Preparation of Gd-Doped TiO2 Nanotube Arrays by Anodization Method and Its Photocatalytic Activity for Methyl Orange Degradation

Gd-doped TiO2 nanotube arrays with 3D ordered and high specific surface (176 m2/g) area are successfully prepared on a Ti foil surface via an anodizing method. The characterizations of Gd-doped TiO2 nanotube arrays are carried out using X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscope (TEM), X-ray photoelectron spectroscopy (XPS), energy dispersive X-ray spectrometer (EDX), optical contact angle measurer, and ultraviolet (UV) fluorescence spectrophotometer, respectively. The results reveal that the Gd-doped TiO2 nanotube arrays form on the Ti foil surface with a length of about 50.5 μm and 100 nm in diameter, and the growth direction grown along the (101) direction of anatase crystal. Under the optimized reaction conditions (60 V, 10 h, 0.01mol/L Gd3+), the crystal phase of Gd-doped TiO2 show an excellent nanotube structure, possess a beneficial photocatalytic performance for methyl orange (MO) (Degradation rate was 95.8%) and an excellent hydrophilic property (The optical contact angle was 4.9°). After doping with Gd, the hydrophilic and photocatalytic properties of the TiO2 nanotubes are further improved.


Introduction
Recently, as an emerging green advanced oxidation technology [1], the photocatalytic oxidation technique is widely used in the treatment of organic wastewater due to its mild reaction condition, simple operation, high efficiency, low energy consumption, and environmentally friendly characteristics [2,3]. Among these photocatalytic materials, TiO2 is an inorganic photosensitive semiconductor material which has the advantages of high photocatalytic activity, good stability, nontoxicity, and low cost. Therefore, TiO2 has been applied to prepare many semiconductor photocatalysts and attracts widespread attention [4][5][6][7][8]. Nano-TiO2 has been prepared into a layered structure, like nanoparticles [9], nanobelts [10], nanowires [11], nanorods [12], nanotubes [13], and nano-flowers [14]. Compared with nanospheres or two-dimensional nanofilms, one-dimensional TiO2 nanotubes have a larger specific surface area [15] and a stronger adsorption capacity [16]. TiO2 nanotube arrays have highly ordered and size-controllable properties [17]. The large surface energy enables them to enhance the interface separation effect of electron-hole pairs and the directional transport efficiency of carriers [18][19][20].
Nowadays, the preparation methods of nano-TiO2 materials mainly include a template synthesis method [21], a hydrothermal synthesis method [22], and an anodizing method [23][24][25][26][27]. Anodizing refers to the process of forming an oxide film on the surface of the anode material as a consequence of electric current reaction in the electrolyte solution. TiO2 nanotube arrays prepared by this anodizing method have a neat arrangement, a large specific surface area, and equivalent aperture. Moreover, a dense TiO2 barrier layer is formed between the nanotube arrays and the metal Ti substrate, the prepared photocatalyst has strong adsorption capacity and has the electrical characteristics of non-linear diodes [28]. Gong [24] prepared TiO2 nanotube arrays with a uniform surface arrangement and high order by anodic oxidation with a hydrofluoric acid solution as the electrolyte. Then, after successive researches on anodic oxidation by Ruan [29] and Macak [30], to enhance the comprehensive properties of the TiO2 nanotube arrays, the secondary anodic oxidation method has been used to produce a more orderly TiO2 nanotube array [31,32].
Unfortunately, as a wide band gap semiconductor material (3.20 eV for anatase and 3.03 eV for rutile [33]), the defects of TiO2 are very obvious. A high recombination rate of photogenerated electron-hole pairs [34], a lower conductively [35] and photoelectric conversion efficiency [36] have become the main obstacles hindering the development and application of TiO2. Therefore, study on the modification of TiO2 nanomaterials has been become a popular direction in the field of photocatalyst research, and the most commonly used method is metal ion doping. Previous results [37,38] show that the surface properties and photocatalytic performances of TiO2 nanomaterials are greatly improved. Additionally, there also is doping-TiO2 research with the use of Ag [39], Pd [40], Pt [41] and Au [42]. As a rare earth element, gadolinium (Gd) has a special 4f electronic structure, which can replace part of Ti in TiO2 allowing TiO2 to expand its spectral response range to sunlight without reducing the photoactivity under ultraviolet light and improve its operating factor of visible light. Moreover, Gd-doped nanomaterials have unique optical and magnetic properties which are easy to recycle and do not cause secondary pollution. Presently, Gd-doped TiO2 nanotubes prepared using an anodization method for photocatalytic degradation of organic wastewater has never been published, to the best of the authors' knowledge.
During this study, the highly ordered and neatly arranged Gd-doped TiO2 nanotube arrays are successfully prepared on a Ti foil substrate by an anodizing method. The originality of this paper has been investigated as follows: (1) TiO2 nanotube arrays are synthesized directly on the Ti foil surface using an anodization method with a beneficial photocatalytic performance and hydrophilic property; (2) The influence of different reaction times and voltages on the growth of TiO2 nanotube arrays are illustrated; (3) The Gd-doped TiO2 nanotube arrays are prepared, and the photocatalytic performance and hydrophilic property are further carried out. (4) The TiO2 nanotube arrays are tightly grown on the surface of the Ti foil, and the Gd has strong magnetic properties, making the catalyst easy to recycle in the application is and environmentally friendly.

Formation and Photocatalytic Mechanism of Gd-Doped TiO2 Nanotube Arrays
An anodization method was applied to prepare Gd-doped TiO2 nanotube arrays. The growth of nanotubes was completed under the combination of anodic oxidation and chemical dissolution, and the chemical reaction occurring during the process is conveyed by the following Equations: Ti 4+ + 2H2O → TiO2 + 4H + TiO2 + 6F -+ 4H + ⇌ TiF6 2− + 2H2O The formation mechanism of TiO2 nanotubes is illustrated in Figure 1. The anodic oxidation current-time (I-t) curve can directly explain the growth processes of TiO2 nanotubes. Figure 1a is the I-t curve of an anodized Ti foil with 10 V in 0.5 wt % HF aqueous solution at room temperature (25 °C). According to the above-mentioned equation and I-t curve, the generation of nanotubes can be divided into three stages. During the initial stage of oxidation (Ⅰ), the metal Ti foil is rapidly oxidized and dissolved in an acid electrolyte containing F − to generate a large amount of Ti 4+ , which corresponds to Equation (1). Regarding the second stage (Ⅱ), Ti 4+ reacts with oxonium ions in the electrolyte to form a dense TiO2 oxide film with the current rapidly decreasing, which corresponds to Equation (2). Meanwhile, as shown in Figure 1b, the TiO2 oxide film interacts with F − in the electrolyte under an electric field, and the TiO2 barrier layer is partially etched to generate micropores, which causes the anode current to rise slowly. Concerning the third stage (Ⅲ), TiO2 nanotubes grow steadily and the anode current is provided by ion migration on both sides of the TiO2 barrier layer, thus forming a relatively stable current curve. The photocatalytic mechanism of TiO2 is demonstrated in the following Equations. When ultraviolet light irradiates the surface of TiO2, electrons in the valence band are stimulated to the conduction band, thereby generating highly active carriers (e − ) and holes (h + ). Dissolved oxygen (O2) on the catalyst surface captures electrons to form a superoxide anion radical (O 2− ), and the holes oxidize H2O and OH − into a highly oxidative hydroxyl radical (OH), which can degrade organic matter. The degradation mechanism of MO by TiO2 in this experiment is how the holes and H2O react to generate a hydroxyl radical and superoxide anion radical, then the azo bond and benzene ring of MO are broken by oxidation and, finally, MO is degraded into CO2, H2O, SO4 2− , NO3 − . planes, respectively. There is a maximum in the peak intensity when the 2θ angle was 25.49° (XRD standard card PDF NO. 78-2486). These results prove that the reaction produces anatase phase TiO2.   (116) crystal planes of anatase TiO2, respectively. The diffraction peak of the titanium substrate weakens as the reaction voltage increases, and the TiO2 diffraction peak of the anatase phase is enhanced. Additionally, we can observe a maximum in the XRD peak intensity at 60 V. When the voltage continuously increases, the generated TiO2 crystal is rapidly dissolved, so the intensity of the corresponding anatase phase TiO2 diffraction peak is slightly reduced. Figure 2d reveals the XRD spectra of TiO2 nanotube samples with different Gd ion concentrations. The diffraction peaks appear at 2θ angles of 25.49°, 37.11°, 37.95°, 48.21°, 54.09°, 55.16°, and 69.02°, corresponding to the anatase phase of TiO2. The presence of XRD diffraction peaks in the crystal structure of the doped phase is not detected in the samples, which proves that the doped Gd 3+ has good dispersity in the crystal lattice of the samples. It can be seen from Figure 2c that the curves and diffraction peak intensities of samples prepared with different Gd 3+ concentrations are generally consistent and each diffraction peak corresponds to the standard anatase TiO2 spectrum chart (standard card PDF NO.78-2486), indicating that doping Gd 3+ with different concentrations has no significant effect on the structure of the sample. Figure 3 shows the SEM images of the samples at different reaction stages. Figure 3a shows the surface morphology of the pure titanium sheet. It can be seen that the surface of the titanium sheet is uneven, but there is no tubular structure. Figure 3b,c are SEM images of the samples after the reaction. It is known that dense TiO2 nanotube arrays have been formed, and the diameter of a single nanotube is about 100 nm. Figure 3d shows a cross-sectional view of the samples where the neatly arranged tubular structure can be clearly seen.  Figure 4 contains SEM images of samples at different reaction times. It is found that different reaction times have little effect on the surface morphology of TiO2 nanotube arrays-nanotubes with 100 nm diameters and neatly arranged surfaces were formed-but different reaction times have a significant effect on the nanotube length of arrays. Shown in Figure 4a, when the reaction time is 6 h, the length of the generated nanotubes is approximately 18.6 μm. When the time increased to 10 h, the tubes grew obviously and reached 50.5 μm. However, when the reaction time continues to increase, the change of tube length is no longer palpable, as seen in Figure 4d,e. This is because the growth of TiO2 nanotubes is accompanied by the increase of H + concentration in the electrolyte; however, the chemical corrosion rate of TiO2 nanotubes is carried out at a faster rate. After comprehensive analysis from the perspective of characterization and economics, the 10 h anodic oxidation reaction time is the best time for TiO2 nanotube arrays.  It can be seen from Figure 6 that the surfaces of the TiO2 nanotube arrays generate at different Gd 3+ concentrations are clear and neat, and the nanotubes are arranged closely and orderly together. The generated nanotubes are approximately 50.5 μm in length and 100 nm in diameter. Gd2O3 is a kind of transparent semiconductor material, thus the existence of doping phase particles cannot be seen intuitively in the SEM diagram. The experimental results indicate that the doped treatment will not destroy the highly ordered tubular structure of the TiO2 nanotube arrays, and the introduction of the rare earth element Gd will not affect the surface morphology and microstructure of the nanotube arrays.    Figure 8 is the XPS spectrum of the TiO2 and Gd-doped TiO2 nanotube arrays. It can be seen that the undoped sample contain three elements of C, Ti, and O, whereas Gd-doped TiO2 samples contain not only Ti, C, and O elements but also certain amounts of the Gd element, indicating that the Gd is successfully doped into the TiO2 samples. Shown in Figure 8b, the characteristic peak of Ti2p3/2 in the undoped samples appears at 458.7 eV, while Gd-doped TiO2 appears at 458.4 eV, and the characteristic peak of Ti2p1/2 in the undoped sample appears at 464.5 eV while Gd-doped TiO2 is 464.2 eV. The distance between the two characteristic peaks of Ti2p3/2 is about 5.8 eV and the peak content of Ti2p3/2 is much higher than Ti2p1/2, indicating that Ti in Gd-doped TiO2 mainly exists in the form of Ti 4+ [45]. The Ti peak of the Gd-doped sample shifts toward the low binding energy compared to the undoped sample. The reason might be that Gd 3+ replaces part of the Ti 4+ doped into the oxide lattice and combines with oxygen, causing an imbalance of the electric charge. The XPS spectrum of O1s is shown in Figure 8c. It can be seen that the characteristic peaks of O1s are wide and asymmetric, indicating that the oxygen is in more than one chemical state. The characteristic peaks of O1s in Gd-doped samples appear at 529.7 eV (Ti-O) and 531.4 eV (O-H). Figure 8d is the XPS spectrum of Gd 4d. The characteristic peaks of Gd-doped samples appear at 142.1eV and 148.1eV, indicating that the Gd element will exist as Gd-O in TiO2, which would increase the carrier lifetime and improve the ability to transfer photogenerated electrons.

EDX
An EDX spectrum of the sample before and after the rare earth element Gd doping is shown in Figure 9. Undoped samples contain four elements of C, O, Cu, and Ti while, after doping, samples contain five elements: C, O, Cu, Ti, and Gd. Among them, C and Cu come from carbon film and copper mesh, and O and Ti form TiO2. The inclusion of Gd in the doped sample indicates that the Gd is successfully doped into the TiO2 lattice structure, which further confirms the successful preparation of the Gd-doped TiO2 nanotube arrays. The specific area and porosity analyzer are used to calculate the specific surface area of the prepared Gd-doped TiO2 nanotube arrays. Shown in Figure 10, Gd-doped TiO2 nanotube arrays with a large BET area and more reactive sites are successfully prepared by anodization under the conditions of 60 V voltage, 10 h and 0.01 mol/L Gd 3+ concentration. Figure 10 shows that the specific surface area of the prepared TiO2 nanotube arrays is 176 m 2 /g, while the TiO2 (P-25) is 56 m 2 /g. Large specific surface area is considered to be an important factor that could improve the photocatalytic performance of the catalyst.  Figure 11 demonstrates that the different reaction conditions have significant effects on the degradation rate of the MO solution. Figure 11a shows that the photocatalytic performance of the prepared TiO2 improves rapidly with the increase of the reaction time, and the degradation rate of the MO is increased from 54.8% to 90.8%. The short reaction time makes the lower crystalline phase content of TiO2 and the shorter length of the TiO2 nanotubes, thus the photocatalytic property of the sample is poor under such a condition. When the reaction time increases, the TiO2 crystal phase content and the specific surface area of the nanotubes is improved, and the sample's degradation rate of the MO become better. Considering the crystalline phase structure, surface morphology and economy, the optimal reaction time is determined to be 10 h. The effect of reaction voltages on photocatalytic performance is roughly consistent with reaction time (Figure 11b). Accompanying the increase in the reaction voltage, the driving force of the anodization reaction is enhanced, and the degradation rate of the MO also is increased obviously. Taking all factors into consideration, 60 V is the best condition for the TiO2 nanotube array to obtain the best photocatalytic performance.  It can be seen that the photocatalytic performance of TiO2 nanotubes is further enhanced by the doping of Gd 3+ . Compared with the non-doped 79.8%, continuously increasing the concentration of Gd 3+ to 0.01 mol/L, the degradation rate of the MO reaches a maximum of 95.8%. It is not easy to further improve the degradation rate on the basis of the original high degradation rate. The enhancement of the photocatalytic performance is due to the doping of the Gd, which helps to generate strong oxidizing hydroxyl groups on the surface of the TiO2 photocatalyst, and reduces the recombination rate of electrons and holes.

Hydrophilicity of TiO2 and Gd-Doped TiO2 Nanotube Arrays
The hydrophilic properties of TiO2 nanotube arrays prepared in different conditions are shown in Figure 12. The optical contact angle of the samples is 6.7° and 6.8° when the reaction time and voltage is 10 h and 60 V, respectively. Figure 12a shows that when the reaction time increases, the length of the generated nanotubes becomes longer, the sample has better crystallinity and a larger specific surface area, which greatly improves the ability for chemically and physically adsorbed water, as well as an excellent hydrophilic property. Similarly, as shown in Figure 12b, when the reaction voltage is low, the driving force obtained by the reaction is small and it is more difficult for ions to pass through the barrier layer to form TiO2 nanotubes, indicating that the hydrophilic property of the sample is worse. However, the excessively high reaction voltage makes the anodic oxidation rate too fast and corrodes the surface of the nanotubes, which causes the hydrophilic performance to decrease.  Figure 12c shows the hydrophilic property of TiO2 nanotube arrays with different Gd 3+ concentrations. It can be observed that, when the doping concentration of Gd 3+ is 0.01 mol/L, the contact angle reaches a minimum value of 4.9°. The doped rare earth oxide is hydrophobic and the hydrophilic property of the samples becomes worse when the concentration of Gd 3+ continues to increase. Compared with the undoped samples (6.7°), it can be seen that the hydrophilic properties of the samples doped with Gd 3+ (4.9°) are significantly better. This is because the samples after doping can lead the small content of the Gd 3+ to enter the Ti 4+ position of the TiO2 crystal lattice, resulting in lattice distortion and causing an imbalance of the electrical charge with generating hydrophilic hydroxyl groups.

Materials
Pure Ti foil was obtained from the Tenghui Titanium Corporation (Jinhua, China). Ammonium fluoride (NH4F) and ethylene glycol (C2H6O2) were purchased from the Xilong Science and Technology Co., Ltd (Shantou, China). Hydrofluoric acid, nitric acid and anhydrous ethanol (C2H5OH) were purchased from the Da Mao chemical reagent Corporation (Tianjin, China). Acetone (C3H6O) was acquired from the Shanghai Yanchen Chemical Industrial Co., Ltd (Shanghai, China). Methyl orange (MO) powders were acquired from the Shanghai Zhanyun Chemical Co., Ltd. (Shanghai, China). The ultrapure water was acquired from a Mili-Q Ultrapure water depurator (Millipore, Bedford, MA, USA).

Preparation of TiO2 Nanotube Arrays
The polished titanium foil was ultrasonicated in an acetone solution, anhydrous ethanol, and ultrapure water for 10 minutes to eliminate surface oil and oxide. Hydrofluoric acid, nitric acid, and deionized water were prepared into a chemical polishing solution at a volume ratio of 1:3:5, the sonicated titanium foil was placed in the solution for 30 seconds, and then washed with pure water. A titanium foil was used as an anode, a graphite sheet as a cathode, and an ethylene glycol solution of ammonium fluoride was used as an electrolytic solution. The reaction was carried out for 10 h after commencing power. The titanium foil was washed with ethanol and annealed at 450 °C for 2 h in a muffle furnace. Afterward, the TiO2 nanotube arrays were generated on the Ti foil surface.

Fabrication of Gd-Doped TiO2 Nanotube arrays
Different concentrations of a Gd nitrate solution (0.025−0.04 mol/L) was prepared. The TiO2 nanotube arrays were immersed in the Gd nitrate solution for 1 h at 30 °C by an impregnation method. Then, the sample was taken out and placed in a muffle furnace and annealed at 500 °C for 2 h, and Gd-doped TiO2 composite nanotube arrays were prepared successfully on the Ti substrate surface. The preparation processes were as follows (Scheme 1): Scheme 1. Preparation processes of Gd-doped TiO2 nanotube arrays

Characterizations
The morphology of TiO2 nanomaterials was observed by an environmental scanning electron microscope (ESEM, FEI Quanta200F, FEI, Hillsboro, USA) and high resolution transmission electron microscope (HRTEM) (200 kV, JEOL, Tokyo, Japan). X-ray diffractometry (XRD, D8 ADVANCE, Bruker, Germany) was performed to identify the phase structure of TiO2 samples. Elemental analysis was measured by X-ray photoelectron spectroscopy (XPS) (ESCALAB250Xi, ThermoFisher Scientific, USA). The specific surface area data of the catalyst was calculated by the Specific Area and Porosity Analyzer (JW-BK132F, Beijing, China). The photocatalytic degradation process of methyl orange was performed in a photochemistry reactor (BL-GHX-V, BILON, Shanghai, China). Hydrophilic properties were calculated by an optical contact angle meter (Zheke, DSA 100, Zheqi Technology, Beijing, China).

Photocatalytic Oxidation of Gd-Doped TiO2 Nanotube arrays
Gd-doped TiO2 nanotube array samples were cut into squares with a 10 mm length and placed in a quartz test tube for the photocatalytic degradation process. To obtain photocatalytic performance data, an MO solution with a concentration of 20 mg/L was prepared as simulated pollutants for photodegradation reactions [46]. The 500 W high-pressure mercury lamp with a main wavelength of 365 nm was used as the luminous source, and the light intensity measured during the reaction was 160 mW/cm 2 . The absorbance of residual MO was detected every 20 min. The photocatalytic performance of the catalyst was evaluated by degradation rate (η). Related calculation is shown in Equation (9): where the initial concentration and absorbance of the catalyst were C and , and the concentration and absorbance at a certain time were and . was the absorbance that was observed under 466 nm ultraviolet (UV) light irradiation. The same method was used to calculate the absorbance of the catalyst at different voltages (30-70 V), different times (6-14 h) and different concentrations (0.0025-0.04 mol/L).

Conclusions
During this study, tightly ordered anatase TiO2 nanotube arrays were successfully prepared using anodizing technology, which was then doped with the rare earth element Gd to fabricate a novel photocatalyst of Gd-doped TiO2 nanotube arrays. Reaction times, voltages, and Gd concentrations were proved to be the vital factors affecting the crystal phase and microstructure of the TiO2 nanotube arrays. When the reaction voltage was 60 V, the reaction time was 10 h and the concentration of Gd 3+ was 0.01 mol/L, the Gd-doped TiO2 nanotube arrays exhibited the larger specific surface area (176 m 2 /g), a beneficial hydrophilic property (optical contact angle was 4.9°), and an excellent photocatalytic performance (degradation rate of MO was 95.8%). The hydrophilic and photocatalytic properties of the TiO2 nanotube arrays were greatly improved after doping with Gd. Gd-doped TiO2 nanotube arrays may be a potential photocatalyst in applications for water pollution.