Journal of Power Sources 269 (2014) 661e670
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
ZnO/PbS core/shell nanorod arrays as efﬁcient counter electrode for quantum dot-sensitized solar cells Xiaohui Song a, Minqiang Wang a, *, Jianping Deng a, Yanyang Ju a, Tiying Xing a, Jijun Ding a, Zhi Yang a, Jinyou Shao b a
Electronic Materials Research Laboratory (EMRL), Key Laboratory of the Ministry of Education, International Center for Dielectric Research, Xi'an Jiaotong University, Xi'an 710049, China State Key Laboratory of Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an 710049, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
ZnO/PbS nanorod arrays were fabricated and used as counter electrodes for QDSSC. ZnO nanorods arrays provide high surface area for loading more PbS catalysts. ZnO/PbS counter electrodes exhibited high catalytic activity and low Rct values. A maximum efﬁciency of 3.06% was achieved with ZnO/PbS(9) counter electrode.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 March 2014 Received in revised form 19 June 2014 Accepted 8 July 2014 Available online 17 July 2014
To improve the electrochemical catalytic activity of the counter electrode (CE) to polysulﬁde electrolyte that commonly applied in quantum dot-sensitized solar cells (QDSSC), combination of catalysts with nanostructured materials is equally important as it is in the case of photoanodes. Here, we design and fabricate a new catalytic electrodes by combining PbS nanoparticles catalyst with ZnO nanorods arrays (NRs) for use as CEs in QDSSC. The ZnO/PbS core/shell CEs are prepared via the combination of chemical bath deposition and successive ionic layer adsorption and reaction methods. Compared with planar CEs, the ZnO NRs framework presents larger surface area for loading more PbS catalysts and easy accessibility of electrolyte. Additionally, the ZnO NRs core offers an excellent electron pathway for shuttling electrons to highly catalytic PbS sites due to the high electrical conductivity of ZnO nanorods. Therefore, the ZnO/ PbS composite CEs exhibit much higher catalytic activity for polysulﬁde electrolyte than conventional Pt CE. As a result, the power conversion efﬁciency of QDSSC with the optimized ZnO/PbS CEs increase by nearly 120% and 78% compared to that with planar Pt and bare PbS CE, respectively. © 2014 Elsevier B.V. All rights reserved.
Keywords: Nanorods array Lead sulﬁde Counter electrode Quantum dot-sensitized solar cell Successive ionic layer adsorption and reaction
1. Introduction Semiconductor nanoparticles, or quantum dots (QDs), are attractive materials for inexpensive optoelectronic devices due to
* Corresponding author. Tel./fax: þ86 29 82668794. E-mail address: [email protected]
(M. Wang). http://dx.doi.org/10.1016/j.jpowsour.2014.07.044 0378-7753/© 2014 Elsevier B.V. All rights reserved.
their fascinating electronic and optical properties, such as sizecontrolled absorption spectra [1,2], high extinction coefﬁcient , and large intrinsic dipole moments . All of these characteristics in QDs are particularly suitable for photon harvesters in solar cells. More importantly, it has been reported that the process of multiple exciton generation (MEG) can occur under speciﬁc excitation conditions of semiconductor QDs [5,6]. If generated charges can be
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
extracted rapidly, the theoretical maximum conversion efﬁciency of a solar cell can reach as high as 44% . Despite certain controversy [8,9], the demonstration of internal quantum efﬁciency higher than 100% stresses the potentiality of QDs in solar cells to produce low-cost and high-performing photovoltaic devices [10,11]. A typical strategy to construct solar cells based on QDs is to use QDs as light absorber to sensitize wide-band gap metal oxide (TiO2, ZnO, SnO2 et al.) nanostructure ﬁlms, and this conﬁguration is analogous to the dye-sensitized solar cell (DSSC), so called quantum dots-sensitized solar cells (QDSSC). Although considerable progress has been achieved in recent years [12e16], the power conversion efﬁciency (PCE) of QDSSC is still much lower than that of DSSC due to inefﬁcient charge separation and transfer at its various interfaces [17e21]. In particular, poor charge transfer at counter electrode/ electrolyte interface is considered to be a major hurdle in attaining a high ﬁll factor and conversion efﬁciency in QDSSC [20,22]. Iodide/ triiodide redox couple, even though working efﬁciently in DSSC, would cause rapid photocorrosion of the QDs . Therefore, an aqueous polysulﬁde (S2/S2 n ) redox electrolyte is usually adopted in QDSSC, which exhibits strong selectivity toward counter electrodes (CEs). Conventional planar Pt electrode, as a standard CE material in DSSC, is inappropriate in QDSSC systems owing to the chemisorption of sulfur compounds on the Pt surface which undermine the electrode performance through poisoning effect . Therefore, alternative catalytic materials for polysulﬁde solution have been widely explored, including carbon materials [24,25], CoS , Cu2S , PbS , and Cu2ZnSnS4  etc. Among these CEs, metal sulﬁde exhibit high electrocatalytic activity. These metal sulﬁde CEs are usually made by exposing Co, Cu, or Pb foils in hydrochloric acid or sulfuric acid and then applying drops of polysulﬁde solution onto the treated metal foil to obtain an interfacial layer of metal sulﬁde. The problem is that such CEs suffer from continual corrosion of metal foils in the presence of polysulﬁde electrolyte, making the electrode chemically and mechanically unstable . To solve this problem, some deposition methods, such as electrodeposition , hydrothermal , and successive ionic layer adsorption and reaction (SILAR) [32,33], have been applied to deposit metal sulﬁde on ITO or FTO glass. However, the adhesion between metal sulﬁde and ITO or FTO glass is usually poor, and the deposition of metal sulﬁde CEs on plain ITO or FTO substrate does not always produce materials with sufﬁciently high speciﬁc surface, leading to thin ﬁlm thickness, low coverage and hence, low catalytic activity of CEs. Therefore, development of a new counter electrode is required to further enhance the PCE of QDSSC. The role of a counter electrode in a QDSSC is twofold: i) to collect electrons from external circuit; and ii) to regenerate the hole scavenger by catalyzing the reduction of the oxidized species in the electrolyte . Thus, high catalytic activity, large surface area, high electrical conductivity, chemical durability, and low cost are some of the basic requirements for the practical counter electrode . Herein, one-dimensional (1-D) ZnO/PbS core/shell nanorods arrays (NRs) were fabricated by a facile and low temperature wet chemical route and applied as an efﬁcient counter electrode for QDSSC. Large area and well-aligned ZnO NRs were ﬁrst hydrothermally grown on a conducting ﬂuorine-doped tin oxide (FTO) substrate. Then, using successive ionic layer adsorption and reaction (SILAR) method, PbS nanoparticles catalysts were deposited onto the ZnO nanorods to form core/shell nanorod structures. The structure of the solar cell based on ZnO/PbS CEs is illustrated in Scheme 1. To the best of our knowledge, this promising 1-D ZnO/PbS electrocatalysts have not been employed in polysulﬁde electrolyte up to date. Compared with planar CEs, ZnO NRs offers an extremely high surface area for the large deposition of PbS catalysts and easy accessibility for an
Scheme 1. Schematic diagram showing the QDSSC architecture composed of ZnO/PbS counter electrode and TiO2/CdSe photoanode with sulfur redox electrolyte (a). The electron transport from QDs into TiO2 and reduction reaction at CE/electrolyte interface are also shown (b).
electrolyte, thus, higher available catalytic surface area can be achieved. Moreover, the vertically aligned ZnO NRs framework acts as an excellent electrical tunnel for fast electron transport from external circuit to highly catalytic PbS catalyst sites where the electrons are used to reduce the oxidized polysulﬁde. Therefore, the combination of the high catalytic PbS catalyst and the high surface area ZnO NRs would greatly enhance the catalytic activity of CE toward the polysulﬁde electrolyte. As a result, the PCE of CdSe QDSSC with the best ZnO/PbS CEs increased by 118.6% and 77.9% compared to that with Pt and bare PbS CE, respectively. Therefore, the resulting nanostructured CEs hold great promise for constructing highly efﬁcient QDSSC.
2. Experimental details 2.1. Synthesis of the ZnO nanorods arrays In this study, ZnO NRs were prepared by a two-step simple chemical method. The ﬁrst step was to coat the FTO substrate with a ZnO seed layer using spin coating technique. The basic precursor solution was prepared by dissolving Zn(CH3COO)2$2H2O and monoethanolamine (MEA) in 2-methoxyethanol as described in our previous research [16,35,36]. Then the resultant precursor was
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
stirred at 60 C for 1 h to yield a clear and homogeneous solution, which was used as the coating solution. The FTO substrate was then spin-coated with precursor solution and dried on a hot plate to evaporate the solvent and remove organic residuals, and the procedure from coating to drying was repeated twice. The second step was to grow ZnO nanorods on the seeded substrate using chemical bath deposition method. The seeded substrate was suspended upside down in a mixed aqueous solution containing 0.04 M Zn(NO3)2 and 0.8 M NaOH at 83 C for 30 min. When the growth of ZnO nanorods was ﬁnished, the samples were rinsed with deionized water repeatedly and were calcined at 450 C for 30 min. 2.2. Fabrication of the ZnO/PbS, Pt, bare PbS, and ZnO/Pt counter electrodes The PbS nanoparticles were deposited on the ZnO NRs surface by the SILAR method. Typically, the ZnO NRs were ﬁrst immersed into a 0.1 M Pb(NO3)2 aqueous solution for 1 min, to allow Pb2þ to adsorb onto the ZnO, and then rinsed with deionized water for 1 min to remove the excess Pb2þ. Successively, the ZnO NRs were dipped into the aqueous solution containing 0.1 M sodium sulﬁde (Na2S) for another 1 min to allow S2 to react with the preadsorbed Pb2þ, leading to the formation of PbS, followed by rinsing with deionized water. The two-step dropping and rinsing procedure represents one SILAR deposition cycle of PbS. The amount of PbS deposited can be increased by repeating the deposition cycle. In order to obtain optimized devices, such an immersion cycle was repeated 3, 6, 9, and 12 times to vary the amount of the PbS nanoparticles assembled on the ZnO surface. The ZnO/PbS CEs prepared with the 3, 6, 9, and 12 SILAR cycles are referred to as ZnO/PbS(3), ZnO/PbS(6), ZnO/PbS(9), and ZnO/PbS(12), respectively. The electrode became darker as the number of SILAR cycles was increased. The Pt CE was prepared by depositing 50 nm Pt ﬁlm on FTO glass using a magnetron sputtering. The FTO glass has a small hole for subsequent electrolyte injection. The bare PbS CE was prepared via a simple chemical bath deposition (CBD) method as previous report . Brieﬂy, a cleaned FTO glass was immersed in an aqueous solution containing 0.17 M of lead nitrate, 0.57 M of sodium hydroxide and 0.1 M of thiourea at room temperature for 30 min. At the end of each immersion time, the withdrawn substrates covered with PbS ﬁlm were rinsed with distilled water, and dried in the air. To fairly compare the catalytic activity of ZnO/PbS with Pt CEs, we prepared the ZnO/Pt CEs by thermal decomposition (TD) of H2PtCl6 that commonly employed in the fabrication of DSSC. A drop of H2PtCl6 solution (10 mM in isopropanol) was spin-coated on ZnO nanorod arrays at room temperature. Subsequently, electrodes were heat-treated at 400 C for 20 min to induce thermal reduction of Pt nanoparticles on ZnO nanorods, and the procedure of coating and heat treatment was repeated ﬁve times to increase the concentration of Pt nanoparticles. 2.3. Preparation of CdSe QDs-sensitized mesoporous TiO2 (TiO2/ CdSe) working electrodes Mesoporous TiO2 ﬁlms were prepared by successive screen printing a 11.0 mm thick transparent layer (DSL 18NR-T, Dyesol) over FTO (15 U square1) glass substrates, followed by sintered at 500 C for 30 min in a mufﬂe-type furnace. The typical area of the TiO2 ﬁlm was approximately 0.25 cm2. The TiO2 ﬁlms were then sensitized with CdSe QDs via the electrodepositon method, following procedures previously described . After deposition, the samples were taken out from the electrolyte, rinsed successively with deionized water, ethanol, and then dried by N2 stream.
2.4. Fabrication of thin-layer liquid-junction QDSSC and symmetric cell for electrochemical impedance measurement The liquid-junction QDSSC were fabricated by sandwiching the as-prepared TiO2/CdSe photoanodes and ZnO/PbS (or Pt) CEs by using a 65 mm hot-melt polymer (Surlyn, DuPont) as the spacer. The inter-electrode space was ﬁlled with a redox liquid electrolyte (1 M Na2S and 1 MS aqueous solution) by vacuum back-ﬁlling through a hole pre-drilled in the counter electrode, and then the hole was sealed using a small piece of hot-melt polymer and a microscope cover-slip. The symmetrical dummy cells were fabricated with two identical ZnO/PbS (or Pt, bare PbS) CEs following the same sandwich-style cell assembly procedure described above, and the electrolyte injection and ﬁnal sealing were carried out in the same way as for QDSSC. An active area of thin layer symmetric cell is 0.36 cm2. 2.5. Characterizations The surface morphology and elemental compositions of the ZnO/PbS, bare PbS, and ZnO/Pt CEs were characterized by using a ﬁeld emission scanning electron microscope (FEI Quanta 250) equipped with an energy-dispersive X-ray (EDS) analysis system. Xray diffraction (XRD) patterns were recorded on a Rigaku D/MAX2400 diffractometer equipped with Cu Ka radiation (l ¼ 1.5406 Å) operating at 40 kV and 100 mA. Optical absorption and reﬂection spectra of the as-prepared ZnO NRs and ZnO/PbS CEs were recorded using a Jasco-V570 UVeViseNIR spectrophotometer equipped with an integrating sphere. Photocurrentevoltage characteristics (JeV) were measured with a Keithley 2400 source meter under illumination from a solar simulator (Sciencetech Inc., SS150), and the light intensity was calibrated with a silicon photodiode. The incident photon-to-current conversion efﬁciency (IPCE) was measured as a function of wavelength from 300 to 800 nm by using a model 7-SCSpec II system (Beijing 7-Star Optical Instruments Co., Ltd.). Electrochemical impedance spectroscopy (EIS) of the CEs was recorded using an electrochemical workstation (CHI 660D, Chenhua, Shanghai) and performed on dummy cells with a symmetric sandwich-like structure between two identical electrodes, that is, CE/electrolyte/CE. The CEs used for testing in this paper are consistent with equal area. The measured frequency ranged from 50 mHz to 100 kHz and the amplitude was set at 10 mV. The spectra were ﬁtted by Zview software. Tafel polarization measurement was carried out on the CHI 660D electrochemical analyzer in the dummy cell used in the EIS measurement with a scan rate of 50 mV s1. 3. Results and discussion Scheme 1 illustrates the structure of the QDSSC based on ZnO/ PbS CEs and the transport pathways of electrons. Upon photoexcitation of the QDs, electrons are injected from an excited state of CdSe QDs into TiO2 and ﬁnally to the FTO, whereas the remaining holes oxidize the polysulﬁde electrolyte through the reaction: nS2 þ hþ /S2 n : Electrons in the FTO move to the cathode via external circuit and participate in reduction reaction at the CEs/ 2 electrolyte interface: S2 n þ e /nS : Therefore, an ideal CE should possess high electrocatalytic activity for the reduction of charge carriers in electrolyte as well as high conductivity. Fig. 1 shows the low and high magniﬁcation ﬁeld-emission scanning electron microscope (FESEM) images of the as-grown ZnO NRs on an FTO substrate. It can be clearly seen that largescale vertical growth of ZnO nanorods has been realized and uniformly distributed throughout the substrate, with slight variations in growth angle resulting from the roughness of the FTO substrate.
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
Fig. 1. Top-view (a and b) and cross-sectional (c and d) FESEM images of the ZnO nanorods arrays at different magniﬁcations.
From the higher magniﬁcation (Fig. 1b) of such arrays, the average diameter of the ZnO nanorod is 95e110 nm, and the sides of the ZnO nanorods are relatively smooth. The nanorods have a hexagonal cross section, as shown in Fig. 1b, demonstrating the axial direction is aligned with the c-axis of the hexagonal ZnO crystal structure. The vertical alignment of ZnO NRs is beneﬁcial for the improvement in the charge transfer of the solar cells. Although the density of ZnO NRs is high, there are still many interstices between the nanorods. Fig. 1c and d shows the low and high magniﬁcation cross-sectional FESEM image of the vertically grown ZnO nanorods. Almost all ZnO nanorods are perpendicular to the substrate and have a uniform length of ~1.7 mm. The length of the ZnO nanorods depends on the growth time. Here, we have kept a 30 min growth time for all the ZnO samples. Fig. 2 shows the low and high magniﬁcation FESEM images of the ZnO/PbS CEs fabricated with different SILAR cycles. It was found that the deposition of PbS QDs did not destroy the morphology of ZnO nanorod. However, an enlarged SEM image revealed that the surface of the nanorods became rough after PbS deposition. When the cycle number was 3, the deposition of PbS nanoparticles on the ZnO nanorods was incomplete and a few sparsely distributed PbS nanoparticles were observed (Fig. 2d). It can be said that the loading amount of PbS nanoparticles is still not enough to form a continuous composite ﬁlm. As the cycle number increased to 9, the size and density of the PbS nanocrystals increased, and the roughened texture of the nanoscale PbS coating on these nanorods is clearly visible. From panel e, it is evidence that the PbS catalysts are uniformly decorated on the surface of the ZnO nanorods. The uniform distribution of PbS nanoparticles arose from the growth mechanism. In general, SILAR is governed by an ion-by-ion growth mechanism that enables growth of a gradual conformal layer . Coating with PbS increases the diameter of the nanorods and hence, the interstices between ZnO nanorods is reduced. After 12
SILAR cycles deposition, thickness of the PbS layer increases, and the PbS nanoparticles tend to agglomerate with increasing particle size, but the nanorods array structure is still visible in the top-view FESEM image (Fig. 2f). In addition, the insets of Fig. 2 show photographs of the ZnO/PbS CEs. It was observed clearly that the colors of ZnO NRs thin ﬁlms gradually changed from white to deep black with the increasing cycle number, conﬁrming more PbS nanoparticles have been coated onto the ZnO surface. This is consistent with the observation of SEM images. To conﬁrm the uniform coverage of the PbS nanoparticles over ZnO nanorods, cross-sectional SEM images of ZnO/PbS CEs for different SILAR cycles were obtained and presented in Fig. 2gei. It can be seen that the length of ZnO nanorod is kept at 1.7 mm. For the sample with 3 cycles of deposition, it can be noticed from Fig. 2g that the surfaces of ZnO nanorod are covered by loosely bound PbS nanoparticles, and there are uncovered sites on the surface of ZnO. These uncovered regions have no catalytic effect toward the electrolyte in QDSSC and therefore can cause degradation of the device performance. By increasing the cycle number to 9, the layer of PbS nanoparticles which covered the nanorod surface became thick and dense as shown in Fig. 2h, and the entire surface of ZnO nanorods was covered by PbS in a conformal way, forming a ZnO/PbS coreeshell structure and providing a large area of catalytic sites, which is a requisite of high catalytic property. With the deposition cycle further increased to 12, a lot of PbS nanoparticles were deposited on the ZnO NRs and the interstices between ZnO nanorods, especially at the bottom of the ZnO NRs, were almost completely ﬁlled by PbS nanoparticles, although the nanorods structure is still visible in the topeview FESEM images. This phenomenon will hinder the penetration of the electrolyte into the CEs and therefore decrease the rate of the redox reactions in electrolyte. As discussed later, this phenomenon can cause a dramatic decrease in photocurrent and efﬁciency for the solar cells. The composition of the ZnO/PbS CEs
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
Fig. 2. Top-view SEM images of the ZnO/PbS prepared with various SILAR cycles. (a and d) 3 cycles, (b and e) 9 cycles, and (c and f) 12 cycles at low and high magniﬁcation. Panel gei show the cross-sectional SEM images of the ZnO/PbS CEs shown in panel aec. The insets show the photographs of the corresponding ZnO/PbS CEs.
with 9 cycles of deposition was identiﬁed by energy dispersive Xray spectroscopy (EDS) measurement, and the results shown in Fig. S1 in the Supporting Information revealed the presence of Pb and S elements. This result conﬁrms the successful deposition of PbS on ZnO ﬁlm. Also, the atomic ratio of Pb/S was also determined to be 9.37/8.43, which is close to 1/1. This result correlated well with the atomic ratio of PbS. As observed in Fig. S2, the CBD deposited PbS CEs are specularly reﬂective, homogeneous, and well adhered to the FTO substrates. Its top-view SEM image is shown in Fig. S3. The PbS thin ﬁlm is deposited uniformly on FTO substrate, covering the entire surface with nearly no empty surface regions, and the average size of PbS crystallites is about 220 nm. With increasing the deposition time the mean size of PbS crystallites increases (not shown here), but the PCE of the assembled QDSSC is reduced, so the deposition time for the bare PbS CE is kept at 30 min. Fig. S4 shows the FESEM image of the obtained ZnO/Pt CE. It reveals that the ZnO nanorods become rough after H2PtCl6 spin-coating, which indicates Pt nanoparticles are satisfactorily coated onto the ZnO surface. To further investigate the phase composition and phase structure of ZnO/PbS and bare PbS CEs, XRD measurements were carried out. Fig. 3a shows the XRD spectra taken from bare ZnO NRs, bare PbS, and ZnO/PbS core/shell CEs prepared with different SILAR cycles. For the bare ZnO NRs, 5 diffraction peaks marked by diamond can be respectively indexed to the (002), (101), (102), (103), and (004) planes of hexagonal phase ZnO (JCPDS ﬁle No. 36-1451) , and no other diffraction peaks besides SnO2, which comes from the FTO substrate, were detected. In comparison with the
standard XRD pattern, the much higher relative intensity of the (002) diffraction peak (Fig. S5) at 34.50 suggests that ZnO nanorods grow along the  orientation. This is conﬁrmed by SEM observation. After deposition of PbS, four new peaks are observed in addition to the diffraction peaks from the ZnO nanorods. These broad peaks are centered at 25.96 , 30.07, 43.06 , and 50.97 which can be indexed respectively to the (111), (200), (220), and (311) planes of the zinc-blend (ZB) PbS (JCPDS No.05-0592) . No peaks for other lead compounds are detected in the spectra. This result clearly shows that the ZnO/PbS CEs had been successfully prepared. The broad peaks suggest that the size of the deposited PbS catalysts is very small. Also, it was observed that the intensity of the main PbS (200) peak became sharper and narrower with increasing number of SILAR cycles, indicating that more PbS nanoparticles were deposited on the ZnO nanorod. The location of the diffraction peaks of bare PbS CE coincide, respectively, with the diffraction signals of PbS in ZnO/PbS CEs, which indicates that CBD deposited PbS ﬁlm has the same crystal structure as that of PbS prepared with SILAR method. The strong intensity and narrow peaks show that PbS has good crystallinity. To further characterize the electrodes, the absorption spectra of ZnO/PbS CEs fabricated with different SILAR cycles were investigated. As observed in Fig. 3b, bare ZnO NRs absorbs only UV wavelengths of less than 380 nm, due to its wide band gap (ca. 3.4 eV) , whereas the deposition of PbS into ZnO ﬁlm extends the absorption spectrum obviously to wavelength exceeding 1000 nm. Furthermore, absorbance increased with additional SILAR cycles, indicating that each cycle increased the amount of deposited
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
Fig. 3. XRD patterns (a) and optical absorption spectra (b) of the bare ZnO NRs, PbS, and as-prepared ZnO/PbS CEs with different SILAR cycles.
PbS nanoparticles. Size quantization effects redshifted the absorption shoulder and absorption onset as the number of SILAR cycles increased, indicating the sequential growth of the PbS nanoparticles. The ability to absorb long-wavelength irradiation makes PbS an ideal CE for harvesting the residual light that penetrates the TiO2/CdSe photoanode. In addition, the PbS nanoparticles prepared with SILAR method is a p-type semiconductor with low band gap, and may provide an auxiliary tandem effect as reported in previous research . Under illumination p-type PbS may results in a positive shift of quasi-Fermi level and enhance the open circuit voltage. We assembled sandwich-style thin-layer liquid-junction QDSSC by employing the as-synthesized Pt, bare PbS, ZnO/Pt, and ZnO/PbS ﬁlm as a counter electrode along with TiO2/CdSe photoanode, as depicted in Scheme 1, to evaluate the electrochemical performance of ZnO/PbS in the presence of the polysulﬁde electrolyte. The JeV curves of QDSSC with various CEs under standard AM1.5 illumination at 100 mW cm2 are shown in Fig. 4a. The detailed photovoltaic parameters including the short-circuit current density (JSC), open-circuit voltage (VOC), ﬁll factor (FF), and photovoltaic conversion efﬁciency (h) are summarized in Table 1. Similar to the reports in literature [20,22,44], the QDSSC based on Pt CE show a very low photoelectrochemical performance with a h of about 1.4%. It has been previously proved that the low performance is due to a very low catalytic property of Pt in polysulﬁde electrolyte resulting
from the strong absorption of S2 on the surface, which reduces the surface activity of Pt . As a result, a relative low FF (0.40) is achieved by the QDSSC based on Pt. PbS CE is well known to exhibit high electrocatalytic activity in polysulﬁde solution, which improves the FF and h of QDSSC [28,43]. Here, using the bare PbS as CE boosts the FF and h to about 0.49 and 1.72%, respectively. For ZnO/ PbS CEs, all parameters (JSC, VOC, FF, and h) show similar trends as SILAR cycles increase. The cell with the ZnO/PbS(3) CE produces the worst performance with a maximum h of 0.26%. Both JSC (2.89 mA cm2) and FF (0.21) are the lowest among the compared samples, which originated from the lack of catalytic activity toward polysulﬁde reduction due to the insufﬁcient amount of PbS catalyst, despite the high electrical conductivity of ZnO NRs. As SILAR cycle of PbS increased to 6, more PbS catalyst was decorated on ZnO support, and the catalytic activity of the counter electrode was enhanced and thus the performance was greatly boosted (1.88%). Further increasing SILAR cycles to 9, the cell with the ZnO/PbS(9) CE exhibited the highest h of 3.06% among the compared cells, with VOC, JSC, and FF values of 0.52 mV, 11.17 mA cm2, and 0.53, respectively. Compared to the device with the bare PbS CE, the cell with ZnO/PbS(9) CE exhibited much higher JSC and h. Considering the ZnO/PbS(9) has lower diffuse reﬂection than Pt and bare PbS in the visible wavelength range (Fig. S6), the superior performance of ZnO/PbS(9) can be attributed to its 1-D nanostructure. The BrunauereEmmetteTeller (BET) surface area of the ZnO nanorod was measured to be 34.717 m2 g1 (Fig. S7). Therefore, it could provide high surface area for large deposition of PbS and easy accessibility for an electrolyte, which would greatly improve the charge-transfer process at the CEs. However, further increasing the number of PbS cycles to 12 reduced cell performance because JSC and FF deteriorated due to the blockage of the interstices between ZnO nanorods, but the overall PCE (2.44%) is still much higher than that with the reference Pt and bare PbS CE. As for ZnO/Pt CE, the JeV curve in Fig. S8 shows that it has an inferior performance to Pt, and this may be due to insufﬁcient coverage of Pt on ZnO nanorod that leads to the direct contact between ZnO and electrolyte. Increasing the deposition amount of Pt might further enhance device performance, but the low FF would limit the PCE of the QDSSC with ZnO/ Pt CE. Through current densityevoltage (JeV) characterization, we found that the ZnO/PbS CE signiﬁcantly outperformed the Pt and bare PbS CE in all of the QDSSC measured except for the ZnO/PbS(3) CE. The signiﬁcantly improved cell performance of the QDSSC with the ZnO/PbS CE, especially for ZnO/PbS(9), was largely due to the increased JSC and FF. The FF of a cell is attenuated by the total series resistance of the cell (Rs,total), which can be deﬁned as
Rs;total ¼ Rs þ Rct þ Rr þ Rw
where Rs is the sheet resistances of the substrate and counter electrode, Rct is the charge transfer resistance at the CE/electrolyte interface, Rr is the electron transport resistance through the photoanode, and Rw is designated as Warburg diffusion impedance within the electrolyte . Due to the same fabrication procedure of the photoanode and the same electrolyte, the Rs, Rr, and Rw can be considered the same. The FF enhancement for ZnO/PbS CEs is mainly a result of its superior electrocatalytic activity to the S2/S2 n shuttle and consequent signiﬁcantly reduced Rct at the CE/electrolyte interface. Signiﬁcant improvement in charge transfer at the CE/electrolyte interface not only reduces internal resistances, but also attenuates recombination rates, and concentration gradients in the electrolyte, which have been proved to affect JSC strongly . Besides, the reduction in electron recombination rates would also shift up the electron Fermi level of TiO2 porous ﬁlm, and hence, VOC could also be increased because VOC is determined by the difference
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
Fig. 4. JeV curves (a) and IPCE spectra (b) of QDSSC assembled with Pt, bare PbS, and ZnO/PbS CEs; Nyquist plots (c) and Bode phase plots (d) of symmetrical dummy cells based on Pt, bare PbS, and ZnO/PbS CEs. Inset of panel c shows Nyquist plots of symmetrical cells assembled with ZnO/PbS(3) CEs.
between the electron Fermi level of TiO2 porous ﬁlm and the redox potential of the S2/S2 n couple [44,45]. The increase in the deposition amount of PbS could improve the charge transfer at the CE/ electrolyte interface. As a result, JSC and FF increased with increasing SILAR cycles of PbS. The reduced performance of the QDSSC with ZnO/PbS(12) CE is due to the blockage of the interstices between ZnO nanorods by PbS aggregates as shown in Fig. 2i, which hinder the inﬁltration and diffusion of the polysulﬁde electrolyte in ZnO/PbS CE, resulting in the lower electro-catalytic capacity of ZnO/ PbS(12) toward the reduction of S2 n as compared to ZnO/PbS(9) CE. The IPCE was also measured to demonstrate the excellent performance of ZnO/PbS(9) as a CE. As shown in Fig. 4b, the IPCE and diffuse reﬂectance absorption spectra of the photoanodes (Fig.S9) shows the excitation of CdSe as the origin of the photocurrent. The ZnO/PbS(9) based QDSSC shows a higher IPCE value in the 430e730 nm range than that of the cells with other CEs, which is in accord with the results of the JeV measurement, demonstrating the excellent performance of ZnO/PbS(9) as a CE for polysulﬁde electrolyte. Beside PbS catalyst, CoS, NiS, and CuS have also been reported to exhibit excellent catalytic activity for polysulﬁde electrolyte and could also be easily deposited by the SILAR method . Therefore, ZnO/CoS, ZnO/NiS, and ZnO/CuS CEs were also prepared following the method used to prepare ZnO/PbS CEs, and their photoelectrochemical performance was also investigated. The JeV curves of QDSSC based on these CEs are presented in Fig. S10 and photovoltaic parameters are tabulated in Table S1. In these experiments, all metal sulﬁde CEs are deposited on ZnO NRs with the same SILAR cycle of 9. It can be readily determined in Table S1 that the photoelectrochemical performance for these CE is much lower than that for the Pt CE. In previous report, the CuS CE exhibited much higher catalytic activity and performance than Pt and even PbS CE [20,44], which is opposite to our result. This can be explained as follows. During the deposition experiment of CoS, CuS, and NiS by the SILAR method, the loading amount of these metal sulﬁde is observed to be much less than that of PbS at the same SILAR cycles, especially for CoS and NiS catalyst. Therefore, the amount of CoS,
CuS, and NiS on ZnO NRs might be insufﬁcient to cover the entire ZnO NRs, which results in the lower performance for the ZnO/CoS, ZnO/NiS, and ZnO/CuS than that for the ZnO/PbS CEs. To further understand the high electrocatalytic activity of ZnO/ PbS CE toward polysulﬁde reduction relative to that of Pt, we carried out EIS spectroscopy using the sandwich type electrochemical cells comprising two identical ZnO/PbS (or Pt, bare PbS) CEs, that is CE/electrolyte/CE. Fig. 4c presents the corresponding Nyquist spectra obtained from various symmetrical dummy cells with the same electrolyte composition, and each cell had an effective area of 0.36 cm2. Experimental curves are represented by symbols while the solid lines are the ﬁtted curves, which are obtained with Zview software by using the equivalent circuit given in Fig. S11. According to previous literature [29,44], we attribute the semicircle to the charge-transfer resistance (Rct, the radius of the arc on the real axis) and the corresponding constant phase angle element (CPE) at the CE/electrolyte interface, and the high-frequency intercept on the real axis (Z0 -axis) can be assigned to the ohmic serial resistance (Rs) mainly related to the FTO glass. By ﬁtting the impedance spectroscopy using the equivalent circuit, the EIS parameters are for different CEs could be extracted and listed in Table 1. It is found that the Rs values are at the same level for the various counter electrode materials, and the ZnO/PbS(6) exhibits a low value. Part of ZnO ﬁlm was dissolved by HCl for the assembly of QDSSC and electrical connection, and this behavior may inﬂuence the conductivity of FTO, which result in the small difference of Rs. The value of Rct decreases as the SILAR cycles of PbS increase from 3 to 9. Since Rct is inversely proportional to the electrocatalytic activity of CEs, it can be determined that the electrocatalytic activity of ZnO/PbS CEs increases as SILAR cycles of PbS increase from 3 to 9, which conﬁrms well the results of JeV curves. The ZnO/PbS(3) CE exhibits the largest Rct value of 53,989 U, which is an order of magnitude larger than the Rct value of the Pt (1613 U), and this value is in accordance with its inferior FF, suggesting the catalytic activity is very poor in polysulﬁde electrolyte due to insufﬁcient deposition amount of PbS. As the SILAR cycle of PbS in ZnO/PbS CE increases to 6 and 9, the loading amount of the PbS catalyst increases, and the Rct value
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
Table 1 Photovoltaic parameters of QDSSC with different CEs. Counter electrode VOC (V) JSC (mA cm2) FF (%) h (%) RS Pt Bare PbS ZnO/PbS(3) ZnO/PbS(6) ZnO/PbS(9) ZnO/PbS(12)
0.49 0.48 0.42 0.50 0.52 0.50
7.09 7.32 2.89 7.62 11.17 9.39
40.30 48.95 21.42 49.34 52.68 51.97
1.40 8.9 1613 251.7 1.72 17.1 778.4 105.4 0.26 13.6 53,989 2124.1 1.88 7.2 686.7 44.1 3.06 8.9 122.8 5.2 2.44 8.4 371.6 27.6
decreased strikingly to 686.7 U and 122.8 U, respectively. The Rct value for ZnO/PbS(9) is 13 fold lower than that of conventional Pt CE (1613 U) under the same measurement conditions, highlighting the superior electrocatalytic activity of the ZnO/PbS(9) towards S2/S2 n redox couples. As a result, the FF of the devices increased from 0.40 to 0.52 and JSC increased from 7.09 to 11.17 mA cm2. The values of Rct for bare PbS CE is 778.4 U, much smaller than that for Pt, but larger than ZnO/PbS(6), suggesting that the catalytic activity follows the order Pt < bare PbS < ZnO/PbS(6) < ZnO/PbS(9). The vertical array nanostructure of ZnO/PbS(9) facilitates efﬁcient mass transfer and provides large surface area for the deposition of PbS catalyst, which hence increases the electrocatalytic activity of ZnO/ PbS(9) for the reduction of S2/S2 n . On further increasing the SILAR cycle of PbS to 12, the Rct of the ZnO/PbS(12) CE becomes larger compared to that of ZnO/PbS(9). The impedance increment originates from the nearly complete ﬁlling of the interstices of the ZnO NRs, which makes only the outermost surface of the electrode active in catalyzing the reduction of polysulﬁde. Fig. 4d shows the EIS spectra with Bode phase plots of symmetrical dummy cells. The frequency of the characteristic response peaks of the Bode plot signiﬁes the electron transfer rate at the CE/electrolyte interface [25,46]. It can be seen in Fig. 4d that the peak shifted from 4.2 Hz for ZnO/PbS(3) and 18.4 Hz for Pt to 117.2 Hz for ZnO/PbS(9) CE, demonstrating a considerably faster electron transfer rate across the ZnO/PbS(9) CE/electrolyte interface.
EIS was employed to characterize the charge transfer kinetics of CdSe QDs-sensitized solar cells, and the obtained Nyquist plots for the frequency range of 500 mHze100 kHz are shown in Fig. 5. The Nyquist plots measured at open circuit voltage VOC (dark condition) consists of two semicircles. The smaller semicircle at high frequency corresponds to the charge-transfer resistance (RCE) at CE/ electrolyte interface. The large circle at lower frequencies was assigned to charge-transfer resistance (Rr) at photoanode/electrolyte interface. In this study, Rr is not discussed as the value is not directly inﬂuenced by the choice of CE materials. The RCE were obtained from the ﬁt of EIS and were summarized in Table 1. The ZnO/PbS(9) CE exhibits the lowest RCE value of 5.2 U, which is two order of magnitude lower than the RCE value of the bare PbS CE (105.4 U), suggesting that ZnO NRs greatly enhances the charge transfer of the ZnO/PbS(9) composite CE. To evaluate the interfacial charge-transfer property of S2/S2 n couple on CE surface, Tafel plots for various CEs were measured. Fig. 5c shows the current density (on a logarithmic scale) as a function of applied voltage of the symmetrical dummy cells. The catalytic activities of the CEs can be evaluated from the exchange current density (J0), which could be obtained by extrapolating the cathodic branch of each curve to the zero overpotential. Thus, the steep gradient of the anodic and cathodic branches was related to J0. As expected, larger slopes were observed for the ZnO/PbS(9) than for the Pt and bare PbS CE, which indicated the larger J0 value for the electrode surface. This result veriﬁed that the electrocatalytic activities of the ZnO/PbS(9) was higher than the activity of Pt, bare PbS, and other ZnO/PbS CEs. To elucidate the results more clearly, we carried out cyclic voltammetry (CV) measurement using a three-electrode cell consisting of Pt or ZnO/PbS as working electrodes, Pt as counter electrode, and SCE as reference electrode. Fig. 5d shows the CV of different CEs carried out in a polysulﬁde solution containing 1 MS and 1 M Na2S. The ZnO/PbS(9) CE shows relatively much higher current density than those observed with Pt and other ZnO/PbS CEs. It is because the reduction rate of S2 is higher on ZnO/PbS(9) electrode n
Fig. 5. (a) Nyquist plots of QDSSC consisting of Pt, bare PbS, and different ZnO/PbS CEs. Inset shows the Nyquist plots of solar cell based on ZnO/PbS(3). (b) Zoom-in view of panel a in the high-frequency region. Inset is the equivalent circuit for ﬁtting EIS. (c) Tafel curves of the symmetrical dummy cells fabricated with various CEs. (d) Cyclic voltammograms of Pt and different ZnO/PbS CEs in polysulﬁde electrolyte.
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
compared to other CEs and hence tends to have a higher current density [22,47]. The EIS, Tafel, and CV results conﬁrm that the catalytic activity towards polysulﬁde redox shuttle of the ZnO/ PbS(9) CE is superior to that of Pt, bare PbS and other ZnO/PbS CE, in accordance with its excellent photovoltaic performance. Fig. 6a shows the photocurrent transient behavior of QDSSC that incorporated ZnO/PbS(6), ZnO/PbS(9), and Pt CE at a bias of 0 V with a switch time of 30 s. It can be seen that all of the cells produced photocurrent promptly upon being illuminated, and the photoresponse was steady and reproducible during ONeOFF cycles of the illumination. This comparison highlights the greater JSC achievable with the ZnO/PbS(9) CE. However, a sharp decrease in the short circuit current is seen, in the initial stage of each illumination period, before stabilizing, and this phenomenon has also been observed by other groups [48,49]. The reason for this photocurrent decay has been investigated and mainly arises from charge recombination at the CdSe/TiO2 heterointerface or at grain boundaries within the mesoporous TiO2 network . The stability of the counter electrode is important to the performance of the QDSSC. Here, the photostability of the cells with ZnO/PbS(6) and ZnO/PbS(9) CEs were measured under continuous illumination with AM1.5 simulated sun light at an applied potential of 0 V. For comparison, the stability of QDSSC with Pt CE was also investigated.
Fig. 6. (a) Measurement of the photocurrent response under chopped simulated AM 1.5G (100 mW cm2) illumination for a QDSSC that incorporated ZnO/PbS(6), ZnO/ PbS(9) or Pt CE. (b) Photostability of QDSSC that based on ZnO/PbS(6), ZnO/PbS(9), and Pt CE under AM 1.5G illumination at 100 mW cm2 for 1 h.
Fig. 6b gives the variation of their JSC with respect to time under the continuous illumination. The sharp decrease of the photocurrent shown in Fig. 6a was also observed in Fig. 6b during the initial excitation period. Then, a relative steady photocurrent density of ~7 mA cm2 and ~11 mA cm2 was obtained for the cell with the ZnO/PbS(6) and ZnO/PbS(9) CE, respectively, during the 1 h illumination period. A slight drop in the value of JSC for ZnO/PbS(6) and ZnO/PbS(9) CEs resulted in a ~5% decrease of initial JSC for both cells. However, the cell with the reference Pt CE exhibited low photostability, and the photocurrent dropped by ~20% during the onehour illumination. The chemisorption of sulfur compounds from the electrolyte reduced the availability of the active catalytic sites; thus, the catalytic activity of the Pt CE decreased gradually over time. It is thus deduced that the short-time stability of the ZnO/PbS composite counter electrode with polysulﬁde electrolyte is better than that of Pt CE. In order to investigate the long-term stability of ZnO/PbS CEs in polysulﬁde electrolyte, the assembled QDSSC incorporated ZnO/ PbS(9) CE were disassembled after 5 (or 10) days of storage and then examined by X-ray diffraction. Fig. S12 shows the XRD patterns of the freshly prepared ZnO/PbS(9) CE and disassembled ZnO/ PbS(9) CEs. Compared to the fresh ZnO/PbS(9) CE, it can be seen that the diffraction peaks of PbS are still remained after 10 days immersion in polysulﬁde electrolyte, but the diffraction peaks of ZnO are greatly reduced after immersion in electrolyte for 5 days and disappeared after 10 days. This indicates that the PbS catalyst is stable in polysulﬁde electrolyte, whereas the ZnO NRs framework is unstable. As we all know, ZnO is an amphoteric oxides and is unstable in acidic or alkaline solution, which might results in the slow dissolution of ZnO nanorods when they are soaked in alkaline polysulﬁde electrolyte (pH 13), leading to the disappearance of the diffraction peaks of ZnO NRs. Therefore, the degradation of ZnO NRs in polysulﬁde electrolyte may hinder the long term stability of the prepared ZnO/PbS CEs. In order to increase the stability of ZnO in photoelectrochemical (PEC) cells, coreeshell structures have be developed to by coating a chemically stable layer on the ZnO surface . SiO2 has been demonstrated to be one of very effective shell materials on ZnO, preventing the generation of Zn2þ/dye agglomerates in DSSC due to the strong interaction between Si4þ and O2 ions . TiO2 modiﬁcation of the ZnO surface is also a representative method to prevent the surface Zn atoms from being dissolved and forming Zn2þ/dye agglomerates . This coreeshell structure can also be applied to ZnO/PbS CEs to prevent the dissolution of ZnO nanorods in alkaline electrolyte, and the ZnO nanorod framework can also be replaced with TiO2 or ITO nanorod that chemically stable material in polysulﬁde electrolyte to improve the stability. The detail studies directed at improving the long-term stability of ZnO/PbS CEs are now in progress in our lab and it is expected that more stable ZnO/PbS CEs will be achieved in the near future. Although the maximum conversion efﬁciency obtained in this work is still less than that of the best liquid-junction QDSSC [12e15], this efﬁciency is achieved with single CdSe QD sensitizer, rather than the cascade CdS/CdSe or CuInS2/CdS co-sensitized structures adopted by most of the high-efﬁciency QDSSC [12,13,50]. Therefore, we believe that there is enormous scope for further improvement by optimizing the ZnO/PbS CEs and TiO2/CdSe photoanodes. Efforts to further enhance the PCE are currently in progress, which include using longer ZnO nanorods, post-annealing treatment to the prepared ZnO/PbS CEs, optimizing the photoanode structure and employment of other sensitizers with longer wavelength absorption edge. These measurements are expected to further improve the photovoltaic performance of QDSSC that incorporated ZnO/PbS CEs.
X. Song et al. / Journal of Power Sources 269 (2014) 661e670
4. Conclusion In summary, a series of ZnO/PbS core/shell nanorod arrays were designed and fabricated on FTO substrates via a low cost wetchemical route and used for the ﬁrst time as new efﬁcient counter electrodes for liquid-junction QDSSC. These electrode composites exhibited synergy between the intrinsically high electrocatalytic PbS nanoparticles, the large surface area, and high electrical conductivity of the ZnO nanorods supports, leading to the remarkable decrease in charge transfer resistance and the enhancement of catalytic activity of CE. As a result, the QDSSC with the optimized ZnO/PbS(9) CE demonstrated an increase of 118.6% and 77.9% in PCE compared to that with planar Pt and PbS CE, respectively. It is believed that the PCE can be further enhanced by optimizing the length and density of ZnO NRs, the material and structure of photoanode, and post-heat treatment to the prepared ZnO/PbS CEs. Therefore, the synthesized ZnO/PbS composites could be used as promising catalytic electrodes with high activity and low cost replacing the incumbent Pt electrode. Acknowledgments The authors gratefully acknowledge ﬁnancial support from Natural Science Foundation of China (NSFC Grant Nos. 61176056 and 91123019). This work has been ﬁnancially supported by NSFC Major Research Program on Nanomanufacturing (Grant No. 91323303), 111 Program (No. B14040) and the open projects from Institute of Photonics and Photo-Technology, Provincial Key Laboratory of Photoelectronic Technology, Northwest University, China. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2014.07.044. References  A. Kongkanand, K. Tvrdy, K. Takechi, M. Kuno, P.V. Kamat, J. Am. Chem. Soc. 130 (2008) 4007e4015.  Z. Yang, M. Wang, Y. Shi, X. Song, Z. Lin, Z. Ren, J. Bai, J. Mater. Chem. 22 (2012) 21009e21016.  W.W. Yu, L.H. Qu, W.Z. Guo, X.G. Peng, Chem. Mater. 15 (2003) 2854e2860.  R. Vogel, P. Hoyer, H. Weller, J. Phys. Chem. 98 (1994) 3183e3188.  M.C. Beard, K.P. Knutsen, P. Yu, J.M. Luther, Q. Song, W.K. Metzger, R.J. Ellingson, A.J. Nozik, Nano Lett. 7 (2007) 2506e2512.  R.D. Schaller, V.I. Klimo, Phys. Rev. Lett. 92 (2004) 186601.  A.J. Nozik, Inorg. Chem. 44 (2005) 6893e6899.  G. Nair, L.-Y. Chang, S.M. Geyer, M.G. Bawendi, Nano Lett. 11 (2011) 2145e2151.  M.T. Trinh, A.J. Houtepen, J.M. Schins, T. Hanrath, J. Piris, W. Knulst, A.P.L.M. Goossens, L.D.A. Siebbeles, Nano Lett. 8 (2008) 1713e1718.  J.B. Sambur, T. Novet, B.A. Parkinson, Science 330 (2010) 63e66.  O.E. Semonin, J.M. Luther, S. Choi, H.-Y. Chen, J. Gao, A.J. Nozik, M.C. Beard, Science 334 (2011) 1530e1533.  P.K. Santra, P.V. Kamat, J. Am. Chem. Soc. 134 (2012) 2508e2511.
 M.A. Hossain, J.R. Jennings, C. Shen, J.H. Pan, Z.Y. Koh, N. Mathews, Q. Wang, J. Mater. Chem. 22 (2012) 16235e16242.  Z. Pan, K. Zhao, J. Wang, H. Zhang, Y. Feng, X. Zhong, ACS Nano 7 (2013) 5215e5222.  H. Huang, L. Pan, C.K. Lim, H. Gong, J. Guo, M.S. Tse, O.K. Tan, Small 9 (2013) 3153e3160.  J. Deng, M. Wang, X. Song, Y. Shi, X. Zhang, J. Colloid Interface Sci. 388 (2012) 118e122.  M.S. Faber, K. Park, M. Caban-Acevedo, P.K. Santra, S. Jin, J. Phys. Chem. Lett. 4 (2013) 1843e1849.  P.V. Kamat, Acc. Chem. Res. 45 (2012) 1906e1915.  V. Chakrapani, D. Baker, P.V. Kamat, J. Am. Chem. Soc. 133 (2011) 9607e9615.  Z. Yang, C.-Y. Chen, C.-W. Liu, H.-T. Chang, Chem. Commun. 46 (2010) 5485e5487.  X. Song, M. Wang, Y. Shi, J. Deng, Z. Yang, X. Yao, Electrochim. Acta 81 (2012) 260e267.  J.G. Radich, R. Dwyer, P.V. Kamat, J. Phys. Chem. Lett. 2 (2011) 2453e2460.  Y.-L. Lee, C.-H. Chang, J. Power Sources 185 (2008) 584e588.  G.S. Paul, J.H. Kim, M.-S. Kim, K. Do, J. Ko, J.-S. Yu, ACS Appl. Mater. Interfaces 4 (2012) 375e381.  F. Hao, P. Dong, J. Zhang, Y. Zhang, P.E. Loya, R.H. Hauge, J. Li, J. Lou, H. Lin, Sci. Rep. 2 (2012) 368.  G. Hodes, J. Manassen, D. Cahen, J. Electrochem. Soc. 127 (1980) 544e549. nez, I. Mora-Sero , L. Macor, N. Guijarro, T. Lana-Villarreal, R. Go mez,  S. Gime L.J. Diguna, Q. Shen, T. Toyoda, J. Bisquert, Nanotechnology 20 (2009) 295204.  Z. Tachan, M. Shalom, I. Hod, S. Rühle, S. Tirosh, A. Zaban, J. Phys. Chem. C 115 (2011) 6162e6166.  Y. Zhang, C. Shi, X. Dai, F. Liu, X. Fang, J. Zhu, Electrochim. Acta 118 (2014) 41e44.  H. Yuan, J. Lu, X. Xu, D. Huang, W. Chen, Y. Shen, M. Wang, J. Electrochem. Soc. 160 (2013) H624eH629.  S.S. Kalanur, S.Y. Chae, O.S. Joo, Electrochim. Acta 103 (2013) 91e95.  S. Yang, A.S. Nair, P. Zhu, S. Ramakrishna, Mater. Lett. 76 (2012) 43e46.  Z. Yang, C.-Y. Chen, C.-W. Liu, C.-L. Li, H.-T. Chang, Adv. Energy Mater. 1 (2011) 259e264. €tzel, Sci. Rep. 3  X. Xu, D. Huang, K. Cao, M. Wang, S.M. Zakeeruddin, M. Gra (2013) 1489.  J. Deng, M. Wang, X. Song, J. Liu, J. Colloid Interface Sci. 588 (2014) 399e405.  J. Ding, M. Wang, J. Deng, W. Gao, Z. Yang, C. Ran, X. Zhang, J. Alloys Compd. 582 (2014) 29e32.  S. Seghaier, N. Kamoun, R. Brini, A.B. Amara, Mater. Chem. Phys. 97 (2006) 71e80.  X. Song, M. Wang, H. Zhang, J. Deng, Zhi Yang, Chenxin Ran, Xi Yao, Electrochim. Acta 108 (2013) 449e457.  M. Seol, H. Kim, W. Kim, K. Yong, J. Mater. Chem. 19 (2009) 5945e5951.  S.S. Mali, H. Kim, P.S. Patil, C.K. Hong, Dalton Trans. 42 (2013) 16961e16967.  X. Wang, H. Zhu, Y. Xu, H. Wang, Y. Tao, S. Hark, X. Xiao, Q. Li, ACS Nano 4 (2010) 3302e3308.  H. Kim, H. Jeong, T.K. An, C.E. Park, K. Yong, ACS Appl. Mater. Interfaces 5 (2013) 268e275.  C.-Y. Lin, C.-Y. Teng, T.-L. Li, Y.-L. Lee, H. Teng, J. Mater. Chem. A 1 (2013) 1155e1162.  H. Chen, L. Zhu, H. Liu, W. Li, J. Phys. Chem. C 117 (2013) 3739e3746.  S. Rühle, M. Shalom, A. Zaban, ChemPhysChem 11 (2010) 2290e2304.  F. Hao, Z. Wang, Q. Luo, J. Lou, J. Li, J. Wang, S. Fan, K. Jiang, H. Lin, J. Mater. Chem. 22 (2012) 22756e22762.  C.J. Raj, K. Prabakar, A.D. Savariraj, H.-J. Kim, Electrochim. Acta 103 (2013) 231e236.  I. Robel, V. Subramanian, M. Kuno, P.V. Kamat, J. Am. Chem. Soc. 128 (2006) 2385e2393.  T.-L. Li, Y.-L. Lee, H. Teng, J. Mater. Chem. 21 (2011) 5089e5098.  R. Zhao, L. Zhu, F. Cai, Z. Yang, X. Gu, J. Huang, L. Cao, Appl. Phys. A 113 (2013) 67e73.  Z. Qin, Y. Huang, Q. Liao, Z. Zhang, X. Zhang, Y. Zhang, Mater. Lett. 70 (2012) 177e180.