Neodymium-Doped TiO 2 with Anatase and Brookite Two Phases : Mechanism for Photocatalytic Activity Enhancement under Visible Light and the Role of Electron

Titanium dioxide (TiO2) doped with neodymium (Nd), one rare earth element, has been synthesized by a sol-gel method for the photocatalytic degradation of rhodamine-B under visible light. The prepared samples are characterized by X-ray diffractometer, Raman spectroscopy, UV-Vis diffuse reflectance spectroscopy, X-ray photoelectron spectroscopy, and Brunauer-Emmett-Teller measurement. The results indicate that the prepared samples have anatase and brookite phases. Additionally, Nd as Nd3+ may enter into the lattice of TiO2 and the presence of Nd3+ substantially enhances the photocatalytic activity of TiO2 under visible light. In order to further explore the mechanism of photocatalytic degradation of organic pollutant, photoluminescence spectrometer and scavenger addition method have been employed. It is found that hydroxide radicals produced by Nd-doped TiO2 under visible light are one of reactive species for Rh-B degradation and photogenerated electrons are mainly responsible for the formation of the reactive species.


Introduction
Dye wastewater discharged into nature mainly by dyestuff, textile industry, and some artificial way causes severe ecological problems.These compounds are highly colored and can heavily contaminate water source.Chemical oxidation method for dye wastewater treatment is too costly and physical adsorption method often results in secondary pollution [1].Many attempts have been carried out to develop efficient biological methods to decolorize these effluents but they have not been very successful.Semiconductor TiO 2 as a photocatalyst has been deeply investigated and can successfully degrade various organic pollutants [2][3][4][5][6].However, the overall efficiency achieved so far with TiO 2based systems is not sufficiently high to enable practical applications.This low efficiency is mainly due to the fast recombination of charge carriers produced by irradiated TiO 2 and thus the low quantum yield in the generation of reactive species for organic degradation [7,8].Moreover, due to its large bandgap (∼3.2 eV), all photo-driven applications of TiO 2 require ultraviolet light excitation.As a consequence, TiO 2 shows photocatalytic activities under only a small fraction (<5%) of solar irradiation [9], limiting its practical applications.Therefore, the modification of TiO 2 to enhance light absorption and photocatalytic activities under visible light has been the subject of recent research [10].
Many factors that influence the photodegradation efficiency of organic pollutants over TiO 2 have been reported during the past 30 years.The main focus is the physical properties of TiO 2 , including crystal phases, crystal facets, crystallinity, particle size, surface area, porosity, and morphology.It is not easy to make a reliable correlation between the structure and photoactivity of TiO 2 because the solid physical properties often affect one another.In general, anatase is considered to be much more active than rutile, while a mixture of anatase and rutile, like Degussa International Journal of Photoenergy P25, is claimed to be more active than anatase.So far, TiO 2 with anatase and brookite phases as a photocatalyst has been rarely studied because brookite TiO 2 has no activity for organic pollutant degradation and it needs higher temperature to prepare.Moreover, there has been little report about brookite TiO 2 preparation at low temperature.
The investigations about introducing foreign species are in progress for improving the photocatalytic activity of TiO 2 and broadening its absorption to solar spectrum.Among them, titania doped with metals or metallic cations such as transition metallic cations, rare metal cations, and noble metal have been widely explored [11,12].On one hand, these doped metallic cations tend to serve as recombination centers.As a result, in most cases, photoexcited charges are recombined by the sites of doping metallic cations [13].On the other hand, it has been reported that suitable amount of metallic cation doping can promote the separation of photogenerated electrons for the improvement of photocatalytic activity [14,15].It is obvious that there is conflict.So, although there are quite a few publications about metal-doped TiO 2 , the mechanism for the improvement of photocatalytic activity has not been clear so far.It needs explore in detail.
In this paper, to explore the mechanism of photodegradation of organic pollutant by rare earth element-doped TiO 2 , neodymium (Nd) ion was selected as a TiO 2 dopant due to its stability to form complexes with various Lewis bases in the interaction of these functional groups with f-orbitals of neodymium metal [16].Thus, incorporation of Nd ion into a TiO 2 matrix could provide an effective method to concentrate the organic pollutant at semiconductor surface.Here, Nd-doped TiO 2 samples were prepared by sol-gel method together with pure TiO 2 for comparison.The characteristics and properties of these photocatalyst were studied.The photocatalytic activity was evaluated by measuring photodegradation efficiency of rhodamine-B (Rh-B) under visible light.The photocatalytic mechanism of Nd-doped TiO 2 was also investigated in detail by photoluminescence spectrometer (PL) [17] and scavenger addition method [18].

Experimental Section
2.1.Materials.The chemicals used in this experiment except P25 were all analytical reagent and purchased from Shanghai Guoyao Chemical Co.Pure TiO 2 , P25, was commercial product from Germany.The used water throughout the whole research was double-distilled water.

Experimental
Procedure.Neodymium-doped TiO 2 samples (Nd-doped TiO 2 ) were prepared by sol-gel method.First, 5 mL of TiCl 4 was dissolved and hydrolyzed with 200 mL distilled frozen water (0 • C).Second, Nd 2 O 3 with certain amount required for doping was suspended into a small amount of ethanol and then added to the above solution to produce transparent Nd 3+ aqueous solution under vigorous stirring.Third, 10 M NaOH aqueous solution was added dropwisely into the transparent TiCl 4 aqueous solution with Nd 3+ to obtain a grey precipitate with an ultimate suspension of pH = 10.
In order to remove residual Na + and Cl − ions, the precipitate was adequately washed with deionized water till the pH value of filtrated water was below 7.Then, the amorphous Nd-doped TiO 2 was well dispersed into 400 mL of water, and chloride acid (20%) was added with corresponding amount.The above suspension was adjusted to pH 1.5, stirred for 4 hours at room temperature, and then had been aged at 70 • C for 24 h in airproof condition.Finally, Nd 3+ -modified TiO 2 sol was formed with uniform, stable, and semitransparent characteristics.The obtained sol can maintain homogenous distribution for a long time without sedimentation and delamination.Powder sample was prepared by aging, gelation, and vacuum-drying treatment of the above sol at 70 • C for 8 hours.TiO 2 sample was also prepared by the same procedure without the addition of the neodymium oxide suspension.The pure TiO 2 and Nd-doped TiO 2 samples were annealed in air at 400 • C for 3 hours and ground to fine particles for different analysis.

Characterization of Photocatalysts.
The crystalline phases of the synthesized TiO 2 and Nd-doped TiO 2 catalyst were analyzed by a Y-2000 diffractometer (D/max 30 kV) using graphite monochromatic copper radiation (Cu Kα) (λ = 0.154178 nm) at a scan rate of 0.06 • 2θ•s −1 at 40 kV, 40 mA over the diffraction angle range of 10 ∼ 60 • .Raman spectroscopy measurements were performed by using a Jobin Yvon Lab RAM HR 800UV micro-Raman system under an Ar + (514.5 nm) laser excitation.UV-Vis diffuse reflectance spectra (DRS) were recorded with a PerkinElmer Lambda 35 Spectrophotometer.X-ray photoelectron spectra (XPS) measurements were performed in a PHI Quantum 2000 XPS system with a monochromatic Al Kα source and charge neutralizer to analyze surface chemicals and evaluate the amount and states of Nd atoms in the prepared samples.Nitrogen adsorption/desorption isotherms and Brunauer-Emmett-Teller (BET) specific surface area (S BET ) were recorded on a Micrometrics ASAP 2010 analyzer (accelerated surface area and porosimetry system).All the samples were treated at 100 • C prior to BET measurements.Barret-Joyner-Halender (BJH) method was used to determine pore size distribution.The formation of hydroxyl radicals ( • OH) on the surface of prepared samples under visible light was detected by a terephthalic acid (TA) probe method [17].The PL spectra of generated 2-hydroxyterephthalic acid (TAOH) were measured by using a PerkinElmer Lambda 55 fluorescence spectrophotometer.

Photocatalytic Activity Measurement.
The photocatalytic activities of the prepared samples were examined by the degradation of Rh-B under visible light.A 350 W Xenon lamp (Lap Pu, XQ) was used as a light source with a 420 nm cutoff filter right above the reactor to provide visiblelight irradiation.The experiments were performed in selfconstructed beaker-like glassware reactor with double walls for cooling system.0.1 g of prepared sample was dispersed in 100 mL Rh-B solution with concentration of 10 μM.Prior to irradiation, the solution was stirred in dark for 30 minutes to reach an adsorption/desorption equilibrium.Then, the solution was exposed to light while stirring.1 mL suspension was collected every 30 minutes and was centrifuged for solid-liquid separation.Rh-B residue concentration was determined by measuring characteristic absorption intensity of Rh-B absorption spectrum over a Shimadzu UV-1700 UVvis spectrophotometer.
Scavenger addition method [18] was used to explore the mechanism of photodegradation of Rh-B over Nd-doped TiO 2 .The whole process was almost the same as that of the above photocatalytic activity measurement except with the presence of different scavengers: sodium oxalate (Na 2 Cr 2 O 4 ) as hole scavenger, KI as hole, and • OH scavenger and K 2 Cr 2 O 6 (Cr (VI)) as electron scavenger.

Results and Discussion
3.1.Characterization of TiO 2 and Nd-Doped TiO 2 Photocatalysts.The degree of crystallinity and crystal structure of the prepared TiO 2 and Nd-doped TiO 2 samples were examined by XRD.The diffractograms recorded are shown in Figure 1.It reveals that the samples had anatase and brookite phases, a mixture of crystal structure.The phase content for the prepared samples is displayed in Table 1.With the increase in Nd doping amount, brookite phase content increased.In this study, the addition of NaOH to reaction solution increased pH value, leading to the increase of OH − concentration.This might promote the formation of brookite phase because its complex consumed the amount of OH − twice as much as rutile's one.However, too much NaOH could result in the formation of amorphous TiO 2 which contained much more OH − than brookite complex [19][20][21][22].The grain sizes of the prepared TiO 2 and Nd-doped TiO 2 powders were calculated using Scherrer equation: where D is crystalline size, λ the wavelength of X-ray radiation (0.1541 nm), K the constant usually taken as 0.89, and β the peak width at half-maximum height and θ diffraction angle.The obtained crystalline sizes are shown in Table 1.There was almost no change for crystallite size when Nd was incorporated (20.32 nm for pure TiO 2 and 20.35 nm for Nd-doped TiO 2 with Ti : Nd = 11 : 1).The phase content of TiO 2 can be calculated from the integrated intensities of anatase (101), rutile (101), and brookite (121) peaks with the following formulas [23]: where  1. Substantially, the enlarged peaks at (101) (JCPDS number: 21-1272) and (121) plane for the three samples (Figures 1(B) and 1(C)) showed a slight shift to smaller angles with Nd incorporation.It indicates that Nd ion could enter into TiO 2 lattice or interstitial site.We know that the difference in ionic radius between neodymium ion (Nd 3+ = 0.11 nm) and titanium ion (Ti 4+ = 0.064 nm) is large.Without calcination treatment in high temperature, neodymium ions introduced by the coprecipitation-peptization method would unlikely enter into TiO 2 crystal structure.Actually, during TiCl 4 hydrolysis and peptization reaction, neodymium and oxygen ions could form Nd-O oxide on the superficial layer of TiO 2 [19,20] particle by chemical bonding process and Nd 3+ ions mainly existed as Nd-O-Ti in Nd-doped TiO 2 mixture [25].So, the conversion from amorphous to well-crystalline structure usually requires high annealing temperature of at least 400 • C.
The XRD studies revealed the anatase and brookite structure of the prepared samples.The formation of the structure was further confirmed by Raman spectroscopy.In Figure 2(A), Raman spectra of pure TiO 2 and Nddoped TiO 2 are shown.It can be seen that the peaks observed at around 144.86, 398.17, 515.33, and 637.86 cm −1 were attributed to anatase TiO 2 .The strongest Eg mode at 144.86 cm −1 arising from the external vibration of the anatase structure was well resolved, which indicates that anatase phase was formed in the as-prepared nanocrystals [24].Its amplified figure is shown in Figure 2(B).It can be found that there was a distinct shift to high wave number after Nd doping, which is also an evidence to verify that Nd may be introduced into the lattice or interstitial site of titania.Additionally, there were three weak peaks at 240.22, 317.01, and 360.44 cm −1 attributed to vibration modes for brookite phase of TiO 2 in Figure 2(A).The absence of the characteristic vibration modes of Nd or Nd 2 O 3 in the Raman spectra suggests that there was no Nd 2 O 3 segregation into TiO 2 .
The UV-Vis diffuse reflectance spectra (DRS) for the doped and the undoped TiO 2 nanoparticles samples are presented in Figure 3. Modification of TiO 2 with Nd significantly affected the light absorption property of the photocatalysts.A red shift of the absorption edge toward the visible region was observed for Nd-doped TiO 2 compared with pure TiO 2 .From XRD and Raman data, we can see that Nd doping led to a little change of TiO 2 crystal structure and there was a little shift for the main peaks.The change of TiO 2 crystal structure may be due to the incorporation of Nd ion into TiO 2 lattice.It can result in new energy level in the bandgap and charge transfer between TiO 2 valence band and Nd ion doping level [21].As a result, the Nddoped TiO 2 had lower bandgap and the presence of Nd in the doped photocatalyst facilitated visible-light absorption.By plotting (αE) 1/2 versus E with α the absorption coefficient, the bandgap was calculated to be 3.   for that with Ti : Nd = 11 : 1.So, with the presence of Nd ion in TiO 2 lattice, the bandgap of the materials decreased distinctly.Additionally, with the increase of Nd doping amount, the bandgap decreased gradually.It indicates that the Nd-doped TiO 2 samples may have the higher ability to absorb visible light.
XPS technique is used to investigate chemical component at the surface of a sample.Figure 4 shows high-resolution XPS spectrum for Nd in the doped TiO 2 sample.Although the peak for Nd was not so distinct, it can be confirmed that Nd was present in the doped sample.Figure 5  energy had 0.49 eV shift for Ti 2p 3/2 and 0.65 eV for Ti 2p 1/2, indicating the presence of Ti 4+ and Ti 3+ .Besides, the graph of doped sample was fitted by three subpeaks with the binding energy of 460.05, 458.86, and 458.27 eV, which indicates the presence of Ti 4+ , Nd-Ti, and Ti 3+ , respectively.So, the complete incorporation of Nd into TiO 2 lattice can be verified.The XPS spectra of the O 1s region of the pure TiO 2 could be fitted into two subpeaks at 529.36 and 531.08 eV (Figure 6(a)), corresponding to the Ti-O bond in TiO 2 and hydroxyl groups on the surface, respectively.However, the O 1s peak for Nd-doped TiO 2 was located at 530.21   surface, including Ti-O in TiO 2 , Nd-O, and hydroxyl group.These hydroxyl groups as hole trapped sites may inhibit the simple recombination of electron hole [29].Surface textural characteristics of the samples pure TiO 2 and Nd-doped TiO 2 are derived from N 2 adsorption analysis.Specific surface area (S BET ) by BET method, total pore volume calculated at P/P 0 = 0.99, and average pore diameter values are presented in Table 1.The adsorption isotherms (Figure 7) of the samples showed type IV behaviour with the typical hysteresis loop.This hysteresis is characteristic of mesoporous materials [30,31].The pore size distributions for the pure TiO 2 and Nd-doped samples are shown in Figure 8, which confirms the mesoporous nature of the samples.It can be seen in Table 1 that the doped sample had lower surface area and narrower average pore size distribution than the undoped one.This might be due to the slight increase of crystalline size for the doped titania mentioned in XRD pattern.

Photocatalytic Activity of Prepared Samples.
The photocatalytic property of titania is known to depend on several factors like crystallinity, phase assemblage, and surface area [32].The photocatalytic activity of neodymium-doped and undoped titania was studied through Rh-B degradation under visible light in comparison with commercial product P25.Figure 9 shows the variations of Rh-B concentration against irradiation time with the presence of photocatalysts.It can be seen that the concentration of Rh-B decreased gradually with the exposure time for the pure and doped TiO 2 samples.Without the presence of photocatalysts, almost no Rh-B could be degraded.The results illustrate that all of the prepared samples had the ability for Rh-B degradation under visible light even for pure TiO 2 and P25.Additionally, with the increase of Nd doping amount, the photocatalytic activity for the Nd-doped TiO 2 increased.However, when the amount of Nd in the doped TiO 2 was too much (Ti : Nd ≥ 6 : 1), the activity was even lower than that of pure TiO 2 and P25.So, there is an optional Nd doping amount, which is in accordance with the published studies [14,33].Nevertheless, the optional Nd doping amount is 0.5% [14] and 1 ∼ 3 wt% [33] in the two papers.The amount calculated by raw reagents for the formation of Nd-doped TiO 2 samples was about 9%, which is much higher.In fact, there should not be so much Nd incorporated into TiO 2 in this case.The measurement of actual Nd amount in TiO 2 is in progress.

Mechanism for Photocatalytic Activity Improvement by
Nd Doping.From Figure 9, we can see that even pure TiO 2 could degrade Rh-B under visible light.It is known that Rh-B can be excited by visible light and inject electrons to the conduction band of TiO 2 .So, the injected electrons react with O 2 molecules adsorbed on TiO 2 surface to yield • O 2 − radical anion and subsequently HO • radical by protonation [34].So, in this case, Rh-B can be degraded in even undoped TiO 2 and P25 systems although their photodegradation efficiency is not high.From Table 1 and Figure 7, we can see that the surface area of Nd-doped samples was lower than that of pure TiO 2 although all of them had mesoporous structure.So, based on dye and O 2 adsorption, there is no advantage for Nd doped samples compared with pure TiO 2 .As discussed in the XRD patterns and XPS spectra, there is Nd which may substitute titanium in the lattice and exist as the state of Nd 3+ although the ion radius of Nd 3+ is much larger than that of Ti 4+ .The doping energy level of Nd 3+ /Nd 2+ is −0.4 eV [35], which is more positive than the potential of conduction band of TiO 2 particles {E cb = −0.5 eV versus NHE (normal hydrogen electrode) at pH = 1} [36].Therefore, after Nd doing, the electrons can be excited from valance band to the Nd 3+ /Nd 2+ doping energy level under visible light.The UV-Vis spectra verifies that the Nd doped TiO 2 had lower bandgap than TiO 2 and was responsive to visible light (Figure 3).Besides, the energy of Nd 3+ /Nd 2+ is below the conduction band of TiO 2 , so that it is easy for Nd 3+ to capture injected electron by Rh-B from the conduction band.The electrons on the Nd 3+ /Nd 2+  It can be seen from Figure 9 that the Nd-doped TiO 2 with Ti : Nd = 6 : 1 shows a negative effect on the photodegradation of Rh-B.In fact, metal ion dopant can act as a mediator of interfacial charge transfer or act as recombination center.So, there is an optimal value for dopant concentration [37].
Here, the doping energy level of Nd 3+ /Nd 2+ has three functions: firstly, it can capture electrons from the conduction band of TiO 2 that injected from the dye, and thus the number of hydroxide radicals is reduced; secondly, electrons can be excited to the Nd 3+ /Nd 2+ energy level from the valence band of TiO 2 , and then the holes leaving on the valance band can degrade dyes; thirdly, Nd 3+ /Nd 2+ energy level can be the recombination center if the dopant concentration is not proper.These three processes compete with each other, so it is critical to find an optimal dopant concentration.In our experiment, from the result of photocatalytic activity, we can deduce that Nd doping concentration of 4 ∼ 9 at.%, resulting in the good photocatalytic performance of Nd-doped TiO 2 .
From the discussed above, we know that Nd doping level exists between bandgap of TiO 2 and is close to the conduction band, which corresponds to the calculation results [14].So, with Nd-doped TiO 2 as a photocatalyst irradiated under visible light, there are photogenerated holes and electrons which participate in the Rh-B degradation.It is well known that in liquid photocatalysis system, hole may oxidize organic pollutant directly or form hydroxide radical for degradation, and photogenerated electron can form • OH through the reaction with adsorbed • O 2 − .In order to explore which, photogenerated hole or electron, is mainly responsible for Rh-B degradation, some scavengers were used to investigate the specific reactive species that may play important roles in this process [18].Na 2 Cr 2 O 4 was used as hole scavenger, KI as the scavenger for hole and • OH, and Cr (VI) for electron.Figure 10 shows the photodegradation efficiencies of Rh-B over Nd-doped TiO 2 (Ti : Nd = 11 : 1) in the presence of the scavengers under visible light.It can be seen that when Na 2 Cr 2 O 4 was used as diagnostic tool for suppressing the hole process, the photocatalytic degradation of Rh-B was inhibited but not down close to zero as shown in Figure 10.From this inhibited effect, it can be deduced that photogenerated holes played a role in the photodegradation of Rh-B under visible light.In the presence of KI as hole and • OH scavenger, the removal efficiency of Rh-B was almost the same as that Na 2 Cr 2 O 4 .It indicates that • OH played a significant role in the system since the inhibition effect of • OH can be regarded to be close to 100% after the deduction of hole effect.The addition of Cr (VI) as electron scavenger resulted in the degradation of only 6% of Rh-B after 90 min, which means that there was almost no activity for the photocatalyst with electron inhibition effect in the system.So, photogenerated electrons played a very important role in the degradation system.We know that the electrons resulted from Nd-doped TiO 2 and excited dye have strong reductive ability and they can reduce the In order to verify that • OH is present in the photodegradation system, the changes of PL spectra of terephthalic acid solution with irradiation time were recorded and the data is shown in Figure 11.It can be seen that a gradual increase in PL intensity at about 430 nm was observed with the increase of irradiation time.However, no PL increase was observed in the absence of visible light or Nd-doped TiO 2 sample.This suggests that the fluorescence is from the chemical reaction between terephthalic acid and • OH formed via photogenerated holes and electrons.Usually, PL intensity is proportional to the amount of produced hydroxyl radical over the photocatalyst [38].So, the amount of formed hydroxyl radical gradually increased for Rh-B degradation under visible light with the increase of time.
Moreover, since the radius of Nd 3+ is much larger than that of Ti 4+ , it will be difficult for all of Nd ions added to enter into TiO 2 lattice.So, the presence of Nd 3+ or Nd metal on nanoparticle surface may also promote the charge separation to improve photocatalytic activity of Nd-doped TiO 2 , which has been described by many published papers [14,[39][40][41][42][43].

Conclusion
In summary, a simple sol-gel method has been used for the preparation of Nd-doped titania nanoparticles with anatase and brookite phases.The prepared samples can International Journal of Photoenergy achieve photocatalytic degradation of dye (Rh-B) under visible light, and Nd-doped TiO 2 has better activity than pure TiO 2 and P25.The enhanced activity is related to the change of crystalline structure of TiO 2 .Partial Nd enters into TiO 2 lattice for the formation of doping level, which makes TiO 2 responsive to visible light.In Rh-B degradation system, hydroxide radicals as one of the reactive species are mainly produced by photogenerated electrons both from dye and Nd-doped TiO 2 excitation and they are responsible for Rh-B degradation.The contribution from hole is not as pronounced as electron for the formation of hydroxide radials.This study may shed light on the improvement of photocatalytic activity of TiO 2 under solar light by using metal doping method.
B are the integrated intensities of anatase (101), rutile (110), and brookite (121) peaks, respectively.The variables k A and k B are two coefficients and their values are 0.886 and 2.721, respectively.The calculated data are shown in Table the weight fractions of anatase, rutile, and brookite, respectively.The other symbols A A , A R , and A

Table 1 :
Summary of the physicochemical properties of TiO 2 and Nd-doped TiO 2 composites.
[26][27][28].It could be fitted into three subpeaks at about 530.21, 531.34, and 532.84 eV (in Figure6(b)), which can be ascribed to Ti-O, Nd-O, and hydroxyl group, respectively[26][27][28].Therefore, three types of oxygen existed on the photocatalyst level can react with O 2 adsorbed on TiO 2 surface to yield • O 2 − radical anion and subsequently HO • radical by protonation, just as the International Journal of Photoenergy electrons generated by Rh-B do.So, with the presence of part Nd in the lattice of TiO 2 , there are more photogenerated electrons which participate in the formation of • O 2 − radical anion and subsequently HO • radical by protonation for dye degradation.