Abstract

We employed high-energy ball-milling technique to fabricate TiO/ heterogeneous nanostructures. XRD proved the existence of TiO/ heterogeneous structures. SEM and HRTEM investigation evidenced that the mean particle size and mean grain size of the as-prepared samples are 23 nm and 13 nm, respectively. UV-Vis spectra exhibited that TiO has enhanced the visible light absorption of and has changed the of . UPS examination indicated that the electron work function (EWF) of TiO is higher than that of . Photocatalytic degradation experiments revealed that an appropriate TiO content can enhance the photocatalytic activity of pure anatase . The best photocatalytic activity of TiO/ heterogeneous nanostructures is even better than that of Au-deposited by keeping high degradation efficiency of 93%. The internal electrical field producing in TiO/ heterogeneous nanostructures was considered to be dominantly responsible for the enhanced photocatalytic activity. Therefore, the substitution of TiO with noble metal in will be widely used in the future due to its low cost. This study also provides a clear direction of enhancing photocatalytic activity of : incorporating a guest compound into with an appropriate content if the compound has much higher electron work function than that of .

1. Introduction

Anatase titanium dioxide () has been proved to be the most effective and suitable photocatalyst for the degradation of environment pollutants due to its low cost, nontoxicity, availability, relatively high efficiency, structural stability, and so forth [1]. However, its low photocatalytic activity limits the application. Therefore, it is a critical issue to enhance its photocatalytic activity to realize its photocatalytic application. From many works, it can be found that the photocatalytic activity of the photocatalysts is strongly dependent on their structures [24]. And it was also reported that nanosized heterogeneous structures can enhance the quantum yield and therefore can enhance the photocatalytic activity of anatase [5, 6]. The reason is that the heterogeneous structures in the interface produce internal fields and the recombination of photogenerated carriers can be suppressed [5, 6]. The most effective approach to suppress charge-carrier recombination in the photocatalysis process is to incorporate noble metal (Au, Ag, or Pt) into [710]. In these systems, the electron work functions (EWF) of the noble metals are much higher than that of . Therefore, larger internal fields can be produced in these systems. In addition, the photogenerated carriers tend to be transported from to the noble metals and easily transfer to the sites needed for the deoxidized reactions [11]. As a result, the photocatalytic activity of can be improved by deposited noble metals. From the above analysis, one can find a route to increase the photocatalytic activity of : incorporating a second compound which has higher EWF than and has high electrical conductivity.

In this work, we added TiO, rather than the noble metals, into commercial anatase to improve the photocatalytic activity of to lower the cost of photocatalyst. TiO has low cost, safety, and metallic properties [12, 13]. It is expected to have higher EWF than like most metals. If anatase is incorporated with TiO, it would have high photocatalytic activity of photocatalysis due to TiO/ heterogeneous structures. In order to mix TiO with particle evenly, a high-energy ball-milling process was adopted [14]. As presumed above, an enhanced photocatalytic activity has been found in the as-prepared TiO- system.

2. Experimental Procedures

2.1. Preparation of Photocatalyst

The planetary ball-milling apparatus used is commercially available (KEQ-0.4 L, Qidonghongchun Company, China). A zirconia () milling vessel and milling balls were used to grind the needed particles. The rotating speed was 800 rpm. The purchased TiO (CP, 99.9%, 1-2 m, Alden Company, China) and (CP, 99.9%, 1-2 m, Alden Company) were independently ball-milled for 60 hrs. Then the as-ball-milled TiO powders and powders were mixed according to the TiO-to- molar ratio: 1/1000, 10/1000, 20/1000, 50/1000, and 100/1000 (denoted as TT1, TT10, TT20, TT50, and TT100, resp.). The mixtures were followed by 60 hr ball milling in the above-mentioned ball-milling machine. In order to isolate the effects of ball milling from that of TiO on photocatalytic activity of the , a pure TiO (denoted as “PTiO”) and pure (denoted as “”) ball-milled for 120 hrs were prepared as comparisons.

2.2. Photocatalyst Characterization

The X-ray diffraction (XRD, X’Pert Pro, Panalytical) was employed to investigate the phases of as-ball-milled powders with slow scanning speed (2° min−1). The scanning electron microscopy (SEM, Zeiss Ultra 55, Carl Zeiss) was used to examine the micrographs of the particles of as-prepared powders. The high resolution electron microscopy (HRTEM, JEM-2100HR, JEOL) was utilized to observe the grain size. UV-Vis diffuse reflection spectra (DRS) were obtained using a spectrophotometer Shimadzu 2550PC and the reflection data were converted to absorbance through the standard Kubelka-Munk method. The X-ray photoelectron spectroscopy/ESCA (Esclab 250, Thermo Fisher Scientific) was obtained by using UPS mode and He (I) irradiation to examine the EWF and valence band maximum (VBM) of as-ball-milled powders. We also measured the electrical properties (ZEM-3, Ulvac-Riko) of as-ball-milled powders by testing the pressed powders (at room temperature, 200 MPa).

2.3. Photocatalytic Experiments

The photocatalytic activity of the samples was evaluated by measuring the degradation rates of methylene orange (MO) solution (10 mg L−1) under ultraviolet-visible (UV-Vis) light irradiation [15]. The MO concentration was monitored by a UV-Vis spectrometer. The photocatalytic reaction was performed in a quartz reactor (XQ350W, Shanghai Lansheng Company). A 350 W xenon-arc lamp, with similar emitting spectrum to the sun, was used as light resource, which was located at a distance of 10 cm from the quartz reactor. A stirring bar, MO aqueous solution, and dispersion solution were placed in a quartz cuvette. Before photogradation, the solution was stirred in dark for 40 mins to achieve adsorTiOn/desorTiOn equilibrium. The irradiation time ranged from 0 to 120 min. At given time intervals (15 mins), about 3.5 mL of MO solution was taken out and centrifuged for 15 min. After removing the nanoparticles by filtration, the MO solution was measured by the UV-Vis spectrophotometer. The concentration of the MO solution was analyzed by checking the absoption peak around 464 nm, which is attributed to the Azo functional group (–N=N–) of MO [15]. According to the Beer-Lambert law, the absorbance of MO solution is proportional to its concentration. The change of MO concentration can thus be evaluated by the intensity change of the absoption peak around 464 nm.

The degradation rate () of MO is expressed in the following equation [16, 17]: where is the concentration of MO after absoption equilibrium in the dark, is the corresponding absorbance intensity of MO after absoption equilibrium in the dark, is the concentration of MO at reaction time (min), and is the corresponding absorbance intensity of MO at reaction time (min).

3. Results and Discussion

3.1. Phases, Grain Size, and Particle Size

Figure 1 showed the XRD patterns of the as-prepared samples. It shows that there is only anatase phase (JCPDS file number 21-1272) in while in TiO-incorporated the majority is anatase and the minority is the incorporated TiO (JCPDS file number 08-0117). And the peak positions of anatase have not been changed by the incorporated TiO. The diffraction peaks at () 25.4°, 37°, 37.9°, 38.7°, 47.9°, 53.9°, 54.9°, 62.6°, 68.8°, 70.3°, 75°, and 76.2° are attributed to (101), (103), (104), (112), (200), (105), (211), (204), (116), (220), (215), and (301) faces of anatase , respectively. The diffraction peaks at () 37.2°, 43.2°, 62.8°, 75.4°, and 79.4° can be ascribed to (111), (200), (220), (311), and (222) faces of TiO. It can be found that the more the TiO contained in TiO-incorporated is, the higher the diffraction peak at 43.2° representing TiO is. Clearly, the incorporated TiO has not changed the peak positions of anatase . TiO and just mixed with each other and formed a large number of heterogeneous structures TiO/ in host powders after the high-energy ball milling.

The grain size of the as-prepared samples (Table 1) was estimated according to the Debye-Scherrer equation [18]. From Table 1, it can be seen that there is no effect of the incorporated TiO on the grain size of anatase . All the grain sizes of the as-prepared samples are 13~14 nm. The grain size observed under HRTEM (Figure 2) is also about 13 nm. The two data agree very well.

SEM investigation (Figure 3) shows that the mean particle size of the as-prepared samples is 23 nm. And all the particle sizes of the as-prepared samples are very close. Therefore, the incorporated TiO has no effects on the particle size of anatase .

3.2. Energy Band Characterization: UV-Vis Absoption, Energy Band Gap (), EWF, and Electrical Conductivity ()
3.2.1. UV-Vis Spectrum and

Figure 4 shows the UV-Vis absoption spectra of the as-prepared samples in the range of 220~800 nm obtained by calculating the absoption coefficient from the obtained DRS according to [19] where is the absolute reflection ratio (%). In our work, we substitute with which is the relative reflection ratio by comparing the diffuse reflection ratios (DR) of the as-prepared samples with that of pure ; that is,

Figure 4 indicates that the A-s of as-prepared samples in visible light (VL) spectrum increase with increasing TiO and increase with increasing wavelength of VL. The increasing absoption with increasing wavelength is characteristic of absorbing properties of free carriers because TiO has metallic properties. But generally, free carriers can strongly absorb infrared light (intraband free-carrier absoption (FCA)) not visible light and are characteristic of increasing absoption with increasing wavelength [20]. Apparently, the intraband FCA has been blue-shifted to the VL range. This implies a great extension of conduction band in the TiO-incorporated . The blue shift of intraband FCA could be related to the internal fields which exist in TiO/ heterogeneous nanostructures. Detailed discussion will be performed in the next text. The absoption of the as-prepared samples in VL suggests that the incorporated TiO can extend the light absoption region of to VL and improve the UV-Vis absoption efficiency of the samples.

Because there is no strong energy dependence of optical absoption coefficient of the as-prepared near absoption edge in Figure 4, the as-prepared samples should have direct band gap [21, 22]. In other people’s works, the nanosized anatase was also found to have direct band gap [23]. Furthermore, anatase is an N-type semiconductor [24, 25]. The optical energy band gap (, eV) can be estimated by using the Kubelka-Munk function (), as expressed by

We substituted with . The band gap (, eV) was the crossing point between the line extrapolated from the linear part and the -axis of the plot of as a function of energy of light (, eV), that is, curve (Figure 5) [23]. Table 2 is the -s of the as-prepared samples (excluding PTiO since it is metallic conduction). It can be seen that all of the -s of the as-prepared samples are larger than 3.4 eV. The maximum is that of TT10, 3.504 eV. The -s of the as-prepared samples would increase with increasing TiO at first and then decrease with increasing TiO. We have not found the exact reason for the -s variation of the as-prepared samples with TiO content. But we guess that the probable reason is that the internal fields from the nanosized TiO/ heterogeneous structures have bent the energy band and energy gap [11].

It is known from the previous work that the optical of is related to the measuring method (absoption spectrum/reflection spectrum), doping, processing technique, material form and particle size, and so forth. Therefore, the reported optical -s of range widely, for example, from 2.99 eV to 3.6 eV [2629]. But the most reported optical is 3.0–3.2 eV [26, 27]. The large in this work mainly results from the small grain size of 13 nm in the as-prepared samples, which has strong size effects [29, 30]. But the variation of with the content of TiO cannot be ascribed to size effect because the grain size and particle size of all the samples are very close. And the processing procedure and parameters are identical among all the samples. The only different factor is the content of the incorporated TiO. Therefore, the internal fields generated due to heterogeneous structures TiO/ should be responsible for it as discussed above.

3.2.2. Electrical Conductivity and EWF

Table 3 gives the and the Seebeck coefficients () of the as-prepared samples. Because some samples could not be detected by ZEM-3 due to too low electricity, their -s and are denoted by “N.” From Table 3, it can be seen that only the of PTiO can be detected by ZEM-3 due to loose contact between particles in the as-prepared samples since the bulk samples for measuring electrical properties are pressed at room temperature with pressure of 200 MPa. The tested of PTiO is 86.5 S m−1 and the is close to 0 with negative sign. From the sign of PTiO, it can be seen that PTiO is a weak N-type semiconductor, suggesting a little bit oxidization in PTiO during high-energy ball milling. On the other hand, previous work presents the fact that pure anatase PTiO2 is a natural N-type semiconductor [24, 25].

Figure 6 is the UPS spectra for PTiO and TT20. The spectra have been referenced to the Fermi level () of the gold reference sample. During UPS experiment, a bias voltage of −5 V has been added. The EWF, which is the minimum energy required to move electrons at the Fermi energy level from inside a metal or semiconductor to its surface with zero-kinetic energy, equals to the difference between the onset of the measured UPS spectra and 5 eV. While the valence band maximum () equals 26.22 eV, the cutoff on the high energy side of UPS spectra. Figure 6(a) combines the UPS spectra of PTiO and those of TT20. Figures 6(b) and 6(c) show the five fitted Gaussian peaks of UPS spectra for PTiO and TT20, respectively. The fitted peaks were based on the expected contributions from the molecular orbitals that comprise the valence bonding states, respectively [31]. Figure 6(a) indicates that valence band of PTiO has been apparently changed after being ball-milled with . This is another proof that the energy band of has been changed by the incorporated TiO. The EWF of PTiO is 3.01 eV, while that of TT20 is 1.54 eV. The of PTiO is 2.67 eV, while that of TT50 is 3.08 eV. Apparently, the EWF of is different from that obtained in other works [31]. It can be attributed to the band bending arising from nanosized TiO/ heterogeneous structures [11, 32]. We guess that the EWF of is less than that of PTiO. In the nanosized heterogeneous TiO/ structures, the is aligned to the level which is between the EWF of PTiO and that of . Therefore, the EWF of TT20 is less than that of TiO and larger than that of . The above-mentioned two proofs that there are internal fields are powerful support for the assumTiOn about the EWF of TiO and . Actually, the presence of a large number of space charge regions has made the crystal field and the EWF of free electrons, of TT20 different from that of pure TiO and that of pure [32].

3.3. Photoactivity

Figure 7 is the photocatalytic degradation of MO solution (10 mg L−1) by the as-prepared samples under UV-Vis irradiation. It can be seen from Figure 7(a) that, at the beginning irradiation for 70 mins, the degradation efficiencies (DE) reduce as the order of TT20, TT10, TT50, and TT1 (very close), , and PTiO. Among the six samples, the PTiO possesses far lower DE than the others. However, the maximum DE of all samples except PTiO reach an identical value 93% after 75 mins. Therefore, we can conclude that the presence of the incorporated TiO has not enhanced the final DE of nanosized but has improved the photocatalytic activity of the nanosized samples. It can be clearly seen that, before 38.5 mins, the PDRs of MO solution by the as-prepared samples reduce as the same order as that of DE, that is, TT20, TT10, TT50, and TT1 (very close), TT100, , and PTiO. But after the irradiation for 38 mins, the order of PDR is turned over as , TT100, TT50, and TT1, and TT20 (PTiO is an exceTiOn and always is close to 0). This phenomenon can be understood from the view of point that the PDR is related to the MO concentration: when the MO solution is low, the larger the MO concentration, the larger the PDR [33, 34]. After 38 mins, the residual MO concentration photocatalyzed by the as-prepared samples (excluding PTiO) reduces as the order of , TT100, TT50, and TT1, and TT20, the same order as that of PDR.

We here explain why TT20 has a maximum DE and PDR before 38 mins. We think there are three dominate mechanisms playing on the DE of . One mechanism is that larger would result in reducing photogenerated carriers while less would increase the recombination possibility of photogenerated electron-hole pairs. The second mechanism is the generated internal electrical fields in the nanosized heterogeneous TiO/ structures [31]. The internal electrical fields are favors in separating photogenerated electron-hole pairs and prolonging the lifetime of photogenerated carriers [31]. But too much TiO might add the recombination sites of photogenerated electron-hole pairs [35]. The third mechanism is that the more the incorporated TiO is, the larger the Vis absoption is. The combined action of the three mechanisms has resulted in that the DE and PDR of TT20 are the maximum among the as-prepared samples.

Although [36] mentioned that nitrogen- (N-) doped TiO with 8 nm has efficient photoactivity under UV and visible irradiation, the nanosized TiO with 13 nm in our work shows no considerable photoactivity under UV-Vis irradiation. The considerable response to UV-Vis in [36] probably results from the dopant N. It was reported that the dopant N can add additional energy levels in energy band gap and improve the photocatalytic properties of [37]. Here we think that similar additional levels appear in N-doped TiO and lead to the UV-Vis response.

TiO has much higher EWF than that of . Thus, it introduces an easy route for carriers to transfer from the inside to the outside of particles to participate in photocatalytic activity. Consequently, an appropriate content of the incorporated TiO can increase the photocatalytic activity of anatase . Furthermore, TiO can enhance the Vis absoption. TiO-incorporated has maximum degradation efficiency of 93%. However this value is less than that in noble metal incorporated , for example, 97% in Ag-incorporated system and 100% in Au-incorporated [7, 9, 38]. But the photocatalytic activity of TiO-incorporated is much higher than that of Ag-incorporated . In [7, 38], Ag-incorporated took 2 hrs to reach DE of 97% and Au-incorporated TiO2 took 50 mins to reach DE of 100%. While in this work, TT20 just took 38 mins to reach DE of 93%. Therefore, the TiO-incorporated has much application value.

4. Conclusion

In this work, we employed high-energy ball-milling method to fabricate TiO/ heterogeneous nanostructures. XRD proved the existence of TiO/ heterogeneous structures. SEM and HRTEM investigation evidenced that the mean particle size and mean grain size of the as-prepared samples are 23 nm and 13 nm, respectively. UV-Vis exhibited that the incorporated TiO has enhanced the Vis absoption of and has changed the of . UPS examination found that the electron work function of TiO is higher than that of . Photocatalytic experiment has revealed that an appropriate TiO content can lift the photocatalytic activity of pure to a level better than that of Au-incorporated by keeping high degradation efficiency of 93%. The internal electrical field produced by TiO/ heterogeneous nanostructures was thought to be dominantly responsible for the enhanced photocatalytic activity. Therefore, the substitution of the noble metal with TiO to enhance the photocatalytic activity of has much application value. This work also gives a clear direction of enhancing photocatalytic activity of by incorporating a second compound into particle with an appropriate content if the second compound has much higher electron work function than that of .

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Grant nos. 51172078 and 51372092) and the Guangzhou Science and Technology Project of China (Grant no. 2013J4100045).