Iron Modified Titanate Nanotube Arrays for Photoelectrochemical Removal of E. coli

This study used iron modified titanate nanotube arrays (Fe/TNAs) to remove E. coli in a photoelectrochemical system. The Fe/TNAs was synthesized by the anodization method and followed by the square wave voltammetry electrochemical deposition (SWVE) method with ferric nitrate as the precursor. Fe/TNAs were characterized by SEM, XRD, XPS, and UV-vis DRS to investigate the surface properties and light absorption. As a result, the iron nanoparticles (NPs) were successfully deposited on the tubular structure of the TNAs, which showed the best light utilization. Moreover, the photoelectrochemical (PEC) properties of the Fe/TNAs were measured by current-light response and electrochemical impedance spectroscopy. The photocurrent of the Fe/TNAs-0.5 (3.5 mA/cm2) was higher than TNAs (2.0 mA/cm2) and electron lifetime of Fe/TNAs-0.5 (433.3 ms) were also longer than TNAs (290.3 ms). Compared to the photolytic (P), photocatalytic (PC), and electrochemical (EC) method, Fe/TNAs PEC showed the best removal efficiency for methyl orange degradation. Furthermore, the Fe/TNAs PEC system also performed better removal efficiency than that of photolysis method in E. coli degradation experiments.


Introduction
Titanium dioxide (TiO 2 ) is one of the most promising photocatalysts because of its excellent properties such as resistance to acid and alkali corrosion, relative low price, nontoxicity, and small environmental footprint. Therefore, it has been widely applied in water treatment [1], atmospheric VOC (volatile organic compounds) removal [2], disinfection [3,4], hydrogen production [5][6][7], CO 2 reduction [8], DSSC (dye-sensitized solar cell) [9,10], and sensing devices [11], etc. The photo-generated holes can react with water molecular to form hydroxyl radicals, which is a powerful oxidant, to oxidize most organic compounds such as dyes, PPCPs (pharmaceutical and personal care products) [6,12], VOCs [2], and PCDD/Fs (polychlorinated dibenzodioxins and polychlorinated dibenzofurans) [13,14]. On the other hand, the photo-generated electrons can react to generate hydrogen, which is a kind of clean energy. Moreover, the photo-generated electrons can also be applied for environmental applications, e.g., to reduce TCE (trichloroethylene) [15] and to react with ammonia borane (NH 3 BH 3 ) to produce hydrogen [16].
The large band gap of 3.2 eV is the main obstacle for TiO 2 applications. Many researchers employed non-metals such as N [17,18], C [19], F [20], and S [21] to dope into TiO 2

Preparation of Fe Modified Titanium Dioxide Nanotube Arrays (Fe/TNAs)
The titanium foil (99.5%, Zhang Jia Ltd., New Taipei City, Taiwan) was cut into the size of 2 × 2.5 cm and then was washed via acetone, ethanol, and DI water sequent to remove dust and impurities. After cleaning, the titanium foil was dried in the air. The pre-treated titanium foil was used to synthesize TNAs by anodization method. The positive electrode was connected to a titanium sheet, while the negative electrode was connected to a platinum (Pt) electrode in the solution of ammonium fluoride (0.5 wt % NH 4 F (98%, Alfa Aesar, Haverhill, MA, USA) and 6 vt % H 2 O) as the etching solution. The titanium foil was etched at a constant voltage of 40 V for 1 h to obtain TNAs. The as-synthesized TNAs was rinsed with DI water, and dried in oven at 40 • C. To increase the crystalline, the TNAs was calcined at 450 • C for 3 h. The iron deposited TNAs were synthesized by square wave voltammetry electrochemical deposition method (SWVE). First, pH of mixed solution of 0.2 M Fe(NO 3 ) 3 ·9H 2 O (98%, Alfa Aesar, Haverhill, MA, USA) and 0.05 M NaBH 4 (99%, Alfa Aesar, Haverhill, MA, USA) was adjusted to 3.5 via 0.5 M HNO 3 (65%, MERCK, Darmstadt, Germany). The TNAs sample was then immersed in the prepared solution for 20 min. The initial and final applied voltage was set as −1.0 and 0.0 V, respectively, with amplitude of 0.005 V and frequency of 2 Hz for the SWVE method. The as-synthesized samples were named as Fe/TNAs-0.2 and Fe/TNAs-0.5 according to the concentration of Fe(NO 3 ) 3 ·9H 2 O as 0.2 M and 0.5 M, respectively. The as-synthesized Fe/TNAs was rinsed by DI water and dried at 40 • C for two hours.

Characterization of Nanotube Arrays
The crystal structure of TNAs and Fe/TNAs was investigated by X-ray diffraction (XRD) (X'Pert Pro MRD, PANalytical, Almelo, The Netherlands) using a Cu Kα source at a wavelength of 0.154 nm. JCPDS PDF card database was selected as the identification of XRD peaks. The morphology was studied using a field-emission scanning electron microscope (Nova NanoSEM 430 FEI, Hillsboro, OR, USA). The specific surface area of TNAs and Fe/TNAs were measured by BET (ASAP 2020 N (S/N: 1195), Micromeritics, Norcross, GA, USA) analysis under 740 mmHg of pressure (P 0 ) and 77.350 K of bath temperature. Raman analysis was conducted by using 532 nm Laser, with 1800 cm −1 grating and 50 s exposure time, operating by Labspec5 software. The X-ray photoelectron spectroscopy (XPS) experiments were conducted on the TNAs and Fe/TNAs using a PHI 5000 Versa Probe system (Physical Electronics, Chanhassen, MN, USA). The binding energy that was obtained from the XPS spectra was calibrated with reference to the C1s peak at 284.8 eV. The UV-vis absorption spectra were measured in diffused reflection mode using an integrating sphere (ISV-922, Jasco, Tokyo, Japan) that was attached to a Jasco V-750 UV-vis DRS spectrometer (V-750, Jasco, Tokyo, Japan).

PEC and Electrochemical Measurements
All PEC experiments were carried out in three-electrode mode at room temperature. A Fe/TNAs (2 cm 2 ) and a Pt wire (99.997%, Alfa Aesar, Haverhill, MA, USA) served as a working electrode and a counter electrode, respectively. An Ag/AgCl (3 M KCl, ΩMetrohm, Herisau, Switzerland) electrode was selected as a reference electrode. A self-designed Htype reactor separated the anode and cathode to enhance the pollutants degradation. These two chambers were connected with a cation-exchanged membrane (Nafion 212, DuPount, Wilmington, DE, USA) to keep the ion balance in the system. A quartz window (7 cm 2 ) on the side of the anode chamber provided high optical quality. For the photocurrent measurement, the operating light source is a 100 W mercury lamp (GGZ100, Shanghai Jiguang, Shanghai, China) at a bias potential of +1.0 V (vs. Ag/AgCl) in the electrolyte of 0.1 M NaCl (99%, Avantor, Radnor, PA, USA), while a platinum (Pt) and Ag/AgCl electrode work as the counter and reference electrode, respectively. Electrochemical impedance spectroscopy (EIS) was performed under an open circuit voltage with frequencies in the range from 10 kHz to 10 mHz with an AC voltage with an amplitude of 5 mV. The potential for I-V, I-t curve measurements and PEC degradation experiments were controlled by an electricity workstation (ΩMetrohm-Autolab B.V. PGSTAT204, Herisau, Switzerland). For the methyl orange (85%, Sigma Aldrich, St. Louis, Missouri, USA) degradation, the PEC, photocatalytic (PC), electrochemical (EC) and photolysis (P) experiments were examined at the illumination using 100 W Hg lamp. Applied voltage in PEC and EC process was +1.0 V (vs. Ag/AgCl). E. coli is a gram-negative bacteria and used as an indicator of fecal pollution in water. E. coli (CCRC 10674) was purchased from the Bioresource Collection and Research Center of the Food Industry Research and Development Institute (Hsinchu City, Taiwan). E. coli K-12 strain CCRC 10674 was grown aerobically in 100 mL of brain-heart infusion at 37 • C for 18 h and the micro-organisms were cultured at 25 • C. Consequently, 1 mL aqueous samples containing 70 CFU/mL initial E. coli concentration were prepared and subjected to treatment. In this study, E. coli was used to evaluate the Fe/TNAs PEC degradation ability. The operated light source is 100 W Hg lamp at a bias potential of +1.0 V (vs. Ag/AgCl), while a reference electrode of Ag/AgCl, counter electrode of Pt wire, and a working electrode of Fe/TNAs-0.5 were employed.

Surface Morphology of TNAs and Fe/TNAs
In this study, SEM was used to identify the morphology of the surface of the material, and to explore the tubular structure of TNAs and Fe/TNAs prepared by the anodic oxidation etching method. Figure 1 shows the SEM images of TNAs and Fe/TNAs, respectively. Figure 1a shows an array of titanium dioxide nanotubes synthesized by anodizing method. The structure is compact and neatly arranged, and the diameter of the tube were approximately in the range of 50-100 nm. Many previous studies have reported a linear correlation between the diameter of titanate nanotube arrays and the applied anode voltage (Marck et al., 2017;Yasuda et al., 2007). Figure 1b shows the cross-sectional view of the un-modified titanium dioxide nanotube arrays with a length about 1.3 µm. Cao et al. (2016) applied anodization method to synthesize titanium dioxide nanotube arrays and found that the length of TNAs increases with the time of anodization [35]. Figure 1c,d shows the iron modified titanate nanotube arrays synthesized (Fe/TNAs) via SWVE method with 0.2 and 0.5 M Fe(NO 3 ) 3 ·9H 2 O precursor, respectively. The structure of the nanotube is not damaged after ferric nitrate modification, and the iron is successfully deposited on the surface of the TNAs in the form of sol particles. The morphology of iron is irregular, and it is evenly distributed and accumulated around the surface of TNAs. Notably, as the concentration of precursor of iron nitrate increases, the iron deposits increased on the TNAs. Previous literature showed a similar deposition morphology, confirming that sodium borohydride can successfully reduce Fe 3+ in ferric nitrate to zero-valent iron Fe 0 [36]. Table 1 shows the results of EDX analyses of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. The measured elements include C, O, Ti, and Fe, individually. The atomic percentage of Fe were 0.00, 0.93, and 7.44% for TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively, illustrating that the Fe were successfully deposited on the surface of TNAs. Notably, as shown in Figure 1d, most surface of the TNAs were covered by Fe, therefore, the Ti content is much less in Fe-TNAs-0.5 (16.51%) than that of TNAs (34.15%) and Fe/TNAs-0.2 (33.84%). Figure S1 shows the BET nitrogen adsorption-desorption isotherms of TNAs and Fe/TNAs. The measured surface area of TNAs and Fe/TNAs were 0.1464 and 0.0029 m 2 /g, respectively.
100 W Hg lamp at a bias potential of +1.0 V (vs. Ag/AgCl), while a reference electrode of Ag/AgCl, counter electrode of Pt wire, and a working electrode of Fe/TNAs-0.5 were employed.

Surface Morphology of TNAs and Fe/TNAs
In this study, SEM was used to identify the morphology of the surface of the material, and to explore the tubular structure of TNAs and Fe/TNAs prepared by the anodic oxidation etching method. Figure 1 shows the SEM images of TNAs and Fe/TNAs, respectively. Figure 1a shows an array of titanium dioxide nanotubes synthesized by anodizing method. The structure is compact and neatly arranged, and the diameter of the tube were approximately in the range of 50-100 nm. Many previous studies have reported a linear correlation between the diameter of titanate nanotube arrays and the applied anode voltage (Marck et al. 2017;Yasuda et al. 2007). Figure 1b shows the cross-sectional view of the un-modified titanium dioxide nanotube arrays with a length about 1.3 μm. Cao et al. (2016) applied anodization method to synthesize titanium dioxide nanotube arrays and found that the length of TNAs increases with the time of anodization [35]. Figure 1c,d shows the iron modified titanate nanotube arrays synthesized (Fe/TNAs) via SWVE method with 0.2 and 0.5 M Fe(NO3)3·9H2O precursor, respectively. The structure of the nanotube is not damaged after ferric nitrate modification, and the iron is successfully deposited on the surface of the TNAs in the form of sol particles. The morphology of iron is irregular, and it is evenly distributed and accumulated around the surface of TNAs. Notably, as the concentration of precursor of iron nitrate increases, the iron deposits increased on the TNAs. Previous literature showed a similar deposition morphology, confirming that sodium borohydride can successfully reduce Fe 3+ in ferric nitrate to zero-valent iron Fe 0 [36]. Table 1 shows the results of EDX analyses of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. The measured elements include C, O, Ti, and Fe, individually. The atomic percentage of Fe were 0.00, 0.93, and 7.44% for TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively, illustrating that the Fe were successfully deposited on the surface of TNAs. Notably, as shown in Figure 1d, most surface of the TNAs were covered by Fe, therefore, the Ti content is much less in Fe-TNAs-0.5 (16.51%) than that of TNAs (34.15%) and Fe/TNAs-0.2 (33.84%). Figure S1 shows the BET nitrogen adsorption-desorption isotherms of TNAs and Fe/TNAs. The measured surface area of TNAs and Fe/TNAs were 0.1464 and 0.0029 m 2 /g, respectively.   showing that the successfully deposited iron [37]. Moreover, with the increase amount of Fe (see Table 1), the diffraction peak intensities corresponding to Fe increases as well. Figure S2 shows the Raman analyses for bare TNAs and Fe/TNAs. In general, anatase phase of titanium dioxide has six Raman active modes 1A 1g , 2B 1g , and 3E g . TNAs and Fe/TNAs show solo peak at 143.1 and 144.5 cm −1 , respectively, that corresponding to the active mode of E g of anatase phase of TiO 2 . The results in Ramana analyses are in good agreement with the XRD observations. Notably, the main E g mode located at around 144 cm −1 broadened and shifted toward higher wavenumber in Fe/TNA (inset plot in Figure S1), indicating that the Fe 3+ are successfully incorporated into the TiO 2 framework, replacing Ti 4+ cations. This results in in consistent to the findings of Fe-TiO 2 NPs [38].   Figure 2 shows the XRD results for TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5. Compared with the standard chart of the Joint Committee for Powder Diffraction Files (JCPDFs: 21-1272), the XRD results show that TNAs exhibits anatase characteristic peaks at 2θ = 25.3°, 37.8° and 54.0°, corresponding to the crystal planes of (101), (004), and (105), respectively. No rutile phase was observed due to the insufficient temperature [5,15]. Notably, Fe/TNAs-0.2 and Fe/TNAs-0.5 show the diffraction peaks of Fe at 2θ = 24.16°, 35.74°, 54.23°, 62.26°, corresponding to the crystal planes of (012), (110), (116), and (214) (JCPDS file number: 00-001-1053), respectively, indicating that the Fe nanoparticles have been deposited on the surface of TNAs. Similar results of XRD analyses for TiO2-Fe2O3 were found in literature, showing that the successfully deposited iron [37]. Moreover, with the increase amount of Fe (see Table 1), the diffraction peak intensities corresponding to Fe increases as well. Figure S2 shows the Raman analyses for bare TNAs and Fe/TNAs. In general, anatase phase of titanium dioxide has six Raman active modes 1A1g, 2B1g, and 3Eg. TNAs and Fe/TNAs show solo peak at 143.1 and 144.5 cm −1 , respectively, that corresponding to the active mode of Eg of anatase phase of TiO2. The results in Ramana analyses are in good agreement with the XRD observations. Notably, the main Eg mode located at around 144 cm −1 broadened and shifted toward higher wavenumber in Fe/TNA (inset plot in Figure S1), indicating that the Fe 3+ are successfully incorporated into the TiO2 framework, replacing Ti 4+ cations. This results in in consistent to the findings of Fe-TiO2 NPs [38]. 20 30 40 50 60 70 80

XPS Analyses
The electron energy spectrometer (XPS), as shown in Figure 3a-e, are used to clarify the composition of elements of Fe/TNAs and to analyze the energy spectrum. It is also used to identify the uniformity and dispersion of surface elements of materials. Figure 3a is the full-energy spectra of Fe/TNAs-0.5, showing that C, O, Ti, and Fe are distributed on the surface of the material. Figure 3b shows the characteristic peak of C 1s at 284.5 eV. The C element was attributed to the addition to the apparatus during the analysis. Figure 3c shows the characteristic peaks of Fe at 710 eV and 725 eV, corresponding to the element states of 2p 3/2 and 2p 1/2 , respectively. Notably, the enhancement in binding energy in the elemental state of 2p 3/2 is due to the fact that Fe 3+ disperses to Ti 4+ in the anatase lattice and forms Fe−O−Ti bonds [39].  [42]. Hence, we confirmed the state of Fe in Fe/TNAs is Fe 3+ instead of Fe 0 as the findings in Figure 3c. Figure 3d shows that O has a characteristic peak at 530 eV, which represents the formation of Ti or Fe oxides [43]. Figure 3e shows that the characteristic peaks of Ti at 458 eV and 464 eV correspond to Ti 2p 3/2 and Ti 2p 1/2 respectively, confirming that Ti exists in the form of Ti 4+ [44]. Table 2 shows the element percentage of Fe/TNAs-0.2 and Fe/TNAs-0.5 by XPS analysis. The Fe content for Fe/TNAs-0.2 and Fe/TNAs-0.5 is 0.9% and 2.9%, respectively, that are less than the results of EDX analyses. This is because the distribution of iron particles on the surface of the titanium sheet is uneven, and the particle size is different. The comparison of element components showed a slight difference between EDX and XPS because of the following reasons: (a) the range of analytical depth for XPS and EDX is within 10 and 100-1000 nm, respectively; (b) the analytical area of samples was larger in XPS than that in EDX; (c) the appearance of the C element can be attributed to extrinsic carbon during the sample fabrication and/or the XPS instrument itself.

XPS Analyses
The electron energy spectrometer (XPS), as shown in Figure 3a-e, are used to clarify the composition of elements of Fe/TNAs and to analyze the energy spectrum. It is also used to identify the uniformity and dispersion of surface elements of materials. Figure 3a is the full-energy spectra of Fe/TNAs-0.5, showing that C, O, Ti, and Fe are distributed on the surface of the material. Figure 3b shows the characteristic peak of C 1s at 284.5 eV. The C element was attributed to the addition to the apparatus during the analysis. Figure 3c shows the characteristic peaks of Fe at 710 eV and 725 eV, corresponding to the element states of 2p3/2 and 2p1/2, respectively. Notably, the enhancement in binding energy in the elemental state of 2p3/2 is due to the fact that Fe 3+ disperses to Ti 4+ in the anatase lattice and forms FeOTi bonds [39].  [42]. Hence, we confirmed the state of Fe in Fe/TNAs is Fe 3+ instead of Fe 0 as the findings in Figure 3c. Figure 3d shows that O has a characteristic peak at 530 eV, which represents the formation of Ti or Fe oxides [43]. Figure 3e shows that the characteristic peaks of Ti at 458 eV and 464 eV correspond to Ti 2p3/2 and Ti 2p1/2 respectively, confirming that Ti exists in the form of Ti 4+ [44]. Table 2 shows the element percentage of Fe/TNAs-0.2 and Fe/TNAs-0.5 by XPS analysis. The Fe content for Fe/TNAs-0.2 and Fe/TNAs-0.5 is 0.9% and 2.9%, respectively, that are less than the results of EDX analyses. This is because the distribution of iron particles on the surface of the titanium sheet is uneven, and the particle size is different. The comparison of element components showed a slight difference between EDX and XPS because of the following reasons: (a) the range of analytical depth for XPS and EDX is within 10 and 100-1000 nm, respectively; (b) the analytical area of samples was larger in XPS than that in EDX; (c) the appearance of the C element can be attributed to extrinsic carbon during the sample fabrication and/or the XPS instrument itself.

UV-Vis DRS Analyses
The absorption coefficient of the thin films was calculated with the following formula: where α is the absorption coefficient, d is the thickness of the film, and for a particular wavelength T is transmission and R is the reflectance characteristics of the film.
The optical band gap is calculated by plotting (αhv) 1/n versus hv (i.e., eV), where α is the absorption coefficient of the material, h is the Planck constant, ν is the photon's frequency, A is a constant, and EG is the bandgap energy. As to the constant n is 0.5 and 1 for indirect and direct band gap materials, respectively. TiO2 has an indirect band gap [38]. As shown in Figure 4, the absorbance wavelength is 420, 530, 730 nm for TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5, respectively. The absorbance wavelength of TNAs red shift to longer wavelength region after Fe nanoparticles modification illustrating the better light utilization. Band gap (EG) can be calculated according to the Tauc formula [45] as shown Equation (2) above. The inset plot in Figure 4 shows the corresponding bandgap is 2.9, 2.3, and 1.7 eV for TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5, individually. The results here are similar to the findings of Fe-TiO2 thin film [46] and Fe-TiO2 nanoparticles [38,47] that applying Tauc formula to investigate the bandgap energy. In general, there is a good correlation

UV-Vis DRS Analyses
The absorption coefficient of the thin films was calculated with the following formula: where α is the absorption coefficient, d is the thickness of the film, and for a particular wavelength T is transmission and R is the reflectance characteristics of the film.
The optical band gap is calculated by plotting (αhv) 1/n versus hv (i.e., eV), where α is the absorption coefficient of the material, h is the Planck constant, ν is the photon's frequency, A is a constant, and E G is the bandgap energy. As to the constant n is 0.5 and 1 for indirect and direct band gap materials, respectively. TiO 2 has an indirect band gap [38]. As shown in Figure 4, the absorbance wavelength is 420, 530, 730 nm for TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5, respectively. The absorbance wavelength of TNAs red shift to longer wavelength region after Fe nanoparticles modification illustrating the better light utilization. Band gap (E G ) can be calculated according to the Tauc formula [45] as shown Equation (2) above. The inset plot in Figure 4 shows the corresponding bandgap is 2.9, 2.3, and 1.7 eV for TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5, individually. The results here are similar to the findings of Fe-TiO 2 thin film [46] and Fe-TiO 2 nanoparticles [38,47] that applying Tauc formula to investigate the bandgap energy. In general, there is a good Nanomaterials 2021, 11, 1944 8 of 17 correlation between light absorption and photocatalytic activity. Stronger light absorption correlated with higher photocatalytic activity. UV-vis DRS spectra results confirmed that Fe nanoparticles were successfully deposited on TNAs. between light absorption and photocatalytic activity. Stronger light absorption correlated with higher photocatalytic activity. UV-vis DRS spectra results confirmed that Fe nanoparticles were successfully deposited on TNAs.  Figure 5 shows the measured photocurrent diagram by using TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5 as the photo-anode in a PEC system. As shown in the figure, the photocurrent was monitored under a 50 s on-off cycle. The photocurrent of three tested photoanodes is about 1-4 μA when the illumination is off. Because there is no light to excite the photoelectrons on the surface of the material, leading to no photocurrent is generated. When the light is on, all materials generate photocurrent as a stable straight line, which means that the photoelectrochemical performance of the photo-anodes tend to a stable state. The photocurrent densities of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 are 2.0, 3.0, and 3.5 mA/cm 2 , respectively. Notably, the photocurrent was 1.75 times higher for Fe/TNAs-0.5 than that of pure TNAs at the same illumination. TNAs modified by iron can excite more photo-generated electrons than TNAs under irradiation, which promotes the increase of photocurrent, indicating that more electron-hole pairs are generated on the surface of the material, resulting to better degrade pollutants.  Figure 5 shows the measured photocurrent diagram by using TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5 as the photo-anode in a PEC system. As shown in the figure, the photocurrent was monitored under a 50 s on-off cycle. The photocurrent of three tested photo-anodes is about 1-4 µA when the illumination is off. Because there is no light to excite the photoelectrons on the surface of the material, leading to no photocurrent is generated. When the light is on, all materials generate photocurrent as a stable straight line, which means that the photoelectrochemical performance of the photo-anodes tend to a stable state. The photocurrent densities of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 are 2.0, 3.0, and 3.5 mA/cm 2 , respectively. Notably, the photocurrent was 1.75 times higher for Fe/TNAs-0.5 than that of pure TNAs at the same illumination. TNAs modified by iron can excite more photo-generated electrons than TNAs under irradiation, which promotes the increase of photocurrent, indicating that more electron-hole pairs are generated on the surface of the material, resulting to better degrade pollutants. Nanomaterials 2021, 11, x FOR PEER REVIEW 9 of 17

Electrochemical Impedance Spectrum
In this study, electrochemical impedance spectrum (EIS) was used to analyze the electron-hole separation ability and electron transfer characteristics of TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5 under light irradiation. The smaller the diameter of the arc refers to the better the separation of electrons and holes, resulting in better photoelectrochemical ability. Figure 6a shows the Nyquist plots of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 by EIS analysis, indicating that the diameter of each semicircle follows the order: TNAs > Fe/TNAs-0.2 > Fe/TNAs-0.5, which demonstrating that the electron mobility of TNAs was enhanced by Fe modification. Figure 6b shows the Bode plot obtained by EIS analysis for three photo-anodes. The maximum frequency obtained from the figure is then fitted with the equation of = 1 2π× to calculate the electron lifetime.  Figure 6a, for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. Table 3 summarizes the result of fitted equivalent circuit for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. Notably, Rs refers to the resistance of the solution and Rp represents the charge transfer at the interface between Fe/TNAs and electrolyte, while the CPE refers to the constant phase element [5]. The Rp was 290.5, 143.33, and 130.97 for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, individually, with +1.0 V (vs. Ag/AgCl) of bias potential. According to the results of Figure 6 and Table 3, we can preliminarily conclude that TNAs modified with iron can effectively reduce the recombination of electron holes, prolong the residence time of electrons or increase the mobility of electrons in the PEC system.

Electrochemical Impedance Spectrum
In this study, electrochemical impedance spectrum (EIS) was used to analyze the electron-hole separation ability and electron transfer characteristics of TNAs, Fe/TNAs-0.2 and Fe/TNAs-0.5 under light irradiation. The smaller the diameter of the arc refers to the better the separation of electrons and holes, resulting in better photoelectrochemical ability. Figure 6a shows the Nyquist plots of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 by EIS analysis, indicating that the diameter of each semicircle follows the order: TNAs > Fe/TNAs-0.2 > Fe/TNAs-0.5, which demonstrating that the electron mobility of TNAs was enhanced by Fe modification. Figure 6b shows the Bode plot obtained by EIS analysis for three photo-anodes. The maximum frequency obtained from the figure is then fitted with the equation of τ el = 1 2π× f max to calculate the electron lifetime. Table 3 shows the electron lifetime was 290.3, 354.7, and 433.3 ms, corresponding to the f max of 0.548, 0.449, and 0.367 Hz in Figure 6a, for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. Table 3 summarizes the result of fitted equivalent circuit for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. Notably, R s refers to the resistance of the solution and R p represents the charge transfer at the interface between Fe/TNAs and electrolyte, while the CPE refers to the constant phase element [5]. The R p was 290.5, 143.33, and 130.97 for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, individually, with +1.0 V (vs. Ag/AgCl) of bias potential. According to the results of Figure 6 and Table 3, we can preliminarily conclude that TNAs modified with iron can effectively reduce the recombination of electron holes, prolong the residence time of electrons or increase the mobility of electrons in the PEC system. Nanomaterials 2021, 11, x FOR PEER REVIEW 10 of 17

Mott-Schottky Plot for TNAs and Fe/TNAs
The Mott-Schottky plot is performed by applying the electrochemical impedance method, which is carried out in the range of −0.6 V to +0.2 V (vs. Ag/AgCl) under frequency of 100 Hz and amplitude of 10 mV. The type and flat band potential of semiconductor were determined by following equation [40,48,49]: where ε is the dielectric constant of TiO 2 (i.e., 31 for anatase) [50,51], ε 0 is the permittivity of vacuum (8.85 × 10 −14 F cm −1 ), A is the area of photo-anode (4 cm 2 ), q is the elementary charge (1.6 × 10 −19 C), N D is the density of dopants (cm −3 ), V is the applied voltage, V fb is the flat band potential, K is the Boltzman constant (1.381 × 10 −23 J K −1 ), and T is the absolute temperature (298 K). In addition, the charge carrier density (N D ) of semiconductor was also calculated from the slope of Mott-Schottky plots, which is following equation [48,52,53]: As shown in Figure 7, both TNAs and Fe/TNAs show the positive slope, confirming that Fe depositing on TNAs does not change its n-type semiconductor property. Moreover, the V fb of TNAs and Fe/TNAs was −0.34 and −0.25 V (vs. Ag/AgCl), individually, which were estimated from the intercept with the X-axis on the linear plot. The positive shift of V fb in Fe/TNAs photo-anode implies the facilitating of photo-generated electron-hole pairs separation and transfer, which is corresponding to EIS results ( Figure 6) [48]. The charge carrier density (N D ) of TNAs (9.04 × 10 21 ) was higher than that of Fe/TNAs (2.52 × 10 21 cm −3 ). This was attributed to the slope of Fe/TNAs photo-anode (1.13 × 10 8 ) which was higher than that of TNAs photo-anode (3.15 × 10 7 ), indicating the significant decrease of charge carrier density of Fe/TNAs. Similar observations were found in Freitas et al. (2014), indicating that the increasing of charge carrier density was attributed to a high level of defects caused by oxygen vacancies [54].
In addition, when photo-anode is contacted with the electrolyte, the electrons on photoanode will transfer to electrolyte spontaneously to form depletion region with positive charge between the interface of photo-anode and electrolyte. The width of depletion region (W) was also derived from the Mott-Schottky plot relationship and is described by following equation [48,54]: The depletion region of TNAs and Fe/TNAs was 0.311 and 0.661 nm, individually. As the result, the oxygen vacancies were decreased with the substitution of Ti 4+ by Fe 3+ at the surface, enlarging the space charge region at TiO 2 -electrolyte interface [48]. Furthermore, the conduction band (E CB ) is determined from the following equation for n-type semiconductor: where K is the Bolzman constant (8.617 × 10 −5 eV K −1 ) and N c is the effective density of states (DOS) at the conduction band edge, which can be calculated by using The m eff of 9 m 0 (for TiO 2 anatase) is used for N c calculations [55], where m 0 is the mass of electron (9.109 × 10 −31 kg), K is the Bolzman constant (1.381 × 10 −23 J K −1 ), T is the absolute temperature (298 K), and h is the Planck constant (6.626 × 10 −34 J s). By applying the equation, N c is determined as 6.71 × 10 20 cm −3 . Therefore, the conduction band −0.27 and −0.22 V for TNAs and Fe/TNAs, respectively. Furthermore, UV-vis DRS results showed that the band gap of TNAs and Fe/TNAs was 2.9 and 1.7 eV, respectively, thus, the position of the valence band was at approximately 2.63 and 1.48 V for TNAs and Fe/TNAs, respectively.
showed that the band gap of TNAs and Fe/TNAs was 2.9 and 1.7 eV, respectively, thus, the position of the valence band was at approximately 2.63 and 1.48 V for TNAs and Fe/TNAs, respectively.

Methyl Orange Removal
Four methyl orange removal processes, namely, PEC, PC, EC and P processes were conducted to evaluate the degradation efficiency of TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5. As shown in Figure 8, PEC process was the most efficient way to degrade MO among three methods studied. The complete removal of 10 ppm of MO was observed after 120, 50 and 30 min for the PEC method by using TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. The order of MO degradation by TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5 is in consistent to the photocurrent densities of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 that are 2.0, 3.0, and 3.5 mA/cm 2 , respectively. According to the EIS analyses, the electron lifetime was 290.3, 354.7, and 433.3 ms for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively, illustrating that the freedom of electrons has increased. This enhanced PEC performance was attributed to the Fe modified as evidence of results of EDX and XPS.
Briefly, the MO degradation efficiency in four methods follows: PEC > PC > P > EC. It was observed that the PEC method provided the most powerful way to degrade MO due to the combination of electrochemical oxidation and photocatalysis [5].

Methyl Orange Removal
Four methyl orange removal processes, namely, PEC, PC, EC and P processes were conducted to evaluate the degradation efficiency of TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5. As shown in Figure 8, PEC process was the most efficient way to degrade MO among three methods studied. The complete removal of 10 ppm of MO was observed after 120, 50 and 30 min for the PEC method by using TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively. The order of MO degradation by TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5 is in consistent to the photocurrent densities of TNAs, Fe/TNAs-0.2, and Fe/TNAs-0.5 that are 2.0, 3.0, and 3.5 mA/cm 2 , respectively. According to the EIS analyses, the electron lifetime was 290.3, 354.7, and 433.3 ms for TNA, Fe/TNAs-0.2, and Fe/TNAs-0.5, respectively, illustrating that the freedom of electrons has increased. This enhanced PEC performance was attributed to the Fe modified as evidence of results of EDX and XPS.
Briefly, the MO degradation efficiency in four methods follows: PEC > PC > P > EC. It was observed that the PEC method provided the most powerful way to degrade MO due to the combination of electrochemical oxidation and photocatalysis [5]. Nanomaterials 2021, 11, x FOR PEER REVIEW 13 of 17 3.9. E. coli Removal by Fe/TNAs PEC System E. coli was selected as the bio-target for investigating the PEC disinfection ability. Figure 8 shows the E. coli removal efficiency was 87.02% by using Fe/TNAs-0.5 as the photo-anode in the PEC system within the reaction time of 60 min. E. coli degradation by photolysis, the same light source (100 W Hg lamp) in the PEC, was conducted for comparison. The images of E. coli removal by TNAs and Fe/TNAs PEC system were demonstrated in Figure S3. As shown in Figure 9, only 33.1% of E. coli was removed in the photolysis procedure. It is well known that UV light is commonly used in traditional water treatment system. According to the literature, inactivation of EHEC O26 with photoreactivation is 2.2 times greater than that without photoreactivation [3]. In addition, with the addition of TiO2 in UV disinfection system, the damage of outer membrane was observed for the cells. TiO2 alone can break down lipopolysaccharide (LPS), the outermost layer of the E. coli cells [4]. The hydroxyl radicals, which is a kind of most strong oxidants, were responsible for E. coli removal in the TNAs PEC system [6]. Compared to photolysis at the same irradiation, PEC system can generate strong oxidants, i.e., hydroxyl radicals [5,6], to degrade E. coli. On the other hand, the recombination of electron-hole pairs is greatly reduced in PEC method as evidence of enhanced electron lifetime (see Table 3), that allowed more photo-generated holes to react with H2O to form OH to removal E. coli. The result is in consistent to the findings of Ma et al. (2021) that Z-scheme g-C3N4/TNAs reduced recombination of photo-generated electron-hole pairs [56]. 3.9. E. coli Removal by Fe/TNAs PEC System E. coli was selected as the bio-target for investigating the PEC disinfection ability. Figure 8 shows the E. coli removal efficiency was 87.02% by using Fe/TNAs-0.5 as the photo-anode in the PEC system within the reaction time of 60 min. E. coli degradation by photolysis, the same light source (100 W Hg lamp) in the PEC, was conducted for comparison. The images of E. coli removal by TNAs and Fe/TNAs PEC system were demonstrated in Figure S3. As shown in Figure 9, only 33.1% of E. coli was removed in the photolysis procedure. It is well known that UV light is commonly used in traditional water treatment system. According to the literature, inactivation of EHEC O26 with photoreactivation is 2.2 times greater than that without photoreactivation [3]. In addition, with the addition of TiO 2 in UV disinfection system, the damage of outer membrane was observed for the cells. TiO 2 alone can break down lipopolysaccharide (LPS), the outermost layer of the E. coli cells [4]. The hydroxyl radicals, which is a kind of most strong oxidants, were responsible for E. coli removal in the TNAs PEC system [6]. Compared to photolysis at the same irradiation, PEC system can generate strong oxidants, i.e., hydroxyl radicals [5,6], to degrade E. coli. On the other hand, the recombination of electron-hole pairs is greatly reduced in PEC method as evidence of enhanced electron lifetime (see Table 3), that allowed more photo-generated holes to react with H 2 O to form OH· to removal E. coli. The result is in consistent to the findings of Ma et al. (2021) that Z-scheme g-C 3 N 4 /TNAs reduced recombination of photo-generated electron-hole pairs [56]. Nanomaterials 2021, 11, x FOR PEER REVIEW 14 of 17

Conclusions
Fe/TNAs were fabricated by the SWVE method to remove the E. coli for the very first time in this study. The SEM results showed that iron NPs were deposited on the surface of the TNAs in the form of sol particles and the tubular structure was not damaged after iron modification. The Fe content on the surface of TNAs was increased with the increasing of the precursor concentration, which demonstrated by EDX and XPS results. XPS also revealed that the binding energy of Fe located at 710 and 725 eV, which represent +3 state of Fe. XRD analyses showed the characteristic peak of Fe at 2θ = 24.16°, 35.74°, 54.23°, and 62.26°, corresponding to the crystal planes of (012), (110), (116), and (214), illustrating the successful loading of iron NPs. Moreover, UV-vis DRS showed that the light absorbance was increased and the bandgap energy was decreased after iron modification to enhance the light utilization. The results of photoelectrochemical performance indicated that Fe/TNAs-0.5 performed the best photocurrent density (3.5 mA/cm 2 ) as well as the electron lifetime of Fe/TNAs-0.5 is 1.49 times greater than that of pure TNAs through the EIS and Bode plot analyses, illustrating that the loading of Fe NPs is advantage to the separation of electron-hole pairs to enhance the photoelectrochemical performance. The deep and shallow trap states in bandgap of TNAs play an essential role in the electron mobility and PEC performances. Therefore, investigation of the effect of the traps of the TNAs is recommended for future study. Compared to the P, PC, and EC methods, PEC was the best method for MO degradation by using Fe/TNAs-0.5 as the photo-anode. Furthermore, the Fe/TNAs-0.5 PEC system also performed better removal efficiency than the P method for E. coli degradation. To sum up, the iron NPs deposited TNAs were successfully synthesized in this study. Depositing Fe NPs on TNAs not only reduces the bandgap energy of the TNAs but also decreases the electron-hole recombination, therefore, leading to the enhancement of the PEC performances in MO and E. coli removal.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: BET analysis of (a) TNAs and (b) Fe/TNAs, Figure S2: Raman spectrum of TNAs and Fe/TNAs, Figure S3: Images of E. coli removal in photolysis and PEC system.

Conclusions
Fe/TNAs were fabricated by the SWVE method to remove the E. coli for the very first time in this study. The SEM results showed that iron NPs were deposited on the surface of the TNAs in the form of sol particles and the tubular structure was not damaged after iron modification. The Fe content on the surface of TNAs was increased with the increasing of the precursor concentration, which demonstrated by EDX and XPS results. XPS also revealed that the binding energy of Fe located at 710 and 725 eV, which represent +3 state of Fe. XRD analyses showed the characteristic peak of Fe at 2θ = 24.16 • , 35.74 • , 54.23 • , and 62.26 • , corresponding to the crystal planes of (012), (110), (116), and (214), illustrating the successful loading of iron NPs. Moreover, UV-vis DRS showed that the light absorbance was increased and the bandgap energy was decreased after iron modification to enhance the light utilization. The results of photoelectrochemical performance indicated that Fe/TNAs-0.5 performed the best photocurrent density (3.5 mA/cm 2 ) as well as the electron lifetime of Fe/TNAs-0.5 is 1.49 times greater than that of pure TNAs through the EIS and Bode plot analyses, illustrating that the loading of Fe NPs is advantage to the separation of electron-hole pairs to enhance the photoelectrochemical performance. The deep and shallow trap states in bandgap of TNAs play an essential role in the electron mobility and PEC performances. Therefore, investigation of the effect of the traps of the TNAs is recommended for future study. Compared to the P, PC, and EC methods, PEC was the best method for MO degradation by using Fe/TNAs-0.5 as the photo-anode. Furthermore, the Fe/TNAs-0.5 PEC system also performed better removal efficiency than the P method for E. coli degradation. To sum up, the iron NPs deposited TNAs were successfully synthesized in this study. Depositing Fe NPs on TNAs not only reduces the bandgap energy of the TNAs but also decreases the electron-hole recombination, therefore, leading to the enhancement of the PEC performances in MO and E. coli removal.