Preparation and characterization of Cu-doped TiO2 nanomaterials with anatase/rutile/brookite triphasic structure and their photocatalytic activity

Pure TiO2 and Cu–doped TiO2 containing different amounts of copper ions with anatase/rutile/brookite triphasic structure were successfully synthesized through a simple hydrothermal method. The obtained samples were characterized by X–ray diffraction (XRD), Raman spectroscopy, scanning electron microscope (SEM), transmission electron microscope (TEM), X–ray photoelectron spectroscopy (XPS), UV−vis diffuse reflectance spectroscopy (UV-DRS), photoluminescence spectroscopy (PL) and Brunauer–Emmett–Teller surface area analyze (BET). Both pure and Cu–doped TiO2 show relatively high photocatalytic activity owing to their considerable surface areas. Moreover, the three–phase coexisting structure and the conversion between Cu2+ and Cu+ ions facilitate the separation of photogenerated electrons and holes, which is favorable for photocatalytic performance. 1%Cu–TiO2 exhibits the highest photocatalytic activity and the degradation degree of rhodamine B (RhB) reaches 93.5% after 30 min, which is higher than that of monophasic/biphasic 1%Cu–TiO2. ·O2− radical is the main active species, and h+ and ·OH species are subsidiary in the degradation process.


Introduction
Degradation of water pollutants by photocatalysts is a promising technology because of its environmental protection and validity [1][2][3]. Compared to other semiconductors, such as ZnO [4,5], SrTiO 3 [6], CdS [7], TiO 2 has been investigated due to its high chemical stability, high efficiency, availability and nontoxicity [8][9][10][11][12]. However, TiO 2 absorbs only ultraviolet light which is less than 5% of sunlight. Besides, the high recombination rate of photoinduced charge carriers is another drawback [1,2]. It is necessary to develop the modification strategies of TiO 2 photocatalyst and metal ions doping is a simple and effective modification way. It has been widely reported that Cu doping can modify the structure of TiO 2 and enhance the photocatalytic activity [13,14].
There are three crystal phases of TiO 2 : anatase, rutile and brookite. The photogenerated electrons can move from the phase with higher conduction band to lower one in mixed crystal TiO 2 , which promotes the transfer of charge carries, improves quantum efficiency and enhances photocatalytic activity [12,15]. For instance, Li et al. [15] successfully synthesized TiO 2 photocatalysts with different proportions of anatase and rutile mixed phases in an acidic hydrothermal system. The photodegradation of rhodamine B (RhB) and methyl orange (MO) results show that the mixed phase TiO 2 samples display better photocatalytic performance, and the catalyst containing 77wt.% anatase and 23wt.% rutile exhibits the highest photocatalytic activity. Cao et al. [16] prepared brookite TiO 2 with different rutile contents by a facile solvothermal method. TiO 2 consisted of 28wt.% rutile and 72wt.% brookite exhibits the highest degradation rate. The researches on TiO 2 with anatase/rutile/brookite triphasic structure have also been reported [10,17]. Based on two-phase mixed TiO 2 , since the conduction band position of brookite is higher than those of anatase and rutile, the three-phase mixed structure can further accelerate the migration of photogenerated electrons at the phase interfaces, which is beneficial to higher photocatalytic performance [10]. Kaplan et al. [17] prepared TiO 2 with anatase, rutile and brookite polymorph phases by combining sol-gel process and hydrothermal treatment. It has been proved that TiO 2 with triphasic structure exhibits better photocatalytic activity than anatase or rutile.
It is widely recognized that the specific surface area of photocatalysts is an important factor to photocatalytic performance. Many studies have been devoted to improving photocatalytic activity by increasing the specific surface area of TiO 2 . The results of Aguilar et al.'s work reveal that the photocatalytic activity is enhanced owing to the increase of specific surface area after Cu doping [18]. It is reported by Zhang et al. [19] that the specific surface area of P25 increases from 55.7 m 2 /g to 78.7 m 2 /g after combining with graphene, consequently, the photocatalytic performance is extremely improved.
The preparation of TiO 2 photocatalysts by hydrothermal method does not need high-temperature calcination, which avoids the grain growth and particle agglomeration during the calcination process. It is beneficial to high surface area and photocatalytic activity. Therefore, in the present study, pure and Cu-doped TiO 2 nanocomposites with anatase/rutile/brookite triphasic structure were prepared by a facile hydrothermal method. The prepared samples were systematically characterized and their photocatalytic activity was assessed through the decomposition of RhB under xenon lamp irradiation. (China). All the chemical reagents used in the present study were used directly without further purification. 5 mL butyl titanate was dissolved into 10 mL ethanol to obtain solution A. 1 mL hydrochloric acid (37%), 1 mL polyethylene glycol (600) and proper amounts copper nitrate trihydrate (0.036 g, 0.072 g, 0.108 g and 0.144 g) were dissolved into 15 mL distilled water to acquire solution B. Solution B was added dropwise into solution A to form a sol.

Characterization
X-ray diffraction (XRD) patterns were recorded with a diffractometer (DX-2700, China). The test voltage was 40 kV, the current is 30 mA, the scanning speed was 0.06°/s, and the scanning angle was 20°-70°. Raman spectra were measured using a Micro-Raman Renishaw spectrometer (Andor SR-500i, Britain) equipped with an argon laser (532 nm). The surface morphologies were determined by a field-emission scanning electron microscope with a working voltage of 5 kV (SEM, FEI-Inspect F50, USA) and a transmission electron microscope with an acceleration voltage of 200 kV (TEM and HRTEM, FEI-Tecnai G2 F20, USA). X-ray photoelectron spectra (XPS) were recorded on a spectrometer using Mg ka at 12 kV and 12 mA (XSAM800, Britain). The optical properties were tested by UV-Vis diffuse reflectance spectra from 200-800 nm (DRS, UV-3600 spectrophotometer, Japan) and photoluminescence spectra (PL, F-4600 spectrophotometer with a 150 W xenon lamp (excitation wavelength 300 nm), Japan). A V-sorb 2800S surface area analyzer (China) was used to measure the BET specific surface area. The pore size distribution and pore volumes were obtained using the Barrett-Joyner-Halenda (BJH) method.

Photocatalytic activity test
The photocatalytic property of the obtained photocatalysts was evaluated through the degradation of RhB in aqueous solution. 100 mL RhB solution (10 mg/L) and 0.1 g photocatalyst powder were added into a beaker, and then stirred 30 min in dark. Afterwards, a 250 W xenon lamp with the emission wavelength from 300 to 800 nm was turned on as light source. The degradation of RhB was tested by measuring the absorbance at 553 nm every 10 min. The degradation degree (D) was determined by the formula as follows: where A 0 and A t are the initial absorbance and absorbance at time ''t'', respectively.

XRD analysis
where K a (0.886) and K b (2.721) stand for correction coefficients. A a , A b and A r represent the peak intensities of anatase (101), brookite (121) and rutile (110) planes, respectively [20,21]. The crystallite size was calculated by Scherrer's formula. For anatase, rutile and brookite, b are the half widths of the diffraction peaks of anatase (101), rutile (110), and brookite (121) planes, respectively [22,23]. The phase composition and the average crystallite size of samples can be calculated and summarized in Table 1. The content of anatase decreases and the content of rutile increases after Cu doping, which shows that the addition of Cu is conducive to the transformation from anatase to rutile. Since Cu 2? /Cu ? ion radius (0.073/0.077 nm) is close to Ti 4? ion radius (0.0605 nm), Cu 2? /Cu ? can enter into TiO 2 lattice to replace Ti 4? ions and does not react with TiO 2 to generate new crystal phase. This is in line with the earlier literatures [1,13,14]. When reaching a certain concentration, the Cu element is dispersed in the form of oxide on the surface of TiO 2 particles. Therefore, the Cu-related diffraction peak can't be detected in XRD patterns.

Raman analysis
To further confirm the crystal structure of the asprepared samples, Raman measurement was employed to distinguish the three structures of TiO 2 and the results are shown in Fig. 2. The intense Raman bands located at 398 cm -1 , 517 cm -1 and 640 cm -1 ascribe to anatase structure [14,17]. The bands centered at 244 cm -1 and 444 cm -1 can be attributed to rutile structure, while the bands centered at 320 cm -1 and 363 cm -1 are for brookite structure [10,16,17]. The Raman spectra confirm that there are anatase/rutile/brookite three phases coexistence in pure TiO 2 and Cu-TiO 2 , which is in line with XRD results.    between surface hydroxyl groups and photoinduced holes, which possess strong oxidation ability. Therefore, more surface hydroxyl groups are beneficial to photocatalytic activity [27,28]. The ratio of surface hydroxyl groups can be calculated from the deconvolution area of the two peaks [28].

DRS analysis
The diffuse reflectance spectroscopy test was carried out to examine the optical property of obtained photocatalysts and the DRS spectra are exhibited in Fig. 5a. The Kubelka-Munk function for diffuse reflectance is F(R) = (1-R) 2 /2R [18]. Because F(R) is proportionality to the absorption coefficient a, it can be obtained according to the Tauc's formula [31][32][33]. The plots of (ahm) 1/2 versus the photon energy (hm) were demonstrated in Fig. 5b. The band gap energy values of pure TiO 2 , 1%Cu-TiO 2 , 2%Cu-TiO 2 , 3%Cu-TiO 2 and 4%Cu-TiO 2 are 3.06 eV, 3.07 eV, 3.02 eV, 2.96 eV and 3.04 eV, separately. It is worth noting that the absorption of 2%Cu-TiO 2 and 3%Cu-TiO 2 in the ultraviolet region is lower than pure TiO 2 , which hinders the absorption of the light source and is not conducive to photocatalytic activity.

PL analysis
Since photoluminescence emission is derived from the recombination of photoinduced holes and electrons, it can provide the accurate data of separation of photoinduced pairs [34][35][36]. The PL spectra of pure TiO 2 and Cu-TiO 2 are shown in Fig. 6 and it is found that the addition of Cu does not bring new emission peak but only changes the peak intensity. The main peaks around 412 nm can be ascribed to the band gap transition [36,37]. The peaks arranged from 450 to 470 nm are attributed to the transition of charge carriers [36,38]. All the Cu-TiO 2 samples show lower peak intensity than pure TiO 2 , which indicates that the recombination of photoinduced electrons and holes is hindered by Cu adding effectively. Remarkably, the PL intensity decreases with the increase of Cu concentration, which implies that high content Cu doping is favorable to suppress the recombination of photogenerated electrons and holes. Figure 7 presents the N 2 adsorption-desorption isotherms and the pore size distribution curves (inset of Fig. 7) of pure TiO 2 and Cu-TiO 2 . All the photocatalysts exhibit the type IV isotherms with the H2 hysteresis loop in the range of high relative pressure, which are the characteristic of mesoporous materials (2-50 nm) [1,10,27,37,39]. The pore size distribution curves indicate that both pure TiO 2 and Cu-TiO 2 have a narrow pore size distribution and the main pore diameters are ranged from 3 to 9 nm. The specific surface area, pore size and pore volume data are listed in Table 2. The specific surface area results reveal that Cu-TiO 2 samples show higher surface areas than pure TiO 2 . Sibu et al. [40] and Adyani et al. [1] believe that the surface textural property is improved owing to the presence of Ti-O-M (M is the doping element) bonds, which may restrain the conformity and rearrangement of primary crystals, leading to increase of surface area [26,33]. .5%, respectively. The photocatalysts show considerable photocatalytic activity, which attributes to relatively high specific surface areas ([ 140 m 2 /g) and the anatase/rutile/brookite triphasic structure. 1%Cu-TiO 2 shows higher photocatalytic efficiency than pure TiO 2 because of its less recombination rate and higher surface area compared to pure TiO 2 . However, as the amount of doping content increases, the degradation degrees slightly decrease in 2%Cu-TiO 2 and 3%Cu-TiO 2 samples. The effect of doping amount on suppressing the recombination of photoinduced pairs is controversial. Several studies show that there is an optimal concentration in doping, and the inhibition of photogenerated electrons and holes is weakened when the concentration exceeds the optimal value [41][42][43]. The authors believe that excessive doping content will generate new recombination centers, which is not conducive to the migration of photogenerated electrons and holes. Correspondingly, it has also been documented that the higher the doping amount, the higher inhibition effect [44][45][46]. The inconsistent results may be caused by different preparation methods and processes. In the present study, when the concentration of Cu/Ti reaches 4%, it is still conducive to the separation of photogenerated electrons and holes. Therefore, the decreased photocatalytic activity of 2%Cu-TiO 2 and 3%Cu-TiO 2 should not be ascribed to the enhancement of recombination rate. On the other hand, DRS results shows that the absorption of 2%Cu-TiO 2 and 3%Cu-TiO 2 in the ultraviolet region is lower than pure TiO 2 , which may be attributed to the fact that the excessive Cu doping content produces more CuO and Cu 2 O clusters on TiO 2 surface, decreasing the light utilization and photocatalytic efficiency [44]. However, if the specific surface area further increases, the disadvantage caused by CuO and Cu 2 O covering TiO 2 surface can be offset. Therefore, the   Figure 8b displays the kinetics fitting curves of ln (C/C 0 ) = -k t of photocatalysts [47]. The apparent first-order rate constants k of pure TiO 2 , 1%Cu-TiO 2 , 2%Cu-TiO 2 , 3%Cu-TiO 2 and 4%Cu-TiO 2 are 0.071 min -1 (R 2 = 0.948), 0.091 min -1 (R 2 = 0.977), 0.083 min -1 (R 2 = 0.971), 0.063 min -1 (R 2 = 0.948) and 0.084 min -1 (R 2 = 0.992), respectively.

The degradation mechanism
ÁO 2 radicals react with nitro-blue tetrazolium (NBT) to form purple precipitates, therefore, the lower NBT absorbance indicates more ÁO 2 radicals formation. 2,3-dihydroxybenzoic acid (2, 3-HBA) which has a special absorption at 510 nm is obtained by the reaction between salicylic acid (SA) and ÁOH radicals. The higher absorbance is, the more ÁOH radicals are produced [48,49]. The experiment results of 1%Cu-TiO 2 are shown in Fig. 9. The photogenerated electrons are excited to the conduction band after illumination and react with O 2 to form ÁO 2 radicals. The formed ÁO 2 radicals react with NBT to produce purple precipitates, resulting in the consumption of NBT and the decrease of NBT absorbance. Figure 9b shows that 2, 3-HBA absorbance increases with the increased time, indicating that ÁOH radicals are formed under illumination.
NBT and SA experiments confirm that ÁO 2 and ÁOH radicals are generated under illumination. Since ÁO 2 radicals come from the reaction of photogenerated electrons with O 2 , and ÁOH radicals originate from the reaction of holes with OH -, the numbers of ÁO 2 and ÁOH radicals can be used to measure the separation of photogenerated charges. Figure 10a shows that the NBT absorbance of pure TiO 2 is higher than that of 1%Cu-TiO 2 , which suggests that 1%Cu-TiO 2 generates more ÁO 2 radicals than pure TiO 2 . Figure 10b shows that the 2, 3-HBA absorbance of pure TiO 2 is lower than that of 1%Cu-TiO 2 , indicating that 1%Cu-TiO 2 produces more ÁOH radicals. It is can be concluded that 1%Cu-TiO 2 shows higher separation rate than pure TiO 2 . Cu doping enhances the quantum efficiency, which is consistent with the PL spectra.
To determine the active species in the process of photodegradation, isopropanol (IPA), benzoquinone (BQ) and ammonium oxalate (AO) were added to capture ÁOH, ÁO 2 and h ? , respectively. The degradation results of 1%Cu-TiO 2 in the presence of different scavengers are shown in Fig. 12. When BQ, AO and IPA were added, the degradation degree of 1%Cu-TiO 2 decreases from 93.5% to 36.0%, 68.2% and 91.4%, respectively. Obviously, the degradation degree decreases significantly with BQ adding and decreases slightly in the presence of AO and IPA, which indicate that O 2 radical is the main active species, and h ? and ÁOH are subsidiary in the degradation process.
Based on the experiment results, the photocatalytic degradation mechanism of 1%Cu-TiO 2 versus RhB is proposed in Fig. 13. When TiO 2 is exposed to light source, electrons in valence band (VB) are excited to conduction band (CB), producing corresponding holes in VB. The photoinduced electrons in anatase and brookite will transfer to rutile, which accelerates the movement of photoinduced electrons, prolongs the lifetime and inhibits the recombination of photoinduced pairs effectively [10,17]. The photogenerated electrons and holes react with O 2 and OHspecies, yielding ÁO 2and ÁOH radicals, respectively [50,51]. In addition, Cu 2? ions capture electrons in CB to form Cu ? ions, suppressing the recombination of photoinduced pairs and forming more radicals [44,52]. The formed ÁO 2 -, ÁOH radicals and h ? species have strong oxidizing properties and degrade RhB into small inorganic molecules [53].

Conclusions
In this work, pure TiO 2 and Cu-TiO 2 were synthesized by a hydrothermal method. XRD, Raman and HRTEM results confirm that the anatase/rutile/ brookite triphasic structure form in both pure TiO 2 and Cu-TiO 2 samples. SEM and TEM images show the particle sizes in the range of 10-20 nm. XPS results confirm that the Ti element exist in the form of Ti 4? and Cu is present in the Cu 2? and Cu ? oxidation states. Cu doping is beneficial to increase the surface hydroxyl group content of TiO 2 . PL spectra show that the recombination of photogenerated electrons and holes is inhibited effectively with Cu addition and the inhibition enhances with the increase of Cu doping concentration. Pure TiO 2 exhibits relatively high BET specific surface area (140.8 m 2 /g) and Cu adding can  further improve the specific surface area. The pore size distribution curves confirm that the obtained samples are mesoporous materials. RhB solution was employed as the target pollutant and the photocatalytic activity was tested under xenon lamp irradiation. ÁO 2 radicals play a major role in the photodegradation process. Triphasic 1%Cu-TiO 2 shows better photocatalytic property than monophasic/biphasic 1%Cu-TiO 2 . The degradation degree of 1%Cu-TiO 2 is 93.5% in 30 min, and the apparent first-order rate constant of 1%Cu-TiO 2 is 0.091 min -1 , suggesting that it is a promising photocatalyst for dye wastewater treatment.