Selective growth of Ti3+/TiO2/CNT and Ti3+/TiO2/C nanocomposite photocatalysts for enhanced visible-light utilization to degrade organic pollutants by lowering the bandgap of TiO2

A convenient route was developed for the selective preparation of two stable nanocomposites, Ti3+/TiO2/CNT (labeled as TTOC-1 and TTOC-3) and Ti3+/TiO2/carbon layer (labeled as TTOC-2), from the same precursor by varying the amount of single-walled carbon nanotubes used in the synthesis. TiO2 is an effective photocatalyst; however, its wide bandgap limits its usefulness to the UV region. As a solution to this problem, our prepared nanocomposites exhibit a small bandgap and wide visible-light (VL) absorption because of the introduction of carbonaceous species and Ti3+ vacancies. The photocatalytic eciency of the nanocomposites was examined via the degradation of methylene blue dye under VL. Excellent photocatalytic activity of 83%, 98%, and 93% was observed for TTOC-1, TTOC-2, and TTOC-3 nanocomposites within 25 min. In addition, the photocatalytic degradation eciency of TTOC-2 toward rhodamine B was 71% and for methyl orange (MO) dye was 28% under similar experimental conditions, after 25 min. Higher reusability and structural integrity of the as-synthesized photocatalyst were conrmed within ve consecutive runs by photocatalytic test and X-ray diffraction analysis, respectively. The resulting nanocomposites provide new insights into the development of VL-active and stable photocatalysts with high eciencies.

and photocatalytic degradation 4,9,10 . Among the methods for the elimination of dyes, photocatalytic degradation and adsorption are acknowledged as e cient, inexpensive, and environmentally friendly techniques. However, despite the cost-effectiveness of adsorption materials used for dye removal, the adsorption process produces large amounts of solid wastes.
Photocatalytic degradation of organic pollutants is advantageous because of its eco-friendly, safety, and low cost [9][10][11][12] . Semiconductor photocatalysts have attracted intensive attention because of their potential applications in dye-sensitized solar cells, pollutant degradation, biocatalysis, and photocatalytic hydrogen evolution. In particular, titanium dioxide (TiO 2 ) nanomaterials have been commonly studied because of their low toxicity, superior photocatalytic activity, and good chemical and biological stability.
As a result, the development of new TiO 2 photocatalytic systems with enhanced visible-light (VL) absorption is critical and a formidable challenge. To date, several techniques have been used to prolong the separation lifetime of e − /h + pairs and improve the VL absorption of TiO 2 . Among them, heteroelement doping is an excellent approach to addressing these challenges. Cations such as Fe 3+ , Mn 3+ , V 4+ , Re 5+ , Os 3+ , Mo 4+ , and Rh 3+ have been used as dopants in TiO 2 9, 12 . Doping of nonmetals, resulting in F/TiO 2 , S/TiO 2 , N/TiO 2 , C/TiO 2 , and B/TiO 2 , has also been reported 12,15 (Figure 1d). The chemical composition of the TTOC-2 nanocomposite was con rmed from its energy-dispersive X-ray spectrum (EDS) (Figure 1c). The inset FE-SEM image in Figure 1c was   Figures S5 and S6).  Table 1. The adsorption-desorption isotherms of the pristine CNTs and the TTOC-1, TTOC-2, and TTOC-3 nanocomposites showed in Supplementary Figures S9a-9d, respectively and discussed in Supplementary Section 4. Visible-light-driven photocatalytic performance of the nanocomposites. The photocatalytic performance of the TTOC-1, TTOC-2, and TTOC-3 nanocomposites and pristine CNTs was evaluated on the basis of MB degradation under irradiation by a Xe lamp. MB was chosen because it is persistent and widely used in various industries. For comparison, a blank reaction without a catalyst (WC) was also conducted. Figure 6a shows the degradation percentage over the illumination time; the rst-order kinetics of the reaction are represented in Figure 6b. The MB degradation percentage was approximately 83%, 98%, and 93% for the TTOC-1, TTOC-2, and TTOC-3 nanocomposites, respectively, after 25 min of VL irradiation. The self-deterioration of MB was trivial under irradiation of VL. In addition, the pristine CNTs showed no noticeable photocatalytic activity. The change in MB concentration under dark conditions was also measured at regular time intervals. During the adsorption-desorption period, the reduction of the MB concentration after 25 min was negligible. The reaction rate of the MB decomposition on the pristine CNTs and the TTOC-1, TTOC-2, and TTOC-3 nanocomposites was 0.0083, 0.07, 0.15, and 0.10 min −1 , which is 166%, 1400%, 3000%, and 2000% greater, respectively, than the rate of the WC reaction (0.005 min −1 ). All of the prepared nanocomposites showed greater photocatalytic activity because of the coexistence of Ti 3+ and carbon species along with TiO 2 . The bandgap of Ti 3+ -TiO 2 differs from that of pure TiO 2 and can utilize a wide wavelength range of VL radiation for exciting the VB electrons. In addition, the carbon species increased the adsorption of pollutants and reduced the e − /h + recombination rate, thus enhancing the photodegradation e ciency. Because of its greater CNT content, TTOC-3 exhibited greater photocatalytic activity than TTOC-1. However, the photocatalytic performance of TTOC-2 was better than that of TTOC-1 and TTOC-3. In TTOC-2, the Ti 3+ stability is improved by the presence of a carbon layer around the Ti 3+ -TiO 2 NPs. The carbon layer also substantially decreases the particle size and simultaneously increases the speci c surface area. A high speci c surface area enhances the photodegradation e ciency because the reaction occurs at the surface. Discoloration images of an MB solution in the presence of the TTOC-1, TTOC-2, and TTOC-3 nanocomposites are presented in Supplementary Figures S10a, S10b, and S10c, respectively. The results con rm the excellent changes in MB concentration within short periods.
In addition, the degradation activity of TTOC-2 toward MO and RhB dyes was investigated under similar experimental conditions. The photodegradation ratio (C t /C 0 ) over the illumination time is plotted in The pH of the mixture affects the solubility of dyes and the surface chemistry of the adsorbent. The pH PZC demonstrates a sample's surface charge. The drift method was used for pH PZC calculation in the pH range between 2 and 12. HCl and NaOH were employed to control the pH of the solution.
Supplementary Figure S12 shows a graph of (pH i − pH f ) vs pH i , where pH i and pH f and the initial and nal pH, respectively. The measured pH PZC (where the nal pH is equal to the initial value) was 9.31, 9.92, and 9.63 for TTOC-1, TTOC-2, and TTOC-3, respectively. These results imply that the nanocomposites are cationic at pH levels below the pH PZC and anionic at pH levels greater than the pH PZC . To verify this speculation, the effect of pH on the degradation of MB by TTOC-2 was evaluated in the range 2 ≤ pH ≤ 12 under similar experimental conditions; the results are presented in Supplementary Figure S13. Superior photodegradation was observed at pH 12, whereas the worst performance was obtained at pH 2 (Supplementary Figure S13a). The photodegradation ratio clearly increases with increasing pH because MB is cationic at pH values greater than 5.8 (pKa = 5.8). In a basic medium, the electrostatic attraction between the cationic MB and the catalyst's negative surface increases. The opposite effect is observed in an acidic medium, and the photocatalytic e ciency is decreased. The results of kinetics studies of the effect of pH on MB degradation are shown in Supplementary Figure S13b. The linear relation between ln(C 0 /C t ) and irradiation time con rms rst-order kinetics. The rate constant of the reaction at pH 2, 5, 9, and 12 was 0.0163, 0.0314, 0.0754, and 0.1454 min −1 , respectively. The rate constant of the reaction at pH 12 was 892% greater than that of the reaction at pH 2. The superior photodegradation of the assynthesized nanocomposites at high pH levels also con rms the presence of negative surface groups when the nanocomposites are in a basic medium. Discoloration images of MB solution at pH 2 (after 25) and pH 12 (after 20 min) are presented in Supplementary Figures S10d and S10e, respectively.
In addition to the e ciency of a photocatalyst, its stability and reusability are also essential parameters for evaluating its performance. Reusability experiments were performed using recovered nanocomposites. The degradation ratio over ve consecutive cycles is presented in Supplementary Figure  S14a. No signi cant changes were observed throughout the runs. The degradation e ciencies after ve cycles were 81.0%, 96.0%, and 90.2% for TTOC-1, TTOC-2, and TTOC-3, respectively. The XRD patterns of the three nanocomposites were collected (Supplementary Figure S14b) after ve consecutive runs and showed no obvious differences from the patterns of the fresh samples. The XRD crystal planes of the samples before reaction and after the reaction are consistent. The low-intensity characteristic peaks of the (100) plane of carbon species (2θ = 44.27°) are also observed. These results indicate excellent reusability and stability of the nanocomposites.
The photodegradation activity of the as-synthesized nanocomposites is compared with that of various reported catalysts in Supplementary Table S1. This comparison demonstrates that our composites show substantial photocatalytic activity under VL. CB and VB edge potential was estimated to illustrate the photocatalytic mechanism of as-synthesized photocatalyst. Proposed mechanism path of photocatalytic performance was discussed in Supplementary Section 5. The schematic reaction mechanism with redox couples and energy band positions is shown in Figure 7.

Conclusions
In summary, we fabricated Ti 3+ /TiO 2 /CNT and Ti 3+ /TiO 2 /C nanocomposites using a straightforward precipitation and calcination process. The amount of CNTs was varied, whereas the loading amount of other precursors was kept constant. The nanocomposites' E g was remarkably low, and their absorbance covers the entire visible-light wavelength range. The average size of the Ti 3+ /TiO 2 particle in all composites was less than 100 nm, resulting in a high speci c surface area. The photocatalytic e ciency of the nanocomposites was tested for the degradation of a MB solution under VL. All of the nanocomposites showed high photocatalytic e ciency within 25 min: 83%, 98%, and 93% for the TTOC-1, TTOC-2, and TTOC-3 nanocomposites, respectively. The photocatalytic e ciency enhancement was attributed to the introduction of Ti 3+ and carbon species onto TiO 2 . Among the nanocomposites, TTOC-2 exhibited the highest activity because of its large Ti 3+ content as a result of the formation of a carbon shell. Similar experiments with TTOC-2 for the degradation of RhB and MO were performed, resulting iñ 71% and ~28% degradation, respectively. The PZC of the nanocomposites revealed a negative nature of their surface at high pH. The effect of pH on MB degradation using TTOC-2 was also demonstrated. With increasing pH value, the photocatalytic activity increased. Moreover, after ve consecutive cycles, no apparent loss of photocatalytic activity was observed and the XRD patterns showed no structural changes, indicating good cycling stability. Therefore, the proposed nanocomposites are suitable for practical application in wastewater treatment because of their high stability and high photocatalytic e ciency. In addition, the selective preparation techniques for the two different nanocomposites might be useful in the preparation of future photocatalysts.

Experimental Methods
Selective preparation of Ti 3+ /TiO 2 /CNT and Ti 3+ /TiO 2 /C nanocomposite. Titanium(IV) isopropoxide (TTIP), sodium borohydride (NaBH 4 ), and ethanol were sourced from Sigma-Aldrich (USA). The singlewalled carbon nanotubes (CNTs, outer diameter: 1-2 nm, length: 5-30 µm) were purchased from US Research 307 Nanomaterials (Houston, USA). MB was obtained from Alfa Aesar (UK). The nanocomposites were prepared using a convenient and facile two-step precipitation and calcination process with TTIP and CNT as precursors. In the rst step, 5 mg of CNTs was dispersed in 25 mL of ethanol using a sonication bath. Five milliliters of TTIP was then applied to the dispersed CNTs under continuous stirring. A 100 mL aqueous mixture of 0.10 g NaBH 4 was then slowly poured into the solution.
The mixture was covered with Al foil and then vigorously stirred for 3 h at 600 rpm on a magnetic stirrer. The precipitate was rinsed with distilled water and dried overnight at 60°C. Subsequently, in the second step, the precipitate was calcined at 550°C for 6 h with a ramp rate of 7.5°C/min to obtain a stable composite. The thus-obtained nanocomposite was labeled as TTOC-1. Here, the NaBH 4 was used to reduce Ti 4+ ions on the surface of TiO 2 to Ti 3+ ions. Three nanocomposites were prepared with different mass loadings of CNTs; the other reaction conditions were unchanged. During dispersion, the volume of ethanol was increased proportionally with increasing amount of CNTs. The obtained products with CNT loadings of 10, 15, and 20 mg were denoted as TTOC-1, TTOC-2, and TTOC-3, respectively. A reaction mechanism is proposed for the preparation of the two different nanocomposites (Supplementary Figure   S1). The precipitation reaction produced three Ti 3+ /TiO 2 /CNT composites with different CNT loadings.
However, after calcination, the CNTs ruptured and produced a carbon layer around the Ti 3+ /TiO 2 -NPs only in the TTOC-2 nanocomposite. The aforementioned analysis indicates that, among the nanocomposites, TTOC-2 exhibits the strongest interaction between the TiO 2 and the CNTs. The reproducibility of the method was checked by repeating the process several times; in each case, the results were identical.
Characterization methods are described in Supplementary Section 1 in detail.
Investigation of photocatalytic activity. The photocatalytic e ciency of the samples was tested via the degradation of MB, RhB, and MO dyes using a 300 W Xe lamp as a solar-light simulator. A 100 mL aqueous solution containing 10 mg of dye was mixed with 0.05 g of catalyst under ultrasonication for 1 h. The mixture was then placed in the dark for 1 h to ensure that adsorption/desorption equilibrium was achieved. The mixture was irradiated under a Xe lamp for 25 min. A 400 nm UV cutoff lter was used to prevent irradiation with UV light. A xed amount of solution was collected at regular time intervals, and the absorbance of the solution was monitored using a UV-vis spectrophotometer within the wavelength range 200-750 nm. Characteristic wavelength peaks at λ MB = 664, λ RhB = 554, and λ MO = 464.5 nm were monitored to evaluate the extent of organic pollutant degradation. After photodegradation, the catalyst was collected, rinsed with distilled water, and dried. The photocatalytic activity test was repeated for ve consecutive cycles with reused samples under identical experimental conditions.
The photodegradation percentage was estimated using equation (1): where C t is the concentration of dye at degradation time t, C 0 is the initial concentration of dye, and is the degradation e ciency.
The rate constant (k) of the degradation reaction was calculated using equation (2):