Cpn

Fa

b

c

a

ARRAA

KTOCPD

1

dpioreaiTtbtnrtv

0h

Electrochimica Acta 112 (2013) 32– 36

Contents lists available at ScienceDirect

Electrochimica Acta

jou rn al hom ep age: www.elsev ier .com/ locate /e lec tac ta

omposition dependence of structural, optical, andhotoelectrochemical properties of nanocrystallineeodymium-doped titania photocatalyst

engjuan Miaoa,b, Zhe Wanga, Bairui Taoa,b, Junhao Chub,∗, Paul K. Chuc

College of Communications and Electronics Engineering, Qiqihar University, Heilongjiang 161006, ChinaNational Laboratory for Infrared Physics, Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, ChinaDepartment of Physics and Material Sciences, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China

r t i c l e i n f o

rticle history:eceived 21 June 2013eceived in revised form 7 August 2013ccepted 7 August 2013vailable online xxx

eywords:iO2:Ndptical properties

a b s t r a c t

Neodymium-doped titanium dioxide (Nd/TiO2) films with different Nd concentration x from 2 to 10% arefabricated on silicon by chemical solution deposition and a subsequent cathodic electrochemical processusing neodymium nitrate solution as the Nd source. The Nd dopant effects on the structural, optical,electrical, and photoelectrochemical properties of the Nd/TiO2 films are investigated by X-ray diffraction,ultraviolet Raman scattering, UV–vis diffuse reflectance spectra (DRS), and electrochemical methods atroom temperature. XRD shows the polycrystalline anatase phase with Nd atoms incorporated into theTiO2 matrix. The grain size decreases with increasing Nd concentrations. The intensity of the Raman-activemode B1g increases with Nd concentration. The diffuse reflection absorption spectra (DRS) indicate that

ompositionhotoelectrochemical behavioregradation efficiency

the absorbance edge of all the Nd/TiO2 samples shifts to the visible region. The Nd dopant decreasesthe band gap of TiO2 consequently enhancing the visible light absorption ability of the photocatalyst.The photoelectrochemical results indicate that in the same electrolyte, Nd can significantly enhancethe photoconversion efficiency of the TiO2 electrode as well as the photocurrent density. The Nd/TiO2

electrode shows higher efficiency in photoelectrocatalytic (PEC) degradation of p-nitrophenol (PNP) thande un

the undoped TiO2 electro

. Introduction

Since the first report by Fujishima and Honda [1], titaniumioxide or titania (TiO2) has been considered one of the efficienthotocatalysts. TiO2 has also been used to convert solar energy

nto electricity in the worldwide efforts to reduce the dependencen fossil fuels [2–4]. Titania has three common crystalline forms:utile, anatase, and brookite [5]. The anatase form usually has morefficient photoactivity than rutile and brookite [6–8]. Nevertheless,s a wide band gap n-type semiconductor, the band gap of anatases 3.2 eV and so the materials cannot absorb sunlight sufficiently.herefore, it is critical to decrease the band gap in order to improvehe efficiency of TiO2 in photocatalytic applications. There haveeen many suggested approaches, for example, modifications ofhe cationic and anionic sublattices, application of organic dyes andoble metals, etc. [9–11]. For instance, in 2001, Asahi et al. incorpo-

ated anions and transition metals into the TiO2 anionic sublatticeo narrow the energy band gap and increase the solar energy con-ersion efficiency [12]. Zhao et al. doped transition metals into TiO2

∗ Corresponding author. Tel.: +86 452 2742787; fax: +86 452 2738748.E-mail address: tbr sir@163.com (J. Chu).

013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.electacta.2013.08.029

der the same conditions.© 2013 Elsevier Ltd. All rights reserved.

to form compound photocatalyst to improve TiO2 photocatalyticactivity [13].

Titanium dioxide doped with rare-earth ions has attracted con-siderable interest recently because of the unique 4f electronicconfiguration and special luminescent and catalytic properties.Multiple electronic configurations can be easily generated and theoxide is polymorphism exhibiting strong adsorption selectivity andgood thermal stability. In particular, there is considerable shift tothe visible range after doping. Xie et al. reported the relationshipbetween the photocatalytic activity and dye sensitization or elec-tronic removal in Re3+ doped TiO2 [14]. Li et al. reported that Ce3+

doped in to TiO2 enhanced the adsorption capacity to pollutants(MBT) and improved the photocatalytic activity in the ultravioletand visible regions [15]. Xu et al. had researched the photocatalyticactivity of the neodymium-doped TiO2 nanotubes neodymium-doped TiO2 nanotubes, however, the lattice vibration of Nd3+ dopedTiO2 is not well understood and it is also important to investigatethe dependence of the Nd concentration on the physical and chem-ical properties of neodymium doped TiO2 films for photocatalytic

activity.

Many techniques such as pulsed laser deposition (PLD),pyrolysis, and the sol–gel method are employed to prepareNd/TiO2 films but the preparation of Nd doped TiO2 films

imica Acta 112 (2013) 32– 36 33

efisiclpd

2

2

spwd

emburwaTgpfi1fipfcst

2

tghbkd1t

2

XJcstsLwad

F. Miao et al. / Electroch

lectrochemically has seldom been reported. In this work, Nd/TiO2lms with Nd concentration varying from 2 to 10% are prepared onilicon by chemical solution deposition and cathodic electrochem-cal process and the structural, optical, and photoelectrochemicalharacteristics are studied systematically. The doped photocata-yst shows significantly enhanced photoelectrochemical and PECroperties compared to the undoped samples under the same con-itions.

. Experimental details

.1. Preparation of TiO2 films

The TiO2 thin films were deposited directly on p-type Si (1 0 0)ubstrates using the sol–gel method. The substrates were cleaned inure ethanol ultrasonically and rinsed several times with deionizedater. The wafers were dried under flowing nitrogen before filmeposition.

Analytically pure titanium butoxide [Ti (OC4H9)4], anhydrousthanol (C2H5OH), and acetic acid (CH3COOH) were the startingaterials. Acetyl acetone and an equimolar amount of titanium

utoxide were added dropwise to the solution to the required vol-me ratio of C2H5OH:CH3COOH (16:1) under vigorous stirring atoom temperature. Caution was exercised because the reactionas rather violent. The acetate and acetyl acetone were used to

djust the pH and stabilize the titanium butoxide, respectively.he solution was stirred for 2 h at 50 ◦C to increase the homo-eneity. The 0.3 M precursor solution was transparent withoutrecipitates even after two months. Before deposition of the TiO2lms, the silicon wafers were cut into chips with dimension of

cm × 1 cm and cleaned using the standard RCA process. The TiO2lms were deposited by spin coating the solution on the Si sam-le at 4000 rpm for 20 s. Finally, the thin films were dried at 180 ◦Cor 200 s, pyrolyzed at 380 ◦C for 240 s to remove residual organicompounds, and annealed at 500 ◦C for 1 h in ambient air. The depo-ition and annealing treatment procedures were repeated eightimes in order to obtain the desired thickness [16].

.2. Preparation of Nd-doped TiO2 films

The Nd-doped TiO2 films were prepared using a cathodic elec-rochemical process. The TiO2 films were used as the cathode andraphite served as the anode. The 0.05 M neodymium nitrate hexa-ydrate (Nd(NO3)3·6H2O) solution was the electrolyte. The platingath was kept at 60 ◦C and rigorously stirred. The voltage wasept at 2 V. After deposition, the samples were rinsed with double-istilled water, dried in air, and then annealed at 500 ◦C in air for

h. Different amounts of Nd were incorporated by varying the elec-rochemical process time [17].

.3. Characterization of Nd-doped TiO2 films

The crystalline structure of the Nd/TiO2 films was determined by-ray diffraction (XRD) using Cu K� radiation (Rigaku, RINT2000,

apan). A vertical goniometer (Model RINT2000) was used and theontinuous scanning mode (2�/�) with an interval of 0.02◦ andcanning rate of 10◦/min was adopted. Ultraviolet Raman scat-ering was performed at room temperature on a micro-Ramanpectrometer with a spectral resolution of 1.5 cm−1 (Jobin-Yvon

abRAM HR 800 UV). The 325 nm (3.82 eV) line of a He-Cd laserith output power of 30 mW was the excitation source. The light

bsorption properties were determined by monitoring the UV–visiffuse reflectance spectra in the wavelength range of 200–800 nm.

Fig. 1. XRD patterns of the Nd/TiO2 films containing 2, 4, 6, 8, and 10% of Nd preparedon Si (1 0 0).

2.4. Photoelectrochemical and photoelectrocatalytic degradation

The photoelectrochemical experiments were carried out in aconventional three-electrode cell controlled by the LK3200A elec-trochemical workstation. The Nd/TiO2 nanocomposite electrodewas the photo-anode. A platinum wire electrode was adopted as thecounter electrode and a saturated calomel electrode (SCE) served asthe reference electrode. The electrolytes consisted of 50 mL of 0.1 MNa2SO4 and 50 mL of a mixture containing 20 mg L−1 of PNP with0.1 M Na2SO4. A 350 W xenon lamp with an illumination intensityof 100 mW cm−2 was the light source. All the experiments wereperformed at 25 ◦C. The photoelectrocatalytic degradation experi-ments were also conducted in the three-electrode cell controlledby the LK3200A electrochemical workstation. The photo-anode,cathode, reference electrode, and electrolytes were the same as inthe photoelectrochemical experiments. The experiments were per-formed with magnetic stirring at room temperature and the pH ofthe solution was adjusted by H2SO4 or NaOH.

3. Results and discussion

The XRD patterns acquired from the Nd/TiO2 films with dif-ferent Nd concentrations are depicted in Fig. 1. All the films arepolycrystalline with strong (1 0 1) diffraction which position shiftsfrom 25.3◦ to 25.5◦ with Nd incorporation. It can be explained bythat the radius of Nd is larger than that of Ti. Besides this salientfeature, there are several weaker diffraction peaks of (0 0 4), (2 0 0),(2 1 1), (2 0 4), and (1 1 6) at about 37.8◦, 48.0◦, 55.1◦, 62.75◦, and68.84◦ but no impurity phases are observed confirming the pureanatase structure. The polycrystalline grains with different orien-tations are formed in the anatase films and Nd atoms exist in theTiO2 matrix. The (1 0 1) peak intensity decreases and peak widthincreases with increasing Nd concentrations implying Nd plays animportant role in the crystalline lattice. On the basis of the (2 0 0)and (0 0 4) diffraction peaks, the lattice constants a = b and c of theNd-doped TiO2 films are estimated to be slightly larger than thetheoretical values of pure anatase (a = b = 3.747 A, c = 9.334 A) as aresult of Nd substitution. The difference in the lattice constant sug-gests that there are different lattice distortions in the anatase filmsdoped with Nd. The average crystalline size r can be calculatedfrom the (1 0 1) diffraction peak according to the Scherrer equa-tion r = K�/ cos �, where K ≈ 1 is the shape factor, � = 1.540 A is

the average wavelength of Cu K� radiation, is the full width athalf-maximum, and � is the diffraction angle. That shown the aver-age grain size is about 31.4 nm and the grain size diminishes withincreasing Nd concentration while the cell parameter is not varied.

34 F. Miao et al. / Electrochimica Acta 112 (2013) 32– 36

Fig. 2. Raman spectra of the Nd/TiO2 films with different Nd contents excited bythe 325 nm line. The dashed lines show the fundamental phonon mode frequency.

(TaPfaftpcT5orb

m3coownnpc

poUfurtriartata

and

Eph = 1240�

, (3)

Fig. 4. UV–vis DRS spectra acquired from the TiO2 film and Nd-doped TiO2 films.

The anatase TiO2 is tetragonal with the space group D4hI41/amd) and the primitive cell contains two units per cell.here are six first order Raman-active modes (Ag + 2B1g + 3Eg)t the � point of the Brillouin zone [18–20]. According to theerdew–Burke–Ernzerbof (PBE) functional, the theoretical Ramanrequencies of Eg, Eg, B1g, A1g, and Eg are 128, 158, 388, 514, 521,nd 638 cm−1, respectively [21]. The 2Eg modes occurring at lowerrequencies (<300 cm−1) cannot be observed due to the experimen-al limitation in the present work. The other four Raman-activehonon modes of the anatase Nd/TiO2 films with different Nd con-entrations are shown in Fig. 2. The four phonon modes of anataseiO2 are at about 393, 515, and 632 cm−1 (superimposed with the15 cm−1 band), which are assigned to the Raman-active modesf the anatase phase with symmetries of B1g, B1g + A1g, and Eg,espectively. The B1g and A1g phonon frequencies are too close toe discerned.

In order to further investigate the lattice vibration andicrostructure, Fig. 3 shows the Raman spectra in the range of

00–750 cm−1 acquired from the Nd/TiO2 films with different Ndontents. The Lorentzian multipeak fitting includes the four first-rder Raman-active phonon modes (2B1g, A1g, and Eg). The intensityf the B1g phonon mode at 515 cm−1 has increasing tendencyith increasing Nd contents. Compared to the Raman spectra ofeodymium oxide, the shoulder structure cannot originate from theeodymium oxide. The change in the fundamental Raman-activehonon modes and appearance of the additional vibration modean be attributed to Nd introduction.

To investigate the influence of Nd3+ on the optical absorptionroperties of TiO2, diffuse reflection absorption spectra (DRS) arebtained from the TiO2 and Nd/TiO2 samples. Fig. 4 shows theV–vis DRS of the undoped TiO2 film and Nd-doped TiO2 with dif-

erent Nd concentrations. The band gap absorption edges of thendoped TiO2 and Nd-doped TiO2 films are around 380 and 410 nm,espectively. The optical absorption of TiO2 for wavelengths shorterhan 380 nm is mainly attributed to the O2+ → Ti4+ charge transferelated to electron excitation from the valence band [22,23]. Theres a red shift of about 30 nm with increasing Nd3+ concentrationnd enhanced absorption intensities in both UV and visible lightegions are also observed. In particular, pure TiO2 has no adsorp-ion in the visible light region (>400 nm) but Nd/TiO2 exhibits fourbsorption peaks at 527, 586, 762, and 809 nm corresponding tohe transitions of 4I9/2 to 2K13/2 and 4G7/2, 2G7/2 and 4G5/2, 4S3/2nd 4F7/2, and 4F5/2 and 2H9/2, respectively [24].

Fig. 3. Five-peak fit of the Raman spectra acquired from the Nd/TiO2 films with (a)2%, (b) 4%, (c) 6%, 8%, and (d) 10% of Nd, respectively. The four first-order and onesecond-order Raman phonon modes are indicated.

The band gaps of Nd/TiO2 are estimated by DRS by plottingthe Kubelka–Munk functions (F(R)) versus the photon energy (Eph)[25,26]. The equations are shown in the following:

A = log(

1R∞

), (1)

F(R) = (1 − R∞)2

2R∞, (2)

F. Miao et al. / Electrochimica Acta 112 (2013) 32– 36 35

F al of 1o −2) ir

wneb4tvq

pmpussNttisqTdPP

�

wsr(tecijpratP

photocatalytic degradation reaction because a large amount ofadsorbed organic molecules promotes the photocatalytic reac-tion. Moreover, a smaller particle size of Nd/TiO2 enhances the

ig. 5. (A) Photocurrent responses in the light on–off process at an applied potentir 20 mg L−1 PNP containing 0.1 M Na2SO4 solution under UV–vis light (100 mW cm

here A is a absorbance, R∞ is the reflectance of samples in infi-ite thick, F(R) is the Kubelka–Munk functions, Eph is the photonnergy, and � is the wavelength. The band gaps are calculated toe 3.22, 3.14, 3.09, 3.05, 2.97, and 2.92 eV for TiO2, 2.0% Nd/TiO2,.0% Nd/TiO2, 5.0% Nd/TiO2, 8.0% Nd/TiO2, and 10% Nd/TiO2, respec-ively. The absorbance edge of Nd/TiO2 shifts significantly to theisible region and Nd doping decreases the band gap of TiO2 subse-uently increasing the visible light absorption of the photocatalyst.

The photoelectrochemical behavior is studied by measuring thehotocurrent and photopotential. As shown in Fig. 5a, upon illu-ination with UV–vis light at an applied potential of 1 V vs. SCE,

rompt generation of photocurrents is observed from both thendoped and Nd-doped TiO2 films. The Nd-doped TiO2 electrodeshow a larger photocurrent density than the undoped one under theame conditions in 0.1 M Na2SO4 or 20 mg L−1 of PNP with 0.1 Ma2SO4 as the electrolyte. The photoelectrochemical behavior of

he undoped and Nd-doped TiO2 electrodes is further evaluated byhe I–V characteristics (Fig. 5b). The photocurrent increases withncreasing applied voltage before leveling off, suggesting that amall external bias is beneficial to the electron transfer conse-uently reducing recombination of photoelectrons and holes iniO2 and Nd/TiO2. The existence of PNP increases the photocurrentensity on both undoped and Nd-doped TiO2 electrodes, becauseNP can capture the photogenerated holes and be decomposed inEC oxidation process [27]:

% = total power output − electrical power inputlight power input

× 100

= jp

[E0

rev − |Eapp|I0

]× 100,

here jp is the photocurrent density in (mA cm−2), E0rev is the

tandard reversible potential (which is 1.23 V for the water splittingeaction at pH = 0). The external applied potential Eapp = Emeas − Eaoc

Emeas is the electrode potential of the working electrode at whichhe photocurrent is measured under illumination, and Eaoc is thelectrode potential of the same working electrode under open cir-uit conditions and same conditions), I0 is the power density of thencident light (mW cm−2) and jp E0

rev is the total power output whilep |Eapp| is the electrical power input. According to this method, thehotoconversion efficiency of fabricated electrodes is calculated. A

emarkable and maximum photoconversion efficiency of 0.468% ischieved from the Nd-doped TiO2 electrode, but it is only 0.254% forhe undoped one in the same electrolyte comprising 20 mg L−1 ofNP and 0.1 M Na2SO4. Our results provide experimental evidence

V, (B) variation of the photocurrent density versus bias potential in 0.1 M Na2SO4

radiation of the TiO2 film and Nd-doped TiO2 films.

that Nd doping increases the photocurrent density and photocon-version efficiency.

PNP degradation is studied by conducting the photoelectrocatal-ysis. Fig. 6 depicts the concentration changes in 50 mL of 20 mg L−1

PNP under varying conditions for 1 h. Under UV–vis light irradia-tion, almost 67% of the PNP is degraded by applying a 0.8 V biasvs. SCE to the Nd-doped TiO2 films, whereas only 44% of the PNPis degraded by the undoped sample, suggesting that Nd3+ dopingenhances the photocatalytic activity of TiO2.

In photoelectrocatalytic processes, the bias potential drawsthe photo-induced electrons to the counter electrode and so thebias potential is a key factor governing the photoelectrocata-lytic efficiency. Here, five bias potentials of 0.0 V, 0.2 V, 0.4 V,0.6 V, and 0.8 V are investigated for 10% Nd/TiO2 photocatalystselectrode. As shown in the inset of Fig. 6, the concentration ofPNP changes with bias potentials. The degradation of PNP can bedescribed by first-order kinetic of −In(Ct/C0) versus reaction time(t): −ln(Ct/C0) = kap t, where kap is the apparent reaction rate con-stant which ascends with increasing bias potential and C0 andCt are the initial and PNP concentrations at time t, respectively[28]. Usually, a large specific surface area influences the rate of

Fig. 6. Photoelectrocatalytic process of PNP under UV–vis light (100 mW cm−2) irra-diation of the Nd-doped electrode. Inset: Degradation of PNP affected by the biaspotential.

3 imica

ps

4

NbttstcfioTtAttfid

A

eSTaF

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

6 F. Miao et al. / Electroch

hotoinduced charge transfer from the bulk of the particles to theurface-absorbed reactants.

. Conclusion

High-quality anatase TiO2 films with different concentrations ofd are synthesized using a nonhydrolytic sol–gel process followedy cathodic electrochemical preparation. The films are polycrys-alline with a strong (1 0 1) diffraction peak. The dependence ofhe lattice vibration in the Nd/TiO2 films on Nd concentrations istudied. The four Raman-active phonon modes are observed andhe shifts of the Raman modes are sensitive to the Nd dopantoncentration. The optical absorption properties of the Nd/TiO2lms are studied by DRS and Nd doping decreases the band gapf TiO2 subsequently increasing the visible light absorption ability.he photoelectrochemical properties of the Nd-doped TiO2 elec-rode are studied and the photocurrent is dramatically enhanced.

maximum photoconversion efficiency of 0.468% is achieved andhe PEC degradation efficiency of PNP by the Nd-doped TiO2 elec-rode is higher compared to the undoped one. The Nd-doped TiO2lms have potential in water splitting, solar cells, as well as PECegradation of pollutants.

cknowledgments

This work was jointly supported by the National Natural Sci-nce Foundation of China (Grant No. 61204127), China Postdoctoralcience Foundation (No. 2012M510898), New Century Excellentalents in Heilongjiang Provincial University (No. 1253-NECT025),nd Hong Kong Research Grants Council (RGC) General Researchunds (GRF) (Nos. 112510 and 112212).

eferences

[1] A. Fujishima, K. Honda, Electrochemical photolysis of water at a semiconductorelectrode, Nature 238 (1972) 37–38.

[2] J.N. Wilson, H. Idriss, Structure sensitivity and photocatalytic reactions of semi-conductors. Effect of the last layer atomic arrangement, J. Am. Chem. Soc. 124(2002) 11284.

[3] J. Zheng, Z. Liu, X. Liu, X. Yan, D. Li, W. Chu, Facile hydrothermal synthesisand characteristics of B-doped TiO2 hybrid hollow microspheres with higherphoto-catalytic activity, J. Alloys Compd. 509 (2011) 3771–3776.

[4] L. Baia, L. Diamandescu, L. Barbu-Tudoran, A. Peter, G. Melinte, V. Danciu, M.Baia, Efficient dual functionality of highly porous nanocomposites based onTiO2 and noble metal particles, J. Alloys Compd. 509 (2011) 2672–2678.

[5] A.F. Wells, Structural Inorganic Chemistry, Clarendon Press, Oxford, 1975.[6] K. Nagaveni, G. Sivalingam, M.S. Hegde, G. Madras, Photocatalytic degradation

of various dyes by combustion synthesized nano anatase TiO2, Appl. Catal. B 45(2003) 23–38.

[7] O. D’Hennezel, D.F. Ollis, Surface prechlorination of anatase TiO2 for enhancedphoto-catalytic oxidation of toluene and hexane, Helv. Chim. Acta 84 (2001)3511–3518.

[

[

Acta 112 (2013) 32– 36

[8] G.S. Li, L.P. Li, J. Boerio-Goates, B.F. Woodfield, High purity anatase TiO2

nanocrystals: near room-temperature synthesis, grain growth kinetics, andsurface hydration chemistry, J. Am. Chem. Soc. 127 (2005) 8659–8666.

[9] M. Radecka, TiO2 for photoelectrolytic decomposition of water, Thin Solid Films451–452 (2004) 98–104.

10] H.P. Maruska, A.K. Gosh, Transition-metal dopants for extending theresponse of titanate photoelectrolysis anodes, Sol. Energ. Mater. 1 (1979)237–247.

11] M. Radecka, M. Wierzbicka, S. Komornicki, M. Rekas, Influence of Cr onphotoelectrochemical properties of TiO2 thin films, Physica B 348 (2004)160–168.

12] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light photocatalysisin nitrogen-doped titanium oxides, Science 293 (2001) 269–271.

13] W.Y. Zhao, W.Y. Fu, H.B. Yang, C.J. Tian, M.H. Li, J. Ding, W. Zhang, X.M. Zhou, H.Zhao, Y.X. Li, Synthesis and photocatalytic activity of Fe-doped TiO2 supportedon hollow glass microbeads, J. Nano-Micro Lett. 3 (1) (2011) 20–24.

14] Y.B. Xie, C.W. Yuan, Rare earth ion modified TiO2 sols for photocatalysis appli-cation under visible light excitation, Rare Metals 23 (2004) 20–26.

15] F.B. Li, X.Z. Li, M.F. Hou, et al., Enhanced photocatalytic activity of Ce3+-TiO2

for 2-mercapto benzothiazole degradation in aqueous suspension for odourcontrol, Appl. Catal. A: Gen. 285 (2005) 181–189.

16] J.Z. Zhang, X.G. Chen, Y.D. Shen, Y.W. Li, Z.G. Hu, J.H. Chu, Synthesis, sur-face morphology, and photoluminescence properties of anatase iron-dopedtitanium dioxide nano-crystalline films, Phys. Chem. Chem. Phys. 13 (2011)13096–13105.

17] J.L. Nie, Y. Mo, B.Z. Zheng, H.Y. Yuan, D. Xiao, Electrochemical fabrication oflanthanum-doped TiO2 nanotube array electrode and investigation of its pho-toelectrochemical capability, Electrochim. Acta 90 (2013) 589–596.

18] J.F. Zhu, F. Chen, J. Zhang, H. Chen, M. Anpo, Fe3+-TiO2 photocatalysts pre-pared by combining sol–gel method with hydrothermal treatment and theircharacterization, J. Photochem. Photobiol. A: Chem. 180 (2006) 196–204.

19] H. Kamisaka, T. Hitosugi, T. Suenaga, T. Hasegawa, K. Yamashita, DFT-basedfirst-principle calculation of Nb-doped anatase TiO2 and its interactionswith oxygen vacancies and interstitial oxygen, J. Chem. Phys. 131 (2009)034702–34712.

20] M. Giarola, A. Sanson, F. Monti, G. Mariotto, M. Bettinelli, A. Speghini, G. Salvi-ulo, Vibrational dynamicsofanataseTiO2: polarized Raman spectroscopy and abinitio calculations, Phys. Rev. B 81 (2010) 174305.

21] M.B. Yahia, F. Lemoigno, T. Beuvier, J.S. Filhol, R.P. Mireille, L. Brohan, M.L.Doublet, Updated references for the structural, electronic, and vibrationalproperties of TiO2(B) bulk using first-principles density functional theory cal-culations, J. Chem. Phys. 130 (2009) 204501.

22] V. Stengl, S. Bakardjieva, N. Murafa, Preparation and photocatalytic activ-ity of rare earth doped TiO2 nanoparticles, Mater. Chem. Phys. 114 (2009)217–226.

23] E. Borgarello, J. Kiwi, M. Gratzel, E. Pelizzetti, M. Viscald, Visible light inducedwater cleavage in colloidal solutions of chromium-doped titanium dioxide par-ticles, J. Am. Chem. Soc. 104 (1982) 2996–3002.

24] Koepke Cz, K. Wisniewski, L. Sikorski, D. Piatkowski, K. Kowalska, M. Naftaly,Upconverted luminescence under 800 nm laser diode excitation in Nd3+-activated fluoroaluminate glass, Opt. Mater. 28 (2006) 129–136.

25] N. Todorova, T. Giannakopoulou, T. Vaimakis, J. Yu, C. Trapalis, Preparation offluorine-doped TiO2 photocatalysts with controlled crystalline structure, Int. J.Photoenergy 10 (2013) 1–9.

26] X. Jiang, L. Yang, P. Liu, X. Li, J. Shen, The photocatalytic and antibacterialactivities of neodymium and iodine doped TiO2 nanoparticles, Colloids Surf.B: Biointerfaces 79 (2010) 69–74.

27] S.U.M. Khan, M. Al-shahry, W.B. Ingler Jr., Efficient photochemical water split-ting by a chemically modified n-TiO2, Science 297 (2002) 2243.

28] X.Z. Li, H.L. Liu, P.T. Yue, Y.P. Sun, Photoelectrocatalytic oxidation of rose bengalin aqueous solution using a Ti/TiO2 mesh electrode, Environ. Sci. Technol. 34(2000) 4401.