Chemical Structure of TiO 2 Nanotube Photocatalysts Promoted by Copper and Iron

TiO 2 nanotubes (TNTs) promoted by copper (5%) (Cu-TNT) and iron (5%) (Fe-TNT)were prepared for visible-light photocatalysis. By X-ray absorption near edge structure (XANES) spectroscopy, it is found that the enhanced photocatalytic degradation of methylene blue (MB) on Cu-TNT and Fe-TNT is associated with the predominant surface photoactive sites A 2 ((Ti=O)O 4 ). By extended X-ray absorption fine structure (EXAFS) spectroscopy, the dispersed copper and iron also cause increases in the Ti–O and Ti–(O)–Ti bond distances by 0.01-0.02 and 0.04-0.05 Å, respectively. The decreased Ti–O bonding energy may lead to an increase of photoexcited electron transport. The copperor-iron promoted TNT can thus enhance photocatalytic degradation of MB under the visible-light radiation.


Introduction
Titanium nanotubes (TNTs) having a high surface area, chemical stability (under alkali or acidic conditions), and sunlight sensitization have been employed for potential applications in photoinduced reactions [1], sensitized electrodes [2,3], and driven water cleavage for hydrogen generation [4].From environmental perspectives, TNTs are widely used in photocatalytic detoxification of hazardous substances [5,6], adsorption of organic pollutants [7], and enrichment of heavy metals [8].
Nanotubular and nanowired TiO 2 have a better efficiency for photocatalytic degradation of rhodamine B and methyl orange under solar illumination than the commercialized nano TiO 2 (P25) [9].However, it was found that the nitrogendoped TiO 2 nanotube (N-TNT) thin film was not very effective for photocatalytic remediation of oil spill on seawater under the visible-light radiation.
Recently, transition metals such as chromium, iron, nickel, copper, and zinc have been considered as promoters for improving the photoactivity of TNTs.Zhang et al. [10] demonstrated that the Cr 2 O 3 /TNT nanocomposite could enhance photocatalytic yield of H 2 under visible-light.By X-ray absorption near edge structure (XANES) and X-ray photoelectron spectroscopy (XPS), Jang and coworkers [11] found that the surface iron species were composed of iron hydrate on TNT (Fe-TNT).Nevertheless, the Fe-TNT had a greater photocurrent generation but poor photocatalytic performances for water splitting and dye degradation.Nickelintercalated TNT had a greater photocatalytic yield of H 2 than the TNT under ultraviolet (UV) illumination [12].Interestingly, the hydrated nickel complex was the active sites for reduction of proton in water to H 2 .The CuO dispersed tubular TiO 2 could enhance H 2 yield [13].A nanosize ZnO decorated TNT was preferable for rhodamine B degradation under UV light [14].
By component-fitted XANES, the catalytic active species such as copper oxide clusters in the channels of ZSM-5 were found playing an important role in catalytic decomposition of heptane and toluene.Speciation of catalytic active sites and reaction paths involved in catalytic degradation of chlorophenols and reduction of NO was also revealed by XANES.Hsiung and coworkers [15] found by XANES that the photoactive species (A 2 ) ((Ti = O)O 4 ) in TiO 2 were responsible for the photocatalytic decomposition of methylene blue (MB).In the present work, copper and iron which are less toxic and easy to obtain were used for preparation of the visible-light photocatalysts (i.e., Cu-TNT and Fe-TNT).To design the effective visible-light Cu-TNT and Fe-TNT photocatalysts, the molecule-scale data such as local bonding environment and oxidation state of the photoactive species especially the A 2 sites are essential.Thus the main objective of the present work was to study nature of the photoactive species in the copper-and-iron promoted TNTs by synchrotron XANES and extended X-ray absorption fine structure (EXAFS).

Experimental
The preparation methods for the TNT have been described [16,17].Typically, a mixture of 0.625 g of TiO 2 (P25, Degussa) and 12.5 mL of a NaOH (5 M) solution was heated at 473 K for 24 h in a Teflon-lined autoclave.The product slurry was filtered and washed with 300 mL of 0.01 M HNO 3 solution.The filtered solid was dried at 333 K for 6 h.To prepare copper-and-iron dispersed TNTs (Cu-and Fe-TNTs), Cu(NO 3 ) 2 (JT Baker), or Fe(NO 3 ) 3 (JT Baker) at a molar ratio of Cu-or Fe-to-Ti of 0.1 was mixed with the TNT powder in 100 mL of H 2 O.A Na 2 CO 3 (JT Baker) solution was used to adjust the pH values of the Cu-TNT and Fe-TNT mixtures to 7 and 5, respectively.The Cu-TNT or Fe-TNT slurry was dried at 373 K for 24 h.
The surface morphologies of the TNT, Cu-TNT, and Fe-TNT were also analyzed by field-emission scanning electron microscopic (FE-SEM) analyzer (XL-40FEG, Philips) coupled with energy dispersive X-ray spectrometry (EDS).Cross-sectional topologies of the TNT, Cu-TNT, and Fe-TNT were also determined by transmission electron microscopy (TEM) (CM-200 TWIN, Philips) functioning with the selected area electron diffractometry (SAED).The sunlight absorption characteristics of the nanosize TiO 2 (Kanto Chemical), TNT, Cu-TNT, and Fe-TNT were measured on a diffuse reflectance ultraviolet-visible (DR UV-Vis) spectrophotometer (Cary 100 Conc, Varian).Nitrogen adsorption/desorption isotherms of the nanosize TiO 2 , TNT, Cu-TNT, and Fe-TNT were studied on a surface area analyzer (COULTER SA3100, Beckman), and their pore size distributions were also obtained by the Barrett-Joyner-Halenda (BJH) calculation.
Photocatalytic degradation of methylene blue (KATAYAMA Chemical) effected by the nanosize TiO 2 , TNT, Cu-TNT, or Fe-TNT was performed on a homemade photoreactor [18].0.01 g of the nanosize TiO 2 , TNT, Cu-TNT, or Fe-TNT photocatalyst was suspended in the MB (20 mg⋅L −1 ) aqueous solution (50 mL) with magnetic stirring and cooling water circulation (Medel-B401D, Firstek Scientific) to maintain the reaction system at 298 K.The photocatalysts suspended MB solution was well stirred in the dark for 1 h prior to the photocatalytic experiments.A 300 W Xenon lamp solar simulator (no.91160A, Newport) combined with an AM 1.5 G filter (no.59044, Oriel) was served as the light source.The power density of the irradiation was fixed at 100 mW⋅cm −2 by the use of a reference solar cell and meter (no.91150, Oriel).The concentration of the MB was measured by a UV-Vis spectrophotometer (Cary 100 Conc, Varian) at the maximum absorbance of 662 nm.
The EXAFS and XANES spectra of the TNT, Cu-TNT, and Fe-TNT photocatalysts and model compounds such as Cu(OH) 2 (Showa), Fe(OH) 3 (Alfa Aesar), and Cu 2+ or Fe 3+ /MCM-41 (prepared by impregnation of Cu(NO 3 ) 3 (JT Baker) or Fe(NO 3 ) 3 (JT Baker) on MCM-41 and dried at 383 K for 8 h) were collected on the Wiggler beam line at the Taiwan National Synchrotron Radiation Research Center.The storage ring was operated at the energy of 1.5 GeV.A Si (111) double-crystal monochromator was used for beam energy selection at the energy resolution (ΔE/E) of 1.9 × 10 −4 (eV/eV).The beam energy was calibrated by the use of the maximum absorption edge of Ti, Cu, and Fe foils at 4966, 8979, and 7112 eV, respectively.The k 3 -weighted (k) oscillations were Fourier-transformed from k to R spaces using the UWXAFS 3.0 program coupled with the FEFF 8.0 code [19].A Bessel window function was activated in the k ranges of 2.5-12, 3.1-13.7,and 3.1-12.1 Å−1 for Ti, Cu, and Fe K-edges, respectively.The  2 0 (many-body factor) was fixed at 0.9 to reduce the parameter variables during the fitting.For the XANES spectra analysis, the Gaussian-Lorentzian calculation was employed for the curve deconvolution.Fittings of the model compounds to the experimental data have errors of ±0.01 Å and ±0.02 Å in bond distances and ±10% and ±25% in coordination numbers (CNs) for the first and second fitting shells, respectively.

Results and Discussion
The TEM images, SAED patterns, and EDS spectra of the TNT, Cu-TNT, and Fe-TNT photocatalysts are shown in Figure 1.The TNTs having a nanotube structure are 100-200 nm in length, 10-20 nm in the opening at both ends, and 3-5 nm in the wall thickness.It seems that the nanosize copper and iron aggregates are formed on the internal surfaces of the TNT.The EDS spectra indicate that about 5.3% of Cu and 4.2% of Fe are dispersed on the Cu-TNT and Fe-TNT, respectively.The DR UV-Vis spectra of the nanosize TiO 2 , TNT, Cu-TNT, and Fe-TNT are shown in Figure 2. The TNTs have intensive absorption edges at 200-400 nm with a very small absorption at 500 nm.Note that the Cu-and Fe-TNT have extended to visible-light absorption range of 400-800 nm.
The TNT possessing a surface area of 213-235 m 2 ⋅g −1 which was determined and calculated using the data obtained from the N 2 adsorption/desorption isotherms is much greater than the nanosize TiO 2 (42 m 2 ⋅g −1 ).In Figure 3, it is clear that the hysteresis loops at / 0 > 0.8 for the TNTs in the N 2 adsorption and desorption isotherms may be associated with the nanotubular conformation.Figure 3 also shows that the pore size distribution of the TNT, Cu-TNT, and Fe-TNT is between 50 and 100 nm while the nanosize TiO 2 has a relatively less pore structure (10-40 nm).The adsorbed N 2 volume of the TNTs is in the range of 550-700 cm 3 ⋅g −1 which is greater than that of the nanosize TiO 2 by 5-7 times.
As the Cu-TNT and Fe-TNT have a visible-light absorption capacity, photocatalytic degradation of MB was determined on a solar simulator.Figure 4 shows photocatalytic degradation of MB effected by the TNT, Cu-TNT, and Fe-TNT.For comparison, the photocatalytic activity of the nanosize TiO 2 was also determined.It seems that the nanosize TiO 2 and TNT are not very active (<25%) in photocatalytic degradation of MB under sunlight illumination for 240 min, while notably the Cu-TNT and Fe-TNT possess greater MB degradation efficiencies, that is, 86% and 92%, respectively.Note that the enhanced photocatalytic activity of the Cu-TNT and Fe-TNT may be associated with the extended visible-light absorption range.
To better understand the photocatalytic sites of the TNT, Cu-TNT, and Fe-TNT, their XANES and EXAFS spectra of the Ti, Cu, and Fe K-edges were measured.Figure 5  which may be associated with the tetrahedral (TiO 4 ), square pyramid ((Ti = O)O 4 ), and octahedral (TiO 6 ) titanium oxide structures, respectively, within the preedge of 4963-4973 eV [15,20].Note that the A 2 species which may account for the photoactive sites are predominant (51-54%) on the Cu-TNT and Fe-TNT.
Table 1 shows speciation data of the TNT, Cu-TNT, and Fe-TNT studied by EXAFS.The k 3 -weighted EXAFS oscillations are Fourier-transformed from k to R spaces using the Bessel function in the k ranges of 2.5-12 (Ti), 3.1-13.7 (Cu), Photon energy (keV) Normalized absorbance (a.u.) Photon energy (keV) Normalized absorbance (a.u.)     [23].It was also observed by in situ XANES and XPS that electron donation (from the photoexcited nanosize TiO 2 thin film) to the surface CuO might occur [24].The component-fitted XANES data suggest that Cu 2+ and Fe 3+

Figure 1 :
Figure 1: Transmission electron microscopic images, selected area electron diffraction patterns, and energy dispersive X-ray spectra of the (a) TNT, (b) Cu-TNT, and (c) Fe-TNT.

Figure 4 :
Figure 4: Photocatalytic degradation of methylene blue on the (a) nanosize TiO 2 , (b) TNT, (c) Cu-TNT, and (d) Fe-TNT under sunlight illumination (AM 1.5 G).The inset shows the spectrum of the sunlight source.

Figure 5 :
Figure 5: The Ti K-edge XANES and the Gaussian-Lorentzian deconvoluted spectra of the (a) Cu-TNT and (b) Fe-TNT.The dotted and solid lines represent fittings and experimental data, respectively.

Figure 6 :
Figure 6: The Cu and Fe K-edges XANES spectra of the (a) Cu-TNT and (b) Fe-TNT, respectively.The empty and filled symbols represent the fitting and fractions of standards calculated by the linear combination algorithm.The experimental data are shown in solid lines.

Table 1 :
Speciation data of the TNT, Cu-TNT, and Fe-TNT studied by EXAFS., it was found that the nanosize TiO 2 dispersed with surface CuO having the Cu-O bond distance of 1.93-1.94Å and CN of 1.2 in the first shell can, to some extent, enhance the photocatalytic yield of H 2 and dye decomposing in sea water experiment