CuO/ZnO Heterojunction Nanorod Arrays Prepared by Photochemical Method with Improved UV Detecting Performance

Abstract

CuO/ZnO heterojunction nanorod arrays were synthesized using a facile photochemical deposition strategy. The morphology of CuO was related to the concentration of Cu²⁺ in the Cu(NO₃)₂ solution, UV illumination time, and the air annealing temperature. A possible reaction mechanism was proposed. In the photochemical deposition process, the OH⁻ was generated in the vicinity of the ZnO nanorod arrays and reacted with Cu²⁺ and NO₃⁻ in the solution to form Cu₂(NO₃)(OH)₃/ZnO heterojunction nanorod arrays firstly, which were converted into CuO/ZnO heterojunction nanorod arrays completely after air annealing at a low temperature. The fabricated CuO/ZnO heterojunction nanorod arrays exhibits a well-defined rectifying characteristic and an improved photo-response performance compared with pure ZnO nanorod arrays.

1. Introduction

ZnO is an attractive oxide semiconductor because of its tunable bandgap, stability at high temperature, and ability to be operated in hostile environments, which makes it one of the most important materials for UV photodetectors [1,2,3,4,5]. Moreover, ZnO nanorod arrays (NRs) are especially attractive due to their excellent properties, such as large ratio of surface area to volume and outstanding charge carrier transport for ensuring high efficiency and sensitivity [6]. However, the UV photodetectors fabricated from pure ZnO NRs often have low photosensitivity due to the photo-generated carrier fast recombination, and a long response time due to the absorption of oxygen molecules [7,8,9].

To solve these problems and make the UV detecting properties of ZnO NRs further improved, many kinds of materials have been explored to modify the surface of ZnO NRs, such as polymers [10], noble metal nanoparticles [11], and oxide semiconductors [12,13,14]. Wang et al. [15] used Ag nanoparticles decorating ZnO NR surfaces to facilitate the separation of photogenerated carriers and improve the photo response performance of the UV photodetectors. Chen et al. [16] modified a 3-aminopropyltrimethoxysilane monolayer on the surface of the ZnO film to reduce the negative influences of adsorbed oxygen molecules and improve the performance of the ZnO UV sensor. Among various oxide semiconductor materials, CuO is a common metal oxide, which is an intrinsic p-type semiconductor and shows a small band gap (1.2 eV) [17]. So far, CuO has been substantially explored for various fields of applications, such as field emitters, photocatalysis, and gas sensing [18,19,20,21,22,23]. Moreover, CuO has the advantage of its large abundance in nature, its low cost, its inability to pollute, and its simple synthesis like ZnO [24]. Previously, CuO/ZnO heterojunction nanostructures have been extensively studied in photocatalytic reactions [25,26,27]. The CuO of CuO/ZnO heterojunctions were often prepared by radio-frequently (RF) magnetron sputtering techniques or electrochemical deposition (ECD) [28,29,30,31]. These methods often need high temperature, high cost, and a long time. Thus, it is of great significance to develop a facile method to prepare a CuO/ZnO heterojunction, which has a low cost, simple process, is environmentally friendly and has a high product yield. The photochemical deposition method is easy, low cost, and its process is controllable. In a photochemical deposition process, the redox reactions of aqueous chemical species conduct on the surface of photocatalytic solids under light illumination and the target products are fabricated [14,15]. Taking these merits into account, the photochemical deposition method is considered to be an outstanding synthesis strategy for noble metal nanoparticles or semiconductor oxides on ZnO NRs. In addition, photodetectors based on CuO/ZnO heterostructures often show low photo-response performance, including low photosensitivity and a long response time. Morphology is an important factor that affects the photo-response performance of a CuO/ZnO based photodetector. Furthermore, the morphology of a CuO /ZnO heterostructure is often dominated by the precursor. In many previous works, CuO in a CuO/ZnO heterostructure fabricated from Cu₂(NO₃)(OH)₃ precursor has not been reported [28,29,30,31]. Therefore, taking the precursor into account, further explorations for improving the photo-response performance of the device is very important.

In this study, we report CuO/ZnO hierarchical nanostructures heterojunction NRs that consist of 1D ZnO nanorods (NRs) and flower-like CuO nanostructures, which were successfully synthesized by a facile photochemical deposition method and a following air annealing process. The impact of Cu²⁺ ion concentration, UV illumination time, and annealing temperature on the fabricated CuO/ZnO NRs were discussed in detail. All the devices based on CuO/ZnO heterojunction NRs display well-defined rectifying characteristics. Their UV photo-response performance was also measured. All the devices exhibit improved photosensitivity and a shorter reset time compared with the device based on pure ZnO NRs. Among them, CuO/ZnO NRs fabricated from a 50 mM Cu²⁺ ion concentration, 20 min UV illumination time, and 300 °C annealing time shows optimum UV detecting performance.

2. Experiment

Preparation of the CuO/ZnO heterojunction NRs: The preparation of CuO/ZnO heterojunction NRs included three steps: (1) synthesis of ZnO NRs, (2) preparation of Cu₂(NO₃)(OH)₃/ZnO heterojunction NRs precursor, and (3) preparation of CuO/ZnO heterojunction NRs through the air-annealing process of a Cu₂(NO₃)(OH)₃/ZnO heterojunction NR precursor.

Synthesis of ZnO NRs: The ZnO NRs were synthesized by a simple hydrothermal method [32]. A magnetron sputtering system was used to deposit an AZO (Al doped ZnO) seed layer (400 nm) on a glass substrate [33]. A Hall measurement shows that the AZO film has a sheet resistance of 3.4 × 10⁻³ Ω cm, suggesting the feasibility of AZO as a seed layer and transparent electrode. The AZO coated substrate was immersed into an equal mole of zinc acetate (Zn(Ac)₂·2H₂O) and hexamethylenetetramine (C₆H₁₂N₄, HMT) solution with a concentration of 0.02 M in a Teflon liner stainless-steel autoclave. Then, the autoclave was kept in a 90 °C oven for 12 h to grow ZnO NRs. After the growth process was finished, the resulting ZnO NRs were taken out and cleaned with deionized water and ethanol several times. The dried ZnO NRs showed a white appearance.

Preparation of Cu₂(NO₃)(OH)₃/ZnO heterojunction NR precursors: The Cu₂(NO₃)(OH)₃/ZnO precursor was deposited by a simple photochemical process. Copper nitrate (Cu(NO₃)₂·3H₂O) was added into deionized water to form a Cu(NO₃)₂ solution with a concentration ranging from 2 to 100 mM. Then, the ZnO NRs were immersed into the above solution and irradiated by UV light (365 nm, 2.0 mW/cm²). The reaction was conducted at room temperature and the reaction time ranged from minutes to 1 h. After the reaction finished, the samples were washed several times and dried in a 60 °C oven. The samples showed a blue appearance.

Preparation of CuO/ZnO heterojunction NRs: The CuO/ZnO heterojunction NRs were prepared by annealing the as-prepared blue Cu₂(NO₃)(OH)₃/ZnO precursors at 200, 300, and 400 °C in air for 2 h. The oven was heated with a ramping rate of 2 °C/min. The color of the final CuO/ZnO heterostructure NRs was black. Ag paste was used to prepare electrode.

Characterization: X-ray diffraction (XRD, a Panalytical X’PERT PRO diffractometer using Cu-Kα radiation, Almero, The Netherlands) was used to characterize the crystalline of the samples. The morphology was measured by scanning electron microscopy (SEM, JSM-6700F, Tokyo, Japan). Transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM) were measured by a JEOL JEM-ARF200F atomic resolution analytical microscope (JEOL, Tokyo, Japan) with an accelerating voltage of 200 kV. Thermogravimetry (TG), and Differential Scanning Calorimetry (DSC) (STA449F3, NETZSCH, Bavaria, Germany) were used to study the decomposition process of Cu₂(NO₃)(OH)₃. The photo-response performance was measured under 365 nm UV light illumination. The light intensity is 2.0 mW/cm², calibrated by a PM100A light power meter (Thorlabs, Shanghai, China) with a UV-enhanced Si detector. The UV light was from the CuO side. Keithley 2420C Source Meter (Keithley, Shanghai, China) was used to record the I-V characteristics curves. Keithley 2000 multimeter (Keithley, Shanghai, China) was used to measure the voltage of a 1 MΩ resistance in series with the photodetector. Because the resistance of the photodetector was large enough, so the current across the 1 MΩ resistance can be used to represent the photocurrent.

3. Results and Discussion

3.1. Characteristics of Materials

In order to determine the effect of Cu²⁺ ion concentration on the photochemical synthesis of the Cu₂(NO₃)(OH)₃/ZnO heterojunction NRs, we examined the morphology of the products prepared from solutions with various Cu²⁺ concentrations of 2, 5, 10, 50, and 100 mM. Figure 1a exhibits the surface and cross section morphologies of pure ZnO NRs synthesized through a hydrothermal reaction. It is obvious that ZnO NRs grow perpendicularly on the AZO coated substrate and have a smooth surface. The ZnO NRs are about 2.0 μm long and the diameter varies from 100 nm to 300 nm. Moreover, the ZnO NRs show a hexagonal cross section, which indicates the hexagonal wurtzite structure and preferential orientation in the c-axis direction. Figure 1b–f shows the surface and cross-section morphologies of the Cu₂(NO₃)(OH)₃/ZnO heterojunction NRs prepared in various Cu²⁺ ion concentration Cu(NO₃)₂ aqueous solutions with 20 min UV illumination. When the Cu²⁺ ion concentration is 2 mM, the surface of ZnO NRs are mostly covered by a thin layer of Cu₂(NO₃)(OH)₃ nanoparticles, which are suitable to act as a Cu₂(NO₃)(OH)₃ seed for further growth. Moreover, some immature flower-like Cu₂(NO₃)(OH)₃ nanostructures formed on the surface of ZnO NRs. These nanostructures are composed of nanosheet-shape building blocks, which have a smooth surface. By further increasing Cu²⁺ ion concentration, the sheet-like Cu₂(NO₃)(OH)₃ nanostructures deposited on the surface of ZnO NRs become more dense, which can be clearly seen in the SEM images. It is noteworthy that the Cu₂(NO₃)(OH)₃ eventually covers the whole surface of ZnO NRs as the Cu²⁺ ion concentration rises to 50 mM. When the Cu²⁺ ion concentration increases to 100 mM, the Cu₂(NO₃)(OH)₃ layers on the ZnO NRs exhibit a larger size than that of the Cu₂(NO₃)(OH)₃ deposited from the solution with 50 mM Cu²⁺ ion concentration.

Further, the effects of UV illumination time on the surface morphology of the Cu₂(NO₃)(OH)₃/ZnO precursor were investigated. Because the surface of the samples prepared from 50 mM Cu²⁺ concentration solution is fully covered by Cu₂(NO₃)(OH)₃ in a short time and the effect of UV illumination time cannot be clearly observed. Thus, we chose the sample prepared from a low Cu²⁺ concentration solution to study the effect of UV illumination time. The reactions are conducted at a fixed Cu²⁺ ion concentration of 2 mM and UV illumination times of 20 min and 40 min to 60 min. With increasing UV illumination time, the sheet-like Cu₂(NO₃)(OH)₃ nanostructures deposited on the ZnO NRs became more dense. The change can be clearly observed in the surface morphology images, as shown in Figure 2a–c. It is noteworthy that the ZnO NR surface can be fully covered with the Cu₂(NO₃)(OH)₃ layer as the UV illumination time rises to 60 min.

We also studied the dependence of the morphology of the Cu₂(NO₃)(OH)₃/ZnO NR nanostructures (50 mM Cu²⁺ ion concentration and 20 min UV illumination time) on annealing temperature. The Cu₂(NO₃)(OH)₃/ZnO NR precursor prepared with 50 mM Cu²⁺ concentration and 20 min UV illumination time was annealed at 200, 300, and 400 °C for 1 h in the air. There is no obvious change from the ZnO NRs after thermal annealing at 400 °C in the air atmosphere (Figure S1, ESI), which is consistent with Dai’s report [14]. The surface morphologies of the as grown Cu₂(NO₃)(OH)₃/ZnO heterojunction NR precursor and the samples annealing at different temperatures are shown in Figure 3. The surface morphology of the sample annealing at 200 °C in Figure 3b is similar to the as-grown Cu₂(NO₃)(OH)₃/ZnO precursor (Figure 3a), except that the individual nanosheet of the building blocks becomes thinner. In addition, the sample annealing at 200 °C shows a gray appearance, which indicates that a few of the Cu₂(NO₃)(OH)₃ nanostructures begin to convert to CuO. With the annealing temperature rising to 300 and 400 °C, the two samples show a black appearance and Cu₂(NO₃)(OH)₃ is converted to CuO. The SEM morphology of the sample annealing at 300 °C, as shown in Figure 3c, shows obvious change because of the decomposition of Cu₂(NO₃)(OH)₃. The nanosheet–shape structure is not smooth and many nanoparticles appear. For the sample annealing at 400 °C, these nanoparticles become denser and the nanosheet-shape nanostructures disappear, which is shown in Figure 3d. Therefore, it can be concluded that the annealing temperature has quite an impact on the morphology of the samples. Among all the samples, the sample annealed at 300 °C not only keeps the nanosheet morphology of the Cu₂(NO₃)(OH)₃ precursor but also shows a good crystalline quality.

All these results show that the growth process of the CuO/ZnO heterojunction is dependent on the Cu²⁺ ion concentration of the Cu(NO₃)₂ solution, UV illumination time, and the annealing temperature.

The XRD pattern was used to analyze the phase structure of the samples before and after annealing (Figure 4a). All the samples show (002) and (004) diffraction peaks of wurtzite ZnO, further confirming the c-axis preferential orientation. The Cu₂(NO₃)(OH)₃/ ZnO NR precursor prepared at room temperature displays two small peaks, which are (001) and (002) crystal plane of Cu₂(NO₃)(OH)₃ (JCPDS card no. 45-0594). The weak intensity results from the small size of the as-synthesized Cu₂(NO₃)(OH)₃ nanostructures in the specific crystal orientation. After annealing in air at 200 °C, the Cu₂(NO₃)(OH)₃ phase was still maintained, and no peaks of CuO were observed. It can be concluded that only a few of the Cu₂(NO₃)(OH)₃ nanostructures are converted to CuO, which cannot be detected by XRD due to its low content. Thus, the appearance changes from blue to gray while the sample stills shows an XRD diffraction peak of Cu₂(NO₃)(OH)₃. With the annealing temperature rising to 300 and 400 °C, the peaks of Cu₂(NO₃)(OH)₃ disappear, and a small peak of CuO (111) crystal plane (JCPDS card no. 48-1548) appears, which reveals that Cu₂(NO₃)(OH)₃ completely converts to CuO. No other impurity diffractions were found.

The TG and DSC curves are measured to study the decomposition of the Cu₂(NO₃)(OH)₃/ZnO NR precursor fabricated with 50 mM Cu²⁺ concentration and a 20 min UV illumination time, as shown in Figure 4b. The decomposition temperature ranges from 218.2 °C to 241.3 °C. The DSC curves shows a single endothermic peak centered at 224.1°C, indicating the one-step decomposition process of Cu₂(NO₃)(OH)₃. Theoretically, the weight loss for the conversion of Cu₂(NO₃)(OH)₃ to CuO is 33.17%. Here the weight loss is only 13.17%, which is due to the presence of ZnO NRs. In the whole process, the weight of ZnO has no change. Combined with the XRD analysis, the resultant is monoclinic CuO crystal. The thermal analysis results of the decomposition reaction are in accordance with many reports [34,35].

Furthermore, the phase transformation and chemical states of the as-synthesized Cu₂(NO₃)(OH)₃/ZnO (50 mM Cu²⁺ ion concentration and 20 min UV illumination time) and CuO/ZnO heterojunction NRs were studied by XPS. The full survey of the two samples confirms the existence of Zn, Cu, O, and C elements (Figure 5a). Figure 5b–d shows the high resolution XPS spectra of Zn 2p, Cu 2p, and O 1s of the two samples. The XPS spectra of Zn 2p shows intense doublet peaks centered at 1022.3 and 1045.5 eV in the CuO/ZnO heterojunction NRs and 1023.4 eV and 1046.8 eV in the Cu₂(NO₃)(OH)₃/ZnO heterojunction NRs, indicating the +2 oxidation state of Zn. Moreover, the Zn 2p_1/2 and Zn 2p_3/2 peaks show a splitting of ~23 eV, consistent with the value in existing literature [36]. In contrast with the Cu₂(NO₃)(OH)₃/ZnO heterojunction, the peaks of Zn 2p in the CuO/ZnO heterojunction shift towards low binding energies, which result from the weak interaction of CuO and ZnO. Furthermore, the O 1s core level spectrum can be resolved into two O 1s peaks in CuO/ZnO heterojunction NRs. The 529.7 eV peak corresponds to the lattice binding of O²⁻ with Cu, while the peak at 531.8 eV is due to oxygen adsorbed on the CuO surface. For the O 1s spectrum of the Cu₂(NO₃)(OH)₃/ZnO heterojunction NRs, there is only one broad peak centered at 532 eV, which should be attributed to surface adsorbed oxygen and oxygen species in the Cu₂(NO₃)(OH)₃ precursor. The Cu 2p spectrum of the CuO/ZnO sample shows two peaks centered at 954 eV (Cu 2p_1/2) and 934 eV (Cu 2p_3/2). The sample also shows two satellites (asterisks) on the higher binding energy side of Cu 2p_1/2 and Cu 2p_3/2, which is consistent with the reported literature [26,37]. Compared with the Cu₂(NO₃)(OH)₃/ZnO sample, the peaks in the Cu 2p spectrum of the CuO/ZnO sample shift towards lower binding energy, indicating the change of the chemical state of the Cu ions.

We further measured the TEM and HRTEM of the Cu₂(NO₃)(OH)₃ (50 mM Cu²⁺ and 20 min UV illumination time) and CuO (300 °C annealing) samples, as shown in Figure 6. The samples were scraped from the substrates and ultrasonic dispersed in ethanol. In this process, Cu₂(NO₃)(OH)₃ was separated from the ZnO nanorod. Figure 6a shows the TEM image of the as-prepared Cu₂(NO₃)(OH)₃ nanosheet, which is consistent with the result of SEM. The crystalline structure of Cu₂(NO₃)(OH)₃ is shown in its HRTEM image (Figure 6b). The atomic spacing is 0.344 nm, in agreement with the (002) plane of Cu₂(NO₃)(OH)₃. In Figure 6c, CuO still keeps the nanoplate structure after 300 °C annealing. It can also be found that there are many pores in the nanoplate. The HRTEM image of CuO is shown in Figure 6d. The sample shows a good crystalline quality and the plane distance of CuO is 0.232 nm, which is in accordance with the (111) plane of CuO.

Combined with above experimental results, we propose a model to explain the growth process of the CuO/ZnO heterojunction NRs. The reactions can be written as

$$2{\mathrm{Cu}}^{2+}+3\mathrm{O}{\mathrm{H}}^{-}+\mathrm{N}{\mathrm{O}}_3^{-}\to \mathrm{C}{\mathrm{u}}_2\left(\mathrm{N}{\mathrm{O}}_3\right){\left(\mathrm{O}\mathrm{H}\right)}_3$$

(1)

$$2\mathrm{C}{\mathrm{u}}_2\left(\mathrm{N}{\mathrm{O}}_3\right){\left(\mathrm{O}\mathrm{H}\right)}_3\to 4{\mathrm{CuO}+2\mathrm{NO}}_2+{\mathrm{O}}_2+3{\mathrm{H}}_2\mathrm{O}.$$

(2)

The electrons in the valence band (VB) of ZnO transfer to the conduction band (CB) when the ZnO NRs were exposed under UV light. These electrons then transfer to the surface of ZnO nanorod and reacted with H₂O or dissolved O₂ in deionized water to generated OH⁻, which generates a large number of OH⁻ ions around ZnO NRs. Cu²⁺ ions gathered around ZnO NRs and formed Cu₂(NO₃)(OH)₃ in the present of NO₃⁻ (reaction (1)). In Dai’s [14] report, photochemical reaction deposited Ni(OH)₂. Here, it was Cu₂(NO₃)(OH)₃ instead of Cu(OH)₂ formed on ZnO NRs in our experiment. Considering the low UV intensity, the formation of Cu₂(NO₃)(OH)₃ can be attributed to the low OH⁻ concentration around ZnO NRs. Cu₂(NO₃)(OH)₃ has lower solubility product (Ksp) than Cu(OH)₂ and is preferred to produce in a low OH⁻ environment in the present of NO₃⁻. In the photochemical reaction process, Cu₂(NO₃)(OH)₃ nanoparticles is initially generated on upper part of ZnO NRs and facilitates subsequent nucleation and growth. Then these nanoparticles grow directionally and formed their own preferred sheet-like structures. Due to the spaces between ZnO NRs, Cu₂(NO₃)(OH)₃ branches are obvious on the side surface of ZnO NRs, shown in the insets of Figure 1. Considering the low UV intensity of the lower part of the ZnO NRs and the gaps between the ZnO NRs, the growth of Cu₂(NO₃)(OH)₃ branches are hindered at the lower part of the ZnO NRs. Moreover, photogenerated electrons have a fastest transport speed along the c axis. Thereby, Cu₂(NO₃)(OH)₃ is mainly generated at the tips of ZnO NRs and forms bigger sheet-like structures. These results mean that the geometrical arrangement of ZnO NRs has a significant effect on the growth of Cu₂(NO₃)(OH)₃ precursor. The sheet-like appearance of Cu₂(NO₃)(OH)₃ is due to its natural layered crystal structure [38]. In the layered structure, octahedral copper forms layers of stoichiometry [Cu₂(OH)₃]⁺. NO₃⁻ ions play a role of charge balance in the positive layers. NO₃⁻ ions and hydroxyl groups are linked by hydrogen bonding in the octahedral copper layers. These Cu₂(NO₃)(OH)₃ sheets would be converted to CuO by thermal decomposition (reaction (2)).

3.2. UV Detecting Performance of CuO/ZnO Heterojunction NRs

The electrical properties and UV photo-response were measured using the sample synthesized in 50 mM Cu(NO₃)₂ aqueous solution and air annealed at 300 °C. Figure 7a displays the structure of the prepared CuO/ZnO heterojunction photodetector. The I-V characteristics of the ZnO and CuO/ZnO devices in dark and under UV light is displayed in Figure 7b,d. The pure ZnO NR device displays linear I-V curves, While the CuO/ZnO device displays a well-defined rectifying characteristic. The semi-log plot of the dark I-V curve is shown in the inset of Figure 7d. Figure S2 exhibits the linear I-V curves of Ag-CuO and AZO-ZnO, indicating good ohmic contacts. Hence, the rectifying behavior is not due to Schottky junctions at the electrodes but the CuO/ZnO p-n junction. Under 365 nm light illumination, both the two devices display UV response. Figure 7c,e show the time-resolved UV photocurrents of the pure ZnO and CuO/ZnO devices, respectively. The rise time (τ_r) is the time for a 90% current rise and the decay time (τ_d) is the time for a 90% current decay. The ZnO device displays a long rise time (9 s) and decay time (190 s). The CuO/ZnO device has a high response speed and the response times are much shorter (see

Table 1. The device performance of pure ZnO and CuO/ZnO UV photodetectors.

Samples	Idark (A)	Iph (A)	Photosensitivity	τr (s)	τd(s)
ZnO	121	174	1.4	9	190
CuO/ZnO	0.002	0.142	71	6	7

), especially the decay time. It is interesting to see that the rising and falling edges of the CuO/ZnO device both consist of a fast process and a slow process, which will be discussed later. The response time of the fast process is less than 1 s, which is remarkably reduced as compared with that of the ZnO device. It is also seen that the fast process dominates the rising and falling edges and accounts for more than 70% of the photocurrent variations, which is large enough for fabricating fast response devices in real applications. For our CuO/ZnO device, the response speed is improved significantly in the reverse bias and the device shows a short response time, but in the forward current, a positive bias reduces the barrier height of CuO/ZnO p/n junction and the effect of oxygen adsorption is still obvious. The response speed is slower than that in the negative bias.

Photosensitivity, (I_ph − I_dark)/I_dark, is one of the important parameters for a photodetector, where I_ph and I_dark are photocurrent and dark current, respectively. All the performance parameters of the two devices are collected in Table 1. The photosensitivity of the flower-like CuO/ZnO NR device has been enhanced almost 50 times compared with the pure ZnO NRs device. According to these results, the photo-response performance of ZnO device has improved effectively by coating CuO on the surface of ZnO.

Also, the photosensitivity of the UV photodetectors is significantly affected by the CuO coverage degree on ZnO NRs which is related to the Cu²⁺ ion concentration during the photochemical deposition. The UV illumination time is fixed at 20 min and the CuO coverage degree is increasing with rising Cu²⁺ ion concentration. Figure 7f shows that the photosensitivity has a maximum value of 71 at –5 V bias when Cu²⁺ concentration is 50 mM, which is the optimal Cu²⁺ concentration in our experiment. As the Cu²⁺ concentration increases to 100 mM, the photosensitivity of the device deceases, because at a higher Cu²⁺ concentration, the CuO nanostructure layer has a larger thickness, which affects the light absorption of the CuO/ZnO device.

Table 2. The device performance of our device and other CuO/ZnO UV photodetectors.

Samples	Photosensitivity	τr (s)	τd (s)
CuO/ZnO NWs [28]	2.4	>10	>10
CuO/ZnO NCs [29]	7.7	11.3	9
CuO/ZnO NWs [30]	61	—	—
CuO/ZnO [31]	2.2	4.2	5.2
This work	71	6	7

lists several parameters of our device and other CuO/ZnO UV photodetectors. Our device shows a high photosensitivity compared with other UV photodetectors based on CuO/ZnO nanostructures.

The principles of ZnO and CuO/ZnO photodetectors are described in Figure 8. For the ZnO device under the dark environment (Figure 8a), oxygen molecule adsorption traps the carriers in ZnO nanorods, forming O₂⁻. Electrons are depleted from a space-charge layer near the surface and drive the nanorod into high resistivity state. When the device is illuminated by UV light (Figure 8b), there are a large number of photo-generated electron-holes pairs in the ZnO nanorods. Holes that are transferred to the surface of the ZnO nanorods react with O₂⁻ and create O₂ desorption. Thus, the carrier concentration of the device increases and the device switches to a low resistivity state. The adsorption and desorption process of oxygen molecules is slow, so the ZnO device has a long rise time and decay time (see Figure 7c). The adsorbed oxygen has been widely considered to explain the long photocurrent decay time in several nanostructured materials [39,40,41]. For the CuO/ZnO device, we can also observe slow rising and falling processes in the I-t curve (Figure 7e). The reason for this result is that the oxygen molecule adsorption and desorption process on the surface of the CuO/ZnO device is similar to that on the ZnO device. The difference is that there is a depletion layer at the interface between CuO and ZnO. Oxygen molecules cannot reach the interface and have little effect on the heterojunction. The very low dark current of the CuO/ZnO device is due to the heterojunction barrier (Figure 7c). Under the illumination of UV light, electrons in the CB of CuO transferred to ZnO and the holes in the VB of ZnO transferred to CuO (Figure 8e). Photogenerated electron-hole pairs are separated effectively under the heterojunction barrier, so the photosensitivity is high. The charge separation of a p-n junction is also very fast, so we can observe a very sharp rising edge in the I-t curve (Figure 7e). The explanation of the falling edge is similar to that of the rising edge.

4. Conclusions

In conclusion, we prepared CuO/ZnO heterojunction NRs using a facile photochemical deposition strategy. Cu(NO₃)(OH)₃/ZnO heterostructure NRs were synthesized as a precursor and then converted into CuO/ZnO heterostructure NRs after air annealing at a low temperature. The morphology of the CuO/ZnO heterostructure NRs is related to the growth conditions, including the Cu²⁺ ion concentration, UV illumination time, and annealing temperature. All the heterostructure devices showed well-defined rectified characteristic, and their UV detecting performances are improved compared with a pure ZnO NRs device. The CuO/ZnO NRs photodetectors fabricated with 50 mM Cu2+ ion concentration, 20 min UV illumination time, and 300 °C annealing temperature exhibit optimum device performance. The photosensitivity is increased by 50 times in contrast with the pure ZnO NR device. Especially, the response time of the fast process in the rising edge and falling edge is less than 1 s, which is remarkably reduced compared with that of the ZnO device. The facile and low-cost synthesis method not only can be used to fabricate CuO and NiO but also can be extended to the synthesis of the other metal oxide nanostructure on the ZnO NR surface.