Enhanced Photocatalytic Performance of Nitrogen-Doped TiO2 Nanotube Arrays Using a Simple Annealing Process

Nitrogen-doped TiO2 nanotube arrays (N-TNAs) were successfully fabricated by a simple thermal annealing process in ambient N2 gas at 450 °C for 3 h. TNAs with modified morphologies were prepared by a two-step anodization using an aqueous NH4F/ethylene glycol solution. The N-doping concentration (0–9.47 at %) can be varied by controlling N2 gas flow rates between 0 and 500 cc/min during the annealing process. Photocatalytic performance of as-prepared TNAs and N-TNAs was studied by monitoring the methylene blue degradation under visible light (λ ≥ 400 nm) illumination at 120 mW·cm−2. N-TNAs exhibited appreciably enhanced photocatalytic activity as compared to TNAs. The reaction rate constant for N-TNAs (9.47 at % N) reached 0.26 h−1, which was a 125% improvement over that of TNAs (0.115 h−1). The significant enhanced photocatalytic activity of N-TNAs over TNAs is attributed to the synergistic effects of (1) a reduced band gap associated with the introduction of N-doping states to serve as carrier reservoir, and (2) a reduced electron‒hole recombination rate.


Materials and Methods
Titanium foil (99.9% purity, 1 cm × 2.5 cm size, 0.4 mm thickness) was used as the substrate for forming TNAs with modified morphologies by a two-step anodic oxidation. Prior to anodization, titanium (Ti) foil was ultrasonically cleaned in acetone, methanol and deionized water (each solvent 10 min), and then dried by a purging N 2 gas. The anodization was carried out using a two-electrode system with the Ti foil as an anode and a stainless steel foil (SS304) as a cathode. All the electrolytes consisted of 0.3 wt % NH 4 F (SHOWA, Tokyo, Japan) in ethylene glycol solution with 2 vol % water. In the first-step anodization, Ti foil was anodized at 50 V for 1 h to form a nanotube layer. Then, the as-grown layer was ultrasonically removed for approximately 15 min until a bright surface appeared. The same Ti foil underwent the second anodization at 40 V for 30 min to form TNAs with a grid-like top-layer structure on the Ti substrates. The prepared sample was then immersed in ethanol solution for 5 h to remove the ethylene glycol contamination. Nitrogen-doping for the prepared TNAs were carried out by thermal annealing process at 450 • C for 3 h under various nitrogen flows: vacuum (~7 Pa), 200, 350, and 500 cc/min. Prior to introducing pure N 2 gas (purity 99.99%), the annealing chamber was evacuated to a pressure of approximately 7 Pa for preventing contamination.
The crystal structures of the pristine and N-doped TNAs were characterized by X-ray diffraction (XRD, Bruker D2, Bruker, Billerica, MA, USA) using Cu Kα radiation (λ = 1.5406 Å) and high-resolution transmission electron microscopy (TEM, JEOL JEM-ARM200F, JEOL Ltd., Tokyo, Japan), operated at 200 kV. The TEM specimen was prepared by crashing the TNAs on TEM grids. X-ray photoelectron spectroscopy (XPS) experiments were conducted at beamline 09A2, NSRRC, Taiwan to determine the chemical bonding and surface N-doping concentration. XPS curve were fitted using the freeware XPSPEAK4.1 with the Shirley background subtraction and assuming a Gaussian-Lorentzian peak shape. Morphologies of the pristine and N-doped TNAs were characterized by scanning electron microscopy (SEM, JEOL JSM-6500, JEOL Ltd.). Fundamental information on the energy levels lying within the band gap and charge carrier trapping and transfer was gathered by photoluminescence (PL) spectra, measured at room temperature (~22 • C). The excitation light source of the PL spectroscopy was an He-Cd laser with a wavelength of 325 nm. The PL signal was dispersed by a Horiba Jobin Yvon IHR-320 single-grating (1800 grooves/nm grating) spectrometer (Horiba, Kyoto, Japan). The performance of pristine and N-doped TNAs was characterized by ultraviolet photoelectron spectroscopy (UPS) using monochromatized He-I radiation at 21.2 eV. The electrochemical impedance spectroscopy (EIS) measurements were performed using an Autolab system with a scan rate of 50 mV·s −1 and a conventional three-electrode test cell. The counter electrode was a platinum sheet, the reference electrode was Ag/AgCl (aqueous 3 M KCl), and TNAs and N-TNAs samples were used as a working electrode. The electrolyte was 0.1 M Na 2 SO 4 aqueous solution under room light irradiation.
The photocatalytic activity in the degradation of methylene blue (MB) of the samples was measured under visible light illumination at 120 mW·cm −2 . The samples were illuminated using a Xenon lamp with a band-pass filter for λ ≥ 400 nm. Prior to illumination, the suspension was magnetically stirred for 20 min in the dark to ensure absorption-desorption equilibrium between the photocatalyst and MB solution. The reaction temperature was kept at 32-33 • C for all samples. After a certain photocatalytic reaction time, the solution was taken for performing UV-Vis absorption spectrum to measure the MB concentration change using the characteristic peak at 654 nm. Figure 1 shows the XRD patterns of the pristine TNAs and N-doped TNAs with various nitrogen contents (see Table 1  . Also, there were no rutile peaks, indicating a pure anatase phase for the TNAs in this study. The XRD patterns of the TNAs and N-TNAs are similar, agreeing with those reported in the literature [4,5,13,22,24]. Notably, the preferred orientation changed from (101) for TNAs to (004) for N-TNAs. The reason is not clear at present, but the change in preferred orientation to (004) for N-TNAs should not be due to N atom incorporation, which usually coincides with the shift of the XRD peak [23,45]; meanwhile, the (004)-dominant orientation was also found in TNAs (without N-doping) prepared by one-step or two-step anodization [43]. Recently, the intrinsic electric fields were found to be responsible for the oriented self-assembly of multilevel branched rutile-type TiO 2 structures [46]. Therefore, for the present case, the unintentional slight difference in electric fields during anodic oxidation processes may play a role in determining the preferred orientation, i.e., (101) or (004).    Figure 2b shows a grid-like top layer, which plays a role as a capping layer to protect the TNAs' structure from bundling and crumpling. Figure 2c,d shows TEM images of typical N-doped TNAs (i.e., 9.47 at %). The TNAs have an average tube diameter of ~70 nm and a wall thickness of ~20 nm. In addition, the lattice fringes with a d-spacing of 0.245 nm can be assigned to the (004) lattice plane of anatase TiO2 (Figure 2d). The (004)-lattice fringes were primarily observed from HRTEM images, further confirming the (004)-preferred orientation for the N-doped TNAs.

Results and Discussion
It has been found that the two-step anodization is a good method for preparing much more uniform TNAs [31,41] and the growth of diverse top-layer morphologies covering on TNAs [42]. The TNAs with a grid-like top-layer structure are an outcome of competition between the electric-field-driven anodic oxidation of Ti to form TiO2, and the electric-field-assisted chemical dissolution of the TiO2 layer [41,47]. The reactions are given below: Anodic reaction: Ti + 2H2O − 4e → TiO2 + 4H + Cathodic reaction: 4H + + 4e → 2H2 Chemical etching (dissolution) reaction: TiO2 + 6F − + 4NH4 + → TiF6 2− + 2H2O.
The anodic oxidation reaction occurs as Ti 4+ ejection and deposition on the surface in the form of TiO2, while the TiF6 2− etching reaction occurs from top to bottom of the as-grown TiO2. The anodic   Figure 2b shows a grid-like top layer, which plays a role as a capping layer to protect the TNAs' structure from bundling and crumpling. Figure 2c,d shows TEM images of typical N-doped TNAs (i.e., 9.47 at %). The TNAs have an average tube diameter of 70 nm and a wall thickness of~20 nm. In addition, the lattice fringes with a d-spacing of 0.245 nm can be assigned to the (004) lattice plane of anatase TiO 2 ( Figure 2d). The (004)-lattice fringes were primarily observed from HRTEM images, further confirming the (004)-preferred orientation for the N-doped TNAs. It has been found that the two-step anodization is a good method for preparing much more uniform TNAs [31,41] and the growth of diverse top-layer morphologies covering on TNAs [42]. The TNAs with a grid-like top-layer structure are an outcome of competition between the electric-field-driven anodic oxidation of Ti to form TiO 2 , and the electric-field-assisted chemical dissolution of the TiO 2 layer [41,47]. The reactions are given below: Anodic reaction: Ti + 2H 2 O − 4e → TiO 2 + 4H + Cathodic reaction: 4H + + 4e → 2H 2 Chemical etching (dissolution) reaction: The anodic oxidation reaction occurs as Ti 4+ ejection and deposition on the surface in the form of TiO 2 , while the TiF 6 2− etching reaction occurs from top to bottom of the as-grown TiO 2 . The anodic oxidation rate is very fast and dominated over the NH 4 F etching rate, resulting in a thin oxide layer in the early stage [31,41]. In the late stage, the deposition rate of TiO 2 at the entrance of nanotubes slows down, while field-induced random dissolution of the surface becomes more significant or dominant to form pore-like structures, which further develop into TNA structures [31,41,47]. At certain relative rates between TiO 2 deposition and dissolution, a layer of interconnected nanopores can be constructed on the top of TiO 2 , as shown in the inset of Figure 2b. Notably, the first anodization step offers highly ordered imprints after ultrasonication removal, which plays the role of a template for the subsequent growth of well-aligned nanotubes [31]. The highly ordered TNAs are believed to have potential applications in fields such as solar cells, photonic crystals, photocatalyst, and hydrogen storage [31,41,42].
Micromachines 2018, 9, x FOR PEER REVIEW 5 of 13 anodization step offers highly ordered imprints after ultrasonication removal, which plays the role of a template for the subsequent growth of well-aligned nanotubes [31]. The highly ordered TNAs are believed to have potential applications in fields such as solar cells, photonic crystals, photocatalyst, and hydrogen storage [31,41,42]. XPS analysis was performed to quantitatively determine the N-doping concentrations for all the TNAs samples. Figure 3a shows the high-resolution XPS N1s core level spectra of the samples. There was no N1s peak for the TNA annealed in a vacuum. Meanwhile, clear N1s peaks appeared in all the TNAs annealed in N2 gas flows, indicating the presence of N species on the sample surfaces. The N1s peaks can be fitted well with three components, namely Ti-O-N linkage at 400.2 eV [48][49][50], O-Ti-N linkage at 398.7 eV [50], and Ti-N linkage at 397.1 eV [51]. Figure 3b shows a typical O1s spectrum of N-doped TNAs (i.e., the 9.47 at % N sample). As expected, the Ti-O and O-N bonds were clearly observed from O1s spectra of the N-doped TNAs at 530.4 eV and 532.2 eV, respectively [48][49][50].  XPS analysis was performed to quantitatively determine the N-doping concentrations for all the TNAs samples. Figure 3a shows the high-resolution XPS N1s core level spectra of the samples. There was no N1s peak for the TNA annealed in a vacuum. Meanwhile, clear N1s peaks appeared in all the TNAs annealed in N 2 gas flows, indicating the presence of N species on the sample surfaces. The N1s peaks can be fitted well with three components, namely Ti-O-N linkage at 400.2 eV [48][49][50], O-Ti-N linkage at 398.7 eV [50], and Ti-N linkage at 397.1 eV [51]. Figure 3b shows a typical O1s spectrum of N-doped TNAs (i.e., the 9.47 at % N sample). As expected, the Ti-O and O-N bonds were clearly observed from O1s spectra of the N-doped TNAs at 530.4 eV and 532.2 eV, respectively [48][49][50]. Figure 3c presents the two Ti2p peaks: Ti2p 1/2 at 459.2 eV and Ti2p 3/2 at 464.9 eV. The XPS results for N-doped TiO 2 agree well with those in [49,50,52]. The quantitative results of XPS analysis are summarized in Table 1. The N doping concentration varied with the N2 gas flows. The N concentration increased from 0 to 9.47 at % with increasing N2 gas flow from 0 (vacuum) to 350 cc/min, but it decreased under a sufficient high N2 gas flow of 500 cc/min. As shown in Figure 3a, the component of Ti-N monotonically increase with increasing N2 gas flows, suggesting the presence of TiN ultra-thin layer on the surfaces of N-TNAs. Thus, it is reasonable to believe that a considerably thick TiN layer was grown during the annealing process at 450 °C under a sufficient high N2 gas flow (i.e., 500 cc/min). The TiN layer can constrain the injection of N atoms into TNAs via thermal diffusion, resulting in the lower N-doping concentration for 500 cc/min than it is for 350 cc/min ( Table 1).
The optical absorption is an important property of a photocatalyst. The UV-visible absorption (Figure 4a) is recorded in the range of 350-750 nm. The band gap (Eg) is determined by plotting of (αhν) 1/2 against the energy (hν), and by extrapolating the straight line to hv axis, as shown in Figure 4b, where α is the absorption coefficient, hν is the photon energy. The Eg of TNAs, with 0, 5.76, 6.60, and 9.47 at % N-TNAs were found to be 3.13 eV, 3.05 eV, 2.91 eV, and 2.95 eV, respectively. The band gap narrowing effect for N-TNAs could be attributed to the presence of N-doped levels (see Figure 4d). It has been reported that N-doping will introduce N2p states near valance band (VB) [53]. Moreover, the incorporation of nitrogen into the TiO2 structure occurs via substitutional and interstitial means [11]. The substitutional doping (Ns) involves oxygen replacement, which reduces the Eg to 3.06 eV [11,54]. Meanwhile, the interstitial doping (Ni) significantly reduces the Eg to (~ 2.46 eV) [11,54]. The present slight Eg reduction of N-TNAs suggests that the substitutional N doping is the dominant mechanism.
To further elucidate band structures of the samples, the work functions of a pristine TNAs and a N-TNAs (9.47 at %) are characterized by UPS, and the results are shown in Figure 4c. In the experiments, the applied bias is 5 V, and the kinetic energy is determined by drawing the tangent of the spectra. The intercept of the straight line with x-axis are found to be 8.9 eV and 9.2 eV for the TNAs and the N-TNAs, respectively (Figure 4c). Thus, the calculated work functions are 3.9 eV for TNAs and 4.2 eV for the N-TNAs. The UPS results were consistent with the results by Sudhagar et al. [45], where the N-ion implanted TiO2 also had larger work function than that of pristine TiO2 (i.e., The quantitative results of XPS analysis are summarized in Table 1. The N doping concentration varied with the N 2 gas flows. The N concentration increased from 0 to 9.47 at % with increasing N 2 gas flow from 0 (vacuum) to 350 cc/min, but it decreased under a sufficient high N 2 gas flow of 500 cc/min. As shown in Figure 3a, the component of Ti-N monotonically increase with increasing N 2 gas flows, suggesting the presence of TiN ultra-thin layer on the surfaces of N-TNAs. Thus, it is reasonable to believe that a considerably thick TiN layer was grown during the annealing process at 450 • C under a sufficient high N 2 gas flow (i.e., 500 cc/min). The TiN layer can constrain the injection of N atoms into TNAs via thermal diffusion, resulting in the lower N-doping concentration for 500 cc/min than it is for 350 cc/min ( Table 1).
The optical absorption is an important property of a photocatalyst. The UV-visible absorption (Figure 4a) is recorded in the range of 350-750 nm. The band gap (E g ) is determined by plotting of (αhν) 1/2 against the energy (hν), and by extrapolating the straight line to hv axis, as shown in Figure 4b, where α is the absorption coefficient, hν is the photon energy. The E g of TNAs, with 0, 5.76, 6.60, and 9.47 at % N-TNAs were found to be 3.13 eV, 3.05 eV, 2.91 eV, and 2.95 eV, respectively. The band gap narrowing effect for N-TNAs could be attributed to the presence of N-doped levels (see Figure 4d). It has been reported that N-doping will introduce N2p states near valance band (VB) [53]. Moreover, the incorporation of nitrogen into the TiO 2 structure occurs via substitutional and interstitial means [11]. The substitutional doping (N s ) involves oxygen replacement, which reduces the E g to 3.06 eV [11,54]. Meanwhile, the interstitial doping (N i ) significantly reduces the E g to (~2.46 eV) [11,54]. The present slight E g reduction of N-TNAs suggests that the substitutional N doping is the dominant mechanism.
To further elucidate band structures of the samples, the work functions of a pristine TNAs and a N-TNAs (9.47 at %) are characterized by UPS, and the results are shown in Figure 4c. In the experiments, the applied bias is 5 V, and the kinetic energy is determined by drawing the tangent of the spectra. The intercept of the straight line with x-axis are found to be 8.9 eV and 9.2 eV for the TNAs and the N-TNAs, respectively (Figure 4c). Thus, the calculated work functions are 3.9 eV for TNAs and 4.2 eV for the N-TNAs. The UPS results were consistent with the results by Sudhagar et al. [45], where the N-ion implanted TiO 2 also had larger work function than that of pristine TiO 2 (i.e., 3.9 vs. 3.7 eV). This reflects the incorporation of N atoms into TiO 2 , which induces the modification of the TiO 2 lattice through substitutional doping of oxygen with nitrogen and interstitial nitrogen [53,55,56]. Figure 4d shows a schematic band diagram of N-TNAs, in which the N-doping states is located just above the VB, and will involve in photocatalytic activity of N-TNAs in the visible range.
Micromachines 2018, 9, x FOR PEER REVIEW 7 of 13 [53,55,56]. Figure 4d shows a schematic band diagram of N-TNAs, in which the N-doping states is located just above the VB, and will involve in photocatalytic activity of N-TNAs in the visible range. The insights into the charge recombination process, the effectiveness of trapping, migration and transfer of charge carriers are revealed by means of PL spectra of the TNAs and N-TNAs. As PL emission mainly results from the recombination of excited electrons and holes, a lower PL intensity may indicate a lower recombination rate of electron-hole pairs and higher separation efficiency under the same test conditions [57,58]. Figure 5a shows the PL spectra of TNAs (0 at % N) and N-TNAs (5.76-9.47 at % N). Clearly, broad PL peaks are observed around 558 nm (~2.22 eV) for the anatase-phase TNAs and N-TNAs, whose origin has been considered to be partially reduced titanium ions, self-trapped excitons, oxygen vacancies and surface states [35,59]. As shown in Figure 5b, the PL emission spectrum of a N-TNAs (9.47 at %) can be fitted by two peaks at 591 nm (2.2 eV) and 524 nm (2.36 eV). The energy levels of oxygen vacancies are located at ~0.5 eV and ~0.8 eV below the conduction band (CB) of TiO2 [60]. Thus, the photogenerated electrons in the CB can fall into the oxygen vacancies through a non-irradiative process, and then they recombine with photogenerated holes in the valence band (VB) to result in the fluorescence emission (see Figure 5c) [60]. Furthermore, as shown in Figure 5a, the PL intensity of N-TNAs systematically decreases with increasing N-doping concentration. The lower PL intensity of N-TNAs can be attributed to the capture of photogenerated holes by surface states (surface nitrogen species) and N states near VB (Figure 5c), agreeing with the PL results in [60]. The PL results further confirm the presence and N The insights into the charge recombination process, the effectiveness of trapping, migration and transfer of charge carriers are revealed by means of PL spectra of the TNAs and N-TNAs. As PL emission mainly results from the recombination of excited electrons and holes, a lower PL intensity may indicate a lower recombination rate of electron-hole pairs and higher separation efficiency under the same test conditions [57,58]. Figure 5a shows the PL spectra of TNAs (0 at % N) and N-TNAs (5.76-9.47 at % N). Clearly, broad PL peaks are observed around 558 nm (~2.22 eV) for the anatase-phase TNAs and N-TNAs, whose origin has been considered to be partially reduced titanium ions, self-trapped excitons, oxygen vacancies and surface states [35,59]. As shown in Figure 5b, the PL emission spectrum of a N-TNAs (9.47 at %) can be fitted by two peaks at 591 nm (2.2 eV) and 524 nm (2.36 eV). The energy levels of oxygen vacancies are located at~0.5 eV and~0.8 eV below the conduction band (CB) of TiO 2 [60]. Thus, the photogenerated electrons in the CB can fall into the oxygen vacancies through a non-irradiative process, and then they recombine with photogenerated holes in the valence band (VB) to result in the fluorescence emission (see Figure 5c) [60]. Furthermore, as shown in Figure 5a, the PL intensity of N-TNAs systematically decreases with increasing N-doping concentration. The lower PL intensity of N-TNAs can be attributed to the capture of photogenerated holes by surface states (surface nitrogen species) and N states near VB (Figure 5c), agreeing with the PL results in [60]. The PL results further confirm the presence and N states, which serve as a hole reservoir to induce excitation under visible light irradiation and improve electron-hole separation efficiency. The photocatalytic activity of TNAs and N-TNAs were studied through photocatalytic degradation of methylene blue (MB) under visible light irradiation. As can be seen in Figure 6a,b, all the N-TNAs possessed higher photocatalytic activity in MB degradation than that of the TNAs, and the activity of N-TNAs increased with increasing N concentration from 5.76 to 9.47 at %. The pseudo-first-order rate constants were determined by fitting the data with the Langmuir-Hinshelwood kinetics rate model [61,62]. The TNAs had a reaction rate constant (k) of 0.115 (h −1 ), and k increased monotonically with increasing N-doping concentration (see Table 1). The N-TNA (9.47 at %) achieved the highest k value up to 0.259 (h −1 ), which was 125% higher than that of the pristine TNAs. The present enhancements in photocatalytic degradation MB for N-TNAs are attributed to the localized nitrogen (N-doping states) that enables N-TNAs to enhance both visible light absorption and electron-hole separation. A local N inter-band induces the optical absorption of TNAs, which can enhance the generation of electron-hole pairs. This study demonstrates that the facile thermal in N2 ambient at an elevated temperature of 450 °C allows introducing certain N-doping into TiO2 lattice to result in the significant enhancement in photocatalytic activity of TiO2 nanomaterials.
To investigate electron transport properties of TNAs and N-TNAs at the solid-liquid interface, the electrochemical impedance spectroscopy (EIS) data were collected under room light illumination. Figure 6c shows the EIS Nyquist plots for TNAs and N-TNAs (9.47 at %). Clearly, the circular arc radius of N-TNAs electrode is much smaller than TNAs electrode, indicating the smaller interface charge transfer resistance (Rct) of the former. Indeed, by fitting with the equivalent circuit in the inset, we yielded Rct = 19,869 Ω for TNAs and Rct = 6482 Ω for N-TNAs. Thus, it suggests that the introduction of N atom is more beneficial to the separation of the photo-induced electrons and holes and faster charge transfer than that of the TNAs electrode at the solid-liquid interface, agreeing with the results in [4,22,63]. The photocatalytic activity of TNAs and N-TNAs were studied through photocatalytic degradation of methylene blue (MB) under visible light irradiation. As can be seen in Figure 6a,b, all the N-TNAs possessed higher photocatalytic activity in MB degradation than that of the TNAs, and the activity of N-TNAs increased with increasing N concentration from 5.76 to 9.47 at %. The pseudo-first-order rate constants were determined by fitting the data with the Langmuir-Hinshelwood kinetics rate model [61,62]. The TNAs had a reaction rate constant (k) of 0.115 (h −1 ), and k increased monotonically with increasing N-doping concentration (see Table 1). The N-TNA (9.47 at %) achieved the highest k value up to 0.259 (h −1 ), which was 125% higher than that of the pristine TNAs. The present enhancements in photocatalytic degradation MB for N-TNAs are attributed to the localized nitrogen (N-doping states) that enables N-TNAs to enhance both visible light absorption and electron-hole separation. A local N inter-band induces the optical absorption of TNAs, which can enhance the generation of electron-hole pairs. This study demonstrates that the facile thermal in N 2 ambient at an elevated temperature of 450 • C allows introducing certain N-doping into TiO 2 lattice to result in the significant enhancement in photocatalytic activity of TiO 2 nanomaterials.
To investigate electron transport properties of TNAs and N-TNAs at the solid-liquid interface, the electrochemical impedance spectroscopy (EIS) data were collected under room light illumination. Figure 6c shows the EIS Nyquist plots for TNAs and N-TNAs (9.47 at %). Clearly, the circular arc radius of N-TNAs electrode is much smaller than TNAs electrode, indicating the smaller interface charge transfer resistance (R ct ) of the former. Indeed, by fitting with the equivalent circuit in the inset, we yielded R ct = 19,869 Ω for TNAs and R ct = 6482 Ω for N-TNAs. Thus, it suggests that the introduction of N atom is more beneficial to the separation of the photo-induced electrons and holes and faster charge transfer than that of the TNAs electrode at the solid-liquid interface, agreeing with the results in [4,22,63].  Table 2 summarizes the synthesis methods of N-doped TiO2-based nanomaterials and their photocatalytic performance in degradation various organic pollutants under light irradiation. The photocatalytic degradation rate of the N-TNAs (9.47 at %, k = 0.26 h −1 ) in this study was twice as high than that of N-TNAs (k = 0.11 h −1 ) [22], N-TiO2 nanowires (k = 0.13 h −1 ) [63], and CdS-Ag/TiO2 nanotubes (k = 0.13 h −1 ) [64]. However, it was almost 2-6 times lower than the k constants of N-TiO2 nanoparticles (k = 0.44 h −1 ) prepared by sol-solvothermal process [58], N-TiO2 nanosheets (k = 0.45 h −1 ) synthesized chemical route and electrospinning technique [52], and N-TNAs prepared anodic oxidation [4]. It is worthy of mentioning that the difference in k-constant results can come from the differences in treated solution and intrinsic properties of the materials such as surface area, crystallinity, N-doping concentration, band structure, etc. Table 2. N-doped TiO2-based nanomaterials, synthesis methods, organic pollutants, N-doping concentration, and reaction rates of the optimal N-TNAs in this study and in the relevant literature.

Conclusions
Modified TNAs and N-TNAs were fabricated by a two-step anodization, followed by thermal annealing at 450 °C in different N2 gas flows for 3 h. TNAs exhibits highly order uniformly and has a grid-like top layer that is grown by the competition between the electric-field-driven anodic oxidation of Ti to form TiO2, and the electric-field-assisted chemical dissolution of the TiO2 layer. The XPS, UV-Vis, UPS, and PL results confirm that N atoms are successfully incorporated into TiO2 lattice to introduce N states (or N-interband) just above the valance band. The N states induce the  Table 2 summarizes the synthesis methods of N-doped TiO 2 -based nanomaterials and their photocatalytic performance in degradation various organic pollutants under light irradiation. The photocatalytic degradation rate of the N-TNAs (9.47 at %, k = 0.26 h −1 ) in this study was twice as high than that of N-TNAs (k = 0.11 h −1 ) [22], N-TiO 2 nanowires (k = 0.13 h −1 ) [63], and CdS-Ag/TiO 2 nanotubes (k = 0.13 h −1 ) [64]. However, it was almost 2-6 times lower than the k constants of N-TiO 2 nanoparticles (k = 0.44 h −1 ) prepared by sol-solvothermal process [58], N-TiO 2 nanosheets (k = 0.45 h −1 ) synthesized chemical route and electrospinning technique [52], and N-TNAs prepared anodic oxidation [4]. It is worthy of mentioning that the difference in k-constant results can come from the differences in treated solution and intrinsic properties of the materials such as surface area, crystallinity, N-doping concentration, band structure, etc. Table 2. N-doped TiO 2 -based nanomaterials, synthesis methods, organic pollutants, N-doping concentration, and reaction rates of the optimal N-TNAs in this study and in the relevant literature.

Conclusions
Modified TNAs and N-TNAs were fabricated by a two-step anodization, followed by thermal annealing at 450 • C in different N 2 gas flows for 3 h. TNAs exhibits highly order uniformly and has a grid-like top layer that is grown by the competition between the electric-field-driven anodic oxidation of Ti to form TiO 2 , and the electric-field-assisted chemical dissolution of the TiO 2 layer. The XPS, UV-Vis, UPS, and PL results confirm that N atoms are successfully incorporated into TiO 2 lattice to introduce N states (or N-interband) just above the valance band. The N states induce the narrowing the band gap effect to yield the enhancement of visible light absorption for N-TNAs. In addition, N-states easily trap holes and thus serve as carrier reservoir to provide more photocarriers for enhancing photocatalytic activity under visible light irradiation. It also enables us to reduce the electron-hole recombination rate. Thanks to these synergetic effects, the N-doped TNAs with 9.47 at % N exhibit superior photocatalytic activity in MB degradation with k = 0.26 h −1 , which accounts for a 125% enhancement as compared to the pristine TNAs. The results of this study demonstrate that the simple annealing process is beneficial for introducing N-doping into TiO 2 and enhanced the photocatalytic activity and applications.

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