- Email: [email protected]
TiO2 nanorod arrays solar cell
G Model
ARTICLE IN PRESS
APSUSC-27841; No. of Pages 8
Applied Surface Science xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell Libo Yu, Zhen Li, Yingbo Liu, Fa Cheng, Shuqing Sun ∗ Department of Chemistry, Tianjin University, Tianjin 300072, PR China
a r t i c l e
i n f o
Article history: Received 24 February 2014 Received in revised form 23 April 2014 Accepted 5 May 2014 Available online xxx Keywords: TiO2 nanorod Mn-doped CdS CdSe Quantum dots sensitized solar cell
a b s t r a c t Vertically oriented single-crystalline one-dimensional TiO2 nanorod arrays was synthesized directly on transparent fluorine-doped tin oxide (FTO) conducting glass substrate by a facile hydrothermal method and was applied as photoanode in CdSe/Mn-doped CdS quantum dots sensitized solar cells (QDSSCs). The effect of coating cycles of QDs on the photovoltaic performance was investigated to find the optimal combination is 10 cycles of Mn-doped CdS and 9 cycles of CdSe, the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance due to the complementary effect in the light absorption of Mndoped CdS and CdSe QDs. The power conversion efficiency of CdSe(9)/Mn-CdS(10)/TiO2 solar cell reached to 2.40% under one sun illumination (AM 1.5 G, 100 mW/cm2 ), which was 46.34% higher than that of CdSe(9)/CdS(10)/TiO2 solar cell without doping of Mn (1.64%). © 2014 Elsevier B.V. All rights reserved.
1. Introduction TiO2 has attracted much attention in recent years due to its wide application in photocatalysis, gas sensing, water photoelectrolysis, and photoelectrochemical cells [1–4]. Nanostructured TiO2 was usually used as photoanode substrate for quantum dots sensitized solar cells (QDSSCs) because of its several advantages, such as suitable conduction band position, stable chemical and physical properties, strong optical absorption, and inexpensive cost. The one-dimensional (1D) TiO2 nanostructures including nanorods, nanowires, and nanotubes have shown excellent photoelectrochemical performance due to their efficient charge separation and transport properties [5–7]. The typical approaches to prepare ordered 1D TiO2 nanostructures for solar cell application include the template-assisted process [8], anodic oxidation [9] and hydrothermal synthesis [10]. An electrochemical anodization is well-known way to synthesize ordered nanotubes on any shape of Ti substrate such as foil or wire. However, the TiO2 nanotube arrays solar cells have to be constructed in a back illuminated system [11] because of the metal titanium foils are opaque, which cannot make full use of light. In order to overcome this problem, front-side illumination is a better way for QDSSC applications. 1D TiO2 nanorod arrays directly grown on FTO glass has recently been a focus of investigation for QDSSCs, and its advantages can be explained as
∗ Corresponding author. Tel.: +86 13920690912. E-mail address: [email protected] (S. Sun).
follows: first, 1D TiO2 nanorod arrays directly fabricated on FTO glass can be used as front-side illuminated photoanode in QDSSCs, enhancing the light harvesting efficiency. Second, 1D TiO2 nanorod arrays synthesized by hydrothermal method is in single-crystalline rutile TiO2 phase, which possess less boundaries for electron to pass, reducing the possibilities of recombination during the diffusion of electrons. Therefore, 1D single-crystalline TiO2 nanorod arrays grown directly on a transparent conductive FTO glass was fabricated by a modified hydrothermal approach in this study, and was used as substrate for sensitization with quantum dots (QDs). Recently, various narrow-band gap semiconductors including CdS, CdSe, CdTe, and PbS have been studied as sensitizers for QDSSCs [12–16]. Among these semiconductor materials, CdS and CdSe have shown much promising as impressive sensitizers due to their appropriate band gap of 2.25 eV and 1.70 eV in bulk, respectively, which can allow the extension of absorption band to the visible region of the solar spectrum [17,18]. Generally, single CdS or CdSe QDs sensitized solar cells presents low power conversion efficiency [19], thus cosensitized structure was proposed to improve the power conversion efficiency of QDSSCs. The cosensitized structure has been proved to be advantageous over single CdS or CdSe; this can be ascribed to the complementary effect [20] of the two kinds of QDs in light absorption and the improvement of light harvesting efficiency. Moreover, it is reported that doping of Mn2+ into QDs can modify intrinsic property of semiconductor QDs [21] and alter the charge separation and recombination dynamics in QDSSCs, which is in favor of the improvement of power conversion efficiency.
http://dx.doi.org/10.1016/j.apsusc.2014.05.023 0169-4332/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
2
In this work, 1D single-crystalline TiO2 nanorod arrays directly grown on FTO glass was synthesized via a modified hydrothermal method. Mn-doped CdS QDs and CdSe QDs were cosensitized onto TiO2 nanorod arrays by in situ successive ionic layer adsorption and reaction (SILAR) to form a front-side illuminated photoanode. Our CdSe/Mn-CdS/TiO2 solar cell exhibited much better photoelectrochemical performance than the CdSe/CdS/TiO2 solar cell, which verified the potential value of CdSe/Mn-doped CdS solar cell based on 1D single-crystalline TiO2 nanorod arrays in designing enhanced efficiency QDSSCs. 2. Experimental 2.1. Materials Titanium butoxide (Ti(OC4 H9 )4 ), cadmium nitrate (Cd(NO3 )2 ·4H2 O), sodium sulfide (Na2 S·9H2 O), sodium chloride (NaCl), sodium hydroxide (NaOH), sulfur powder (S), manganese acetate (Mn(CH3 COO)2 ·4H2 O), concentrated hydrochloric acid (HCl, 36.5–38 wt%), selenium powder (Se), sodium sulfite (Na2 SO3 ), copper sulfate (CuSO4 ·5H2 O) and thiourea (H2 NCSNH2 ) were purchased from Tianjin Chemical Reagents Co. Ltd. All chemicals were used directly in experiments without further purification. Deionized water (DI water, resistivity of 18.2 M cm) was obtained from MilliQ ultra-pure water system (Millipore, USA). 2.2. Preparation of TiO2 nanorod arrays on FTO glass The 1D single-crystalline TiO2 nanorod arrays on transparent conductive FTO glass electrode was prepared using the hydrothermal method reported by Liu and Aydil [22] with slight modifications. Briefly, 25 ml of DI water and 5 ml of saturated NaCl aqueous solution were mixed with 30 ml of concentrated hydrochloric acid. After stirring 5 min under ambient conditions, 1 ml of titanium butoxide was added dropwise to the mixture to obtain a clear transparent solution, and then was transferred to a Teflon-lined stainless steel autoclave, filling the 80% volume of the autoclave. One piece of FTO substrate (F: SnO2 , 14 /square) ultrasonically cleaned for 30 min in a mixed solution of DI water, acetone, and 2-propanol (volume ratios of 1:1:1), was placed at an angle against the wall of the Teflon liner with the conductive side facing down. The hydrothermal synthesis was conducted at the temperature of 150 ◦ C for 12 h in an electric oven. After synthesis, the autoclave was cooled to room temperature under flowing water. The product was taken out, rinsed thoroughly with DI water and ethanol respectively, and dried in ambient air. 2.3. Sensitization of Mn-doped CdS/CdSe QDs onto TiO2 nanorod electrodes by SILAR In situ successive ionic layer adsorption and reaction (SILAR) method [23] is used to assemble Mn-doped CdS and CdSe QDs onto the TiO2 nanorod photoanode. For Mn-doped CdS QDs, the TiO2 nanorod arrays electrode was dipped into a mixed solution of ethanol and DI water with volume ratio of 1:1 containing 0.1 M Cd(NO3 )2 and 0.075 M Mn(CH3 COO)2 for 5 min, and then dipped for another 5 min into the same mixed solution containing 0.1 M Na2 S. Following each dipping, the electrode was rinsed with ethanol for 2–3 min to remove excess precursors and dried at 150 ◦ C for 10 min before the next dipping. The two-step dipping procedure is termed as one SILAR cycle and the incorporated amount of Mndoped CdS can be increased by repeating the SILAR cycles. For CdSe QDs, Na2 SeSO3 is used as Se source, which is prepared by refluxing Se (0.3 M) in an aqueous solution of Na2 SO3 (0.6 M) at 70 ◦ C for about 7 h. The SILAR process of CdSe is similar to that of CdS except
that a longer time (15 min) is required for dipping the sample in the Na2 SeSO3 solution. We prepared several different SILAR cycles of CdSe/Mn-CdS/TiO2 electrodes to investigate the optimal combination of Mn-doped CdS and CdSe. As a control experiment, we also prepared CdSe/CdS/TiO2 electrode without doping of Mn2+ . 2.4. Solar cell fabrication The QD-sensitized TiO2 nanorod photoanode and Cu1.8 S/CuScoated FTO (the synthesis details can be found in the literature [24]) counter electrode were assembled in sandwich fashion using 60 m thick Surlyn film as spacer. 0.1 M S, 1 M Na2 S, and 0.1 M NaOH [25] in the co-solvent of water/methanol (3:7 by volume) was used as polysulfide electrolyte. The utilization of the co-solvent is to reduce the surface tension of the electrolyte, which is in favor of electrolyte penetration into TiO2 nanorod arrays and scavenging holes, thus improve the overall efficiency. The active area of the cell was 0.16 cm2 . 2.5. Characterization The morphology of the samples was characterized by fieldemission scanning electron microscope (FE-SEM) operating at 5.0 kV. Transmission electron microscope (TEM) and highresolution TEM (HR-TEM) investigations were carried out by JEOL JEM-2100F microscope. To prepare TEM samples, the QDs sensitized TiO2 nanorods was detached from the FTO substrate and dispersed in ethanol, and then the solution was dropped onto a copper grid. Energy dispersive spectroscopy (EDS) was used to analyze the composition of the samples. The X-ray diffraction (XRD) spectra of the samples were recorded by a Bruker D8 Advance X-ray diffrac˚ from 10◦ to 90◦ at a tometer using Cu K␣ radiation ( = 1.5416 A) scan rate of 2.4◦ min−1 . Diffuse reflectance absorption spectra of bare TiO2 nanorod arrays and QD-sensitized TiO2 nanorod arrays were recorded in the range from 300 to 800 nm using a Hitachi U-3010 spectroscopy. 2.6. Photovoltaic measurements The photocurrent–voltage (I–V) curves were measured by Oriel I–V test station under illumination of a solar simulator (Pecell-L, Japan) calibrated by standard silicon cell at one sun (AM 1.5 G, 100 mW/cm2 ). The active illuminated area of the QDSSC was fixed to 0.16 cm2 . The incident photon to current conversion efficiency (IPCE) measurements were performed with a monochromator to select the illumination wavelength, a 500 W xenon arc lamp (Oriel) served as a light source. 3. Results and discussion 3.1. Structure and morphology characterization of single-crystalline TiO2 nanorod arrays The phase structure of the as-prepared TiO2 nanorod arrays on FTO glass was characterized by X-ray diffraction patterns. As shown in Fig. 1, only the rutile structure of SnO2 (JCPDS no. 41-1445) can be indexed in the XRD patterns of bare FTO glass. It is worth noting that as-prepared TiO2 nanorod arrays show a tetragonal rutile structure (JCPDS no. 21-1276) after the hydrothermal reaction. The successful growth of TiO2 nanorods can be ascribed to small lattice mismatch between FTO and rutile TiO2 [26], which promote the epitaxial nucleation of TiO2 on FTO substrate. Compared with the intensity of (1 0 1) and (1 1 2) diffraction peaks in XRD patterns of as-prepared TiO2 , the enhanced (0 0 2) peak indicates that the nanorods were well crystallized and grew in [0 0 1] direction with the growth axis parallel to the substrate surface.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
Fig. 1. XRD patterns of bare FTO glass and TiO2 nanorod arrays.
Fig. 2a shows typical field-emission scanning electron microscopy (FE-SEM) image of the TiO2 nanorod arrays grown at 150 ◦ C for 12 h with saturated NaCl aqueous solution. The top facets of the nanorods are square, which shows the expected growth habit of the tetragonal crystal. The top surfaces of the nanorods appear to contain many step edges, which are the substrates for growth of the nanorod, while the side surfaces are smooth. Fig. 2b is a cross-sectional view of the corresponding sample, showing that the TiO2 nanorods are nearly perpendicular to the FTO glass and the length is about 3 m. In addition, a certain space can be identified among the TiO2 nanorods. The existence of space among TiO2 nanorods is beneficial for the penetration of electrolyte and the deposition of the QDs. Fig. 2c and d is the top view and crosssectional view of TiO2 nanorod arrays synthesized at 150 ◦ C for 12 h without saturated NaCl aqueous solution. It can be obviously seen that the density of TiO2 nanorods increased, and nanorods are connected together to form a continuous film. Furthermore, the side surfaces of the TiO2 nanorods coalesced with each other and reduced internal surface area of the TiO2 nanorod arrays. This will result in deficient quantum dots loading and light harvesting as previously reported [6,27]. However, the length of TiO2 nanorods synthesized without NaCl solution is unchanged, which is still about 3 m, indicating that the saturated NaCl solution has no influence on longitude growth of TiO2 nanorods. The effect of saturated NaCl aqueous solution to the formation of TiO2 nanorods can be interpreted as follows: Cl− can preferentially adsorb onto the (1 1 0) crystal plane of TiO2 , which retard the growth rate of (1 1 0) crystal plane [28,29], resulting in the suppression of lateral growth. Energy dispersive spectroscopy (EDS) was applied to determine the composition of the nanostructure, which is shown in Fig. 2e. It can be seen that Ti and O are dominant elements (atomic ratio of 1:2), which confirms that the nanorod arrays were mainly composed of TiO2 . Other elements (C came from titanium butoxide, Si and Sn came from FTO glass) can also be identified by EDS spectra, but these elements are in a relatively low level. 3.2. Characterization of CdSe/Mn-CdS/TiO2 nanorod arrays After sensitization with Mn-doped CdS and CdSe QDs, the ordered TiO2 nanorod arrays structure was retained, as shown in Fig. 3a. It is obvious that the surface of TiO2 nanorod became rougher than that of bare TiO2 nanorod (Fig. 2a), which means that the QDs had covered on the surface of TiO2 nanorod after SILAR. The composition of the sensitized TiO2 electrode was analyzed by the EDS spectrum in Fig. 3b. The Ti, O, Cd, S, Se and Mn peaks can be
3
observed in the EDS spectrum of the QDs sensitized TiO2 nanorods. What is more, quantitative analysis from the EDS reveals that the atomic ratio of Cd versus S plus Se is close to 1, indicating that the deposition of Mn-doped CdS and CdSe QDs are likely to be stoichiometric. This result confirmed the successful deposition of QDs onto the TiO2 nanorods. There was no other element can be distinguished in this EDS, suggesting the CdSe/Mn-CdS/TiO2 nanorod arrays has a pure element composition. The detailed microscopic characterization of the CdSe/MnCdS/TiO2 nanorod structure was performed by transmission electron microscope (TEM) and high-resolution TEM (HR-TEM). Fig. 3c is the TEM image of CdSe/Mn-CdS/TiO2 nanorod, it can be seen that the surface of the TiO2 nanorod was covered by QDs. The HR-TEM image of CdSe/Mn-CdS/TiO2 nanorod is used to further verify the successful deposition of QDs onto TiO2 nanorods. As shown in Fig. 3d, clear lattice fringes corresponding to well crystallized TiO2 nanorod, Mn-doped CdS, and CdSe QDs can be distinguished. The lattice spacing of 0.322 nm belonging to the (1 1 0) plane of rutile TiO2 (JCPDS no. 21-1276) can be observed. Among the crystal planes of rutile TiO2 , the (1 1 0) plane possesses the lowest surface energy [30]. According to the crystal growth theory, the planes with lower surface energies generally grow slowly and tend to survive during the growth [31]. Therefore, the (1 1 0) plane can be obviously observed in the HRTEM photographs. Connecting to the TiO2 , the lattice fringe of 0.335 nm, which ascribes to the CdS (1 1 1) (JCPDS no. 10-0454) is identified, and close to the CdS layer, outer layer crystallites with lattice spacing of 0.351 nm corresponding to the (0 0 2) plane of CdSe (JCPDS no. 08-0459) can also be observed. These results demonstrated that Mn-doped CdS and CdSe QDs have been successfully deposited on TiO2 nanorods.
3.3. Optical absorption spectra of the QDs sensitized TiO2 nanorod photoanode The UV–vis diffuse reflectance absorption spectra were used to record the variation of light absorption for the bare TiO2 nanorod photoanode and QD-sensitized TiO2 nanorod photoanodes. As shown in Fig. 4, the absorption edge of bare TiO2 nanorods (spectrum a, Fig. 4) appears at about 405 nm and the maximum peak appears at about 300 nm. No significant absorption in visible-light region can be seen due to the large band gap (3.0 eV) of rutile TiO2 [32]. Compared with the bare TiO2 nanorods electrode, the absorption edges, obtained from the intersection of the sharply decreasing region of a spectrum with its baseline, were enlarged to 520 nm for CdS/TiO2 (spectrum b, Fig. 4) electrode, and 670 nm for CdSe/TiO2 (spectrum d, Fig. 4) electrode. The band gaps of CdS and CdSe QDs were calculated to be 2.39 and 1.85 eV, which are higher than the values of bulk CdS (2.25 eV) and CdSe (1.70 eV) due to the quantum size effect, this indicates that the sizes of CdS and CdSe on the TiO2 nanorod are within the scale of QDs. According to the empirical equations given by Yu et al. [33], the sizes of CdS and CdSe particles were estimated to be 8.796 and 9.713 nm, respectively. It is worth noting that the optical absorption spectrum of Mn-doped CdS/TiO2 (spectrum c, Fig. 4) is broader than that of the CdS/TiO2 electrode. This enlarged absorption range could be ascribed to a slightly broader size distribution in doped sample which is caused by the incorporation of Mn into the crystal lattice of CdS [34]. For the cosensitized electrodes, the absorption range of both undoped and Mn-doped CdS/CdSe cosensitized TiO2 electrodes is broader than that of CdS/TiO2 and CdSe/TiO2 electrodes. Furthermore, the absorption onset of the CdSe/Mn-CdS/TiO2 (spectrum f, Fig. 4) shows a slightly red-shift compared to CdSe/CdS/TiO2 (spectrum e, Fig. 4). These results indicate that the light absorption range can be modulated with doping Mn in QDs, providing its application value to improve the performance of QDSSCs.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
4
Fig. 2. FE-SEM images of TiO2 nanorod array film: (a) top view and (b) cross-sectional view with saturated NaCl aqueous solution; (c) top view and (d) cross-sectional view without saturated NaCl aqueous solution; (e) the EDS spectra of TiO2 nanorod array directly grown on FTO glass.
3.4. Photoelectrochemical characterization The incident photon to current conversion efficiency (IPCE) spectra of QDSSCs with the Mn-doped CdS/TiO2 nanorods, CdSe/TiO2 nanorods, and CdSe/Mn-CdS/TiO2 nanorods are shown in Fig. 5. It can be seen that the onsets of IPCE around 600 nm for Mn-doped CdS/TiO2 solar cell, 700 nm for CdSe/TiO2 solar cell and 750 nm for CdSe/Mn-CdS/TiO2 solar cell. The value of IPCE is determined by light harvesting efficiency, charge injection
efficiency, and charge collection efficiency [35]. The conduction band edge of CdS was higher than CdSe, which resulted in the charge injection and collection efficiency of CdS were higher than that of CdSe, thus the IPCE of Mn-doped CdS/TiO2 nanorods cell can reach to 40%, which is higher than CdSe/TiO2 nanorods cell (28%). It is worth noting that the 45% of IPCE value can be obtained for the CdSe/Mn-CdS/TiO2 nanorods cell, which is higher than Mn-doped CdS or CdSe sensitized TiO2 nanorods cell, the possible explanation may be as follows: in short wavelength region ( < 520 nm), the
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
5
Fig. 3. (a) FE-SEM image of CdSe/Mn-CdS/TiO2 nanorod electrode; (b) EDS spectra of corresponding electrode; (c) and (d) TEM and HR-TEM images of CdSe/Mn-CdS/TiO2 nanorod electrode.
higher IPCE of CdSe/Mn-CdS/TiO2 solar cell came from the contributions of Mn-doped CdS and CdSe QDs in the light harvest; while in long wavelength region (520 nm < < 800 nm), the higher IPCE of CdSe/Mn-CdS/TiO2 cell indicates that the introduction of a CdS layer between TiO2 and CdSe is helpful for the collection of excited electrons from CdSe to TiO2 [20].
Fig. 4. Diffuse reflectance absorption spectra of bare TiO2 nanorod photoanode and the TiO2 photoanodes sensitized by CdS, Mn-doped CdS, CdSe, CdSe/CdS, and CdSe/Mn-doped CdS.
The J–V characteristics of the assembled QDSSCs were measured by Oriel I–V test station under one sun (=100 mW/cm2 AM 1.5 G solar illumination) with the active area of 0.16 cm2 . Fig. 6a shows the performance variation of TiO2 nanorod solar cells sensitized by different SILAR cycles of Mn-doped CdS QDs. It is found that both the short circuit current density (Jsc ) and the open circuit voltage (Voc ) increased gradually with the increase of SILAR cycles at initial stage, and the optimum values of these two parameters (7.53 mA/cm2 , 0.36 V) could be obtained at 10 SILAR cycles
Fig. 5. IPCE spectra of QDSSCs assembled with the Mn-doped CdS/TiO2 , CdSe/TiO2 , and CdSe/Mn-CdS/TiO2 photoanodes.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8 6
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx Table 1 Photovoltaic performance parameters of QDSSCs assembled with different photoanodes. Samples
Jsc (mA/cm2 )
Voc (V)
FF
(%)
Mn-doped CdS(4)/TiO2 Mn-doped CdS(7)/TiO2 Mn-doped CdS(10)/TiO2 Mn-doped CdS(13)/TiO2 CdSe(0)/Mn-CdS(10)/TiO2 CdSe(3)/Mn-CdS(10)/TiO2 CdSe(6)/Mn-CdS(10)/TiO2 CdSe(9)/Mn-CdS(10)/TiO2 CdSe(12)/Mn-CdS(10)/TiO2 CdSe(15)/Mn-CdS(10)/TiO2
5.03 6.79 7.53 5.85 7.53 9.02 11.81 13.38 12.58 10.29
0.34 0.35 0.36 0.35 0.36 0.38 0.39 0.42 0.41 0.39
0.41 0.43 0.44 0.41 0.44 0.41 0.43 0.43 0.43 0.43
0.70 1.03 1.19 0.83 1.19 1.41 1.97 2.40 2.23 1.73
Jsc and Voc increased gradually to maximum as CdSe cycles varying from 0 to 9. The increase of Jsc can be attributed to the extending of light absorption range and the increase of absorbance with augmented amount of CdSe QDs loading. The improvement of Voc resulted from quasi-Fermi level of the electrons in semiconductor photoanode tends to more negative. However, when SILAR cycles of CdSe exceeded 9 cycles, it can be seen that both the Jsc and Voc values began to drop, resulting in the decrease of . The excessive SILAR cycles can lead to the aggregation of QDs, which may increase the possibilities of the recombination between the electrons and holes when the photogenerated electrons diffuse across the QD layers to the TiO2 film, resulting in an eventual decrease in the overall efficiency. In addition, the excessive SILAR cycles may also hinder the diffusion of the electrolyte, which limit the efficiency of charge separation and charge extraction. The power conversion efficiency () and short circuit current (Jsc ) as a function of SILAR cycles of CdSe was summarized in Fig. 6c, the value of Jsc and the increased first (from 3 to 9 cycles) and then decreased (from 9 to 15 cycles). The optimum value of the was obtained at 9 cycles of CdSe, thus it can be concluded that the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance, which Jsc , Voc and were 13.38 mA/cm2 , 0.42 V, and 2.40%, respectively. The short circuit current (Jsc ), open circuit voltage (Voc ), fill factor (FF), and power conversion efficiency () of all the solar cells are summarized in Table 1. The performance of CdSe/Mn-CdS/TiO2 nanorod solar cell is superior to Mn-doped CdS/TiO2 nanorod solar cell and CdSe/TiO2 nanorod solar cell, which is shown in Fig. 7. For the cell fabricated with CdSe(9)/Mn-CdS(10)/TiO2 photoanode, the Jsc and Voc were 13.38 mA/cm2 and 0.42 V, respectively, resulting in a very high of 2.40%, which is 101.68% higher than that of Mn-doped CdS(10)/TiO2 nanorod solar cell (Jsc of 7.53 mA/cm2 , Voc of 0.36 V, and of 1.19%) and 247.83% higher than that of CdSe(9)/TiO2 nanorod solar cell
Fig. 6. (a) and (b) J–V curves for Mn-doped CdS/TiO2 nanorods and CdSe/MnCdS/TiO2 nanorods solar cells measured under AM 1.5 G condition, respectively. (c) Power conversion efficiency and photocurrent density as a function of SILAR cycles of CdSe QDs.
of Mn-doped CdS. Then further increase of SILAR cycles of Mndoped CdS QDs (13 cycles) lead to the reduction of the Jsc and Voc . Therefore, it can be concluded that 10 SILAR cycles is optimal for Mn-doped CdS/TiO2 nanorods solar cells. Based on this, different SILAR cycles of CdSe on Mn-doped CdS/TiO2 nanorods were investigated to seek the best cycles for cosensitization system. Fig. 6b displays the J–V characteristics for different SILAR cycles of CdSe onto Mn-doped CdS(10)/TiO2 nanorod solar cells. Obviously, both
Fig. 7. J–V curves of the QDSSCs assembled with different photoanodes measured under AM 1.5 G condition.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
7
Fig. 8. Schematic diagram of the photoelectrical conversion configuration of CdSe/Mn-CdS/TiO2 solar cell.
(Jsc of 5.54 mA/cm2 , Voc of 0.34 V, and of 0.69%). The full utilization of light is the most important factor for enhanced power conversion efficiency of QDSSCs. The band gap of CdS in bulk (2.25 eV) limits its absorption range below the wavelength of approximately 550 nm, which leads to the of Mn-doped CdS(10)/TiO2 nanorod solar cell in a low level. Although CdSe (1.70 eV in bulk) has a wider absorption range than CdS, its electron injection efficiency is not as good as CdS [36,37], also resulting in a low for CdSe(9)/TiO2 nanorod solar cell. Therefore, CdS/CdSe cosensitized TiO2 nanorods can make full use of their respective advantages, such as extension of light absorption range and increase in absorbance, resulting in the enhancement of power conversion efficiency as presented in Fig. 7. In order to evaluate the effect of Mn-doping on the cell performance, we prepared CdSe(9)/CdS(10)/TiO2 cell as a control group. As shown in Fig. 7, after Mn-doping, the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance and the power conversion efficiency can reach to 2.40%, which is 46.34% higher than CdSe(9)/CdS(10)/TiO2 cell (10.06 mA/cm2 , 0.40 V, 1.64%), the increase of power conversion efficiency can be attributed to the midgap states created by Mn2+ in CdS QDs [21], which may cause the electrons to get trapped and screen them from charge recombination with holes, leading to the improvement of charge collection efficiency, thus improve the cell performance. The photoelectrical conversion configuration of CdSe/MnCdS/TiO2 solar cell is shown in Fig. 8, the configuration of the solar cell consisted of TiO2 nanorod film, CdSe/Mn-doped CdS QDs sensitizer, polysulfide electrolyte, and Cu1.8 S/CuS/FTO counter electrode. Because TiO2 nanorod arrays were directly grown on transparent FTO glass, it can be used as a front-side light irradiation photoanode. As shown in the schematic diagram, under illumination, photons are captured by QDs, yielding electron–hole pairs. The electrons are rapidly injected into the TiO2 nanorod arrays film, simultaneously, the remaining holes are scavenged to the counter electrode via a hole transporting redox couple comprising of polysulfide electrolyte. From the working mechanism of QDSSC mentioned above, it can be found that the amount of light harvesting and adsorbed QDs and electron transport are the key factors affecting the performance of QDSSC. The Mn-doped CdS/CdSe consensitzed structure extended the scope of light absorption, increasing the amount of light harvesting. The TiO2 nanorods synthesized by addition saturated NaCl to the grown solution possessed high aspect ratio and high internal surface, which are advantageous for efficient quantum dots loading and electrolytes filling. In addition, the single-crystalline one-dimensional TiO2 nanorod arrays offered a direct electrical pathway for photogenerated electrons, which can increase the electron transport rate. Thus, it can be concluded that
our CdSe/Mn-CdS/TiO2 nanorod solar cell is of great potential for designing high efficiency solar cells. 4. Conclusion In summary, 1D single-crystalline TiO2 nanorod array was successfully synthesized on FTO glass. The CdSe/Mn-doped CdS QDs were assembled onto TiO2 nanorod film by SILAR and used as photoanode in solar cells. The effect of SILAR cycles on the performance of QDSSCs was investigated, and the optimal combination was obtained with Mn-doped CdS(10)/CdSe(9). The short circuit photocurrent density of 13.38 mA/cm2 , the open circuit potential of 0.42 V, and the power conversion efficiency of 2.40% were achieved with CdSe(9)/Mn-CdS(10)/TiO2 solar cell, which was 46.34% higher than CdSe(9)/CdS(10)/TiO2 solar cell (1.64%). The acceptable power conversion efficiency of our CdSe(9)/Mn-CdS(10)/TiO2 solar cell has demonstrated that its significance in designing high efficiency QDSSCs. Acknowledgment This work was supported by Key Project of Tianjin Sci-Tech Support Program (No. 08ZCKFSH01400). References [1] J. Wang, T. Zhang, D. Wang, R. Pan, Q. Wang, H. Xia, J. Alloys Compd. 551 (2013) 82–87. [2] S. Wang, G. Xia, H. He, K. Yi, J. Shao, Z. Fan, J. Alloys Compd. 431 (2007) 287–291. [3] F. Shao, J. Sun, L. Gao, S. Yang, J. Luo, ACS Appl. Mater. Interfaces 3 (2011) 2148–2153. [4] A.M. More, T.P. Gujar, J.L. Gunjakar, C.D. Lokhande, O.-S. Joo, Appl. Surf. Sci. 255 (2008) 2682–2687. [5] M. Yang, B. Ding, S. Lee, J.-K. Lee, J. Phys. Chem. C 115 (2011) 14534–14541. [6] Z.-J. Zhou, J.-Q. Fan, X. Wang, W.-H. Zhou, Z.-L. Du, S.-X. Wu, ACS Appl. Mater. Interfaces 3 (2011) 4349–4353. [7] S. Cheng, W. Fu, H. Yang, L. Zhang, J. Ma, H. Zhao, M. Sun, L. Yang, J. Phys. Chem. C 116 (2012) 2615–2621. [8] F. Xu, Y. Wu, X. Zhang, Z. Gao, K. Jiang, Micro Nano Lett. 7 (2012) 826–830. [9] J. Zhang, R.-G. Du, Z.-Q. Lin, Y.-F. Zhu, Y. Guo, H.-Q. Qi, L. Xu, C.-J. Lin, Electrochim. Acta 83 (2012) 59–64. [10] S. Feng, J. Yang, M. Liu, H. Zhu, J. Zhang, G. Li, J. Peng, Q. Liu, Thin Solid Films 520 (2012) 2745–2749. [11] S.-H. Hsu, S.-F. Hung, S.-H. Chien, J. Power Sources 233 (2013) 236–243. [12] K. Prabakar, H. Seo, M. Son, H. Kim, Mater. Chem. Phys. 117 (2009) 26–28. [13] B. Mukherjee, Y.R. Smith, V. Subramanian, J. Phys. Chem. C 116 (2012) 15175–15184. [14] H. Gholap, R. Patil, P. Yadav, A. Banpurkar, S. Ogale, W. Gade, Nanotechnology 24 (2013) 195101. [15] S.D. Sung, I. Lim, P. Kang, C. Lee, W.I. Lee, Chem. Commun. 49 (2013) 6054–6056. [16] S.S. Kalanur, Y.J. Hwang, O.-S. Joo, J. Colloid Interface Sci. 402 (2013) 94–99.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
[17] M. Li, Y. Liu, H. Wang, H. Shen, W. Zhao, H. Huang, C. Liang, J. Appl. Phys. 108 (2010) 094304. [18] G. Zhu, L. Pan, T. Xu, Z. Sun, ACS Appl. Mater. Interfaces 3 (2011) 3146–3151. [19] Y. Lai, Z. Lin, D. Zheng, L. Chi, R. Du, C. Lin, Electrochim. Acta 79 (2012) 175–181. [20] Y.-L. Lee, C.-F. Chi, S.-Y. Liau, Chem. Mater. 22 (2010) 922–927. [21] P.K. Santra, P.V. Kamat, J. Am. Chem. Soc. 134 (2012) 2508–2511. [22] B. Liu, E.S. Aydil, J. Am. Chem. Soc. 131 (2009) 3985–3990. [23] Y.-L. Lee, Y.-S. Lo, Adv. Funct. Mater. 19 (2009) 604–609. [24] S.S. Kalanur, S.Y. Chae, O.S. Joo, Electrochim. Acta 103 (2013) 91–95. [25] S. Rühle, S. Yahav, S. Greenwald, A. Zaban, J. Phys. Chem. C 116 (2012) 17473–17478. [26] C. Wang, Z. Jiang, L. Wei, Y. Chen, J. Jiao, M. Eastman, H. Liu, Nano Energy 1 (2012) 440–447. [27] C. Xu, P.H. Shin, L. Cao, J. Wu, D. Gao, Chem. Mater. 22 (2010) 143–148.
[28] K. Kakiuchi, E. Hosono, H. Imai, T. Kimura, S. Fujihara, J. Cryst. Growth 293 (2006) 541–545. [29] X. Feng, J. Zhai, L. Jiang, Angew. Chem. Int. Ed. 44 (2005) 5115–5118. [30] U. Diebold, Appl. Phys. A: Mater. Sci. Process. 76 (2003) 681–687. [31] M. Gao, Y. Li, M. Guo, M. Zhang, X. Wang, J. Mater. Sci. Technol. 28 (2012) 577–586. [32] M. Xu, P. Da, H. Wu, D. Zhao, G. Zheng, Nano Lett. 12 (2012) 1503–1508. [33] W.W. Yu, L. Qu, W. Guo, X. Peng, Chem. Mater. 15 (2003) 2854–2860. [34] M.A. Malik, P. O’Brien, N. Revaprasadu, J. Mater. Chem. 11 (2001) 2382–2386. [35] S.W. Jung, J.-H. Kim, H. Kim, C.-J. Choi, K.-S. Ahn, Curr. Appl. Phys. 12 (2012) 1459–1464. [36] M. Wang, J. Jiang, J. Shi, L. Guo, ACS Appl. Mater. Interfaces 5 (2013) 4021–4025. [37] J. Wang, T. Zhang, Q. Wang, D. Wang, R. Pan, H. Xia, J. Wuhan Univ. Technol. Mater. Sci. Ed. 27 (2012) 876–880.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
ARTICLE IN PRESS
APSUSC-27841; No. of Pages 8
Applied Surface Science xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell Libo Yu, Zhen Li, Yingbo Liu, Fa Cheng, Shuqing Sun ∗ Department of Chemistry, Tianjin University, Tianjin 300072, PR China
a r t i c l e
i n f o
Article history: Received 24 February 2014 Received in revised form 23 April 2014 Accepted 5 May 2014 Available online xxx Keywords: TiO2 nanorod Mn-doped CdS CdSe Quantum dots sensitized solar cell
a b s t r a c t Vertically oriented single-crystalline one-dimensional TiO2 nanorod arrays was synthesized directly on transparent fluorine-doped tin oxide (FTO) conducting glass substrate by a facile hydrothermal method and was applied as photoanode in CdSe/Mn-doped CdS quantum dots sensitized solar cells (QDSSCs). The effect of coating cycles of QDs on the photovoltaic performance was investigated to find the optimal combination is 10 cycles of Mn-doped CdS and 9 cycles of CdSe, the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance due to the complementary effect in the light absorption of Mndoped CdS and CdSe QDs. The power conversion efficiency of CdSe(9)/Mn-CdS(10)/TiO2 solar cell reached to 2.40% under one sun illumination (AM 1.5 G, 100 mW/cm2 ), which was 46.34% higher than that of CdSe(9)/CdS(10)/TiO2 solar cell without doping of Mn (1.64%). © 2014 Elsevier B.V. All rights reserved.
1. Introduction TiO2 has attracted much attention in recent years due to its wide application in photocatalysis, gas sensing, water photoelectrolysis, and photoelectrochemical cells [1–4]. Nanostructured TiO2 was usually used as photoanode substrate for quantum dots sensitized solar cells (QDSSCs) because of its several advantages, such as suitable conduction band position, stable chemical and physical properties, strong optical absorption, and inexpensive cost. The one-dimensional (1D) TiO2 nanostructures including nanorods, nanowires, and nanotubes have shown excellent photoelectrochemical performance due to their efficient charge separation and transport properties [5–7]. The typical approaches to prepare ordered 1D TiO2 nanostructures for solar cell application include the template-assisted process [8], anodic oxidation [9] and hydrothermal synthesis [10]. An electrochemical anodization is well-known way to synthesize ordered nanotubes on any shape of Ti substrate such as foil or wire. However, the TiO2 nanotube arrays solar cells have to be constructed in a back illuminated system [11] because of the metal titanium foils are opaque, which cannot make full use of light. In order to overcome this problem, front-side illumination is a better way for QDSSC applications. 1D TiO2 nanorod arrays directly grown on FTO glass has recently been a focus of investigation for QDSSCs, and its advantages can be explained as
∗ Corresponding author. Tel.: +86 13920690912. E-mail address: [email protected] (S. Sun).
follows: first, 1D TiO2 nanorod arrays directly fabricated on FTO glass can be used as front-side illuminated photoanode in QDSSCs, enhancing the light harvesting efficiency. Second, 1D TiO2 nanorod arrays synthesized by hydrothermal method is in single-crystalline rutile TiO2 phase, which possess less boundaries for electron to pass, reducing the possibilities of recombination during the diffusion of electrons. Therefore, 1D single-crystalline TiO2 nanorod arrays grown directly on a transparent conductive FTO glass was fabricated by a modified hydrothermal approach in this study, and was used as substrate for sensitization with quantum dots (QDs). Recently, various narrow-band gap semiconductors including CdS, CdSe, CdTe, and PbS have been studied as sensitizers for QDSSCs [12–16]. Among these semiconductor materials, CdS and CdSe have shown much promising as impressive sensitizers due to their appropriate band gap of 2.25 eV and 1.70 eV in bulk, respectively, which can allow the extension of absorption band to the visible region of the solar spectrum [17,18]. Generally, single CdS or CdSe QDs sensitized solar cells presents low power conversion efficiency [19], thus cosensitized structure was proposed to improve the power conversion efficiency of QDSSCs. The cosensitized structure has been proved to be advantageous over single CdS or CdSe; this can be ascribed to the complementary effect [20] of the two kinds of QDs in light absorption and the improvement of light harvesting efficiency. Moreover, it is reported that doping of Mn2+ into QDs can modify intrinsic property of semiconductor QDs [21] and alter the charge separation and recombination dynamics in QDSSCs, which is in favor of the improvement of power conversion efficiency.
http://dx.doi.org/10.1016/j.apsusc.2014.05.023 0169-4332/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
2
In this work, 1D single-crystalline TiO2 nanorod arrays directly grown on FTO glass was synthesized via a modified hydrothermal method. Mn-doped CdS QDs and CdSe QDs were cosensitized onto TiO2 nanorod arrays by in situ successive ionic layer adsorption and reaction (SILAR) to form a front-side illuminated photoanode. Our CdSe/Mn-CdS/TiO2 solar cell exhibited much better photoelectrochemical performance than the CdSe/CdS/TiO2 solar cell, which verified the potential value of CdSe/Mn-doped CdS solar cell based on 1D single-crystalline TiO2 nanorod arrays in designing enhanced efficiency QDSSCs. 2. Experimental 2.1. Materials Titanium butoxide (Ti(OC4 H9 )4 ), cadmium nitrate (Cd(NO3 )2 ·4H2 O), sodium sulfide (Na2 S·9H2 O), sodium chloride (NaCl), sodium hydroxide (NaOH), sulfur powder (S), manganese acetate (Mn(CH3 COO)2 ·4H2 O), concentrated hydrochloric acid (HCl, 36.5–38 wt%), selenium powder (Se), sodium sulfite (Na2 SO3 ), copper sulfate (CuSO4 ·5H2 O) and thiourea (H2 NCSNH2 ) were purchased from Tianjin Chemical Reagents Co. Ltd. All chemicals were used directly in experiments without further purification. Deionized water (DI water, resistivity of 18.2 M cm) was obtained from MilliQ ultra-pure water system (Millipore, USA). 2.2. Preparation of TiO2 nanorod arrays on FTO glass The 1D single-crystalline TiO2 nanorod arrays on transparent conductive FTO glass electrode was prepared using the hydrothermal method reported by Liu and Aydil [22] with slight modifications. Briefly, 25 ml of DI water and 5 ml of saturated NaCl aqueous solution were mixed with 30 ml of concentrated hydrochloric acid. After stirring 5 min under ambient conditions, 1 ml of titanium butoxide was added dropwise to the mixture to obtain a clear transparent solution, and then was transferred to a Teflon-lined stainless steel autoclave, filling the 80% volume of the autoclave. One piece of FTO substrate (F: SnO2 , 14 /square) ultrasonically cleaned for 30 min in a mixed solution of DI water, acetone, and 2-propanol (volume ratios of 1:1:1), was placed at an angle against the wall of the Teflon liner with the conductive side facing down. The hydrothermal synthesis was conducted at the temperature of 150 ◦ C for 12 h in an electric oven. After synthesis, the autoclave was cooled to room temperature under flowing water. The product was taken out, rinsed thoroughly with DI water and ethanol respectively, and dried in ambient air. 2.3. Sensitization of Mn-doped CdS/CdSe QDs onto TiO2 nanorod electrodes by SILAR In situ successive ionic layer adsorption and reaction (SILAR) method [23] is used to assemble Mn-doped CdS and CdSe QDs onto the TiO2 nanorod photoanode. For Mn-doped CdS QDs, the TiO2 nanorod arrays electrode was dipped into a mixed solution of ethanol and DI water with volume ratio of 1:1 containing 0.1 M Cd(NO3 )2 and 0.075 M Mn(CH3 COO)2 for 5 min, and then dipped for another 5 min into the same mixed solution containing 0.1 M Na2 S. Following each dipping, the electrode was rinsed with ethanol for 2–3 min to remove excess precursors and dried at 150 ◦ C for 10 min before the next dipping. The two-step dipping procedure is termed as one SILAR cycle and the incorporated amount of Mndoped CdS can be increased by repeating the SILAR cycles. For CdSe QDs, Na2 SeSO3 is used as Se source, which is prepared by refluxing Se (0.3 M) in an aqueous solution of Na2 SO3 (0.6 M) at 70 ◦ C for about 7 h. The SILAR process of CdSe is similar to that of CdS except
that a longer time (15 min) is required for dipping the sample in the Na2 SeSO3 solution. We prepared several different SILAR cycles of CdSe/Mn-CdS/TiO2 electrodes to investigate the optimal combination of Mn-doped CdS and CdSe. As a control experiment, we also prepared CdSe/CdS/TiO2 electrode without doping of Mn2+ . 2.4. Solar cell fabrication The QD-sensitized TiO2 nanorod photoanode and Cu1.8 S/CuScoated FTO (the synthesis details can be found in the literature [24]) counter electrode were assembled in sandwich fashion using 60 m thick Surlyn film as spacer. 0.1 M S, 1 M Na2 S, and 0.1 M NaOH [25] in the co-solvent of water/methanol (3:7 by volume) was used as polysulfide electrolyte. The utilization of the co-solvent is to reduce the surface tension of the electrolyte, which is in favor of electrolyte penetration into TiO2 nanorod arrays and scavenging holes, thus improve the overall efficiency. The active area of the cell was 0.16 cm2 . 2.5. Characterization The morphology of the samples was characterized by fieldemission scanning electron microscope (FE-SEM) operating at 5.0 kV. Transmission electron microscope (TEM) and highresolution TEM (HR-TEM) investigations were carried out by JEOL JEM-2100F microscope. To prepare TEM samples, the QDs sensitized TiO2 nanorods was detached from the FTO substrate and dispersed in ethanol, and then the solution was dropped onto a copper grid. Energy dispersive spectroscopy (EDS) was used to analyze the composition of the samples. The X-ray diffraction (XRD) spectra of the samples were recorded by a Bruker D8 Advance X-ray diffrac˚ from 10◦ to 90◦ at a tometer using Cu K␣ radiation ( = 1.5416 A) scan rate of 2.4◦ min−1 . Diffuse reflectance absorption spectra of bare TiO2 nanorod arrays and QD-sensitized TiO2 nanorod arrays were recorded in the range from 300 to 800 nm using a Hitachi U-3010 spectroscopy. 2.6. Photovoltaic measurements The photocurrent–voltage (I–V) curves were measured by Oriel I–V test station under illumination of a solar simulator (Pecell-L, Japan) calibrated by standard silicon cell at one sun (AM 1.5 G, 100 mW/cm2 ). The active illuminated area of the QDSSC was fixed to 0.16 cm2 . The incident photon to current conversion efficiency (IPCE) measurements were performed with a monochromator to select the illumination wavelength, a 500 W xenon arc lamp (Oriel) served as a light source. 3. Results and discussion 3.1. Structure and morphology characterization of single-crystalline TiO2 nanorod arrays The phase structure of the as-prepared TiO2 nanorod arrays on FTO glass was characterized by X-ray diffraction patterns. As shown in Fig. 1, only the rutile structure of SnO2 (JCPDS no. 41-1445) can be indexed in the XRD patterns of bare FTO glass. It is worth noting that as-prepared TiO2 nanorod arrays show a tetragonal rutile structure (JCPDS no. 21-1276) after the hydrothermal reaction. The successful growth of TiO2 nanorods can be ascribed to small lattice mismatch between FTO and rutile TiO2 [26], which promote the epitaxial nucleation of TiO2 on FTO substrate. Compared with the intensity of (1 0 1) and (1 1 2) diffraction peaks in XRD patterns of as-prepared TiO2 , the enhanced (0 0 2) peak indicates that the nanorods were well crystallized and grew in [0 0 1] direction with the growth axis parallel to the substrate surface.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
Fig. 1. XRD patterns of bare FTO glass and TiO2 nanorod arrays.
Fig. 2a shows typical field-emission scanning electron microscopy (FE-SEM) image of the TiO2 nanorod arrays grown at 150 ◦ C for 12 h with saturated NaCl aqueous solution. The top facets of the nanorods are square, which shows the expected growth habit of the tetragonal crystal. The top surfaces of the nanorods appear to contain many step edges, which are the substrates for growth of the nanorod, while the side surfaces are smooth. Fig. 2b is a cross-sectional view of the corresponding sample, showing that the TiO2 nanorods are nearly perpendicular to the FTO glass and the length is about 3 m. In addition, a certain space can be identified among the TiO2 nanorods. The existence of space among TiO2 nanorods is beneficial for the penetration of electrolyte and the deposition of the QDs. Fig. 2c and d is the top view and crosssectional view of TiO2 nanorod arrays synthesized at 150 ◦ C for 12 h without saturated NaCl aqueous solution. It can be obviously seen that the density of TiO2 nanorods increased, and nanorods are connected together to form a continuous film. Furthermore, the side surfaces of the TiO2 nanorods coalesced with each other and reduced internal surface area of the TiO2 nanorod arrays. This will result in deficient quantum dots loading and light harvesting as previously reported [6,27]. However, the length of TiO2 nanorods synthesized without NaCl solution is unchanged, which is still about 3 m, indicating that the saturated NaCl solution has no influence on longitude growth of TiO2 nanorods. The effect of saturated NaCl aqueous solution to the formation of TiO2 nanorods can be interpreted as follows: Cl− can preferentially adsorb onto the (1 1 0) crystal plane of TiO2 , which retard the growth rate of (1 1 0) crystal plane [28,29], resulting in the suppression of lateral growth. Energy dispersive spectroscopy (EDS) was applied to determine the composition of the nanostructure, which is shown in Fig. 2e. It can be seen that Ti and O are dominant elements (atomic ratio of 1:2), which confirms that the nanorod arrays were mainly composed of TiO2 . Other elements (C came from titanium butoxide, Si and Sn came from FTO glass) can also be identified by EDS spectra, but these elements are in a relatively low level. 3.2. Characterization of CdSe/Mn-CdS/TiO2 nanorod arrays After sensitization with Mn-doped CdS and CdSe QDs, the ordered TiO2 nanorod arrays structure was retained, as shown in Fig. 3a. It is obvious that the surface of TiO2 nanorod became rougher than that of bare TiO2 nanorod (Fig. 2a), which means that the QDs had covered on the surface of TiO2 nanorod after SILAR. The composition of the sensitized TiO2 electrode was analyzed by the EDS spectrum in Fig. 3b. The Ti, O, Cd, S, Se and Mn peaks can be
3
observed in the EDS spectrum of the QDs sensitized TiO2 nanorods. What is more, quantitative analysis from the EDS reveals that the atomic ratio of Cd versus S plus Se is close to 1, indicating that the deposition of Mn-doped CdS and CdSe QDs are likely to be stoichiometric. This result confirmed the successful deposition of QDs onto the TiO2 nanorods. There was no other element can be distinguished in this EDS, suggesting the CdSe/Mn-CdS/TiO2 nanorod arrays has a pure element composition. The detailed microscopic characterization of the CdSe/MnCdS/TiO2 nanorod structure was performed by transmission electron microscope (TEM) and high-resolution TEM (HR-TEM). Fig. 3c is the TEM image of CdSe/Mn-CdS/TiO2 nanorod, it can be seen that the surface of the TiO2 nanorod was covered by QDs. The HR-TEM image of CdSe/Mn-CdS/TiO2 nanorod is used to further verify the successful deposition of QDs onto TiO2 nanorods. As shown in Fig. 3d, clear lattice fringes corresponding to well crystallized TiO2 nanorod, Mn-doped CdS, and CdSe QDs can be distinguished. The lattice spacing of 0.322 nm belonging to the (1 1 0) plane of rutile TiO2 (JCPDS no. 21-1276) can be observed. Among the crystal planes of rutile TiO2 , the (1 1 0) plane possesses the lowest surface energy [30]. According to the crystal growth theory, the planes with lower surface energies generally grow slowly and tend to survive during the growth [31]. Therefore, the (1 1 0) plane can be obviously observed in the HRTEM photographs. Connecting to the TiO2 , the lattice fringe of 0.335 nm, which ascribes to the CdS (1 1 1) (JCPDS no. 10-0454) is identified, and close to the CdS layer, outer layer crystallites with lattice spacing of 0.351 nm corresponding to the (0 0 2) plane of CdSe (JCPDS no. 08-0459) can also be observed. These results demonstrated that Mn-doped CdS and CdSe QDs have been successfully deposited on TiO2 nanorods.
3.3. Optical absorption spectra of the QDs sensitized TiO2 nanorod photoanode The UV–vis diffuse reflectance absorption spectra were used to record the variation of light absorption for the bare TiO2 nanorod photoanode and QD-sensitized TiO2 nanorod photoanodes. As shown in Fig. 4, the absorption edge of bare TiO2 nanorods (spectrum a, Fig. 4) appears at about 405 nm and the maximum peak appears at about 300 nm. No significant absorption in visible-light region can be seen due to the large band gap (3.0 eV) of rutile TiO2 [32]. Compared with the bare TiO2 nanorods electrode, the absorption edges, obtained from the intersection of the sharply decreasing region of a spectrum with its baseline, were enlarged to 520 nm for CdS/TiO2 (spectrum b, Fig. 4) electrode, and 670 nm for CdSe/TiO2 (spectrum d, Fig. 4) electrode. The band gaps of CdS and CdSe QDs were calculated to be 2.39 and 1.85 eV, which are higher than the values of bulk CdS (2.25 eV) and CdSe (1.70 eV) due to the quantum size effect, this indicates that the sizes of CdS and CdSe on the TiO2 nanorod are within the scale of QDs. According to the empirical equations given by Yu et al. [33], the sizes of CdS and CdSe particles were estimated to be 8.796 and 9.713 nm, respectively. It is worth noting that the optical absorption spectrum of Mn-doped CdS/TiO2 (spectrum c, Fig. 4) is broader than that of the CdS/TiO2 electrode. This enlarged absorption range could be ascribed to a slightly broader size distribution in doped sample which is caused by the incorporation of Mn into the crystal lattice of CdS [34]. For the cosensitized electrodes, the absorption range of both undoped and Mn-doped CdS/CdSe cosensitized TiO2 electrodes is broader than that of CdS/TiO2 and CdSe/TiO2 electrodes. Furthermore, the absorption onset of the CdSe/Mn-CdS/TiO2 (spectrum f, Fig. 4) shows a slightly red-shift compared to CdSe/CdS/TiO2 (spectrum e, Fig. 4). These results indicate that the light absorption range can be modulated with doping Mn in QDs, providing its application value to improve the performance of QDSSCs.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
4
Fig. 2. FE-SEM images of TiO2 nanorod array film: (a) top view and (b) cross-sectional view with saturated NaCl aqueous solution; (c) top view and (d) cross-sectional view without saturated NaCl aqueous solution; (e) the EDS spectra of TiO2 nanorod array directly grown on FTO glass.
3.4. Photoelectrochemical characterization The incident photon to current conversion efficiency (IPCE) spectra of QDSSCs with the Mn-doped CdS/TiO2 nanorods, CdSe/TiO2 nanorods, and CdSe/Mn-CdS/TiO2 nanorods are shown in Fig. 5. It can be seen that the onsets of IPCE around 600 nm for Mn-doped CdS/TiO2 solar cell, 700 nm for CdSe/TiO2 solar cell and 750 nm for CdSe/Mn-CdS/TiO2 solar cell. The value of IPCE is determined by light harvesting efficiency, charge injection
efficiency, and charge collection efficiency [35]. The conduction band edge of CdS was higher than CdSe, which resulted in the charge injection and collection efficiency of CdS were higher than that of CdSe, thus the IPCE of Mn-doped CdS/TiO2 nanorods cell can reach to 40%, which is higher than CdSe/TiO2 nanorods cell (28%). It is worth noting that the 45% of IPCE value can be obtained for the CdSe/Mn-CdS/TiO2 nanorods cell, which is higher than Mn-doped CdS or CdSe sensitized TiO2 nanorods cell, the possible explanation may be as follows: in short wavelength region ( < 520 nm), the
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
5
Fig. 3. (a) FE-SEM image of CdSe/Mn-CdS/TiO2 nanorod electrode; (b) EDS spectra of corresponding electrode; (c) and (d) TEM and HR-TEM images of CdSe/Mn-CdS/TiO2 nanorod electrode.
higher IPCE of CdSe/Mn-CdS/TiO2 solar cell came from the contributions of Mn-doped CdS and CdSe QDs in the light harvest; while in long wavelength region (520 nm < < 800 nm), the higher IPCE of CdSe/Mn-CdS/TiO2 cell indicates that the introduction of a CdS layer between TiO2 and CdSe is helpful for the collection of excited electrons from CdSe to TiO2 [20].
Fig. 4. Diffuse reflectance absorption spectra of bare TiO2 nanorod photoanode and the TiO2 photoanodes sensitized by CdS, Mn-doped CdS, CdSe, CdSe/CdS, and CdSe/Mn-doped CdS.
The J–V characteristics of the assembled QDSSCs were measured by Oriel I–V test station under one sun (=100 mW/cm2 AM 1.5 G solar illumination) with the active area of 0.16 cm2 . Fig. 6a shows the performance variation of TiO2 nanorod solar cells sensitized by different SILAR cycles of Mn-doped CdS QDs. It is found that both the short circuit current density (Jsc ) and the open circuit voltage (Voc ) increased gradually with the increase of SILAR cycles at initial stage, and the optimum values of these two parameters (7.53 mA/cm2 , 0.36 V) could be obtained at 10 SILAR cycles
Fig. 5. IPCE spectra of QDSSCs assembled with the Mn-doped CdS/TiO2 , CdSe/TiO2 , and CdSe/Mn-CdS/TiO2 photoanodes.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8 6
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx Table 1 Photovoltaic performance parameters of QDSSCs assembled with different photoanodes. Samples
Jsc (mA/cm2 )
Voc (V)
FF
(%)
Mn-doped CdS(4)/TiO2 Mn-doped CdS(7)/TiO2 Mn-doped CdS(10)/TiO2 Mn-doped CdS(13)/TiO2 CdSe(0)/Mn-CdS(10)/TiO2 CdSe(3)/Mn-CdS(10)/TiO2 CdSe(6)/Mn-CdS(10)/TiO2 CdSe(9)/Mn-CdS(10)/TiO2 CdSe(12)/Mn-CdS(10)/TiO2 CdSe(15)/Mn-CdS(10)/TiO2
5.03 6.79 7.53 5.85 7.53 9.02 11.81 13.38 12.58 10.29
0.34 0.35 0.36 0.35 0.36 0.38 0.39 0.42 0.41 0.39
0.41 0.43 0.44 0.41 0.44 0.41 0.43 0.43 0.43 0.43
0.70 1.03 1.19 0.83 1.19 1.41 1.97 2.40 2.23 1.73
Jsc and Voc increased gradually to maximum as CdSe cycles varying from 0 to 9. The increase of Jsc can be attributed to the extending of light absorption range and the increase of absorbance with augmented amount of CdSe QDs loading. The improvement of Voc resulted from quasi-Fermi level of the electrons in semiconductor photoanode tends to more negative. However, when SILAR cycles of CdSe exceeded 9 cycles, it can be seen that both the Jsc and Voc values began to drop, resulting in the decrease of . The excessive SILAR cycles can lead to the aggregation of QDs, which may increase the possibilities of the recombination between the electrons and holes when the photogenerated electrons diffuse across the QD layers to the TiO2 film, resulting in an eventual decrease in the overall efficiency. In addition, the excessive SILAR cycles may also hinder the diffusion of the electrolyte, which limit the efficiency of charge separation and charge extraction. The power conversion efficiency () and short circuit current (Jsc ) as a function of SILAR cycles of CdSe was summarized in Fig. 6c, the value of Jsc and the increased first (from 3 to 9 cycles) and then decreased (from 9 to 15 cycles). The optimum value of the was obtained at 9 cycles of CdSe, thus it can be concluded that the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance, which Jsc , Voc and were 13.38 mA/cm2 , 0.42 V, and 2.40%, respectively. The short circuit current (Jsc ), open circuit voltage (Voc ), fill factor (FF), and power conversion efficiency () of all the solar cells are summarized in Table 1. The performance of CdSe/Mn-CdS/TiO2 nanorod solar cell is superior to Mn-doped CdS/TiO2 nanorod solar cell and CdSe/TiO2 nanorod solar cell, which is shown in Fig. 7. For the cell fabricated with CdSe(9)/Mn-CdS(10)/TiO2 photoanode, the Jsc and Voc were 13.38 mA/cm2 and 0.42 V, respectively, resulting in a very high of 2.40%, which is 101.68% higher than that of Mn-doped CdS(10)/TiO2 nanorod solar cell (Jsc of 7.53 mA/cm2 , Voc of 0.36 V, and of 1.19%) and 247.83% higher than that of CdSe(9)/TiO2 nanorod solar cell
Fig. 6. (a) and (b) J–V curves for Mn-doped CdS/TiO2 nanorods and CdSe/MnCdS/TiO2 nanorods solar cells measured under AM 1.5 G condition, respectively. (c) Power conversion efficiency and photocurrent density as a function of SILAR cycles of CdSe QDs.
of Mn-doped CdS. Then further increase of SILAR cycles of Mndoped CdS QDs (13 cycles) lead to the reduction of the Jsc and Voc . Therefore, it can be concluded that 10 SILAR cycles is optimal for Mn-doped CdS/TiO2 nanorods solar cells. Based on this, different SILAR cycles of CdSe on Mn-doped CdS/TiO2 nanorods were investigated to seek the best cycles for cosensitization system. Fig. 6b displays the J–V characteristics for different SILAR cycles of CdSe onto Mn-doped CdS(10)/TiO2 nanorod solar cells. Obviously, both
Fig. 7. J–V curves of the QDSSCs assembled with different photoanodes measured under AM 1.5 G condition.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
7
Fig. 8. Schematic diagram of the photoelectrical conversion configuration of CdSe/Mn-CdS/TiO2 solar cell.
(Jsc of 5.54 mA/cm2 , Voc of 0.34 V, and of 0.69%). The full utilization of light is the most important factor for enhanced power conversion efficiency of QDSSCs. The band gap of CdS in bulk (2.25 eV) limits its absorption range below the wavelength of approximately 550 nm, which leads to the of Mn-doped CdS(10)/TiO2 nanorod solar cell in a low level. Although CdSe (1.70 eV in bulk) has a wider absorption range than CdS, its electron injection efficiency is not as good as CdS [36,37], also resulting in a low for CdSe(9)/TiO2 nanorod solar cell. Therefore, CdS/CdSe cosensitized TiO2 nanorods can make full use of their respective advantages, such as extension of light absorption range and increase in absorbance, resulting in the enhancement of power conversion efficiency as presented in Fig. 7. In order to evaluate the effect of Mn-doping on the cell performance, we prepared CdSe(9)/CdS(10)/TiO2 cell as a control group. As shown in Fig. 7, after Mn-doping, the CdSe(9)/Mn-CdS(10)/TiO2 solar cell exhibited the best performance and the power conversion efficiency can reach to 2.40%, which is 46.34% higher than CdSe(9)/CdS(10)/TiO2 cell (10.06 mA/cm2 , 0.40 V, 1.64%), the increase of power conversion efficiency can be attributed to the midgap states created by Mn2+ in CdS QDs [21], which may cause the electrons to get trapped and screen them from charge recombination with holes, leading to the improvement of charge collection efficiency, thus improve the cell performance. The photoelectrical conversion configuration of CdSe/MnCdS/TiO2 solar cell is shown in Fig. 8, the configuration of the solar cell consisted of TiO2 nanorod film, CdSe/Mn-doped CdS QDs sensitizer, polysulfide electrolyte, and Cu1.8 S/CuS/FTO counter electrode. Because TiO2 nanorod arrays were directly grown on transparent FTO glass, it can be used as a front-side light irradiation photoanode. As shown in the schematic diagram, under illumination, photons are captured by QDs, yielding electron–hole pairs. The electrons are rapidly injected into the TiO2 nanorod arrays film, simultaneously, the remaining holes are scavenged to the counter electrode via a hole transporting redox couple comprising of polysulfide electrolyte. From the working mechanism of QDSSC mentioned above, it can be found that the amount of light harvesting and adsorbed QDs and electron transport are the key factors affecting the performance of QDSSC. The Mn-doped CdS/CdSe consensitzed structure extended the scope of light absorption, increasing the amount of light harvesting. The TiO2 nanorods synthesized by addition saturated NaCl to the grown solution possessed high aspect ratio and high internal surface, which are advantageous for efficient quantum dots loading and electrolytes filling. In addition, the single-crystalline one-dimensional TiO2 nanorod arrays offered a direct electrical pathway for photogenerated electrons, which can increase the electron transport rate. Thus, it can be concluded that
our CdSe/Mn-CdS/TiO2 nanorod solar cell is of great potential for designing high efficiency solar cells. 4. Conclusion In summary, 1D single-crystalline TiO2 nanorod array was successfully synthesized on FTO glass. The CdSe/Mn-doped CdS QDs were assembled onto TiO2 nanorod film by SILAR and used as photoanode in solar cells. The effect of SILAR cycles on the performance of QDSSCs was investigated, and the optimal combination was obtained with Mn-doped CdS(10)/CdSe(9). The short circuit photocurrent density of 13.38 mA/cm2 , the open circuit potential of 0.42 V, and the power conversion efficiency of 2.40% were achieved with CdSe(9)/Mn-CdS(10)/TiO2 solar cell, which was 46.34% higher than CdSe(9)/CdS(10)/TiO2 solar cell (1.64%). The acceptable power conversion efficiency of our CdSe(9)/Mn-CdS(10)/TiO2 solar cell has demonstrated that its significance in designing high efficiency QDSSCs. Acknowledgment This work was supported by Key Project of Tianjin Sci-Tech Support Program (No. 08ZCKFSH01400). References [1] J. Wang, T. Zhang, D. Wang, R. Pan, Q. Wang, H. Xia, J. Alloys Compd. 551 (2013) 82–87. [2] S. Wang, G. Xia, H. He, K. Yi, J. Shao, Z. Fan, J. Alloys Compd. 431 (2007) 287–291. [3] F. Shao, J. Sun, L. Gao, S. Yang, J. Luo, ACS Appl. Mater. Interfaces 3 (2011) 2148–2153. [4] A.M. More, T.P. Gujar, J.L. Gunjakar, C.D. Lokhande, O.-S. Joo, Appl. Surf. Sci. 255 (2008) 2682–2687. [5] M. Yang, B. Ding, S. Lee, J.-K. Lee, J. Phys. Chem. C 115 (2011) 14534–14541. [6] Z.-J. Zhou, J.-Q. Fan, X. Wang, W.-H. Zhou, Z.-L. Du, S.-X. Wu, ACS Appl. Mater. Interfaces 3 (2011) 4349–4353. [7] S. Cheng, W. Fu, H. Yang, L. Zhang, J. Ma, H. Zhao, M. Sun, L. Yang, J. Phys. Chem. C 116 (2012) 2615–2621. [8] F. Xu, Y. Wu, X. Zhang, Z. Gao, K. Jiang, Micro Nano Lett. 7 (2012) 826–830. [9] J. Zhang, R.-G. Du, Z.-Q. Lin, Y.-F. Zhu, Y. Guo, H.-Q. Qi, L. Xu, C.-J. Lin, Electrochim. Acta 83 (2012) 59–64. [10] S. Feng, J. Yang, M. Liu, H. Zhu, J. Zhang, G. Li, J. Peng, Q. Liu, Thin Solid Films 520 (2012) 2745–2749. [11] S.-H. Hsu, S.-F. Hung, S.-H. Chien, J. Power Sources 233 (2013) 236–243. [12] K. Prabakar, H. Seo, M. Son, H. Kim, Mater. Chem. Phys. 117 (2009) 26–28. [13] B. Mukherjee, Y.R. Smith, V. Subramanian, J. Phys. Chem. C 116 (2012) 15175–15184. [14] H. Gholap, R. Patil, P. Yadav, A. Banpurkar, S. Ogale, W. Gade, Nanotechnology 24 (2013) 195101. [15] S.D. Sung, I. Lim, P. Kang, C. Lee, W.I. Lee, Chem. Commun. 49 (2013) 6054–6056. [16] S.S. Kalanur, Y.J. Hwang, O.-S. Joo, J. Colloid Interface Sci. 402 (2013) 94–99.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023
G Model APSUSC-27841; No. of Pages 8 8
ARTICLE IN PRESS L. Yu et al. / Applied Surface Science xxx (2014) xxx–xxx
[17] M. Li, Y. Liu, H. Wang, H. Shen, W. Zhao, H. Huang, C. Liang, J. Appl. Phys. 108 (2010) 094304. [18] G. Zhu, L. Pan, T. Xu, Z. Sun, ACS Appl. Mater. Interfaces 3 (2011) 3146–3151. [19] Y. Lai, Z. Lin, D. Zheng, L. Chi, R. Du, C. Lin, Electrochim. Acta 79 (2012) 175–181. [20] Y.-L. Lee, C.-F. Chi, S.-Y. Liau, Chem. Mater. 22 (2010) 922–927. [21] P.K. Santra, P.V. Kamat, J. Am. Chem. Soc. 134 (2012) 2508–2511. [22] B. Liu, E.S. Aydil, J. Am. Chem. Soc. 131 (2009) 3985–3990. [23] Y.-L. Lee, Y.-S. Lo, Adv. Funct. Mater. 19 (2009) 604–609. [24] S.S. Kalanur, S.Y. Chae, O.S. Joo, Electrochim. Acta 103 (2013) 91–95. [25] S. Rühle, S. Yahav, S. Greenwald, A. Zaban, J. Phys. Chem. C 116 (2012) 17473–17478. [26] C. Wang, Z. Jiang, L. Wei, Y. Chen, J. Jiao, M. Eastman, H. Liu, Nano Energy 1 (2012) 440–447. [27] C. Xu, P.H. Shin, L. Cao, J. Wu, D. Gao, Chem. Mater. 22 (2010) 143–148.
[28] K. Kakiuchi, E. Hosono, H. Imai, T. Kimura, S. Fujihara, J. Cryst. Growth 293 (2006) 541–545. [29] X. Feng, J. Zhai, L. Jiang, Angew. Chem. Int. Ed. 44 (2005) 5115–5118. [30] U. Diebold, Appl. Phys. A: Mater. Sci. Process. 76 (2003) 681–687. [31] M. Gao, Y. Li, M. Guo, M. Zhang, X. Wang, J. Mater. Sci. Technol. 28 (2012) 577–586. [32] M. Xu, P. Da, H. Wu, D. Zhao, G. Zheng, Nano Lett. 12 (2012) 1503–1508. [33] W.W. Yu, L. Qu, W. Guo, X. Peng, Chem. Mater. 15 (2003) 2854–2860. [34] M.A. Malik, P. O’Brien, N. Revaprasadu, J. Mater. Chem. 11 (2001) 2382–2386. [35] S.W. Jung, J.-H. Kim, H. Kim, C.-J. Choi, K.-S. Ahn, Curr. Appl. Phys. 12 (2012) 1459–1464. [36] M. Wang, J. Jiang, J. Shi, L. Guo, ACS Appl. Mater. Interfaces 5 (2013) 4021–4025. [37] J. Wang, T. Zhang, Q. Wang, D. Wang, R. Pan, H. Xia, J. Wuhan Univ. Technol. Mater. Sci. Ed. 27 (2012) 876–880.
Please cite this article in press as: L. Yu, et al., Enhanced photoelectrochemical performance of CdSe/Mn-CdS/TiO2 nanorod arrays solar cell, Appl. Surf. Sci. (2014), http://dx.doi.org/10.1016/j.apsusc.2014.05.023