Synthesis of Self-Gravity Settling Faceted-Anatase TiO2 with Dominant {010} Facets for the Photocatalytic Degradation of Acetaminophen and Study of the Type of Generated Oxygen Vacancy in Faceted-TiO2

In this study, faceted TiO2, predominately exposed with {010} facets (T-{010}), was synthesized with a two-step hydrothermal reaction and used for the degradation of acetaminophen (ACE) in an aqueous solution. T-{010} showed considerable photocatalytic reactivity, and its easy-settling (gravity-settling, ~97% of T-{010} settled after 30 min) property demonstrated acceptable reusability. A solid-state chemical reduction approach (NaBH4) at a mild temperature (300 °C) was used for generation of an oxygen vacancy in T-{010} and P25 (commercial TiO2). The oxygen vacancy concentrations of the samples were investigated by electron paramagnetic resonance (EPR). It was also found that NaBH4 reduction induced the generation of both surface and subsurface Ti3+ on colored P25, but only surface Ti3+ species were formed on colored T-{010}. The prepared colored TiO2 samples were successfully used for photocatalytic degradation of ACE in an aqueous solution under visible light illumination.


Introduction
Today, photocatalysis is a promising technology for the degradation of organic pollutants in water and wastewater [1,2]. The synthesis of engineered, faceted-TiO 2 has been widely pursued due to its high selectivity and photocatalytic reactivity [3][4][5]. Faceted-TiO 2 with a large percentage of high-energy {001} facets, has attracted attention due to its higher surface energy and photocatalytic activity in comparison to other low-index facets ({010}, {101}) [6]; however, for the synthesis of faceted-TiO 2 with a large percentage of high-energy {001}, some toxic and harmful fluoride ions (such as HF, TiF 4 , HBF 4 ) should be used as capping agents [7]. The surface energies of the {010}, {001}, and {101} facets are theoretically calculated and reported as 0.53 J m −2 , 0.90 J m −2 , and 0.44 J m −2 , respectively. As can be seen, the surface energy of {010} is slightly higher than that of {101} and much lower than that of {001} [8]. Recently, it was reported that faceted-TiO 2 with a high percentage of {010} facets showed a close photocatalytic activity in comparison to the {001} facet [4,9]. Despite the fact that these technologies are proven to be effective for the degradation of organic pollutants in the aqueous phase, their energy consumption is substantial due to the utilization of UV lamps (photon provider in the process) [10]. Solar irradiation is a renewable and clean energy source. The major component of the solar radiation that reaches the Earth's surface is in the range of visible light (400-600 nm); thus, visible-light active technology can be counted on as a priority source for developing photocatalytic systems [11].
Black TiO 2 was initially prepared by Chen and co-workers using a hydrogen thermal treatment that exhibited extremely high rates of photocatalytic activity for water splitting and organic pollutant degradation under solar or visible light irradiation [12,13]. Since then, black TiO 2 has attracted enormous attention from researchers and scientists. Several strategies, including thermal treatment at high temperatures and pressures, plasma-assisted hydrogenation, and chemical reduction, have been widely used for the preparation of black TiO 2 [14]. Among the chemical reductants, recently, NaBH 4 has been widely used for the preparation of black TiO 2 . Kang et al. produced black TiO 2 nanotubes in 0.1 M NaBH 4 solution for 10-60 min at room temperature [15]. In another work, black TiO 2 was prepared by mixing TiO 2 and NaBH 4 powders, with subsequent calcination at 350 • C for 1 h in a tubular furnace under a N 2 gas atmosphere [16].
In this study, faceted TiO 2 predominately exposed with {010} facets (T-{010}) was synthesized with a two-step hydrothermal reaction and used for the degradation of acetaminophen (ACE) in an aqueous solution. P25, as a commercial form of TiO 2 , has been widely used for photocatalytic degradation of organic pollutants, but the main challenge of P25 application is its recovery and separation from aqueous media. One of the novelties of this paper is the demonstrated synthesis of a photocatalyst with a photocatalytic activity close to P25. Its separation from aqueous media is easily possible by gravity-settling. A solid-state chemical reduction approach (NaBH 4 ) at a mild temperature of 300 • C was utilized for generation of an oxygen vacancy in T-{010} and P25 (for application under visible light). In this study, for the first time, the type and concentration of the generated oxygen vacancy (Ti 3+ ) in faceted-TiO 2 ({010}-TiO 2 ) and P25 were studied. The oxygen vacancy concentration of the samples was investigated by electron paramagnetic resonance (EPR). The synthesized colored TiO 2 samples were also used for photocatalytic degradation of ACE in aqueous solution under visible light illumination.

Synthesis of T-{010}
T-{010} was synthesized through the following double-step hydrothermal processing: (i) Hydrothermal reaction of P25 with KOH solution at 200 • C for 48 h in a Teflon-lined stainless steel autoclave, followed by washing of the forming precipitate of potassium titanate with deionized water (D.I. water), followed by drying in an oven, and (ii) reaction of the potassium titanate with water at 200 • C for 24 h in the autoclave, followed by washing of the product with D.I. water, followed by drying in an oven at 60 • C [17].

Experimental Procedure and Sample Preparation
The solution was kept in a 40 mL beaker, which was located 10 cm under a UVC/lamp. A magnetic stirrer was used for mixing the photocatalysts in the acetaminophen (ACE) solution at a rotational speed of 300 rpm. Before the photocatalytic experiments, the mixture of the MB solution and the photocatalyst was stirred in the dark for 30 min. The UV light intensity and the initial concentration of ACE were 5 mW/cm 2 and 1 ppm, respectively. For the photocatalytic experiments under AM 1.5 G solar light illumination, a xenon arc lamp (150 W) equipped with a UV cut-off filter (L > 420 nm) was used. All experiments were done under the ambient temperature. To evaluate the experimental results, all photocatalytic experiments were repeated twice, and the average was taken and used. After completion of the photodegradation experiments for the required time, 1 mL of each experimental sample was taken, and its concentration was analyzed by Liquid chromatography-mass spectrometrymass spectrometry (LC/MS/MS).
The Langmuir-Hinschelwood (L-H) model was used to describe the kinetic of the ACE photodegradation by P25 and T-{010} under UV light illumination: where r, C and t are the rate of the destruction of the ACE, the concentration of the solution, and reaction time, respectively. K and k r represent the equilibrium constant for sorption of ACE and the limiting rate constant of the reaction under the experimental conditions, respectively. At low concentrations, the KC value is much lower than 1, and thus this term can be neglected in Equation (1). By neglecting this term, Equation (1) converts to a first-order kinetics model (Equation (2)): where k app is the apparent first-order rate constant. By integrating from Equation (2) and using boundary condition C = C o at t = 0, it gives: where C o is ACE initial concentration in aqueous solution.

Sedimentation of P25 and T-{010}
Sedimentation of P25 and T-{010} has been studied on a time-resolved UV-vis spectrophotometer according to other works [18]. The concentrations of P25 and T-{010} are represented at their specific absorbance wavelengths at 326 nm. The absorbance was measured versus time and the normalized concentration C t /C o was calculated, where C t is the concentration at time t, and C o is the absorbance value.

Preparation of Colored T-{010} (BT-{010}) and P25 (BP25) by NaBH 4 Reduction
At ambient temperature, 2.0 g of photocatalyst (P25 or T-{010}) powder and 0.8 g of NaBH 4 were mixed together and then ground using a mortar for 10 min. The mixture was transferred to an alumina crucible and heated at 300 • C for 50 min under an Ar atmosphere. After furnace cooling down to room temperature to eliminate unreacted NaBH 4 , the obtained sample was washed several times with distilled water and ethanol, and subsequently was dried in an oven overnight.

Instruments
The X-ray diffractometry (XRD) of samples was performed by Cu−Kα radiation by a Shimadzu lab-X XRD-6000 X-ray Diffraction instrument. The surface morphologies of the T-{001} were evaluated using a field emission scanning electron microscope (FESEM, JEOL JSM-6700F, Tokyo, Japan). Transmission electron microscopy (TEM, JEM-2010) was used for the detection of the formed disordered layers in the samples. An ultraviolet-visible spectrophotometer (UV-vis, Shimadzu, Kyoto, Japan, UV-3600) was utilized for ultraviolet-visible diffuse reflectance spectra in the range of 200-800 nm with BaSO 4 as the reference. The samples band gaps (E g ) were calculated by Equation (4): where h and c are Planck's constant (6.626 × 10 −34 J·s) and the speed of light (3 × 10 8 m/s), respectively; L is the extrapolated wavelength (nm) at which the absorbance value reaches the instrument limit. Electron paramagnetic resonance (EPR) spectra (JEOL FA200 SPECTROMETER (X-Band with LN2 VT accessories)) were used at room temperature, at a microwave frequency of 9.85 GHz, for confirmation of high spin Ti 3+ , as well as oxygen vacancy.       The structures of the samples were further investigated using HRTEM (Figure 3). In terms of P25 (Figure 3a), the interplanar distances of anatase (101) planes and rutile (110) were measured to be 0.35 nm and 0.324 nm, respectively. Similarly, the lattice fringe spacing of the (010) crystal planes for T-{010} was measured to be 0.38 nm. The found (measured) values of the interplanar distance of various facets are in good agreement with other reports [19,20]. for T-{010} was measured to be 0.38 nm. The found (measured) values of the interplanar distance of various facets are in good agreement with other reports [19,20]. The specific surface area, measured using the nitrogen adsorption and desorption isotherms, is 47 m 2 ·g −1 for P25 and 32 m 2 ·g −1 for T-{010}. The photoactivity of the photocatalyst is largely governed by the electronic structure and the redox ability in a photocatalyst, whereas the separation efficiency of photo-generated electron-hole pairs is mainly dominated by valence band and conduction band energy levels [19]. In light of these facts, investigations into the electronic structures of P25 and T-{010} were carried out using UV-vis absorption spectra as shown in Figure S1, in which a redshift was observed at the absorption edge of P25 as compared with {010}. Furthermore, the band gaps calculation based on Equation (4) gave 3.10 and 3.29 eV for P25 and T-{010}, respectively.

Photocatalytic Degradation
To study the effect of the {010} facet on the photocatalytic activity of faceted-TiO2, T-{010} was used for the photocatalytic degradation of acetaminophen (ACE) under UV illumination and dark conditions. Furthermore, P25 with an average particle size of 25 nm was used for ACE photocatalytic degradation as a reference for comparison ( Figure 4). As can be seen, in the dark condition (without light illumination), C/Co change by time is too low, which shows that adsorption does not play an important role in the removal of ACE from the aqueous solution. It was found that P25 exhibits a slightly higher rate of photocatalytic activity than t T-{010}. As is known, photocatalytic activity is highly dependent on surface area, occurring because of the photocatalytic reaction on the surface of the TiO2. The lower photocatalytic activity of T-{010} can be ascribed to its smaller surface area in comparison to P25. To better understand the surface photocatalytic activity, kapp was normalized with respect to the surface area of the samples. The normalized results shown in Table S1 indicate that T-{010} shows better photocatalytic activity than P25. It was found that the photocatalytic activity strongly depended on the surface energy and the relative number of Ti 4+ sites with five-fold coordination environments, both of which were greater on the {010} facets (T-{010} sample) compared to the {101} (P25 contains {101} facet) [21,22].
However, due to the hard settling nature of P25, >97% of P25 is lost after gravity settling and cannot be reused in the subsequent cycle. This is entirely different for T-{010}, owing to its excellent sedimentation property, as 96.3% of the material could be separated after 30 min of gravity-settling ( Figure 5). Although slightly lower ACE degradation efficiency was observed in comparison to P25, reusability of T-{010} was good. Figure S2 clearly shows the effect of settling gravity (after 20 min) on P25 and T-{010}; as can be seen, a considerable ratio of T-{010} is settled and the color of the aqueous solution after gravity-settling is clear. Therefore, T-{010} is a much more promising material compared to P25 for practical and cost-effective application in large scale. To find the reasons for the better settling of T-{010} compared to P25, zeta potentials of samples were determined and are presented in Figure S3. As can be seen, the isoelectric points of these two samples are very close The specific surface area, measured using the nitrogen adsorption and desorption isotherms, is 47 m 2 ·g −1 for P25 and 32 m 2 ·g −1 for T-{010}. The photoactivity of the photocatalyst is largely governed by the electronic structure and the redox ability in a photocatalyst, whereas the separation efficiency of photo-generated electron-hole pairs is mainly dominated by valence band and conduction band energy levels [19]. In light of these facts, investigations into the electronic structures of P25 and T-{010} were carried out using UV-vis absorption spectra as shown in Figure S1, in which a redshift was observed at the absorption edge of P25 as compared with {010}. Furthermore, the band gaps calculation based on Equation (4) gave 3.10 and 3.29 eV for P25 and T-{010}, respectively.

Photocatalytic Degradation
To study the effect of the {010} facet on the photocatalytic activity of faceted-TiO 2 , T-{010} was used for the photocatalytic degradation of acetaminophen (ACE) under UV illumination and dark conditions. Furthermore, P25 with an average particle size of 25 nm was used for ACE photocatalytic degradation as a reference for comparison ( Figure 4). As can be seen, in the dark condition (without light illumination), C/C o change by time is too low, which shows that adsorption does not play an important role in the removal of ACE from the aqueous solution. It was found that P25 exhibits a slightly higher rate of photocatalytic activity than t T-{010}. As is known, photocatalytic activity is highly dependent on surface area, occurring because of the photocatalytic reaction on the surface of the TiO 2 . The lower photocatalytic activity of T-{010} can be ascribed to its smaller surface area in comparison to P25. To better understand the surface photocatalytic activity, k app was normalized with respect to the surface area of the samples. The normalized results shown in Table S1 indicate that T-{010} shows better photocatalytic activity than P25. It was found that the photocatalytic activity strongly depended on the surface energy and the relative number of Ti 4+ sites with five-fold coordination environments, both of which were greater on the {010} facets (T-{010} sample) compared to the {101} (P25 contains {101} facet) [21,22].
However, due to the hard settling nature of P25, >97% of P25 is lost after gravity settling and cannot be reused in the subsequent cycle. This is entirely different for T-{010}, owing to its excellent sedimentation property, as 96.3% of the material could be separated after 30 min of gravity-settling ( Figure 5). Although slightly lower ACE degradation efficiency was observed in comparison to P25, reusability of T-{010} was good. Figure S2 clearly shows the effect of settling gravity (after 20 min) on P25 and T-{010}; as can be seen, a considerable ratio of T-{010} is settled and the color of the aqueous solution after gravity-settling is clear. Therefore, T-{010} is a much more promising material compared to P25 for practical and cost-effective application in large scale. To find the reasons for the better settling of T-{010} compared to P25, zeta potentials of samples were determined and are presented in Figure S3. As can be seen, the isoelectric points of these two samples are very close together. It can be demonstrated that, based on the values of isoelectric points, the samples show similar behaviors in the solution. It can be concluded that the larger particles size of the T-{010} is the main reason for the better settling of this sample.       Photocatalyst recycling and reusability are two important parameters for their application at the industrial scale. The results of the investigation on recycling and reusability of the photocatalysts for ACE photocatalytic degradation in the present work are presented in Figure 6. After each cycle, the Photocatalyst recycling and reusability are two important parameters for their application at the industrial scale. The results of the investigation on recycling and reusability of the photocatalysts for ACE photocatalytic degradation in the present work are presented in Figure 6. After each cycle, the photocatalysts were collected and reused for the next cycle of photocatalytic degradation, without any purification. As can be seen, the photocatalytic degradation percentages of T-{010} and BT-{0101} after five cycles did not significantly change. For T-{010}, the photocatalytic degradation percentage was reduced by less than 4%, whereas, for BT-{0101}, this reduction was less than 8%.
Water 2018, 10, x 7 of 10 photocatalysts were collected and reused for the next cycle of photocatalytic degradation, without any purification. As can be seen, the photocatalytic degradation percentages of T-{010} and BT-{0101} after five cycles did not significantly change. For T-{010}, the photocatalytic degradation percentage was reduced by less than 4%, whereas, for BT-{0101}, this reduction was less than 8%.

Photocatalytic Activity of BT-{010} and BP25
The NaBH4 reduction resulted in the change of color of both samples (T-{010} and P25) from white to dark blue ( Figure S4). The UV-vis diffuse reflection spectrograms of BT-{010} and BP25 are shown in Figure S5. The colored and normal TiO2 (T-{010} and P25) show a similar absorbance trends in the UV region; however, they indicate distinctive absorbance behaviors in the visible light region. As can be seen, due to surface oxygen vacancies generation, both BT-{001} and BP25 samples exhibit a stronger absorbance in the visible light region, which is indicative of surface oxygen vacancies playing an important role in the visible light absorbance [16]. For this system, the apparent quantum efficiency (Φ) was calculated based on the ACE degradation versus the flux of the incident photons. Under visible range excitation, the colored samples show higher apparent quantum efficiencies (Φ) in the range of 450-650 nm ( Figure S6).
Due to the high sensitivity of EPR for characterization of the surface and bulk Ti 3+ , lowtemperature EPR was used to evaluate the concentration and distribution of defects in the normal and colored TiO2 samples. Values of the g factor were used as the key parameter to distinguish the location of Ti 3+ . Generally, surface Ti 3+ has lower values of g factors in comparison with the bulk [23]. T-{010} and P25 exhibited no signals for EPR analysis. As shown in Figure 7, BT-{010} indicated a strong EPR signal at g = 1.996, corresponding to the surface Ti 3+ [24,25]. BP25 (rich in {101} facets) showed a strong response at an average g value of 2.003, which was related to the bulk Ti 3+ . The EPR analyses show that the NaBH4 reduction caused generation of the surface and subsurface Ti 3+ (oxygen vacancy) in the BP25, and only surface Ti 3+ (oxygen vacancy) in the BT-{010}.

Photocatalytic Activity of BT-{010} and BP25
The NaBH 4 reduction resulted in the change of color of both samples (T-{010} and P25) from white to dark blue ( Figure S4). The UV-vis diffuse reflection spectrograms of BT-{010} and BP25 are shown in Figure S5. The colored and normal TiO 2 (T-{010} and P25) show a similar absorbance trends in the UV region; however, they indicate distinctive absorbance behaviors in the visible light region. As can be seen, due to surface oxygen vacancies generation, both BT-{001} and BP25 samples exhibit a stronger absorbance in the visible light region, which is indicative of surface oxygen vacancies playing an important role in the visible light absorbance [16]. For this system, the apparent quantum efficiency (Φ) was calculated based on the ACE degradation versus the flux of the incident photons. Under visible range excitation, the colored samples show higher apparent quantum efficiencies (Φ) in the range of 450-650 nm ( Figure S6).
Due to the high sensitivity of EPR for characterization of the surface and bulk Ti 3+ , low-temperature EPR was used to evaluate the concentration and distribution of defects in the normal and colored TiO 2 samples. Values of the g factor were used as the key parameter to distinguish the location of Ti 3+ . Generally, surface Ti 3+ has lower values of g factors in comparison with the bulk [23]. T-{010} and P25 exhibited no signals for EPR analysis. As shown in Figure 7, BT-{010} indicated a strong EPR signal at g = 1.996, corresponding to the surface Ti 3+ [24,25]. BP25 (rich in {101} facets) showed a strong response at an average g value of 2.003, which was related to the bulk Ti 3+ . The EPR analyses show that the NaBH 4 reduction caused generation of the surface and subsurface Ti 3+ (oxygen vacancy) in the BP25, and only surface Ti 3+ (oxygen vacancy) in the BT-{010}.  BT-{010} showed higher activity than BP25, owing to surface Ti 3+ that is more effective for a photocatalytic process (some of the generated Ti 3+ in BP25 is located in the subsurface and is less effective).   BT-{010} showed higher activity than BP25, owing to surface Ti 3+ that is more effective for a photocatalytic process (some of the generated Ti 3+ in BP25 is located in the subsurface and is less effective).   Figure 8 shows the photocatalytic degradation of the ACE, in aqueous solution, by numerous pristine and colored TiO2 samples under visible light illumination for 3 h. P25 and T-{010} exhibited lower photocatalytic degradation rates. It could be verified that the generation of Ti 3+ (oxygen vacancy) sites in the colored samples resulted in an extension of visible light absorbance and an improvement of charge carrier transfer. BT-{010} showed higher activity than BP25, owing to surface Ti 3+ that is more effective for a photocatalytic process (some of the generated Ti 3+ in BP25 is located in the subsurface and is less effective).  Electrochemical impedance spectroscopy (EIS) is a useful method for investigating the charge transfer in photocatalysis. Figure S7 shows the Nyquist plot of P25, T-{010}, BP25, and BT-{010} photocatalysts under standard light illumination. The semicircle shape of Nyquist curves at high frequencies is an indication of a photo-generated charge transfer between the aqueous phase and the photocatalyst, and the diameter of the semicircle corresponds to the charge transfer resistance (R ct ). The lower R ct values for BP25 and BT-{010} and the higher values for the P25 and T-{010} photocatalysts indicate the positive impact of oxygen vacancy on the reduction of carrier recombination and the improvement of charge transfer [26].

Conclusions
The {010} anatase TiO 2 facets were successfully synthesized via a two-step hydrothermal process. The synthesized faceted-TiO 2 was used for the photocatalytic degradation of MB in an aqueous solution, and the photocatalytic performance of the synthesized photocatalyst was compared with P25. Although slightly lower ACE degradation efficiency was observed in comparison to P25, the reusability of T-{010} was good. A considerable part of the T-{010} settled, and the color of the aqueous solution after gravity-settling was clear. The results showed that T-{010} is a much more promising and cost-effective material compared to P25 for practical application at large scale. It was also found that NaBH 4 reduction induced the generation of both surface and subsurface Ti 3+ on the colored P25, but only surface Ti 3+ species were formed on the colored T-{010}. The synthesized colored TiO 2 samples were successfully used for photocatalytic degradation of ACE in the aqueous solution under visible light illumination; BT-{010} also showed a higher photocatalytic activity than other photocatalysts.  Table S1: Summary of k app and normalized k app results, Figure S5: The diffuse reflectance UV-Vis spectra of colored TiO 2 , Figure S6: Wavelength-dependent apparent quantum yield for the photocatalytic degradation of ACE, Figure S7: Nyquist characterization of the samples.