Fabrication and correlation between photoluminescenceand photoelectrochemical properties of vertically alignedZnO coated TiO2 nanotube arrays

Hua Cai a, Qinghu You a, Zhigao Hu b, Zhihua Duan b, Yong Cui c, Jian Sun a,Ning Xu a, Jiada Wu a,n

a School of Information Science and Engineering, Fudan University, Shanghai 200433, Chinab Key Laboratory of Polar Materials and Devices, Ministry of Education, East China Normal University, Shanghai 200241, Chinac Shanghai Institute of Laser and Plasma, Shanghai 201800, China

a r t i c l e i n f o

Article history:Received 28 September 2013Received in revised form14 January 2014Accepted 20 January 2014Available online 5 February 2014

Keywords:Titania nanotubeZinc oxide coatingHeterogeneous nanostructurePhotoluminescencePhotoelectrochemical property

a b s t r a c t

We report on a study on the correlation between the photoluminescence and the photoelectrochemicalproperties of ZnO–TiO2 heterogeneous nanostructure composed of anatase TiO2 nanotubes and wurtziteZnO coatings. Vertically aligned ZnO coated TiO2 nanotube (ZnO/TiO2 NT) arrays have been fabricated byatomic layer deposition of ZnO coatings on electrochemical anodization formed TiO2 nanotubes. Theobtained ZnO/TiO2 NT arrays show higher photoresponse to shorter wavelength light than to longerwavelength light. A reduction in photoluminescence and an enhancement in photoelectrochemicalactivity are observed for annealed ZnO/TiO2 NT arrays. ZnO/TiO2 NTs with thinner ZnO coatings ratherthan thicker ZnO coatings have better photoelectrochemical properties. Compared with bare TiO2 NTarrays, an increase in photocurrent of about 50 percent is obtained for the arrays of 450 1C annealed ofZnO/TiO2 NTs with 10-cycle deposition of ZnO coatings under visible illumination with cutoff of 420 nm.

& 2014 Elsevier B.V. All rights reserved.

1. Introduction

By virtue of their high surface-to-volume ratio and superioroptical, electronic and photoelectrochemical properties, nanos-tructured TiO2 and ZnO have recently attracted much attentionbecause of their potential applications in photovoltaic processes,photocatalytic reactions, etc. [1–3]. Both metal oxides have alsoadvantages of low cost, stability, nontoxicity, and ease of avail-ability. Among various nanostructured TiO2, highly ordered tubu-lar nanostructures i.e. TiO2 nanotubes (NTs) are of particularinterest because of their larger surface area which increases thedensity of active sites for surface reactions and improves inter-facial carrier transfer rates [4–6]. The aligned NT structure alsodramatically improves charge transport properties, which greatlycontributes to photoelectrochemical performance. TiO2 NT arrayshave therefore been extensively studied as electrodes in photo-catalysts [2,7–9] and solar cells [10–12]. ZnO nanomaterials havealso been demonstrated to be promising when used in photo-catalysts [13–15] and photovoltaics [16–18]. Due to their largeband gaps (TiO2: �3.2 eV, ZnO: �3.4 eV), however, both TiO2 and

ZnO normally only work under UV irradiation, thus restrictingtheir practical applications.

The good compatibility between TiO2 and ZnO allows them tobe composed to form a heterogeneous nanostructure. It has beenproved that heterostructures composed of nanostructured TiO2

and ZnO can improve the quantum efficiency of photocatalysts andphotovoltaics due to the combination of the high reactivity of TiO2

and the large binding energy of ZnO. Covering of a ZnO coating onTiO2 surface also increases the mobility of charge carriers [(TiO2:0.5 cm2 V�1 s�1)o(ZnO: 200 cm2 V�1 s�1)]. The process of elec-tron and hole transfer between the corresponding conduction andvalence bands can be improved and a better charge separation ofphotogenerated carriers can be achieved when compared withcatalysts from a single metal oxide. In addition, an extension forspectral range of photoresponse can be expected due to thestaggered band alignment of the heterostructures [19–21].The improved electron and hole conductivity, the enhancedcharged carries separation and the extended photoresponse range,all facilitate an increase of solar conversion efficiencies and aremost favorable for photocatalytic reactions and photovoltaicprocesses [22–25]. Much work has been therefore devoted tonano-scaled ZnO–TiO2 composites, in particular, to one-dimensional hetero-nanostructures such as ZnO modified TiO2 orTiO2 modified ZnO nanotubes (NTs) and nanowires [23–25].

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/solmat

Solar Energy Materials & Solar Cells

0927-0248/$ - see front matter & 2014 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.solmat.2014.01.033

n Corresponding author.E-mail address: jdwu@fudan.edu.cn (J. Wu).

Solar Energy Materials & Solar Cells 123 (2014) 233–238

In the present work, we fabricated vertically aligned ZnOcoated TiO2 (ZnO/TiO2) NTs by electrochemical anodization of Tifoil followed by atomic layer deposition (ALD) of ZnO on the TiO2

NTs. After a detailed morphological and structural characteriza-tion, the correlation between the photoluminescence (PL) and thephotoelectrochemical properties was studied by measuring PL andphotocurrents to understand the suppressed radiative recombina-tion and the enhanced charge separation of photogeneratedcarries in the nano-heterojunction composed of TiO2 NTs andZnO coatings as well as to obtain better visible-light photore-sponse and higher photoelectrochemical activity.

2. Experimental procedure

2.1. Sample fabrication

Two processes were applied to the fabrication of ZnO/TiO2 NTarrays. Vertical aligned TiO2 NT arrays were first formed on Ti foil(99.99% in purity, 0.1 mm in thickness) by electrochemical anodi-zation which was performed in a two-electrode electrochemicalcell using a Ti foil as the working electrode and a graphite sheet asthe counter electrode separated by 1.3 cm. Before anodization, theTi foil was chemically polished in a mixture of de-ionized water,HNO3 and HF with the ratio of 5:4:1 (in volume) for 30 s to removethe surface oxide layer and then rinsed with de-ionized water. Thepolished Ti foil was anodized in 0.5 wt% HF electrolyte at aconstant voltage of 20 V for 40 min to form TiO2 NT arrays. Thearrays were then annealed in a furnace at 450 1C for 3 h inatmospheric air to convert the formed amorphous TiO2 intoanatase. The annealed TiO2 NTs were used as the template forthe deposition of ZnO coatings by ALD (TFS200, BENEQ) usingdiethylzinc (Zn(C2H5)2, DEZ) as the metal precursor for Zn and de-ionized water (H2O) as the reactant. DEZ was kept at 18 1C anddelivered into the deposition chamber with a high purity N2

carrier gas flow. The ZnO coatings were typically deposited byseveral DEZ–H2O cycles consisting of the sequence: 0.5 s DEZpulse, 2 s N2 purge, 0.5 s H2O pulse, 2 s N2 purge. A one-cycledeposit of ZnO was approximately 0.2 nm thick when depositedon a polished plane substrate. The deposition temperature and the

working pressure were kept at 200 1C and �1 Torr, respectively.The obtained ZnO coated TiO2 NT arrays were rinsed withde-ionized water and annealed at different temperatures for30 min in atmospheric air.

2.2. Structural characterization

The surface morphology of the prepared ZnO/TiO2 NT arrayswas examined by field-emission scanning electron microscopy(FESEM) with a Hitachi S-4800 microscope. The crystal structurewas characterized by X-ray diffraction with a Rigaku D/max-γ BX-ray diffractometer using a rotating anode operating at 40 kV and300 mA and Ni-filtered Cu Kα radiation (λ¼ 0.15406 nm). Thestructure was also characterized through the analysis of vibra-tional modes by measuring Raman scattering spectra. Thesemeasurements were performed with a Jobin-Yvon LabRAM HR800 UV micro-Raman spectrometer using 488 nm Arþ laser beamand 325 nm He–Cd laser beam to excite the samples.

2.3. Photoluminescence and photocurrent measurements

PL measurements were performed at room temperature byexciting the samples at normal incidence with a CW 325 nm He–Cd laser. PL spectra were recorded by collecting the emittedluminescence at normal direction with a 0.5 m spectrometer(Acton Research, Spectra Pro 500i) and an intensified charge-coupled device (ICCD, Andor Technology, iStar DH720) attached onthe exit port of the spectrometer.

The photoelectrochemical properties were studied by measuringphotocurrents in a three-electrode cell with 0.5 M Na2SO4 electro-lyte using a CHI electrochemical analyzer (CHI 660A Instruments).With an active area of 1 cm2, TiO2 or ZnO/TiO2 NT arrays were usedas the working electrode and subjected to the irradiation of visiblelight with Ag/AgCl as the reference electrode and Pt foil as thecounter electrode. A 500 W xenon lamp was used as the lightsource to provide two wavelength ranges of visible light by cuttingultraviolet light using high-pass filters with cutoff of 380 or 420 nm.The light intensity on the working electrode was 100 mW/cm2.

Fig. 1. (a) Top-view and (b) cross-sectional SEM images of bare TiO2 NTs, (c) top-view and (d) cross-sectional SEM images of ZnO/TiO2 NTs with 50 cycles of ZnO coating.

H. Cai et al. / Solar Energy Materials & Solar Cells 123 (2014) 233–238234

3. Results and discussion

3.1. Morphology

Highly ordered and vertically aligned TiO2 NT arrays wereformed on Ti foil by electrochemical anodization, as shown inFig. 1(a) and (b). The bare TiO2 NTs have an average diameter of�60 nm and a wall thickness of �15 nm. FESEM examinationreveals that the ZnO coatings deposited by ALD cover uniformlythe TiO2 NTs, both on the top and the inner walls of TiO2 NTs, evendeep down into the tubes, forming ZnO coated TiO2 NT arrays.Fig. 1(c) and (d) shows the planar and cross-sectional FESEMimages of the as-fabricated ZnO/TiO2 NT arrays fabricated bydepositing 50 cycles of ZnO on TiO2 NTs. It can be seen that theZnO coated NTs have smaller diameters and thicker walls than thebare TiO2 NTs. Unless otherwise specified, the results described inthis paper are obtained with the samples fabricated by depositing50 cycles of ZnO on electrochemically anodized TiO2 NTs. Nochanges in morphology were observed for the ZnO/TiO2 NT arraysafter annealing.

3.2. Structure

XRD was employed to characterize the crystal structure of ZnO/TiO2 NT arrays as well as that of bare TiO2 NT arrays. Fig. 2(a)illustrates the XRD patterns of the as-fabricated and annealedZnO/TiO2 NT arrays together with that of the bare TiO2 NT arraysfor comparison. The XRD pattern of the bare TiO2 NT arraysexhibits three strong diffraction peaks indexed to the (101),(102) and (103) diffractions of Ti and three weak ones indexed tothe (002), (110) and (112) diffractions of Ti, respectively. They areall diffracted by Ti foil and are denoted by T in Fig. 2 (JCPDS 44-1294). Besides, a prominent peak with 2θ at 25.221 and full widthat half-maximum (FWHM) of 0.181 exists, is identified as a (101)diffraction of anatase TiO2 (JCPDS 21-1272), denoted as A in Fig. 2.Additionally, a very weak diffraction near 27.401 is resolved, andcan be assigned to the (110) diffraction of rutile TiO2, denoted as Rin Fig. 2. The TiO2 NT arrays formed by electrochemical anodiza-tion are therefore composed of nanocrystallites with almosttetragonal anatase phase. After covered by ZnO, additional weakdiffraction peaks appear, as indicated in XRD pattern 2 obtainedfor the as-fabricated ZnO/TiO2 NT arrays. They can all be ascribedto hexagonal wurtzite ZnO corresponding to the (100), (002), (101)and (110) planes and are denoted by W in Fig. 2. The diffractionangles are somewhat larger than the standard values for bulk ZnO(JCPDS: 36-1451). Fig. 2(b) displays the magnified XRD patterns inthe 2θ range from 30.01 to 38.51. The multi diffractions of XRDpattern reveal a non-oriented growth of ZnO on the surface of TiO2

NTs. The larger diffraction angles suggest imperfection in the crystalstructure of ZnO. After annealing, the intensity of the diffractionsascribed to wurtzite ZnO generally increases with a reduced peakwidth as the annealing temperature increases, especially the (100)and (101) peaks, indicating an improvement in the crystallinity ofZnO after annealing. Compared with the as-fabricated ZnO/TiO2 NTarrays, in addition, nearly all the diffraction angles of ZnO exhibit asmall shift towards smaller values. The 2θ values of the ZnO (100)and (101) peaks shift from 32.171 and 36.461 for the as-fabricatedsample to 31.641 and 36.101 for the annealed sample at 450 1C,respectively. They are smaller than the standard values for bulkZnO, suggesting the presence of tensile strain in ZnO most probablydue to the larger lattice constants of anatase TiO2 than those ofwurtzite ZnO. Using the well-known X-ray diffraction formula forhexagonal structure, the lattice constants of ZnO were determinedto be a¼ 0.326 nm and c¼ 0.522 nm, respectively. The above XRDresults indicate that the ZnO/TiO2 NT arrays fabricated by atomiclayer deposition of ZnO on electrochemically anodized TiO2 NTs are

composed of tetragonal anatase TiO2 nanotubes and hexagonalwurtzite ZnO coatings.

Raman scattering measurement confirmed the tetragonal ana-tase structure of TiO2 and the hexagonal wurtzite structure of ZnOfor the fabricated ZnO/TiO2 NT arrays. Tetragonal structure has sixRaman active fundamental vibrational modes (A1gþ2B1gþ3Eg)[26,27]. For single crystal anatase TiO2, these modes are locatedat 144 cm�1 (Eg), 197 cm�1 (Eg), 399 cm�1 (B1g), 513 cm�1 (A1g),519 cm�1 (B1g), and 639 cm�1 (Eg) [28,29]. At low temperatures,the assignment for the modes at 513 and 519 cm�1 is not possiblebecause of the overlap between the two modes in an experimentalRaman spectrum [30]. Wurtzite ZnO has six Raman active phononmodes E2 (low), E2 (high), A1 (TO), A1 (LO), E1 (TO), and E1 (LO)[31–33]. Spectrum 1 in Fig. 3(a) was taken from the bare TiO2 NTarrays with the excitation by 488 nm Arþ laser light. Thisspectrum exhibits five distinct peaks located at 144, 197, 392,514, and 633 cm�1. These peaks are close to those found in thebulk anatase TiO2 [28,29] and can be assigned as characteristicRaman modes Eg, Eg, B1g, A1g/B1g, and Eg, respectively, as shown inFig. 3(a) where (A) denotes the anatase phase.

Fig. 3(a) also exhibits Raman spectra of ZnO/TiO2 NT arrays.After being covered by ZnO coatings, all the anatase TiO2 modes

Fig. 2. (a) XRD patterns of bare TiO2 NTs (1), as-fabricated ZnO/TiO2 NTs (2), 350 1Cannealed ZnO/TiO2 NTs (3), and 450 1C annealed ZnO/TiO2 NTs (4). (b) MagnifiedXRD patterns ranging from 2θ¼30.01 to 38.51 for bare TiO2 NTs (1), as-fabricatedZnO/TiO2 NTs (2), 350 1C annealed ZnO/TiO2 NTs (3), and 450 1C annealed ZnO/TiO2

NTs (4). Here T, A, R and W denote the diffractions from Ti, anatase TiO2, rutile RTiO2, and wurtzite ZnO, respectively.

H. Cai et al. / Solar Energy Materials & Solar Cells 123 (2014) 233–238 235

are still well resolved for the as-fabricated and annealed samples,only resulting in intensity decrease. However, the vibrationalmodes ascribed to ZnO could not be recognized. Under theexcitation with 325 nm laser light, the Raman spectrum takenfrom the bare TiO2 NT arrays still features the anatase TiO2 modes.An additional weak and broad band centered around 820 cm�1 isresolved, and can be attributed to B2g vibrational mode of rutileTiO2 [B2g (R)] [29]. For ZnO coated TiO2 NT arrays, in contrast, threestrong scattering peaks predominate the Raman spectra in addi-tion to those scattered from TiO2. The strongest peak at 574 cm�1

is attributed to the polar optical mode associated with longitudinaloptical (LO) phonons of ZnO [A1 (LO)], while the other two peaksat 1147 and 1720 cm�1 are overtones of A1 (LO) and result from2-phonon and 3-phonon scattering processes [A1 (2LO) and A1

(3LO)] [32,34]. The efficient Raman scattering from ZnO is due tothe resonant excitation by 325 nm laser light whose photonenergy is resonant with the electronic interband transition energyof wurtzite ZnO. Although the Raman scattering from TiO2 isgreatly suppressed due to the coverage of ZnO on TiO2 NTs underthe excitation by 488 or 325 nm laser light, the above Ramanscattering measurements provide further evidence for the forma-tion of ZnO/TiO2 NTs composed of tetragonal anatase TiO2 nano-tubes and hexagonal wurtzite ZnO coatings.

3.3. Photoluminescence

Room-temperature PL spectra of the samples are shown inFig. 4. As expected, no obvious ultraviolet (UV) luminescencecorresponding to the near-band-edge (NBE) emission of TiO2 wasobserved by optically exciting the bare TiO2 NT arrays with 325 nmlaser light because of its indirect-band energy structure [35].

And almost no visible luminescence can be recognized, suggestingthe absence of oxygen vacancies [35]. In contrast, UV luminescencecentered at about 382 nm (equivalent photon energy of 3.246 eV)was observed from the as-fabricated ZnO/TiO2 NTs with the samelight excitation. This UV luminescence is emitted from the ZnOcoatings and corresponds to the room-temperature free exciton ofNBE emission of ZnO [34,36,37]. For the ZnO/TiO2 NT arrays afterannealing at 350 1C for 30 min in air, the intensity of the UV NBEemission decreases as compared with that of the sample withoutannealing. The decrease in UV emission suggests a less efficientradiative recombination of photogenerated electrons and holesbecause of charge separation of photogenerated carriers in a typeII nano-heterojunction constructed fromwurtzite ZnO and anataseTiO2. The interface between ZnO and TiO2 was improved duringthe post-fabrication annealing, which is favorable for the separa-tion of photogenerated electrons and holes, resulting in thesuppression of radiative recombination of photogenerated carriersand hence the reduction in PL intensity. For the ZnO/TiO2 NT arraysafter annealing at 450 1C, the UV NBE emission becomes strongerthan that of the sample annealed 350 1C, which is probably due tothe improvement in the crystal structure of ZnO by high-temperature annealing. However, the UV emission from the ZnOcoatings is still weaker than that of the as-fabricated ZnO/TiO2 NTarrays, suggesting that an enhancement of charge separation canalso be expected mainly due to the improved interface betweenZnO and TiO2. It is worthwhile noting that for the prepared ZnO/TiO2 NT arrays either annealed or not, the visible defect-relatedemission is very weak, indicating that few intrinsic defects such asoxygen vacancies exist in ZnO and TiO2 [35,38].

3.4. Photocurrent

The photoelectrochemical properties of the fabricated ZnO/TiO2

NT arrays were investigated by measuring the photocurrents ofZnO/TiO2 NT arrays used as the electrode under intermittentvisible illumination and comparing with that of TiO2 NT arraysunder the same illumination. Fig. 5(a) and (b) illustrates transientphotocurrent densities of TiO2 NT arrays (curve 1) and ZnO/TiO2

NT arrays (curves 2–4) over several on–off cycles of visibleillumination with the cutoff wavelengths of 380 and 420 nm,respectively. By comparing both figures, it can be seen that theprepared TiO2 NT arrays and ZnO/TiO2 NT arrays respond to thelight with wavelength longer than 420 nm less effectively than tothe light with wavelength shorter than 420 nm. Covering a TiO2

nanotube with a ZnO coating constructs a nano-heterojunctionwith type II staggered band alignment, in which the spatialseparation of the charge carriers will reduces the radiative

Fig. 3. Raman scattering spectra obtained by exciting bare TiO2 NTs (1),as-fabricated ZnO/TiO2 NTs (2), 350 1C annealed ZnO/TiO2 NTs (3), and 450 1Cannealed ZnO/TiO2 NTs (4) with 488 nm laser light (a) and 325 nm laser light (b).

Fig. 4. Room-temperature PL spectra of bare TiO2 NTs (1), as-fabricated ZnO/TiO2

NTs (2), 350 1C annealed ZnO/TiO2 NTs (3), and 450 1C annealed ZnO/TiO2 NTs (4).

H. Cai et al. / Solar Energy Materials & Solar Cells 123 (2014) 233–238236

recombination of the photogenerated electrons and holes [19–21].As a result, an enhanced photoresponse can be expected. However,it can be seen from Fig. 5 that the coverage of ZnO on the walls ofTiO2 NT arrays does not result in an obvious enhancement in thephotoresponse to visible illumination, but in a decrease of thephotocurrent density, observations which were not expected. Onlyfor the annealed ZnO/TiO2 NT arrays, a slight increase of photo-current density is observed. With the illumination by visible lightwith 380 nm cutoff, the photocurrent density increases from�3.8 μA/cm2 for the bare TiO2 NT arrays to �5.2 μA/cm2 for the

450 1C annealed ZnO/TiO2 NT arrays. The increase in photocurrentdensity is however less (from �1.4 μA/cm2 to �1.6 μA/cm2) underthe illumination by visible light with 420 nm cutoff. As describedabove, post-fabrication annealing improved the interface betweenZnO and TiO2 as well as the structures of ZnO and TiO2, which isfavorable for the charge separation of photogenerated carriers andthe suppression of radiative recombination of photogeneratedelectrons and holes. However, it seems that the enhancement inphotoresponse by the coverage of ZnO on the walls of TiO2 NTs isnot remarkable. One of the factors hampering the enhancement inphotoresponse is probably due to a too large thickness of the ZnOcoatings on the walls of the TiO2 NTs, which presents a long wayfor the generated electrons to migrate to the TiO2 surface and thegenerated holes to migrate to the ZnO surface, and hence isunfavorable for charge separation.

In order to effectively enhance the photoresponse, we fabri-cated ZnO/TiO2 NT arrays with thinner ZnO coatings on the wallsof TiO2 NTs (10 cycles). This configuration exhibits a significantenhancement in photoresponse of ZnO/TiO2 NT arrays as com-pared with that of the bare TiO2 NT arrays. Fig. 6(a) and (b)compares the photocurrent densities of the ZnO/TiO2 NT arraysfabricated by depositing 10 cycles of ZnO (approximately 2 nm) onthe walls of TiO2 NTs with those of the bare TiO2 NT arrays underthe same illumination as for the measurements illustrated in Fig. 5.Although the photocurrent densities of the as-fabricated ZnO/TiO2

NT arrays are smaller or equal to those of the bare TiO2 NT arraysunder the irradiation by 380 nm or 420 nm cutoff visible light, thephotocurrent densities of the annealed ZnO/TiO2 NT arraysincrease significantly. The enhanced photoelectrochemical proper-ties should be attributed to the improved interface between ZnOand TiO2 as well as the improved crystal structures of ZnO andTiO2 due to annealing. The former facilitates the charge separationof photogenerated carriers, while the latter increases the mobilityof photogenerated carriers. Under the irradiation by 420 nm cutoffvisible light, in particular, the photocurrent density increases from�1.4 μA/cm2 for the bare TiO2 NT arrays to �2.2 μA/cm2 for the450 1C annealed ZnO/TiO2 NT arrays. The photoresponsivity of theZnO/TiO2 nano-heterostructures has thus a nearly 1.5-fold increasecompared with that of bare TiO2 NTs, revealing an enhanced chargeseparation and a reduced radiative recombination of photoexcitedelectron�hole pairs in ZnO/TiO2 nano-heterostructure as comparedwith TiO2 NT. The transient photocurrent density curves illustratedin Fig. 6 also demonstrate that the vertically aligned ZnO/TiO2 NTarrays have a fast photoresponse speed and good photostabilitywhen used as electrodes. More important practically is the sig-nificant enhancement in photoresponse to the light with longerwavelengths, which reveals an extended region of photoresponseand a better match with the solar spectrum due to the presence ofZnO coatings on TiO2 NTs.

4. Conclusion

In summary, we have measured the photoluminescence andthe photocurrents of ZnO coated TiO2 NT arrays to study thecorrelation between the photoluminescence and the photoelec-trochemical properties of ZnO/TiO2 heterogeneous nanostructures.Vertically aligned ZnO/TiO2 NT arrays composed of tetragonalanatase TiO2 NTs and hexagonal wurtzite ZnO coatings werefabricated by atomic layer deposition of ZnO coatings on electro-chemically anodized TiO2 NTs. The crystal structures of both ZnOcoatings and TiO2 NTs and the interface between ZnO and TiO2

were improved by post-fabrication annealing. With a thin ZnO coatingon TiO2 NT, the formed type II staggered band alignment constructedfrom the band structures ZnO and TiO2 facilitated the charge separa-tion of photogenerated carries, subsequently suppressed the radiative

Fig. 5. Transient photocurrent densities of bare TiO2 NTs (1), as-fabricated ZnO/TiO2 NTs (2), 350 1C annealed ZnO/TiO2 NTs (3), and 450 1C annealed ZnO/TiO2 NTs(4) under visible illumination with 380 nm cutoff (a) and 420 nm cutoff (b).

Fig. 6. Transient photocurrent densities of bare TiO2 NTs (1), and ZnO/TiO2 NTswith 10-cycle deposition of ZnO coatings: as-fabricated (2), 350 1C annealed (3),and 450 1C annealed ZnO/TiO2 NTs (4) under visible illumination with 380 nmcutoff (a) and 420 nm cutoff (b).

H. Cai et al. / Solar Energy Materials & Solar Cells 123 (2014) 233–238 237

recombination of photogenerated carries and enhanced the photo-electrochemical properties. A reduction in photoluminescence and anincrease in photocurrent were observed for annealed ZnO/TiO2 NTarrays due to the improvement in the interface between ZnO and TiO2.The increased photocurrent also results from the improved crystalstructures of ZnO and TiO2. Compared with bare TiO2 NTs, thephotoelectrochemical activity increased to about 1.5-fold for 450 1Cannealed ZnO/TiO2 NTs with 10-cycle deposition of ZnO coatings onTiO2 NTs under visible illumination with 420 nm cutoff.

Acknowledgments

This work is supported by the National Basic Research Programof China (Contract no. 2012CB934303) and the National NaturalScience Foundation of China (Contract no. 11275051). Acknowl-edgment is also made to the Doctoral Fund of Chinese Ministry ofEducation (Contract no. 20110071110020).

References

[1] A. Hagfeldt, G. Boschloo, L.C. Sun, L. Kloo, H. Pettersson, Chem. Rev. 110 (2010)6595–6663.

[2] W. Krengvirat, S Sreekantan, A.F.M. Noor, N. Negishi, G. Kawamura, H. Muto,A. Matsuda, Mater. Chem. Phys. 137 (2013) 991–998.

[3] J.H. Lee, J.H. Shin, J.Y. Song, W.F. Wang, R. Schlaf, K.J. Kim, Y. Yi, J. Phys. Chem.C 116 (2012) 26342–26348.

[4] L.K. Tsui, T. Homma, G. Zangari, J. Phys. Chem. C 117 (2013) 6979–6989.[5] M. Hattori1, K. Noda, T. Nishi, K. Kobayashi, H. Yamada, K. Matsushige, Appl.

Phys. Lett. 102 (2013) (043105-1–043105-3).[6] H. Li, Ji.W. Cheng, S.W. Shu, J. Zhang, L.X. Zheng, C.K. Tsang, H. Cheng,

F.X. Liang, S.T. Lee, Y.Y. Li, Small 9 (2013) 37–44.[7] X.B. Chen, S.H. Shen, L.J. Guo, S.S. Mao, Chem. Rev. 110 (2010) 6503–6570.[8] H. Wender, A.F. Feil, L.B. Diaz, C.S. Ribeiro, G.J. Machado, P. Migowski, D.

E. Weibel, J. Dupont, S.R. Teixeira, ACS Appl. Mater. Interfaces 3 (2011)1359–1365.

[9] A. Kubacka, M. Fernandez-García, G. Colon, Chem. Rev. 111 (2012) 1555–1614.[10] J.R Jennings, A. Ghicov, L.M Peter, P. Schmuki, A. B Walker, J. Am. Chem. Soc.

130 (2008) 13364–13372.

[11] D. Kuang, J. Brillet, P. Chen, M. Takata, S. Uchida, H. Miura, K. Sumioka,S.M. Zakeeruddin, M. Gratzel, ACS Nano 2 (2008) 1113–1116.

[12] Y.Z. Liu, M. Misra, ACS Nano 4 (2010) 2196–2200.[13] O.A. Fouad, A.A. Iamail, Z.I. Zaki, R.M. Mohamed, Appl. Catal. B 62 (2006)

144–149.[14] M.A. Hasnat, M.M. Uddin, A.J.F. Samed, S.S. Alam, S. Hossain, J. Hazard. Mater.

147 (2007) 471–477.[15] R. Ullah, J. Dutta, J. Hazard. Mater. 156 (2008) 194–200.[16] P. Suri, R.M. Mehra, Sol. Energy Mater. Sol. Cells 91 (2007) 518–524.[17] P. Ravirajan, A.M. Peiro, M.K. Nazeeruddin, M. Graetzel, D.D.C. Bradley,

J.R. Durrant, J. Nelson, J. Phys. Chem. B 110 (2006) 7635–7639.[18] J. Huang, Z.G. Yin, Q.D. Zheng, Energy Environ. Sci. 4 (2011) 3861–3877.[19] P. Reiss, M Protiere, L. Li, Small 5 (2009) 154–168.[20] J. van Embden, J Jasieniak, D.E. Gomez, P. Mulvaney, M. Giersig, Aust. J. Chem.

60 (2007) 457–471.[21] Z.M. Wu, Y. Zhang, J.J. Zheng, X.G. Lin, X.H. Chen, B.W. Huang, H.Q. Wang,

K. Huang, S.P. Li, J.Y Kang, J. Mater. Chem. 21 (2011) 6020–6026.[22] V. Manthina, Juan Pablo Correa Baena, G.l. Liu, A.G. Agrios, J. Phys. Chem. C 116

(2012) 23864–23870.[23] Y.L. Xie, Z.X. Li, Z.G. Xu, H.L. Zhang, Electrochem. Commun. 13 (2011) 788–791.[24] F.X Xiao, ACS Appl. Mater. Interfaces 4 (2012) 7055–7063.[25] R. Liu, W.D. Yang, L.S. Qiang, H.Y. Liu, J. Power Sources 220 (2012) 153–159.[26] S.P.S. Porto, P.A. Fleury, T.C. Damen, Phys. Rev. 154 (1967) 522–526.[27] M. Mikami, S. Nakamura, O. Kitao, H. Arakawa, Phys. Rev. B 66 (2002) (155213-

1–155213-6).[28] T. Ohsaka, F. Izumi, Y. Fujiki, J. Raman Spectrosc. 7 (1978) 321–324.[29] A. Orendorz, A. Brodyanski, J. Losch, L.H. Bai, Z.H. Chen, Y.K. Le, C. Ziegler,

H. Gnaser, Surf. Sci. 601 (2007) 4390–4394.[30] V. Swamy, A. Kuznetsov, L.S. Dubrovinsky, R.A. Caruso, D.G. Shchukin,

B.C Muddle, Phys. Rev. B 71 (2005) (184302-1–184302-11).[31] T.C. Damen, S.P.S. Porto, B. Tell, Phys. Rev. 142 (1966) 570–574.[32] J.F. Scott, Phys. Rev. B 2 (1970) 1209–1211.[33] C.Y. Geng, Y. Jiang, Y. Yao, X.M. Meng, J.A. Zapien, C.S. Lee, Y. Lifshitz, S.T. Lee,

Adv. Funct. Mater. 14 (2004) 589–594.[34] Q. Li, K. Gao, Z.G. Hu, W.L. Yu, N. Xu, J. Sun, J.D. Wu, J. Phys. Chem. C 116 (2012)

2330–2335.[35] Y.H. Chang, C.M. Liu, C. Chen, H.E. Cheng, J. Electrochem. Soc. 159 (2012)

D401–D405.[36] W. Shan, W. Walukiewicz, J.W. Ager , K.M. Yu, H.B. Yuan, H.P. Xin, G. Cantwell,

J.J. Song, Appl. Phys. Lett., 191911-1–191911-3 (2005).[37] N. Xu, Y. Cui, Z.G. Hu, W.L. Yu, J. Sun, N. Xu, J.D. Wu, Opt. Express 20 (2012)

14857–14863.[38] K. Vanheusden, W.L Warren, C.H. Seager, D.R. Tallant, J.A. Voigt, B.E. Gnade,

J. Appl. Phys. 79 (1996) 7983–7990.

H. Cai et al. / Solar Energy Materials & Solar Cells 123 (2014) 233–238238