Chemical bath deposition of Cu₂O quantum dots onto ZnO nanorod arrays for application in photovoltaic devices

Abstract

Cu₂O quantum dots (QDs) decorated ZnO nanorod arrays (ZNAs) were fabricated using a facile hydrothermal method followed by a chemical bath deposition (CBD) process. The surface morphology, crystal structure and photovoltaic behaviors of the heterostructure films were investigated. The results indicate that the Cu₂O QDs decorated ZNAs can be used as a good light absorber improving the visible spectral absorption. In addition, the photo-induced electrons can easily transfer to ZnO, leading to an increase in the photovoltaic performance. When the number of CBD cycles was 10, an optimal photovoltaic performance could be obtained under simulated sunlight illumination (AM 1.5G, 100 mW cm⁻²) with a photocurrent of 3.21 mA cm⁻², an open circuit photovoltage of 0.65 V and a conversion efficiency of 1.17%. Moreover, for improving the photovoltaic stability, a protective layer was prepared on the Cu₂O QDs by a simple process of heat treatment in ambient air at 100 °C for 2 h. The results demonstrate that the passive CuO layer could be an effective protective layer to increase the photovoltaic stability.

---

1. Introduction

At present silicon-based solar cells dominate the market of photovoltaics, but the silicon purification process is energy-intensive and very costly. Thus, it is important to develop low cost, efficient and reliable photovoltaic materials for managing the growing global energy demand and reducing greenhouse gas emissions.¹ Metal oxides are considered to be potentially interesting candidates in this respect because the range of electronic and optical properties they support is truly exceptional.² Over the past ten years and more, ZnO have been deeply studied and widely used in various fields, such as light emitting diodes, photocatalysts, gas sensors, solar cells and field emission devices because it is photostable, non-toxic, low-cost and various morphologies are readily available.^3,4 However, a major drawback of ZnO is it can only absorb a small portion (∼5%) of the solar spectrum in the ultraviolet region because of its wide band gap of ∼3.2 eV (λ = 380 nm).^5,6 For extending the absorption range to the visible light region, various approaches have been utilized, including doping transition-metal ions into ZnO or sensitizing ZnO with a dye or another low band gap semiconductor.^7,8 In the sensitization system, ZnO acts as an electron transporter for its high electron mobility and the dye or low band gap semiconductor acts as a light harvester for their wide photo-absorption region. When compared with other methods, sensitization systems have received more interest because of their synergetic effects, which not only extend the absorption range to the visible light region but also decrease the probability of charge recombination under irradiation. To date, several semiconductors have been used as sensitizers for widening the absorption range, such as CdSe, CdS, PbS, WO₃ and Cu₂O.^9,10 Among these materials, Cu₂O is of interest on account of its many advantages, including a narrow band gap of approximately 2.1 eV, non-toxicity, low cost and the abundance of its starting material, i.e. copper. Hence, a Cu₂O/ZnO heterostructure should be a promising material structure, which will not only lead to a functional integration of the properties of both Cu₂O and ZnO but also to novel interface effects and phenomena.¹¹

To date many photovoltaic devices based on a Cu₂O/ZnO hetero-junction have been reported.^12–14 However, to synthesize these devices often needs expensive equipment (e.g. high temperature equipment, vacuum equipment and laser equipment) an aqueous method, such as hydrothermal growth, electrodeposition and chemical bath deposition, seems as a promising method for preparing low cost solar cells because it is simple, inexpensive and does not require complicated equipment. Chemical bath deposition (CBD) has been successfully employed to deposit quantum dots of metal chalcogenides, metal oxides and metal sulfides on ZnO or TiO₂ nanostructures.^15–17 Moreover, control can be exercised on the morphology of the deposits by adjusting the precursor concentration, growth temperature and time or using templates and surfactants.^18–20 For example, high density perpendicularly orientated ZnO rod arrays have been obtained using nanostructured thin films as seeding layers.²¹ Cu₂O nanorod thin films have been successfully grown on FTO glass substrate using the CBD technique with the assistance of CTAB.²²

Herein, we report the synthesis of hetero-nanostructured films comprising of ZnO nanorod arrays (ZNAs) decorated with Cu₂O QDs on indium-doped tin oxide (ITO) glass substrates using a hydrothermal method followed by a CBD process. Moreover, a CuO passive layer was fabricated on the Cu₂O QDs as a protective layer by heat treatment. The detailed synthetic process and characterization of the films are reported. The effects of the amount of Cu₂O QDs and CuO layer on the photoelectrochemical properties of the resultant cells were investigated. Through extensive experimentation, we found that the Cu₂O QDs could be an effective visible light sensitizer for ZNAs, which demonstrates that Cu₂O/ZnO hetero-nanorod arrays (Cu₂O QDs decorated ZNAs) can be used as a promising photoelectrode. In addition, the thin layer of CuO on Cu₂O QDs can be used as an effective protecting layer inhibiting the degradation of Cu₂O from occurring in the photovoltaic process.

2. Experimental details

2.1. Preparation of the Cu₂O/ZnO hetero-nanorod arrays

First, vertically aligned pure ZNAs were prepared on transparent ITO/glass substrates (15 Ω square⁻¹) using a facile hydrothermal method reported previously.²³ In brief, a ZnO seed layer (ca. 10 nm) coat was prepared on the ITO glass substrate (ca. 15 mm × 25 mm) by a spin coating method using a zinc oxide colloid (4.15 g zinc acetate dihydrate and 1.16 g monoethanolamine in 50 mL ethanol). Then, the substrate with the seed coat was vertically immersed in a 40 mL equimolar aqueous solution of zinc nitrate and hexamethylenetetramine at a concentration of 0.10 M in a Teflon kettle and heated on a hot plate at 148 °C for 3 h.

Then, the Cu₂O QDs were deposited on the ZNAs using a CBD process. The typical CBD process involved dipping the ZNAs into a colorless mixed aqueous solution of 20 mL of 1.0 M CuSO₄ and 80 mL of 1.0 M Na₂S₂O₃ (Solution I) for 2 s at room temperature, rinsing it with DI water, and then dipping it into a 0.5 M NaOH aqueous solution (Solution II) for another 2 s at 70 °C and rinsing it again with DI water. The two-step dipping procedure is considered as one CBD cycle. The incorporated amount of Cu₂O can be increased by repeating the CBD cycle.

2.2. Characterization

The morphology of the heterostructure nanorod arrays was characterized by scanning electron microscopy (SEM, Supra 55, Zeiss, Oberkochen, Germany) and transmission electron microscopy (TEM, Tecnai F30 G2, FEI, Hillsboro, OR, USA). X-ray diffractometer (XRD, X'Pert MPD, Philips, Amsterdam, The Netherlands), energy-dispersive X-ray spectrometer (EDS, INCA Xstream-2, Oxford, Abingdon, Oxfordshire, UK) attached to the SEM and X-ray photoelectron spectrometer (XPS, K-Alpha, Thermo Scientific, Atlanta, GA, USA) were used to characterize the phase and chemical composition of the heterostructure nanorod arrays. Ultraviolet-visible spectra of the heterostructure films were recorded using a UV-vis spectrophotometer (UV-3150, Shimadzu, Tokyo, Japan).

2.3. Photovoltaic performance and EIS measurements

Sandwich cells were assembled for investigating the photovoltaic performance of the heterostructure nanorod arrays. The Pt counter-electrode was prepared by dripping a drop of 5 mM H₂PtCl₆ isopropyl alcohol solution onto an ITO glass substrate (ca. 15 mm × 25 mm), followed by heating at 385 °C for 30 min. The heterostructure nanorod arrays were used as the photoanode, and the counter-electrode was cohered together by a Surlyn film. The space between the two electrodes was controlled at ∼3 mm. The electrolyte comprised of 0.2 M Na₂SO₄ and 0.1 M NaCH₃COO (pH 7) according to a previous report.²⁴ A Keithley 2410 source meter was used to measure the photocurrent at a scan rate of 10 mV s⁻¹. Sunlight was simulated with a 500 W xenon lamp (Spectra Physics). The light intensity was adjusted to 1 sun condition (AM 1.5G, 100 mW cm⁻²) using a NREL-certified Si reference cell. The effective area of the cell was 0.5 cm². The incident photon-to-electron conversion efficiency (IPCE) under monochromated light (350–800 nm) was measured using an IPCE system (QTest Station 1000, Crowntech, Macungie, PA, USA). A 150 W tungsten halogen lamp was used as the light source to generate a monochromatic beam. A silicon solar cell was used as the standard during calibration. The electrochemical impedance spectroscopy (EIS) measurements were carried out on an electrochemical workstation (CHI600E, CH Instruments, Shanghai, China) under the same simulated sunlight illumination. The impedance spectra were recorded with the help of CHI 13.07 software under an AC perturbation signal of 10 mV over the frequency range of 1 MHz to 0.01 Hz at 0.3 V in the dark.

3. Results and discussion

3.1. SEM and EDS analysis

Fig. 1a shows the SEM image of pure ZNAs grown on an ITO/glass substrate using the hydrothermal method. It can be appreciated that all of the nanorods show hexagonal-rod morphology with smooth surface and planar top ends, which results from their wurtzite structure.²³ The majority of nanorods are vertically oriented to the substrate plane. Moreover, the axial direction was aligned with the c-axis of the hexagonal ZnO crystal structure.⁸ The nanorods have a mean diameter of 207 nm, which was obtained by measuring 30 individual nanorods and has a standard deviation of 66. Fig. 1b–e shows the SEM images of the heterostructure nanorod arrays obtained at different CBD cycles. The insets in Fig. 1b–e are the corresponding larger magnification images of the different heterostructure nanorod arrays. It can be clearly seen that the sides of the ZnO nanorods were rather rough and possess uniform nanoparticles. Moreover, we can estimate that the particle size of the Cu₂O quantum dots obtained at different CBD cycles was 11 nm (2 cycles), 28 nm (5 cycles), 55 nm (10 cycles) and 120 nm (20 cycles), respectively. By comparing the SEM images of Fig. 1b–e, it can be found that as the number of CBD cycles increases, the amount of nanoparticles increases and the nanorod shape becomes slightly round from hexagonal. Fig. 1f shows the typical energy-dispersive spectrum (EDS) taken on the nanoparticles. Only the elements O, Zn and Cu can be observed, which indicates that these nanoparticles could be copper oxide. The atomic ratio of Cu, Zn and O was estimated to be 7.02%, 49.39% and 43.59%, respectively. This information is not enough to conclude that the nanoparticles are comprised of only Cu₂O. It may also contain some CuO. Thus, a further characterization was required.

Fig. 1 SEM image of pure ZNAs (a) and the Cu₂O/ZnO hetero-nanorod arrays obtained at different CBD cycles: 2 cycles (b), 5 cycles (c), 10 cycles (d) and 20 cycles (e). Insets in (b), (c), (d) and (e) are the corresponding larger magnification images. The typical EDS spectrum of the Cu₂O/ZnO hetero-nanorod arrays (f).

3.2. Phase and chemical composition

The phase and chemical composition of the heterostructure nanorod arrays were further examined using X-ray diffraction and energy dispersive X-ray spectrometry. Fig. 2A shows the X-ray diffractogram of the heterostructure nanorod arrays obtained at different CBD cycles. For the pure ZNAs grown on ITO substrates (0 cycles), all the diffraction peaks agree well with the hexagonal zincite phase (JCPDS 36-1451). The enhanced (0 0 2) diffraction peak at 2θ of 34.42 also indicates that the as-prepared ZNAs are oriented with respect to the substrate. For the heterostructure nanorod arrays obtained at CBD cycles of 10 and 20 (d and e patterns in Fig. 2A), in addition to the ZnO diffraction peaks, the Cu₂O (111), (200), (220) and (311) peaks emerged at 2θ of 36.50°, 42.38°, 61.50° and 73.69°, respectively, which agree well with the crystal planes of cubic Cu₂O (JCPDS 65-3288). The Cu₂O (111) peak (2θ = 36.50°) was very close to the ZnO (101) peak (2θ = 36.25°) and they are overlapped in the pattern. When compared with the ZnO peaks, the Cu₂O peaks are very weak, which reveals that the Cu₂O deposited onto ZnO were of very small size or amount.²⁵ For the heterostructure nanorod arrays obtained at CBD cycles of 2 and 5 (b and c patterns in Fig. 2A), no evident Cu₂O diffraction peaks were observed. There may be two reasons for the absence of the Cu₂O peaks, the first reason may be that only few QDs were attached to the ZnO nanorods, the second reason may be that the Cu₂O QDs are highly dispersed on the ZnO surface.¹⁷

Fig. 2 (A) X-ray diffractogram of the Cu₂O/ZnO hetero-nanorod arrays with different CBD cycles: (a) 0 cycles, (b) 2 cycles, (c) 5 cycles, (d) 10 cycles and (e) 20 cycles. (B) XPS survey spectrum of the Cu₂O/ZnO hetero-nanorod arrays. (C) The high resolution XPS Cu 2p scan spectrum of the Cu₂O/ZnO hetero-nanorod arrays.

Taken together, from Fig. 2A we can see that the intensities of the Cu₂O characteristic peaks increase with an increase in the number of CBD cycles, suggesting the increased amount of the Cu₂O QDs.

Then, X-ray photoelectron spectroscopy (XPS) was employed to confirm the presence of the heterostructure nanorod arrays (Fig. 2B and C). The entire survey of the heterostructure nanorod arrays surface presents all the elements detected, as shown in Fig. 2B. Because only C (from the atmosphere), O, Zn and Cu were detected, it suggests that only copper oxide or zinc oxide was formed on the surface, this agrees well with the EDS results. In the Cu 2p spectrum (Fig. 2C), there are only two peaks located at 932.3 eV (Cu 2p_3/2) and 952.1 eV (Cu 2p_1/2) for the heterostructure nanorod arrays, indicating the existence of Cu⁺.²⁵ Furthermore, the Cu 2p_1/2 and Cu 2p_3/2 satellite peaks for Cu²⁺ were not observed, which provides powerful evidence for the successful coating of only Cu₂O on the surface of ZnO nanorods.²⁶

3.3. TEM images and Cu₂O growth mechanism

The bright-field transmission electron microscopy (TEM) image in Fig. 3a is a representative ZnO nanorod decorated with an ensemble of Cu₂O QDs. The grey spots on the surface of the nanorod should be Cu₂O QDs, the size is about 5–7 nm in diameter. These tiny Cu₂O QDs can aggregate to form larger nanoparticles as can be seen in the SEM images. Fig. 3b shows a high-resolution transmission electron microscopy (HRTEM) image of the circle area in Fig. 3a. The larger crystallite appearing in the left region of the image is identified to be ZnO. The lattice spacing measured for this crystalline plane is 5.21 Å corresponding to the (001) plane of hexagonal ZnO and the growth direction of the nanorods was dominantly (001). The attached QDs appear as randomly oriented crossed fringe patterns on the edge of the nanorods. In addition, the observed lattice fringe of 2.45 Å corresponds to the (111) crystal plane of the cubic phase of Cu₂O (JCPDS 65-3288).

Fig. 3 Bright-field TEM (a) and HRTEM (b) images of the representative Cu₂O/ZnO hetero-nanorod.

The formation of Cu₂O QDs on the ZnO nanorods via CBD can be explained by following a series of reactions.^22,27,28

2Cu²⁺ + 4S₂O₃²⁻ → 2[Cu(S₂O₃)]⁻ + [S₄O₆]²⁻ (1)

[Cu(S₂O₃)]⁻ ↔ Cu⁺ + S₂O₃²⁻ (2)

2Cu⁺ + 2OH⁻ → Cu₂O + H₂O (3)

During the CBD process, in the colorless Solution I, the copper thiosulfate complex can be formed via eqn (1), in which CuSO₄ and Na₂S₂O₃ were introduced as a precursor for Cu₂O and the reducing agent, respectively. Cu⁺ cations were formed by the dissociation equilibrium (eqn (2)). When the as-drawn ZNAs from Solution I were immersed in Solution II (hot NaOH solution), the adherent Cu⁺ cations react with OH⁻ anions to form Cu₂O QDs on the surface of the ZNAs and the chemical reaction in eqn (3) occurs. As the number of CBD cycles increases, the amount of Cu₂O QDs increases gradually.

3.4. UV-vis absorption spectroscopy

Fig. 4 shows the UV-vis absorption spectra of ZNAs and the Cu₂O/ZnO hetero-nanorod arrays with different numbers of CBD cycles. The maximum absorbance peak of the ZNAs occurs at around 380 nm and almost has no absorbance for visible-light due to its large energy gap (∼3.2 eV). When compared with the ZNAs, the Cu₂O/ZnO hetero-nanorod arrays show a significant red-shift in the peak maxima at 550 nm and exhibit broad absorption bands from 200 to 550 nm, indicating the effective photo-absorption ability for the heterostructure. In addition, the absorbance of the spectra increases as the number of CBD cycles increased. Furthermore, the absorption edge shifts slightly to red with an increase in the number of CBD cycles, indicating the growth of the Cu₂O particles. The band gap for the Cu₂O bulk was 2.2 eV at room temperature. However, when the particle size was reduced to a quantum-size, the number of atoms will significantly decrease resulting in a larger interval in the electron energy levels. Hence, the electrons no longer possess continuous energy bands. The electronic energy level will change from a quasi-continuous phase to a split phase and the band gap will become wider, leading to a blue shift in the absorption.^17,29 Therefore, it is easy to understand the slight red shift of the absorption edge when the number of CBD cycles was increased.

Fig. 4 UV-vis absorption spectra of pure ZNAs and the Cu₂O/ZnO hetero-nanorod arrays with different numbers of CBD cycles.

3.5. Photovoltaic performance

Fig. 5 shows the characteristics of the photocurrent density versus applied voltage (J–V curve) for the Cu₂O/ZnO heterostructure nanorod arrays based solar cells under simulated sunlight. Table 1 lists the photovoltaic parameters obtained from the J–V curves, including short circuit current (J_sc), open circuit potential (V_oc), fill factor (FF) and the total power conversion efficiency (η). For the bare ZNAs photoelectrode, the photovoltaic performance can almost be neglected due to the infinitesimally small J_sc (only 0.20 mA cm⁻²). For the Cu₂O/ZnO heterostructure nanorod arrays with CBD 10 cycles electrode, J_sc was about 3.21 mA cm⁻², which is more than 16 times higher than that of the bare ZNAs photoelectrode. The optimum power conversion efficiency was 1.17% for the Cu₂O/ZnO heterostructure photoelectrodes. The comparison between the photovoltaic behavior of bare ZNAs and the Cu₂O/ZnO heterostructure photoelectrodes confirms the superior performance for producing higher photocurrents and power conversion efficiencies for the latter over those of the former. The enhanced photovoltaic properties can be ascribed to the following probable reasons: First, the extension of absorption spectrum into the visible region by the Cu₂O/ZnO heterostructure photoelectrodes. When compared with bare ZNAs photoelectrode, these photoelectrodes have an intense absorption in the visible region, which significantly increased the utilization rate of solar energy. Second, the reduction in the recombination rate of photo-induced electron–hole pairs is because of the heterojunction structure. As shown in Fig. 6, the conduction band position of Cu₂O was higher than that of ZnO, once incident photons are absorbed by the Cu₂O QDs, photo-generated electrons in the conduction band of Cu₂O will quickly transfer into that of the ZnO to decrease its energy level. In this manner, photo-generated electron–hole pairs were separated quickly and transported in their respective phases to the opposing electrodes, leading to enhance the photovoltaic efficiency.

Fig. 5 Photocurrent density versus applied voltage curves for the solar cells based on the Cu₂O/ZnO hetero-nanorod arrays with different numbers of CBD cycles.

Table 1 Parameters obtained from the J–V curves for the solar cells based on the different photoelectrodes

Electrode	Jsc (mA cm^−2)	Voc (V)	FF (%)	η (%)
0 cycle	0.20	0.29	30.67	0.01
2 cycles	1.43	0.50	33.36	0.24
5 cycles	2.11	0.58	45.62	0.56
10 cycles	3.21	0.65	56.46	1.17
20 cycles	2.93	0.61	42.79	0.77

---

Fig. 6 Energy band alignment diagram of the ZnO/Cu₂O/CuO heterojunction.

Furthermore, it can be easily found that for a certain applied potential, all the photovoltaic parameters increase sharply as CBD cycles in the early stage indicating that a higher incorporated amount of Cu₂O can induce a higher power output. However, the photovoltaic parameters were found to slightly decrease when the CBD cycles were increased to 20 cycles. This phenomenon may be attributed to the fact that more CBD cycles would cause conglomeration and growth of the Cu₂O crystal nucleus, and the oversized Cu₂O particles will lose the dominance as QDs (large QD extinction coefficients and generating multiple electron–hole pairs); moreover, excess Cu₂O nanoparticles can act as a potential barrier for charge transfer.¹⁷ Fig. 7 shows the IPCE values of the different Cu₂O/ZnO heterostructure nanorod arrays based solar cells as a function of wavelength. The IPCE curve indicates the light response of the photovoltaic device, which is directly related to the photocurrent density. The IPCE presents a similar threshold to UV-vis absorbance. The quantum efficiencies of the different Cu₂O/ZnO heterostructure photoanodes vary in the range of 400–700 nm corresponding to the effect of the Cu₂O QDs CBD cycles. The results are in accordance with the J–V curves.

Fig. 7 IPCE curves for the solar cells based on the Cu₂O/ZnO hetero-nanorod arrays with different numbers of CBD cycles.

It was noted that the power conversion efficiencies measured for the Cu₂O/ZnO QDs sensitized solar cell here were less than the highest value for a solid state heterojunction solar cell previously reported in the literature (5.38%).¹⁴ This can be attributed, at least in part, to the high density of interface states at the Cu₂O/ZnO heterojunction, resulting from the aqueous synthesis, which facilitate unwanted charge recombination.³⁰ Another reason presumably is due to the inappropriate electrolyte.

3.6. EIS spectra

To further understand the effect of the Cu₂O QDs on the charge transfer kinetics and internal resistance of the Cu₂O QDs sensitized solar cells, EIS measurements were conducted and the results are shown in Fig. 8. As can be seen from the Nyquist plots, all the plots show two partially overlapped semicircles in the high-to-medium frequency region and a large arc in the low frequency region. The corresponding equivalent circuit for this cell system is depicted in the inset (I) of Fig. 8.³¹ The first small semicircle in the high-frequency range can be assigned to the charge transfer resistance (R_CE) and double-layer capacitance (C_CE) at the CE/electrolyte interface.^32,33 The second large semicircle in the medium frequency region was due to the charge transfer resistance of the recombination process (R_r) and the interfacial capacitance (C_μ) at the ZnO/Cu₂O/electrolyte interface.³³ The right larger arc appearing in the low frequency region was attributed to the Warburg impedance (Z_d) of the redox couple in the electrolyte.³⁴ The R_r, C_μ and calculated electron lifetime (τ_n) of the different cells determined from EIS analysis fitted results are listed in Table 2. We can find that the R_r of the different cells increases as the number of CBD cycles increases.

Fig. 8 Nyquist plots for the cells based on the Cu₂O/ZnO hetero-nanorod arrays with different numbers of CBD cycles. The insets are the equivalent circuit used to fit the Nyquist plots (I) and the magnified plots of the first semicircle (II).

Table 2 Parameters obtained from the fitting results of the electrochemical impedance spectra

Electrode	Rr (kΩ)	Cμ (μF)	τn (ms)
2 cycles	82.1	1.1	90.3
5 cycles	103.9	1.5	155.9
10 cycles	203.8	3.2	652.2
20 cycles	214.5	1.8	386.1

---

The larger value of R_r indicates the retarded backward reaction of injected electron transfer at the ZnO/Cu₂O/electrolyte interface, i.e., reduced interfacial recombination resulting in high efficiency.³⁵ As a poor loading of QDs would cause the direct exposure of ZnO to the electrolyte, which might be responsible for the comparative low R_r for the photoanode with less CBD cycles. The C_μ was related to the density of states and the surface recombination sites. An increase in the capacitance could indicate passivation of the surface recombination sites.³⁶ Therefore, larger C_μ indicates that a greater number of photo-exited electrons can be transferred into the conduction band of ZnO indicating the reduction of charge recombination. As shown in Table 2, the value of C_μ increases as the number of CBD cycles increases at the beginning, and then decreases and the photoanode with 10 CBD cycles shows the largest C_μ value. This variation tendency in C_μ was consistent with the η results presented in Table 1. The electron lifetime (τ_n) was another parameter that can reflect the charge recombination rate and the electron-transfer rate, and it can be calculated from τ_n = R_rC_μ.^16,37 It can be seen that the photoanode with 10 CBD cycles has the largest electron lifetime, which again indicates a reduced charge-transfer rate. Despite the low charge-transfer resistance of the photoanode with 10 CBD cycles compared with that of the photoanode with 20 CBD cycles, the electron lifetime of the photoanode with 10 CBD cycles was larger than that of the photoanode with 20 CBD cycles, and this result is in good agreement with the larger C_μ value. For the photoanode with 20 CBD cycles, the abundant Cu₂O QDs covered the total surface of the ZnO nanorods and some aggregated on the surface of the ZnO nanorods due to greater number of CBD cycles. The larger aggregates can result in a longer transport path length for the photo-induced electron–hole pairs in the Cu₂O sensitizer before being separated and collected, and therefore act as a potential barrier for charge transfer. Thus, the charge recombination process in Cu₂O was enhanced.^37,38

3.7. The effect of the CuO protective layer

According to the previous studies, we know that the stability of Cu₂O could be enhanced due to the fast charge transfer and blocking of anodic currents when paired with other oxide semiconductors such as ZnO or TiO₂.^8,39 When the heterostructure photoanode was exposed to weak monochromated light illumination, no significant dissolution or phase transformation of Cu₂O happens. However, under intense illumination, Cu₂O was found to be no longer stable, even when it paired with other oxide semiconductors.²⁴ As shown in Fig. 9, the η of the cell based on the as-prepared Cu₂O/ZnO heterostructure with 10 CBD cycles was decreased as the illumination time prolonged. It can be found that the η was decreased by as much as 36% after 40 min of illumination from 1.17% to 0.75%. Hence, there is a tremendous requirement to improve the photochemical stability of the Cu₂O/ZnO heterostructure electrode.

Fig. 9 Power conversion efficiencies (η) of the as-prepared and heat-treated photoelectrodes at different illumination times.

Thimsen and co-workers reported that the Cu₂O photoelectrode could be highly stabilized by depositing a protective layer of Al/ZnO/TiO₂ on the top of Cu₂O using atomic layer deposition (ALD).⁴⁰ The protective oxide layer can prevent the Cu₂O electrode from being in contact with the electrolyte without changing the energy band position. CuO is a relatively stable phase of copper oxide, and therefore it could be used as an effective protection layer to minimize the degradation of Cu₂O under the illumination conditions. Wang's group reported a two-step electrodeposition-anodization process to prepare a CuO nanowire mounted Cu₂O composite photoelectrode with improved stability.⁴¹ The presence of CuO can improve the charge separation of Cu₂O, which suppresses the redox reactions of Cu₂O.⁴²

Here, as for our Cu₂O/ZnO heterostructure photoelectrode, CuO coating layer has been fabricated on the Cu₂O QDs by a simple heat treatment process in ambient air at 100 °C for 2 h. For further characterization of the heat-treated Cu₂O/ZnO photoelectrode, XRD, TEM and XPS analyses were carried out. The XRD and TEM results show no difference between the as-prepared and heat-treated Cu₂O/ZnO heterostructure nanorod arrays photoelectrodes. This can be ascribed to the small amount of CuO formed on the surface of the Cu₂O QDs. XPS is a powerful technique used to study the surface elemental composition, chemical state and electronic state of the elements. Fig. 10 shows the Cu 2p XPS depth profile spectra of the heat-treated Cu₂O/ZnO hetero-nanorod arrays with 10 CBD cycles. A clear difference in the shapes can be observed for the surface and underneath scan spectra, being in agreement with those previously reported and discussed in the literature for Cu⁺ and Cu²⁺ species.^26,43 For the Cu 2p spectrum obtained after an argon ion sputter etching process for 10 s (below spectrum in Fig. 10), it is well consistent with the spectrum in Fig. 2C, only the binding energy peaks related to Cu 2p_3/2 and Cu 2p_1/2 of Cu₂O emerged. Thereby, it indicates that the core of the QDs was still composed of Cu₂O. In the Cu 2p surface scan spectrum obtained before etching, there are two main peaks located at around 934.4 and 954.0 eV, which are assigned to the binding energy of Cu 2p_3/2 and Cu 2p_1/2, respectively. Moreover, we can find that the Cu 2p_1/2 and 2p_3/2 peaks are relatively broad, which seems to be influenced by the presence of other copper species. For acquiring a more precise knowledge about this, we fitted the peaks using a Gaussian/Lorentzian mixed function. From the peak-fit for Cu 2p, we find two other small peaks (1 and 5) with binding energy of 932.5 and 952.8 eV, which are assigned to Cu 2p_3/2 and Cu 2p_1/2 of Cu₂O, respectively.^41,43 Moreover, there are a series of extra shake-up satellite peaks observed on a higher binding energy side, 940.6 (peak 3), 943.3 (peak 4) and 962.2 eV (peak 7) for Cu 2p_3/2 and Cu 2p_1/2, respectively, which can be ascribed to an unfilled Cu 3 d⁹ shell and indicates the presence of CuO on the surface.^42,43

Fig. 10 Cu 2p XPS depth profile spectra of the heat-treated Cu₂O/ZnO hetero-nanorod with 10 CBD cycles.

Based on the above-described results that show the strong XPS peaks for Cu²⁺ and relatively low peaks for Cu⁺, it indicates a protection layer of CuO formed on the top of Cu₂O. Because the sputtering time is very short, the CuO layer should be very thin. Although the layer of CuO was thin, the photoelectrochemical stability can be improved significantly.

As shown in Fig. 9, the η of the cell based on the heat-treated Cu₂O/ZnO hetero-nanorod arrays with 10 CBD cycles was decreased from 1.10% to 0.99% after 40 min of illumination; it only decreases by 10%. The photoelectrochemical stability was increased obviously under the same conditions. This result indicates that CuO can act as an effective protecting layer for the Cu₂O/ZnO photoelectrodes enhancing the photoelectrochemical stability of the assembled solar cells. When compared with the as-prepared Cu₂O/ZnO with 10 CBD cycles, the initial η was slightly decreased. As shown in Fig. 6, the conduction band of CuO was lower than that of Cu₂O. The photo-generated electrons in the conduction band of Cu₂O were able to transfer into the conduction band of CuO to decrease its energy level. This goes against the electron transfer to ZnO and increases the charge recombination, leading to a decreased photovoltaic efficiency.

4. Conclusions

Cu₂O quantum dots can be successfully prepared on ZNAs forming a Cu₂O/ZnO heterostructure system using a chemical bath deposition technique. It is a promising model system for photovoltaic, photoelectrochemical and photocatalytic applications because it comprises of relatively inexpensive materials and can be fabricated at low energy intensity. In addition, it is capable of using a significant fraction of the solar spectrum. The Cu₂O quantum dots have an effect on the photovoltaic performance. Moreover, it is shown that heat treatment promotes the photovoltaic stability of Cu₂O/ZnO under simulated sunlight illumination due to the CuO protective layer formed at the surface of Cu₂O inhibiting the corrosion process.

Acknowledgements

The work was supported by the National Natural Science Foundation (51172187), the SPDRF (20116102130002), the 111 Program (B08040) of MOE, the Xi'an Science and Technology Foundation (CX12174, XBCL-01-08), the Shaanxi Province Science Foundation (2013KW12-02), the SKLP Foundation (KP201421) and the Fundamental Research Funds for the Central Universities (3102014JGY01004) of China.

Notes and references

1. A. S. Zoolfakar, R. A. Rani, A. J. Morfa, S. Balendhran, A. P. O'Mullane, S. Zhuiykov and K. Kalantar-zadeh, J. Mater. Chem., 2012, 22, 21767.
2. M. Deo, S. Mujawar, O. Game, A. Yengantiwar, A. Banpurkar, S. Kulkarni, J. Jog and S. Ogale, Nanoscale, 2011, 3, 4706.
3. Y. Wang, S. Li, H. Shi and K. Yu, Nanoscale, 2012, 4, 7817.
4. J.-X. Sun, Y.-P. Yuan, L.-G. Qiu, X. Jiang, A.-J. Xie, Y.-H. Shen and J.-F. Zhu, Dalton Trans., 2012, 6756.
5. X. Zhang, J. Qin, Y. Xue, P. Yu, B. Zhang, L. Wang and R. Liu, Sci. Rep., 2014, 4, 4596.
6. S. S. Mali, H. Kim, P. S. Patil and C. K. Hong, Dalton Trans., 2013, 16961.
7. Y. Wang, G. She, H. Xu, Y. Liu, L. Mu and W. Shi, Mater. Lett., 2012, 67, 110.
8. X. W. Zou, H. Q. Fan, Y. M. Tian and S. J. Yan, CrystEngComm, 2014, 16, 1149.
9. W. Jia, H. L. Dong, J. F. Zhao, S. H. Dang, Z. X. Zhang, T. B. Li, X. G. Liu and B. S. Xu, Appl. Phys. A: Mater. Sci. Process., 2012, 109, 751.
10. S. Balachandran and M. Swaminathan, Dalton Trans., 2013, 5338.
11. M. Deo, D. Shinde, A. Yengantiwar, J. Jog, B. Hannoyer, X. Sauvage, M. More and S. Ogale, J. Mater. Chem., 2012, 22, 17055.
12. T. Minami, Y. Nishi, T. Miyata and J.-i. Nomoto, Appl. Phys. Express, 2011, 4, 062301.
13. Y. Nishi, T. Miyata and T. Minami, J. Vac. Sci. Technol., A, 2012, 30, 04D103.
14. T. Minami, Y. Nishi and T. Miyata, Appl. Phys. Express, 2013, 6, 044101.
15. C.-H. Chang and Y.-L. Lee, Appl. Phys. Lett., 2007, 91, 053503.
16. Y. Choi, M. Seol, W. Kim and K. Yong, J. Phys. Chem. C, 2014, 118, 5664.
17. H. Chen, W. Y. Fu, H. B. Yang, P. Sun, Y. Y. Zhang, L. R. Wang, W. Y. Zhao, X. M. Zhou, H. Zhao, Q. A. Jing, X. F. Qi and Y. X. Li, Electrochim. Acta, 2010, 56, 919.
18. A. M. Lockett, P. J. Thomas and P. O'Brien, J. Phys. Chem. C, 2012, 116, 8089.
19. G. Mu, R. V. Gudavarthy, E. A. Kulp and J. A. Switzer, Chem. Mater., 2009, 21, 3960.
20. Q. Li, J. Bian, J. Sun, J. Wang, Y. Luo, K. Sun and D. Yu, Appl. Surf. Sci., 2010, 256, 1698.
21. D. S. Boyle, K. Govender and P. O'Brien, Chem. Commun., 2002, 80.
22. A. A. Aref, L. Xiong, N. Yan, A. M. Abdulkarem and Y. Yu, Mater. Chem. Phys., 2011, 127, 433.
23. X. W. Zou, H. Q. Fan, Y. M. Tian and S. J. Yan, Mater. Lett., 2013, 107, 269.
24. L.-k. Tsui and G. Zangari, Electrochim. Acta, 2014, 128, 341.
25. S. Zhang, S. Zhang, F. Peng, H. Zhang, H. Liu and H. Zhao, Electrochem. Commun., 2011, 13, 861.
26. J. Morales, L. Sánchez, F. Martín, J. R. Ramos-Barrado and M. Sánchez, Thin Solid Films, 2005, 474, 133.
27. Q. Pan, M. Wang and Z. Wang, Electrochem. Solid-State Lett., 2009, 12, A50.
28. L. Xiong, H. Yu, G. Yang, M. Qiu, J. Chen and Y. Yu, Thin Solid Films, 2010, 518, 6738.
29. W. W. Yu and X. G. Peng, Angew. Chem., Int. Ed., 2002, 41, 2368.
30. K. P. Musselman, A. Marin, A. Wisnet, C. Scheu, J. L. MacManus-Driscoll and L. Schmidt-Mende, Adv. Funct. Mater., 2011, 21, 573.
31. F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo and A. Hagfeldt, Sol. Energy Mater. Sol. Cells, 2005, 87, 117.
32. B.-L. He, B. Dong and H.-L. Li, Electrochem. Commun., 2007, 9, 425.
33. J. W. Ondersma and T. W. Hamann, J. Phys. Chem. C, 2010, 114, 638.
34. H.-J. Kim, D.-J. Kim, S. S. Rao, A. D. Savariraj, K. Soo-Kyoung, M.-K. Son, C. V. V. M. Gopi and K. Prabakar, Electrochim. Acta, 2014, 127, 427.
35. R. Zhou, Q. Zhang, E. Uchaker, L. Yang, N. Yin, Y. Chen, M. Yin and G. Cao, Electrochim. Acta, 2014, 135, 284.
36. I. Zarazua, T. Lopez-Luke, J. Reyes-Gomez, A. Torres-Castro, J. Z. Zhang and E. De la Rosa, J. Electrochem. Soc., 2013, 161, H68.
37. I. Mora-Sero, S. Gimenez, F. Fabregat-Santiago, R. Gomez, Q. Shen, T. Toyoda and J. Bisquert, Acc. Chem. Res., 2009, 42, 1848.
38. G. Hodes, J. Phys. Chem. C, 2008, 112, 17778.
39. L. k. Tsui, L. Wu, N. Swami and G. Zangari, ECS Electrochem. Lett., 2012, 1, D15.
40. A. Paracchino, V. Laporte, K. Sivula, M. Gratzel and E. Thimsen, Nat. Mater., 2011, 10, 456.
41. Z. Zhang and P. Wang, J. Mater. Chem., 2012, 22, 2456.
42. P. Wang, Y. H. Ng and R. Amal, Nanoscale, 2013, 5, 2952.
43. M. Yin, C. K. Wu, Y. Lou, C. Burda, J. T. Koberstein, Y. Zhu and S. O'Brien, J. Am. Chem. Soc., 2005, 127, 9506.

---