TiO₂ nanoparticle/ZnO nanowire hybrid photoanode for enhanced quantum dot-sensitized solar cell performance

Abstract

A hybrid structured TiO₂/ZnO photoanode was composed of highly ordered ZnO nanowires (NWs) and small TiO₂ nanoparticles (NPs) filling the gaps among ZnO NWs. In this way, ZnO NWs provide a direct pathway to facilitate electron collection and transport and increase light scattering and trapping. TiO₂ NPs provide a large specific surface to effectively adsorb quantum dots (QDs). In experiments, the density of ZnO NW arrays was first controlled by regulating the ratio of TiO₂ NPs to ZnO NPs and the concentration of ZnO NPs in the hybrid NP seed precursor, and then the TiO₂ paste was successfully use to fill the gaps among NWs by a large centrifugal force. Using TiO₂ NP/ZnO NW films as photoanodes, we fabricated CdSe QD-sensitized solar cells and tested their photovoltaic (PV) performances. The results demonstrated a remarkably enhanced short-circuit current density (J_sc) of 7.9 mA cm⁻² and power conversion efficiency (η) of 1.55%, and this η is enhanced by 19.2% and 30.3%, respectively, compared to those of a TiO₂ NP device with an η of 1.3% and a ZnO NW device with an η of 1.19%.

---

Introduction

In recent years, photoelectrochemical (PEC) cells have attracted worldwide attention as cheap alternatives to conventional devices for solar energy conversion. Generally, the sequence of physical processes in PEC cells is as follows: (i) the incident photons are captured, and electron–hole pairs are generated; (ii) the electrons and holes separate quickly into respective transfer materials; and (iii) the carriers are transported to opposing electrodes. The QD-sensitized solar cell (QDSSC) is one type of PEC cell. In order to improve efficiency of QDSSCs, much effort has been made. Apart from the search for lower band gap QDs (PbSe, PbTe),^1,2 electrolyte³ and counter electrode,⁴ a large number of research studies focused on nanostructured photoanodes.^5,6 We have reported a novel structure (subsectionally sensitized cells using double-layer ZnO nanorods).⁷ Conventional QDSSCs are fabricated using porous NP films. The NP films provide a very high surface area to ensure sufficient sensitizer loading. However, the major drawback of the NP films is high charge recombination loss due to the electron trapping and scattering at the grain boundary and inefficient light scattering ability within small-sized NPs (10–20 nm) in the disordered network.^5,8 To overcome this obstacle, a 1D nanostructure has been widely investigated in photoelectrodes for solar energy conversion devices. They can offer a continuous pathway for photogenerated electrons to transport along the long axis of one dimension and a large decrease in grain boundaries as well as affording low reflectance due to light scattering and trapping. In addition, the electron diffusion coefficient in a single crystal is more than two orders of magnitude higher than that in 0D NPs.⁹ Despite the advantageous characteristics noted above, the efficiency of QDSSCs using a 1D nanostructured photoanode remains low because the specific surface area of 1D nanostructure films is roughly an order of magnitude lower than that of NP films. There is considerable free space between adjacent 1D nanostructures, which results in limited sensitizer loading.^10,11

Herein, we report a hybrid structured TiO_2/ZnO photoanode composed of highly ordered ZnO NWs and small TiO₂ NPs filling the gaps among ZnO NWs. This hybrid structure not only retains the high surface area of NP films but also maintains the advantages of a 1D nanostructure, including the light scattering and trapping and the higher electron diffusion coefficient. In this structure, TiO₂ NPs can effectively adsorb QDs, and ZnO NWs act as highway that breadth-wise receives photogenerated electrons, reduces the interface resistance of NPs, and then rapidly transports the electron to an indium tin oxide (ITO) substrate. In this paper, the commercial P₂₅ powder is used owing to its low cost and unique physical characteristics, and ZnO NWs are used due to their well-developed synthetic methods and excellent electrical properties.¹² The resulting hybrid structured films were subsequently exploited as photoanodes in QDSSCs, exhibiting a markedly enhanced short-circuit current density (J_sc) of 7.9 mA cm⁻² and power conversion efficiency (η) of 1.55%. This η is enhanced by 19.2% and 30.3% compared to those of a TiO₂ NPs device with an η of 1.3% and a ZnO NW device with an η of 1.19%, which can be attributed primarily to the synergistic effect of higher sensitized loading due to the presence of a large surface area, superior light scattering ability and trapping, and a higher electron diffusion coefficient.

Experimental section

Preparation of TiO₂ sol and TiO₂ NP/ZnO NP hybrid colloids

Preparation of TiO₂ sol. First, 3.4 ml of tetra-n-butyl titanate was added to 8.3 ml of diethanolamine. Then a mixture of 1.8 ml deionized water and 100 ml ethanol was added. After stirring for 1 h, the mixture was aged for 24 h at 30 °C.

Preparation of TiO₂ NP/ZnO NP hybrid colloids. First, 1.64 g of zinc acetate and 0.96 g of KOH were dissolved in 84 ml and 46 ml of methanol, respectively. Then KOH solution was added to zinc acetate solution in a flask equipped at room temperature. The system reacted under constant stirring for 2 h at 60 °C. The obtained QDs were separated by centrifugation, washed thoroughly in methanol and dispersed in methanol (50 mg ml⁻¹).

After adding 3.6 ml of tetra-n-butyl titanate to 7.2 ml of ethanol, 7.2 ml of acetic acid glacial and 18 ml of deionized water were added successively. After stirring for 1 h, the mixture reacted in the sealed container for 12 h at 200 °C. As for ZnO NPs, TiO₂ NPs were separated, washed and then dispersed (50 mg ml⁻¹).

Next 0.9, 0.5 and 0.1 ml of the as-prepared 50 mg ml⁻¹ ZnO NP colloids were diluted using 1 ml of methanol each, and then 0.1, 0.5 and 0.9 ml of the as-prepared 50 mg ml⁻¹ TiO₂ NP colloids were added to 1.9, 1.5 and 1.1 ml of diluted ZnO NP colloid, respectively. The ratios of ZnO to TiO₂ were 22.5 : 2.5, 11.5 : 11.5 and 2.5 : 22.5 (mg ml⁻¹ : mg ml⁻¹), and the effect of the ratio on the density of ZnO NWs was investigated. For investigating the effect of the concentration of ZnO NPs on ZnO NWs, we diluted the hybrid colloids to 0.625, 0.156 and 0.039 mg ml⁻¹ under the best ratio (2.5 : 22.5), respectively.

Preparation of hybrid photoelectrodes

The detailed strategy for synthesizing TiO₂ NP/ZnO NW hybrid photoelectrodes is illustrated in Fig. 1. The TiO₂ sol, serving as glue to provide a strong interaction between NPs and the substrate, was first spin-coated on the ITO substrate, as we can see from Fig. 1a. After the spin-coating of TiO₂ sol, immediately, the TiO₂ NP/ZnO NP hybrid colloids were spin-coated (2000 rpm, 30 s, once; Fig. 1b). We already used ZnO NP sol alone to prepare seed films, and the density of NWs decreased by decreasing the concentration of ZnO NPs. The findings were that the island-like NW film arose (very dense or very sparse in some locations) when the concentration was below 5 mg ml⁻¹, and the NWs were still very dense when the concentration was over 5 mg ml⁻¹ (see Fig. S1, ESI†). In this paper, ZnO NPs were spaced by TiO₂ NPs, and the density of ZnO NWs was regulated by the ratio of ZnO NPs to TiO₂ NPs and the concentration of ZnO NPs. Then the wet films were annealed at 450 °C for 0.5 h in air, and ZnO seed films were obtained.

Fig. 1 The detailed strategy for the synthesis of TiO₂ NP/ZnO NW hybrid photoelectrodes.

Fig. 1c shows the growth of the ZnO NWs, which was achieved by performing a procedure similar to that in our previous paper.⁷ In the previous experiment, ZnO seeds were not regulated, and thus, the NWs easily fused when longer. Herein, the purpose of introducing TiO₂ NPs was to space ZnO NPs, so that ZnO NWs only grew in the position where ZnO NPs were seeded. Density-controlled synthesis of ZnO NWs was completed at a certain ratio of ZnO/TiO₂ and concentration of ZnO NPs. Another precondition of the filling paste is the preparation of TiO₂ paste with good fluidity. TiO₂ paste was prepared according to a method id the reference literature.¹³ It was difficult to fill the gaps among NWs with paste by printing pressure and gravity, and therefore, the samples were stuck on the inner wall of the centrifuge and the gaps were filled by a large centrifugal force (5000 r min⁻¹, 10 min) after the ZnO arrays were covered with paste by screen printing as shown in Fig. 1d. After the process from printing to filling was repeated for two cycles, the samples were annealed at 450 °C for 0.5 h in air, and thus, the hybrid photoelectrodes were obtained.

QDSSC fabrication and physical characterization

The deposition of CdSe QDs was achieved by the successive ion layer adsorption and reaction (SILAR) method.^14,15 Briefly, 40 ml of 30 mM SeO₂ water solution was reduced with KBH₄ as a Se²⁻ source in an Ar atmosphere. Moreover, 40 ml of Cd(NO₃)₂ solution (30 mM) as a Cd²⁺ source was prepared. The SILAR process was repeated for 12 cycles.

A 60 μm-thick hot-melt ionomer film (Surlyn) under heating (120 °C) was then sandwiched between the sensitized photoanode and a 100 nm thick Pt counter electrode. A polysulfide electrolyte of 0.5 M Na₂S, 2 M S, and 0.2 M KCl in methanol–water (7 : 3 by volume) was injected via the predrilled hole in Pt.

The crystalline phase of photoelectrodes was identified by X-ray diffraction (XRD) using a D/max-2400 X-ray diffraction spectrometer (Rigaku) with Cu Ka radiation and operated at 40 kV and 100 mA from 20 to 70 °C, and the scanning speed was 15 °C min⁻¹ at a step of 0.02 °C. The morphologies of the samples were analyzed using field emission scanning electron microscopy (FESEM, HITACHI S-4800). The transmittance spectra and absorption spectra of films were recorded at 300–800 nm with a JascoV-570 UV-Vis-NIR photospectrometer. The I–V characteristics were tested under 100 mW cm⁻² AM 1.5G simulated sunlight (100 W Xe source). I (V) was recorded by a Keithley 2400 source meter (active area is 0.25 cm²). It should be noted that the solar simulator was previously calibrated by using a Si reference cell equipped with a KG-5 filter. The external quantum efficiency (EQE) system uses a lock-in amplifier (SolarCellScan 100, Zolix, China) to record the short circuit currents under chopped monochromatic light.

Results and discussion

SEM analysis

In this paper, there are two challenges that we must settle in the preparation of this novel hybrid structure, one is the density-controlled synthesis of ZnO NWs and the other is the injection of TiO₂ paste. The former one is more difficult to handle. The paste can be well injected only when the gaps among the NWs are large enough. The gap is dominated by the density of NWs and eventually regulated by the amount of ZnO NPs that are spin-coated on the ITO substrate. Finally, we regulated the density of NWs by the ratio of ZnO NPs to TiO₂ NPs and the concentration of ZnO NPs. The results are shown in Fig. 2. Fig. 2a–c shows cross-sectional SEM images of NWs synthesized using the different ratio of hybrid colloids (22.5 : 2.5, 11.5 : 11.5 and 2.5 : 22.5) as the seed precursor. It can be seen that the density of NWs decreased as the radio decreased, but this is not obvious. With the low ratio of 2.5 : 22.5, the density of NWs is still high, which can be seen in Fig. 2c. It is possible that the concentration of ZnO NPs is still high and the density of seeds exploitable for NW growth is similar for the three samples although the ratio of 2.5 : 22.5 is low (i.e. the lower ZnO–TiO₂ ratio should allow a higher density). The NWs are in a mess or not observed on substrates when the ratio (ZnO–TiO₂) is lower than 2.5–22.5 (e.g. 1.25–23.75 or 0.25–24.75), and thus, we regulate the density of NWs by decreasing the concentration of ZnO NPs under the ratio of 2.5–22.5. Fig. 2c–f shows the cross-sectional SEM images of NWs grown on seeds prepared by 2.5, 0.625, 0.156 and 0.039 mg ml⁻¹ ZnO NP hybrid colloids, respectively. It can be seen that the density of NWs obviously decreased with the decrease in the concentration of ZnO NPs. The gap was large enough for TiO₂ paste to be injected when the concentration was below 0.1 mg ml⁻¹. As we can see from the results above, compared with the ratio, the concentration plays a more important role in determining the density of NWs. Previously the authors used pure ZnO NPs as the seed precursor to synthetize NWs. Experimental results show the seed films were poor when the concentration of pure ZnO NPs was below 5 mg ml⁻¹. Thus, in this experiment, TiO₂ plays two roles, it not only helps to regulate the density of NWs, but also contributes to lower the concentration of ZnO NPs.

Fig. 2 Cross-sectional SEM images of NWs synthesized using different ratios of ZnO NPs to TiO₂ NPs [a-22.5 : 2.5, b-11.5 : 11.5 and c-2.5 : 22.5 (mg ml⁻¹ : mg ml⁻¹)] as the seed precursor and different concentrations of ZnO NPs (d-0.625, e-0.156 and f-0.039 mg ml⁻¹) as the seed precursor under 2.5 : 22.5, respectively.

ZnO NWs were synthesized by a chemical bath deposition (CBD) method.⁶ Briefly, the seed films were immersed in a 100 ml mixed aqueous solution of 0.04 M Zn(NO₃)₂ and 0.8 M NaOH at 80 °C for 1 h. The length of NWs varied in the range of 3.27–3.44 μm, and the difference is not significant as we can see from Fig. 2. With the decrease in the ratio from 22.5 : 2.5 to 11.5 : 11.5 to 2.5 : 22.5, the diameter of NWs changes from 130 to 150 to 170 nm, indicating an inconspicuous change. Nevertheless, with the decrease in the concentration of ZnO NPs from 0.625 to 0.156 to 0.039 mg ml⁻¹, the diameter of NWs increased obviously from 180 to 220 to 380 nm. There are different explanations for this phenomenon.

Lee et al. thought the growth of ZnO microrods could be considered as a two-stage growth: the growth process of ZnO rods, which bundle together, and the growth steps via the coalescence mode. Therefore, the bundle of ZnO nanorods could be further transformed to microrods.¹⁶ Guo et al. proposed two density-dependent growth regimes: an independent growth regime, where the spacing between adjacent NWs is larger than 2λs (the diffusion length of Zn adatoms on the substrate surface), and competitive growth regime where NWs compete for Zn adatoms due to overlapping surface collection areas.¹⁷ Kang et al. thought the low density ZnO NWs grown at the edge of patterned seeds have larger diameters because of a greater precursor chemical source supply than is present for the high density ZnO NWs in the central regions of patterned seeds.¹⁸ We think this can be better explained using the comprehensive view of Guo and Wook, in which low density NWs (the independent growth regimes) have a greater precursor chemical source supply compared with high density NWs (the competitive growth regimes), and single NW obtains more Zn(OH)₄²⁻ complex for the radial growth of the NW per unit of time. When NWs are the competitive growth regimes, the small gap decreases the growth solution around NWs, and individual NWs get less Zn(OH)₄²⁻ complex for radial growth.

Mechanism analysis

In this way, the presence of ZnO NWs played two major roles, i.e., promoting the electron transfer and increasing the light scattering. Fig. 3a and b illustrates different electron diffusion processes in a bare TiO₂ NP and the obtained TiO₂ NP/ZnO NW hybrid films, respectively. In bare TiO₂ NP film, electrons transport via a zigzag pathway through the NPs, and a large number of grain boundaries result in the increase in charge recombination. In the TiO₂ NP/ZnO NW hybrid film, these NWs supply the highway within TiO₂ NPs, and the generated electrons diffuse breadth-wise to ZnO NWs through a small quantity of grain boundaries of TiO₂ NPs and length – ways transport to ITO in one direction. In addition, Fig. 3a and b also illustrates the sunlight transmission processes in a bare TiO₂ NP film and the obtained TiO₂ NP/ZnO NW hybrid films, respectively. In bare TiO₂ NPs film, sunlight transmission is linear and there is no light scattering. There is also some light transmission out of the film and waste. In the TiO₂ NP/ZnO NW hybrid film, these NWs supply the centers for light scattering, and the optical path is increased by the multiple reflections, so the sunlight is sufficiently absorbed.

Fig. 3 Schematic illustration of electron (e−) diffuse transport and light transmission within the NP film (a) and NP/NW film (b). Cross-sectional SEM images of (c) bare TiO₂ NP and (d) TiO₂ NP/ZnO NW hybrid films. (Blue spheres, red shells, and green pillars represent TiO₂ NPs, CdSe QDs, and ZnO NWs, respectively).

The controlled synthesis of ZnO NWs has been described in the section ‘SEM analysis’. In addition, the injection of TiO₂ paste is also a crucial step. Fig. 3c and d shows cross-sectional SEM image of the bare TiO₂ NP film and TiO₂ NP/ZnO NW hybrid film. The bare TiO₂ NP film was porous and homogeneous as shown in Fig. 3c. From Fig. 3d, TiO₂ paste was successfully injected into the gaps among ZnO NWs, but it did not completely reach the bottom. A large gap existed due to the tilt of NWs, which was detrimental to the cell. We will improve the verticality of NWs in subsequent work.

XRD and optic characteristics of films

Fig. 4a shows XRD patterns of the TiO₂ NP/ZnO NW hybrid film and pure ZnO NW film prepared on ITO substrates. As indicated in the XRD patterns, all diffraction peaks of the hybrid film and the pure ZnO NW film can be well indexed to hexagonal wurtzite ZnO (JCPDS no. 36-1451) and anatase TiO₂ (JCPDS no. 21-1272). Apparently, after injecting TiO₂ paste, the intensities of all ZnO peaks decreased due to the coverage of TiO₂.

Fig. 4 (a) XRD patterns of ZnO NW and TiO₂ NP/ZnO NW films on ITO substrates; (b) transmittance spectra of bare TiO₂ NP film and TiO₂ NP/ZnO NW films without CdSe.

The transmissivity of the films was studied to investigate the scattering effect of the ZnO NWs. Fig. 4b shows the transmittance spectra of TiO₂ NP/ZnO NW hybrid film and TiO₂ NP film without CdSe, and the films have similar thickness. Apparently, the hybrid film had lower transmissivity (<20%) in the visible wavelength range of 400–800 nm than the bare TiO₂ NP film (>36%), suggesting the combination of ZnO NWs and TiO₂ NPs led to the higher light harvesting efficiency and most of the light was restricted within the photoanodes. This evidence confirmed the light scattering effect of the ZnO NWs in the TiO₂ NP film.

Performance of cells

Introducing ZnO NWs into TiO₂ NP films greatly reduces the transmission of light (Fig. 4b), which indicates ZnO NWs have a high scattering effect to injected light. The scattering promotes the increase of absorption spectra and photon-to-electron conversion efficiency. Fig. 5a–c shows the absorption spectra of cells in the visible light, EQE and I–V curve of cells.

Fig. 5 (a) UV-vis absorption spectra, (b) photocurrent–voltage curves and (c) external quantum efficiency (EQE) of ZnO/CdSe, TiO₂/CdSe and TiO₂/ZnO/CdSe solar cells. (6, 9 and 12 are the numbers of CdSe deposition cycles).

UV-vis absorption spectra in Fig. 5a show the absorption intensity and absorption onset with an increase in the number of CdSe deposition cycles. The absorption onset has a red shift from 650 nm to 750 nm with an increase in deposition cycles from 6 to 12, and the absorption intensities are same. TiO₂/CdSe, ZnO/CdSe and TiO₂/ZnO/CdSe films have the same absorption onset of 750 nm when the number of cycles is 12 times. But the absorption intensities of ZnO/CdSe and TiO₂/ZnO/CdSe films in the wavelength range of 350–650 nm are higher than that of TiO₂/CdSe film due to the scattering effect of ZnO NWs. The absorption spectrum of TiO₂/ZnO/CdSe film is similar to that of ZnO/CdSe film. The specific surface area of 1D nanostructured films is roughly an order of magnitude lower than that of NP films when they have the same thickness, as was mentioned in the Introduction, and thus, the adsorption of sensitizer is relatively low and the light absorption is also weak. However, here the absorption of ZnO NW films was enhanced, and we think the key reason is that the light scattering effect of NWs is very good and injected light is fully absorbed after multiple reflections. From the analysis of absorption spectra, a 1D nanostructure can improve light absorption of cells via the scattering effect.

Using TiO₂/CdSe, ZnO/CdSe and TiO₂/ZnO/CdSe films, we fabricated solar cells to evaluate their PV performances. The I–V characteristics for QDSSCs were measured at an illumination of one sun in Fig. 5b. In order to obtain a better understanding of the effect of the amount of CdSe and the structure of photoanodes on the PV performance, Table 1 summarizes the open-circuit voltage (V_oc), short-circuit current density (J_sc), fill factor (FF), and η of cells. First, we investigated the effect of the number of CdSe deposition cycles on the efficiency of cells using bareTiO₂ NP film as the photoanode. From Fig. 5b and Table 1, the V_oc, J_sc, FF, and η increased with an increase in deposited CdSe. The V_oc in electrochemical QDSSCs corresponds to the difference between the quasi-Fermi level of the semiconductor E_fn with respect to the dark value (E_f0), which equals the electrolyte redox energy (E_f0 = E_redox). Therefore, it can be written as:

V_oc = (E_fn − E_f0)/e = (k_BT/e)ln(n/n₀) (1)

where k_BT, e, and n₀ are the thermal energy, the positive elementary charge, and the concentration in the dark, respectively.^19,20 The adsorption of QDs increased with an increase in the number of deposition cycles, which helped to produce more photons under the same testing conditions. More electrons were obtained by capturing more photons and transferred to the conduction band (CB) of the semiconductor. This induced the shift of the Fermi level into the CB and the increase of V_oc. Moreover, more photo-generated carriers were transported in the external circuit and J_sc increased from 2 to 2.91 to 5.1 mA cm⁻². From Table 1, FF increased from 52.3% to 53.9% to 59.1%, and the reason is that QDs were more dense on the surface of TiO₂ with the increase in the number of deposition cycles, which reduced the contact between electrolyte and TiO₂, inducing the recombination of carriers. The photovoltaic conversion efficiency of QDSSCs was estimated, and the η improved from 0.41% to 0.67% to 1.3% with the increase in CdSe due to the increase in J_sc, V_oc and FF. Secondly, we carried out a comparison of TiO₂/CdSe, ZnO/CdSe, and TiO₂/ZnO/CdSe cells, and the best efficiencies are shown in Fig. 5b and Table 1. The films of TiO₂, ZnO and TiO₂/ZnO had the same thickness. It was seen that the ZnO/CdSe cell has a high V_oc of 0.61 V, and the TiO₂/CdSe cell has a low V_oc of 0.43 V. The difference in V_oc between them was due to the electron mobility of optical electrode materials. The electron mobility of ZnO is higher than that of TiO₂, which leads to rapid transfer of photoelectrons to the CB. In addition, the high electron concentration in CB induced the shift of the Fermi level into the CB and led to a larger energy band gap. Thus, the V_oc was increased.¹⁰ The TiO₂/CdSe cell has a high J_sc due to the large specific surface area. However, compared with the TiO₂/CdSe cell, the TiO₂/ZnO/CdSe cell shows improvement in V_oc owing to the introduction of ZnO NWs. Here, there are two important factors affecting current density: one is that the large specific surface area can increase the adsorption of the sensitizer and improve the light absorption, and the other is that the scattering effect of ZnO NWs improves the effective use of light. Therefore, the large specific surface area of TiO₂ NPs and the light scattering effect of ZnO NWs fully improved the J_sc of the TiO₂/ZnO/CdSe cell. In addition, ZnO NWs supplied the highway for photon-generated electrons and accelerated electron collection. However, it is also seen that the FF of the TiO₂/CdSe cell is the highest, and whether the ZnO material affects the FF remains unclear. Further investigations will be carried out in future works. Overall, the TiO₂/ZnO/CdSe cell yielded 19.2% and 30.3% enhancements in power conversion efficiency (1.55%) over the optimized TiO₂/CdSe and ZnO/CdSe cells.

Table 1 Photovoltaic parameters obtained from the I–V curves using TiO₂/CdSe, ZnO/CdSe and TiO₂/ZnO/CdSe cells

Cells	Voc (V)	J (mA cm^−2)	FF (%)	η (%)
TiO2/CdSe-6	0.39	2.00	52.3	0.41
TiO2/CdSe-9	0.41	2.92	53.9	0.67
TiO2/CdSe-12	0.43	5.10	59.1	1.30
ZnO/CdSe-12	0.61	4.64	42.3	1.19
TiO2/ZnO/CdSe-12	0.48	7.90	40.9	1.55

---

Fig. 5c shows the external quantum efficiency (EQE) spectra of representative CdSe sensitized solar cells based on different photoanodes. In general, the EQE for wavelengths close to the quantum dot absorption edge is relatively lower than for short wavelengths. This is mainly due to the optical losses (i.e. lower absorbance for wavelengths close to the absorption onset, Fig. 5a). Strong enhancement of EQE was observed when ZnO NWs were introduced TiO₂ NPs films, and because ZnO NWs improved the scattering of longer wavelength light, further improvement at longer wavelengths (500–600 nm) is realized. It also was shown that the EQE of the ZnO NW cell is higher than that of the TiO₂ NP cell in range of 500–600 nm. The increased photocurrent and EQE are likely to be related to the change in absorption characteristics.

Conclusion

In conclusions, we have reported the fabrication and characterization of a TiO₂ NP/ZnO NW hybrid photoanode-based CdSe-sensitized solar cell. SEM images of ZnO NWs show a density controlled synthesis of ZnO NWs. The density of NWs was controlled by regulating the ratio of ZnO NPs to TiO₂ NPs and the concentration of ZnO NPs in the hybrid seed precursor. The gap was large enough for TiO₂ paste to fill when the ratio of ZnO NPs to TiO₂ NPs and concentration of ZnO NPs were 1 : 9 and below 1 mg ml⁻¹, respectively. The hybrid photoanode had low transmission, high absorption and an EQE in the visible wavelength of 400–800 nm compared to bare TiO₂ NP film, suggesting the combination of ZnO NWs and TiO₂ NPs led to the higher light harvesting efficiency of the photoanode. CdSe QDSSCs based on TiO₂ NP film, ZnO NW film and TiO₂ NP/ZnO NW hybrid film were fabricated, and J–V curves exhibited a markedly enhanced J_sc of 7.9 mA cm⁻² and η of 1.55% for the TiO₂/ZnO cell. This η was a 19.2% and 30.3% enhancement as compared to those of a TiO₂ NP device with a η of 1.3% and a ZnO NW device with a η of 1.19%.

Acknowledgements

This work has been partially supported by the NSFC Major Research Plan on Nanomanufacturing (Grant no. 91323303). The authors gratefully acknowledge financial support from the Natural Science Foundation of China (Grant no. 61176056 and 91123019), the 111 program (no. B14040) and the open projects from Institute of Photonics and Photo-Technology, Provincial Key Laboratory of Photoelectronic Technology, Northwest University, China.

References

1. Z. H. Lin, M. Q. Wang, L. Y. Zhang, Y. H. Xue, X. Yao, H. W. Cheng and J. T. Bai, J. Mater. Chem., 2012, 22, 9082.
2. Z. Yang, M. Q. Wang, Y. H. Shi, X. H. Song, Z. H. Lin, Z. Y. Ren and J. T. Bai, J. Mater. Chem., 2012, 22, 21009.
3. J. D. Fan, Y. Hao, A. Cabot, E. M. J. Johansson, G. Boschloo and A. Hagfeldt, ACS Appl. Mater. Interfaces, 2013, 5, 1902.
4. Z. Yang, C.-Y. Chen, C.-W. Liu, C.-L. Li and H.-T. Chang, Adv. Energy Mater., 2011, 1, 259.
5. Y. Bai, H. Yu, Z. Li, R. Amal, G. Q. Lu and L. Z. Wang, Adv. Mater., 2012, 24, 5850.
6. S. Wooh, H. Yoon, J.-H. Jung, Y.-G. Lee, J. H. Koh, B. Lee, Y. S. Kang and K. Char, Adv. Mater., 2013, 25, 3111.
7. J. P. Deng, M. Q. Wang, X. H. Song, Y. H. Shi and X. Y. Zhang, J. Colloid Interface Sci., 2012, 388, 118.
8. M. D. Ye, D. J. Zheng, M. Q. Lv, C. Chen, C. J. Lin and Z. Q. Lin, Adv. Mater., 2013, 25, 3039.
9. L. Forro, O. Chauvet, D. Emin, L. Zuppiroli, H. Berger and F. Levy, J. Appl. Phys., 1994, 75, 633.
10. J. P. Deng, M. Q. Wang, J. Liu and X. H. Song, J. Colloid Interface Sci., 2014, 418, 277.
11. K. Zhu, N. R. Neale, A. Miedaner and A. J. Frank, Nano Lett., 2007, 7, 69.
12. M. Seol, H. Kim, Y. Tak and K. Yong, Chem. Commun., 2010, 46, 5521.
13. X. H. Song, M. Q. Wang, J. P. Deng, Z. Yang, C. X. Ran, X. Y. Zhang and X. Yao, ACS Appl. Mater. Interfaces, 2013, 5, 5139.
14. J. P. Deng, M. Q. Wang, X. H. Song and J. Liu, J. Alloys Compd., 2014, 588, 399.
15. H. J. Lee, M. K. Wang, P. Chen, D. R. Gamelin, S. M. Zakeeruddin, M. Grätzel and M. K. Nazeeruddin, Nano Lett., 2009, 9, 4221.
16. J.-H. Lee, M.-H. Hon, Y.-W. Chung and I.-C. Leu, J. Am. Ceram. Soc., 2009, 92, 2192.
17. D. L. Guo, L. H. Tan, Z. P. Wei, H. Y. Chen and T. Wu, Small, 2013, 9, 2069.
18. H. W. Kang, J. Yeo, J. O. Hwang, S. Hong, P. Lee, S. Y. Han, J. H. Lee, Y. S. Rho, S. O. Kim, S. H. Ko and H. J. Sung, J. Phys. Chem. C, 2011, 115, 11435.
19. A. Zaban, M. Greenshtein and J. Bisquert, ChemPhysChem, 2003, 4, 860.
20. J. Bisquert, A. Zaban and P. Salvador, J. Phys. Chem. B, 2002, 106, 8774.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra05033a

---