TiO2 lattice. In Figure  4c,d, four peaks centered at 529.5, 531.5, 532.1, and 532.9 eV, corresponding to lattice oxygen (LO), Ti-O in Ti2O3, a surface adsorbed -OH group, and adsorbed H2O (AO) [44,45] can be observed in O1s spectra, respectively. In particular, the presence of peaks centered at 531.5 eV is believed to be caused by partially reduced Ti atoms in TiO2, due to the connection with neighboring oxygen vacancies (Vo), which suggests the existence of Vo [46,47]. Moreover, the increase in its peak intensity indicates that the Vo content in the samples increases with the increasing content of the B-dopant [29]. Moreover, the content of Vo can be obtained by using the CasaXPS analysis software based on the area-calculation of the fitted peaks, and the result is shown in Figure 4c,d. As can be seen, the ratio of the area under the VO XPS peak (visible at around 531.5 eV) in 10% B-TiO2 is the highest, signifying the presence of the highest amount of oxygen vacancies. It can be deduced that the boron in the crystal lattices may reach the highest amount of incorporation, and this sample may XPS analysis was applied to find out the concentrations of defect states (oxygen vacancy) of TiO 2 after the introduction of B (Figure 4). In the survey spectra (Figure 4a), besides the C calibration element, B, O, and Ti elements are detected for all the samples. Figure 4b shows the XPS spectra of B 1s , in which a broad peak, centered at around 191 eV ascribed to a substitutional B that occupies O sites, can be observed for B-doped samples [36]. This result proves the successful incorporation of B into the TiO 2 lattice. In Figure 4c,d, four peaks centered at 529.5, 531.5, 532.1, and 532.9 eV, corresponding to lattice oxygen (L O ), Ti-O in Ti 2 O 3 , a surface adsorbed -OH group, and adsorbed H 2 O (A O ) [44,45] can be observed in O 1s spectra, respectively. In particular, the presence of peaks centered at 531.5 eV is believed to be caused by partially reduced Ti atoms in TiO 2 , due to the connection with neighboring oxygen vacancies (Vo), which suggests the existence of Vo [46,47]. Moreover, the increase in its peak intensity indicates that the Vo content in the samples increases with the increasing content of the B-dopant [29]. Moreover, the content of Vo can be obtained by using the CasaXPS analysis software based on the area-calculation of the fitted peaks, and the result is shown in Figure 4c,d. As can be seen, the ratio of the area under the V O XPS peak (visible at around 531.5 eV) in 10% B-TiO 2 is the highest, signifying the presence of the highest amount of oxygen vacancies. It can be deduced that the boron in the crystal lattices may reach the highest amount of incorporation, and this sample may show distinct photodegradation activity due to the adjusted band structure. Furthermore, the Ti 2p spectra are fitted into four peaks, i.e., 458.3 eV (Ti 4+ 2p 3/2), 457.3 eV (Ti 3+ 2p 3/2), 464.0 eV (Ti 4+ 2p 1/2) and 462.3 eV (Ti 3+ 2p 1/2), respectively [48]. Moreover, it can be seen that Ti 2p peaks (Figure 4e,f) slightly move towards the higher binding energy region from pure-TiO 2 to 10% B-TiO 2 , then shift towards the lower binding energy region from 10% B-TiO 2 to 14% B-TiO 2 , indicating the variation of Vo contents with a high electron-attracting effect [41,49]. The presence of Ti 3+ peaks and the shift further confirm the presence of Vo in all samples and the highest amount of Vo in 10% B-TiO 2 .  [48]. Moreover, it can be seen that Ti2p peaks (Figure 4e,f) slightly move towards the higher binding energy region from pure-TiO2 to 10% B-TiO2, then shift towards the lower binding energy region from 10% B-TiO2 to 14% B-TiO2, indicating the variation of Vo contents with a high electron-attracting effect [41,49]. The presence of Ti 3+ peaks and the shift further confirm the presence of Vo in all samples and the highest amount of Vo in 10% B-TiO2.

Photodegradation Activities of the Catalysts
The RhB photodegradation of the catalysts under simulated sunlight illumination is shown in Figure 5, with commercial P25 and RhB solutions lacking the presence of catalysts as references. No obvious removal of RhB can be observed for the solution without the presence of catalysts. Ten percent B-TiO2 shows the highest RhB degradation rate of

Photodegradation Activities of the Catalysts
The RhB photodegradation of the catalysts under simulated sunlight illumination is shown in Figure 5, with commercial P25 and RhB solutions lacking the presence of catalysts as references. No obvious removal of RhB can be observed for the solution without the presence of catalysts. Ten percent B-TiO 2 shows the highest RhB degradation rate of 94.8% after light illumination for 90 min. The increase in boron doping from 4% B-TiO 2 to 10% B-TiO 2 prompted photodegradation activity. However, for the 12% B-TiO 2 and 14% B-TiO 2 samples, a significantly decreased degradation efficiency can be observed. From the analysis of degradation kinetics in Figure 5b, the degradation rate with 10% B-TiO 2 was 0.033 min −1 , 2.13 times higher than commercial P25. As is known, the chemical stability of photocatalysts plays a crucial role in the application [50]. The cycling test result of the 10% B-TiO 2 sample is shown in Figure 5c. No obvious deduction (less than 2%) can be observed after five cycles, suggesting that the photocatalyst is relatively stable. From the XRD analysis in Figure 5d, by comparing the position and shape of diffraction peaks, no detectable differences can be seen between the as-prepared and cycled 10% B-TiO 2 , indicating a well-preserved crystalline structure of the catalyst after multiple photocatalytic cycles. Moreover, a few studies on RhB photodegradation performance of TiO 2 -based photocatalysts are summarized in Table 1. The table shows that the degradation efficiency of RhB over the catalyst prepared in this work was markedly improved, indicating that the prepared B-doped TiO 2 is a promising photocatalyst for RhB degradation.
Nanomaterials 2023, 13, 794 9 of 18 tensity was apparently higher than that of pure-TiO2, indicating more abundant •OH radicals during light reactions. Based on the band structure configuration of the samples, •OH radicals may not be produced from a direct reaction between photogenerated charge carriers and adsorbed H2O. Photogenerated electrons can be captured by adsorbed O2 and H + in the solution to form H2O2, which would further decompose into the •OH radicals [62,63]. Therefore, these results further confirm the decrease in the energy-band gap and the obvious positive shift of the VB position of 10% B-TiO2. Similarly, quartet signals with peak intensities of 1:1:1:1 attributed to DMPO-•O2 − adducts were detected for both samples [64] (Figure 6c). The slightly weaker signal intensity further confirms the positive shift of band positions for 10% B-TiO2.    , and e − , respectively. Relative impact of the active species was calculated from rate constants and is presented in Figure S2. As can be seen, for both samples, except for e − , the other three active species are found to be critical during the reactions. To further clarify the free active radicals present under simulated sunlight illumination, the characterization of EPR was analyzed and shown in Figure 6. For both samples, •O 2 − and •OH radical are detected after light irradiation. As shown in Figure 6a, characteristic EPR signals at g values of 1.998 (pure-TiO 2 ) and 1.997 (10% B-TiO 2 ), respectively, are detected. Additionally, the higher signal intensity for 10% B-TiO 2 can be obviously observed. According to the literature [60], this signal results from unpaired electrons appearing at oxygen vacancies. Therefore, the above results confirm the presence of Vo in both samples and the higher Vo content for 10% B-TiO 2 , which is consistent with the XPS analysis. In Figure 6b, four EPR signals with peak intensities of 1:2:2:1, which can be assigned to DMPO-•OH adducts, were observed for both samples [61]. Additionally, for 10% B-TiO 2 , the signal intensity was apparently higher than that of pure-TiO 2 , indicating more abundant •OH radicals during light reactions. Based on the band structure configuration of the samples, •OH radicals may not be produced from a direct reaction between photogenerated charge carriers and adsorbed H 2 O. Photogenerated electrons can be captured by adsorbed O 2 and H + in the solution to form H 2 O 2 , which would further decompose into the •OH radicals [62,63].

Photodegradation Mechanism of the Catalysts
The UV-Vis DRS spectra of the catalysts are shown in Figure 7. The light absorption edges show no obvious shift in Figure 7a, with different amounts of B-doping. Furthermore, based on the following Equation (1), the band gaps of the samples were obtained: where n = 1, because TiO2 belongs to the direct transition semiconductor [65]. The results are shown in the inset image in Figure 7a. There are two band-gap values from all asprepared samples based on the absorption edge. The inherent band gap (Eg) is 2.33 eV, and the impurity band gap (ET) is from 1.97 to 2.28 eV [66]. Furthermore, the CB potentials of the samples were calculated, according to the fact that the CB edge of the n-type semiconductor is 0.10 eV lower than the value of the flat band potential (Efb) [66,67]. The Efb value, as shown in Figure 7c, can be obtained through the following Equation (2) [68] based on the Mott-Schottky analysis ( Figure S3).
Efb (vs. NHE) = Efb (vs. Ag/AgCl) + EAgCl + 0.059 × pH (2) Thus, according to ECB = EVB − Eg, the energy band structures of the catalysts were obtained ( Figure 7d). As can be seen, the VBM of the sample can be adjusted from 1.38 to 1.68 eV, and the CMB value can be adjusted from −0.95 to 0.65 eV, confirming that the electronic band structure of the prepared catalyst can be effectively dominated by adjusting the content of the B dopant.
From the PL spectra (Figure 7b), the samples show similar emission peaks at around 400 nm. The 10% B-TiO2 sample shows the lowest intensity, indicating an enhanced charge separation for 10% B-TiO2 with optimized boron doping [69]. Furthermore, the TP response and EIS results are shown in Figure 7e,f. The same trend (firstly increasing and then decreasing) of the TP response and the radicals in EIS can be observed with increasing B-dopant. Especially, the 10% B-TiO2 sample shows the highest photocurrent response and the minimum radicals in EIS, which can result from the highest content of oxygen vacancy defects [70]. The above results are consistent with the XPS and photocatalysis degradation studies. In general, it can be concluded that B-doping can adjust the content of oxygen vacancies, and then mediate the band structure to enhance the photocatalytic performance.

Photodegradation Mechanism of the Catalysts
The UV-Vis DRS spectra of the catalysts are shown in Figure 7. The light absorption edges show no obvious shift in Figure 7a, with different amounts of B-doping. Furthermore, based on the following Equation (1), the band gaps of the samples were obtained: where n = 1, because TiO 2 belongs to the direct transition semiconductor [65]. The results are shown in the inset image in Figure 7a. There are two band-gap values from all asprepared samples based on the absorption edge. The inherent band gap (Eg) is 2.33 eV, and the impurity band gap (E T ) is from 1.97 to 2.28 eV [66]. Furthermore, the CB potentials of the samples were calculated, according to the fact that the CB edge of the n-type semiconductor is 0.10 eV lower than the value of the flat band potential (E fb ) [66,67]. The E fb value, as shown in Figure 7c, can be obtained through the following Equation (2) [68] based on the Mott-Schottky analysis ( Figure S3). E fb (vs. NHE) = E fb (vs. Ag/AgCl) + E AgCl + 0.059 × pH Thus, according to E CB = E VB − Eg, the energy band structures of the catalysts were obtained ( Figure 7d). As can be seen, the VBM of the sample can be adjusted from 1.38 to 1.68 eV, and the CMB value can be adjusted from −0.95 to 0.65 eV, confirming that the electronic band structure of the prepared catalyst can be effectively dominated by adjusting the content of the B dopant.
From the PL spectra (Figure 7b), the samples show similar emission peaks at around 400 nm. The 10% B-TiO 2 sample shows the lowest intensity, indicating an enhanced charge separation for 10% B-TiO 2 with optimized boron doping [69]. Furthermore, the TP response and EIS results are shown in Figure 7e,f. The same trend (firstly increasing and then decreasing) of the TP response and the radicals in EIS can be observed with increasing B-dopant. Especially, the 10% B-TiO 2 sample shows the highest photocurrent response and the minimum radicals in EIS, which can result from the highest content of oxygen vacancy defects [70]. The above results are consistent with the XPS and photocatalysis degradation studies. In general, it can be concluded that B-doping can adjust the content of oxygen vacancies, and then mediate the band structure to enhance the photocatalytic performance. Considering the results of XPS and photocatalysis degradation, the reason for the samples with similar band gaps but obviously different photocatalysis performances can Considering the results of XPS and photocatalysis degradation, the reason for the samples with similar band gaps but obviously different photocatalysis performances can be explained as follows. Firstly, all the prepared catalysts including pure-TiO 2 are identified to be A-TiO 2 /R-TiO 2 nano-structured heterojunction with the introduction of oxygen vacancy defects. The oxygen vacancies result from the B-dopant or the oxygen-deficit environment during heat-treatment, accompanied by the introduction of defect energy level into the band structure, thus leading to a decrease of the photo-excitation energy and a red-shift of the absorption spectrum edge. Furthermore, defect-energy level served as a springboard to enhance the charge density and the charge separation efficiency. Moreover, the surfaceoxygen vacancies worked as reactive sites to obtain quick surface reactions during the photocatalytic process. Secondly, the electronic band structure of the catalyst was mediated, and the optimized band structure was markedly positively shifted, thus more abundant radicals can be produced and participate in the following reactions. Furthermore, according to the literature [71], for B-doped R-TiO 2 , the CB has a decline of approximately 0.24 eV, and the value is approximately 0.48 eV for B doped A-TiO 2 . Moreover, for both B-doped phases, the VB has little shift compared with the pure phases. Combined with the energy-band diagram of pure A-TiO 2 (CB = +0.012 eV, VB = +2.712 eV) and pure R-TiO 2 (CB = −0.380 eV, VB = +2.070 eV) [72,73], if the charge transfer pathway follows the type II mode, the reduced reduction potential, which was lower than that of •O 2 − /O 2 , would be incapable of causing •O 2 − production, thus the direct Z-scheme A-TiO 2 /R-TiO 2 heterojunction mechanism is confirmed.
Photodeposition of Ag nanoparticles is known as a convenient method to track the electron-transfer direction. Generally, the Ag nanoparticles are selectively reduced on the site where the photogenerated electron flows to, thus can be used to determine whether the formed heterojunction is Z-scheme or not [74]. Therefore, in order to disclose whether the B doped R-TiO 2 /A-TiO 2 was a Z-type heterojunction, the charge-transfer tracking experiments, i.e., the e − trapping experiment using AgNO 3 as the scavenger, were carried out by loading Ag onto the TiO 2 using the photodeposition method. Figure 8 presents the TEM, HRTEM, and EDS mapping results of the charge-transfer tracking experiment. From the EDS results, uniformly distributed Ag, B, O, and Ti elements can be observed. Moreover, based on the TEM and HRTEM results, the Ag nanoparticles were uniformly distributed, while isolated on R-TiO 2 and apart from A-TiO 2 . These results demonstrate that the photogenerated electrons were left on the R-TiO 2 , and the photogenerated electrons from A-TiO 2 could be recombined with the holes from R-TiO 2 , thus forming the Z-type heterojunction.
According to the literatures [71,75], for A-TiO 2 and R-TiO 2 with a partially-reduced surface that leads to the remaining oxygen vacancies, the work function of R-TiO 2 (approximately 4.3 eV) is lower than that of A-TiO 2 (approximately 4.7 eV). Therefore, the work function of B-doped R-TiO 2 (R-phase) is deduced to be smaller than that of B-doped A-TiO 2 (A-phase), both of which are doped to form oxygen vacancies. Therefore, considering the higher CB and VB positions of R-phase and its smaller work function, when the R-phase and the A-phase were in intimate contact, the electrons transferred from the R-phase to the A-phase to get the Fermi levels equilibrated and the internal electric field was formed with the direction from the R-phase to the A-phase. At the same time, the downward and upward band bending was generated at the R-phase/A-phase interface (Figure 9b). Under illumination, electron-hole pairs are produced in both phases. The above formed internal electric field would promote the electrons in the CB of A-phase to recombine with the holes in the VB of R-phase (Figure 9c). Meanwhile, the other charge transfer pathways would be effectively suppressed by the band bending and the built-in electric field. Consequently, the photogenerated electrons in the CB of the R-phase and the holes in the VB of the A-phase are spatially separated and participate in the following redox reactions to produce radicals and degrade organic pollutants.

Conclusions
In this work, we prepared a direct Z-scheme heterojunction photocatalyst of B-doped A-TiO2/R-TiO2 with adjustable band structure via B-doping, for which the photocatalytic performances were enhanced by the defect energy level caused by oxygen vacancies, positively shifted-band potential, and synergistically, the Z-type charge transfer pathway. The band structure and surface oxygen-vacancies content can be tailored by controlling the amount of B-dopant. Furthermore, the Z-scheme transfer path between A-TiO2 and R-TiO2 has promoted the charge carrier separation and improved the redox ability of the sample by retaining the higher redox potential. Moreover, 10% of the B-doping achieved the highest photodegradation performance in this study. The method is applicable to other non-metal ions for enhancing photocatalytic performance. Author Contributions: Conceptualization, X.L. and C.L.; methodology, X.L. and C.L.; software, W.W., L.C. and C.X.; validation, X.L. and C.L.; formal analysis, C.X. and C.L.; investigation, W.W., L.C. and C.X.; resources, C.L. and Y.W.; data curation, X.L. and C.L.; writing-original draft preparation, W.W. and C.X.; writing-review and editing, C.X., X.L. and C.L.; visualization, W.W.; supervision, C.L. and Y.W.; project administration, C.L.; funding acquisition, C.L. and Y.W. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.

Conclusions
In this work, we prepared a direct Z-scheme heterojunction photocatalyst of B-doped A-TiO 2 /R-TiO 2 with adjustable band structure via B-doping, for which the photocatalytic performances were enhanced by the defect energy level caused by oxygen vacancies, positively shifted-band potential, and synergistically, the Z-type charge transfer pathway. The band structure and surface oxygen-vacancies content can be tailored by controlling the amount of B-dopant. Furthermore, the Z-scheme transfer path between A-TiO 2 and R-TiO 2 has promoted the charge carrier separation and improved the redox ability of the sample by retaining the higher redox potential. Moreover, 10% of the B-doping achieved the highest photodegradation performance in this study. The method is applicable to other non-metal ions for enhancing photocatalytic performance. Author Contributions: Conceptualization, X.L. and C.L.; methodology, X.L. and C.L.; software, W.W., L.C. and C.X.; validation, X.L. and C.L.; formal analysis, C.X. and C.L.; investigation, W.W., L.C. and C.X.; resources, C.L. and Y.W.; data curation, X.L. and C.L.; writing-original draft preparation, W.W. and C.X.; writing-review and editing, C.X., X.L. and C.L.; visualization, W.W.; supervision, C.L. and Y.W.; project administration, C.L.; funding acquisition, C.L. and Y.W. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.