I. INTRODUCTION
Recently, transparent optoelectronic devices have attracted attention due to novel applications such as light-emitting diodes,Reference Ohta, Kawamura, Orita, Hirano, Sarukura and Hosono1 solar cells, thin-film transistors,Reference Nomura, Ohta, Ueda, Kamiya, Hirano and Hosono2,Reference Hoffman, Norris and Wager3 and ultraviolet (UV) sensors.Reference Yang, Hullavarad, Nagaraj, Takeuchi, Sharma, Venkatesan, Vispute and Shen4 UV sensors have been widely used in various commercial and military applications, including secure space-to-space communications, pollution monitoring, water sterilization, flame sensing, and early missile-plume detection.Reference Monroy, Calle, Pau, Munoz, Omnes, Beaumont and Gibart5 Among various materials suggested for the realization of transparent electronic devices, ZnO has been strongly considered to be a promising functional material because it has a direct band gap of 3.37 eV, a large exciton binding energy of 60 meV at room temperature, and a high transparency in the visible range with native n-type.Reference Pan, Dai and Wang6 In particular, well-aligned, one-dimensional (1D) ZnO arrays are considered to induce high-performance devices due to the increased surface area and quantum effect.Reference Gu, Kuskovsky, Yin, OBrien and Neumark7 Excellent vertical alignment of ZnO nanowires on large area is very important for the fabrication of optoelectronic devices using nanostructures as an active layer, because it allows the use of semiconductor processing and metal interconnection on the top of all nanowires. However, the ZnO nanowires with intrinsically n-type characteristics require the development of p-type oxide materials to fabricate all-oxide p–n junctions with rectifying properties. The reports on optoelectronic devices consisting of all-oxide p–n junctions have been rare because of the absence of p-type oxide materials with reliable characteristics. Up to now, most studies on the ZnO-nanowire-based optoelectronic devices have used p-Si and p-GaN layers as p-type semiconductors. Schottky,Reference Liang, Sheng, Liu, Huo, Lu and Shen8 p–n junction, p–i–n junction, and metal–semiconductor–metalReference Razeghi and Rogalski9 structures have been proposed for semiconductor UV sensors. In the case of photodetectors based on ZnO, the p–n junction is more useful structure regarding the response time and sensing performance compared to Schottky contact photodetectors, due to the formation of a depletion region at the interface. As an alternative to p-type ZnO, which has unstable properties, Cu2O, SnO, and NiO have been proposed due to their relatively high stability. However, Cu2O oxide thin films are not appropriate for transparent optoelectronic devices, because they have a narrow band edge of 2.1 eV, and are not fully transparent. In the case of tin-based oxides, the fabrication of p-type SnO phases is difficult since the most stable tin oxide phase (SnO2) normally has n-type characteristics. Comparatively, NiO is a representative oxide that shows intrinsic p-type characteristics and high stability. The band gap of NiO is approximately 3.6 eV, and it has been demonstrated that it can be a potential candidate for the fabrication of deep-ultraviolet photodetectors when combined with MgO (NixMg1−xO).Reference Mares, Boutwell, Wei, Scheurer and Schoenfeld10,Reference Mares, Boutwell, Scheurer, Falanga and Schoenfeld11 In addition, NiO thin films are expected to show high transparency at visible wave lengths. Nevertheless, since the NiO has lower electrical mobility and emission properties, it has not been studied as an active semiconductor oxide material. According to previous research, it is well known that stoichiometric NiO with insulating characteristics has been highly regarded as a candidate for resistive switching memory devices.Reference Kim, Seo, Ahn, Suh, Lee, Park, Yoo, Baek, Kim and Ryu12 Furthermore, the NiO has been studied as an electrode for lithium ion batteries and electrochromic applications. The high p-type stability of NiO is very attractive for use as a hole injection layer in transparent optoelectronic devices. If transparent photodetectors can be fabricated with p-type oxide and transparent glass substrates, a simple device process is expected, as well as high UV sensibility, and the cutoff of visible spectrum.
It is well known that ZnO nanowires can be synthesized on crystalline or amorphous substrates.Reference Kim, Mahanta, Kong, Cho and Kim13 In the present work, we chose glass substrate rather than Si for transparent devices with no response in the visible region. The NiO films were directly deposited by magnetron sputtering on the top of ZnO nanowire arrays grown by metal organic chemical vapor deposition (MOCVD) for the formation of p–n junctions. The final structure consisted of an indium tin oxide (ITO) electrode/p-NiO/n-ZnO nanowire/AZO bottom electrode/glass substrate. The UV sensing properties of the fabricated devices were compared with those of the ZnO-film-based device.
II. EXPERIMENTAL DETAILS
The all-oxide UV sensors were fabricated with the structure shown in Fig. 1(a). We selected ZnO doped with 5 at.% Al (AZO) as the bottom electrode, because of its high transparency and electrical conductivity. The AZO layer can also function as a seed layer for the vertical alignment of nanowires.Reference Zhang and Que14 The AZO bottom electrode was deposited by atomic layer deposition on corning glass at 200 °C with a thickness of 250 nm. The ZnO nanowires were synthesized by MOCVD technique as reported elsewhere.Reference Kim, Jung, Kwon and Cho15,Reference Mohanta, Kim, Cho, Chua and Tripathy16 The tip shape and length of the ZnO nanowires were found to be moderately uniform. The synthesis of the ZnO nanowires was carried out at 400 °C and 1 Torr for 30 min using diethylzinc (DEZn) and high-purity oxygen gas as precursor sources. The flow rate of Ar carrier gas was set to 10 sccm. The flow rates of the DEZn source and reactive O2 gas were 6.1 and 2230 μmol/min.
(a) Schematic drawing of the ITO/NiO/ZnO nanowires/AZO/glass heterostructure UV sensor. (b) Tilt-view and (d) plane-view SEM images of the all-oxide heterostructure, and (c) optical microscopy image showing transparency.
After the growth of the ZnO nanowires, the NiO layer was directly deposited on the top of the ZnO nanowires using a radio-frequency (RF) magnetron sputtering system. A high purity (99.99%) NiO (4 inches) was used as the sputtering target. Before the NiO growth, presputtering was carried out for 10 min to clear the surface contamination, and the base pressure of the chamber was kept below 5 × 10−6 Torr. During the deposition, the RF power was 150 W and the working pressure was maintained at 15 mTorr. The O2 flow was kept at 30 sccm and sputtering was performed for 180 min at 400 °C.
The ITO top electrode was then deposited by RF magnetron sputtering on the NiO. The sputtering target was 4-inch ITO (In2O3:SnO2 = 90:10) with 99.99% purity. During the deposition, the chamber pressure was maintained at 5 mTorr and the RF power was 150 W. The gas flows were kept at 20 sccm (Ar) and 0.1 sccm (O2), and the growth time was 30 min at 300 °C. In addition, ZnO-film-based UV sensors were fabricated as a reference to compare UV sensing properties. The ZnO film was deposited at 600 °C by RF magnetron sputtering. The electrical resistivity and the carrier concentration of the ZnO film were 10−1 Ω cm and 1.86 × 1018/cm3, respectively, based on Hall measurements.
Morphology observation of the samples was conducted using a field-emission scanning electron microscope (FE-SEM, JSM-7600F, Japan) operated at 10 kV. UV sensor devices based on a p–n diode were used for current–voltage (I–V) measurements by an HP4145B semiconductor parameter analyzer. The structural properties of the p-NiO/n-ZnO heterostructure were investigated by x-ray diffraction (XRD, Bruker AXS D8 Discover). The light source used was a 100 W UV lamp with a wave length of 365 nm. The transmittance of the ITO/NiO/ZnO/AZO/glass structure was evaluated by a UV/VIS/NIR spectrophotometer (Varian Cary 5000).
III. RESULTS AND DISCUSSION
Various methods for the synthesis of 1D ZnO nanostructures have been suggested previously, including thermal evaporation,Reference Yao, Chan and Wang17 chemical bath deposition,Reference Hari, Baumer, Tennyson and Bumm18 MOCVD,Reference Kim, Kong, Cho, Park and Lee19 and hydrothermal processing.Reference Liu and Zeng20 Vertically arrayed 1D ZnO nanowires have been fabricated by MOCVD on various substrates, such as Si, GaN, glass, and sapphire.Reference Kong, Choi, Cho, Kim, Baek and Lee21 It was reported that the electrical properties of the NiO thin films were significantly changed by oxygen partial pressure during the deposition and thermal annealing, changing the electrical conductivity of the films from semiconductors to insulators. Thus, the deposition of the NiO layers with p-type semiconducting properties should be performed after the synthesis of the ZnO nanowires for the formation of the p–n junction. In contrast, the opposite structure of the ZnO nanowires on NiO provides nonconducting NiO layers, which cannot be used as a hole-injection or transport layer. We chose NiO films deposited by magnetron sputtering on the top of the ZnO nanowires as p-type layers. The deposition conditions of the NiO films have been reported in a previous study.Reference Lee, Kwon, Kong, Lee and Cho22
Figure 1 schematically demonstrates the device structure of the UV photosensors, which are expected to show transparent characteristics in the visible region. The AZO bottom electrode was deposited by ALD on corning glass, and the ZnO nanowires were vertically grown on the bottom electrode by MOCVD. The p-NiO layer was deposited on the top of the ZnO nanowires by magnetron sputtering. The electrical resistivity of the AZO bottom layer was low to be at 10−4 Ω cm according to Hall measurements. The AZO film was suitable as an electrode in our devices because of its good electrical resistivity and high transparency. In particular, the AZO layers have preferred orientation axis along the [0001] direction, which assists regular vertical alignment of the nanowires.Reference Kong, Mohanta, Kim and Cho23 The length of the vertical ZnO nanowires was 600–700 nm. Next, mushroom-shaped NiO layers were formed on only the top region of the ZnO nanowires by the shadow effect of sputtering, resulting in the formation of the p–n junction in the top region, as shown in Fig. 1(b). Finally, an ITO top electrode with low electrical resistivity was coated onto the NiO layers by sputtering. This process provides a large area interconnection for the NiO grains isolated on the nanowires. We choose ITO as a p-type electrode due to the high electron affinity and high band gap energy. As shown in Fig. 1(d), the surface morphology is quite smooth, despite the deposition on the nanowires.
Figure 2 shows the XRD patterns for the p-NiO/n-ZnO nanowires without the ITO layer. The XRD pattern corresponding to the p-NiO/n-ZnO film was also shown in Fig. 2(b) for comparison. The dominant peaks at 34.5° correspond to the (0002) plane of wurtzite ZnO crystal with cell parameters of a = 3.245 Å and c = 5.195 Å. An additional peak at ∼37° that originated from the (111) diffraction of the NiO layer indicates the formation of crystallized NiO layers on the ZnO nanowires or ZnO film. No other diffraction peaks from Ni and Ni-related compounds were observed in the XRD patterns of both the structures. The deposited NiO layer showed preferentially oriented growth despite being deposited on the ZnO nanowires.
X-ray diffraction patterns of (a) p-NiO/n-ZnO-nanowire-based sample and (b) p-NiO/n-ZnO-film-based sample.
To evaluate the degree of transparency of the deposited oxide layers, optical transmittance spectra from the p-NiO/n-ZnO heterostructure were characterized, as shown in Fig. 3. The transmittance for the nanowire arrays/AZO/glass is nearly 90% in the UV and visible regions. The corresponding absorption edge of the structure was 370 nm. In contrast, the full structure of ITO/p-NiO/n-ZnO nanowire/AZO/glass showed a reduced transmittance value of about 60%. The transmittance of the full structure including NiO was decreased by approximately 20% compared to the sample without the NiO layer. The reason could be explained due to the formation of polycrystalline surface in the top region consisting of ITO/NiO layers. The ITO/NiO layer exhibited polycrystallinity with many grains due to the underlying nanowires, which induced optical scattering responsible for the decrease of transmittance. But still our ZnO nanowire photosensor with the structure of ITO electrode/p-NiO/n-ZnO nanowire/AZO/glass substrate showed semitransparent characteristics with an average transparency of 60%. The semitransparent characteristics observed for our device were supported by Fig. 1(c) in which we can clearly distinguish the characters written on a paper behind the p–n heterojunction device.
Transmittance measurements of ZnO nanowires/AZO and ITO/NiO/ZnO nanowires/AZO deposited on glass substrates.
To confirm the formation of a p–n heterojunction with rectifying characteristics, we fabricated p-NiO/n-ZnO diodes using photolithography and etching processes. As shown in Figs. 4(a) and 4(b), the I–V characteristics of the transparent p-NiO/n-ZnO nanowires and p-NiO/n-ZnO film exhibit rectification behavior under dark conditions. The deposition of the ITO on NiO was already confirmed to induce linear ohmic characteristics. Thus, this behavior indicates that a p–n junction was successfully formed. The semiconductor photosensors based on p–n diodes have been operated under reverse bias for amplification of the signal. Figure 4(a) shows the current change of the ZnO-nanowire-based UV sensor under dark and UV-illumination (365 nm) conditions, where the bias voltage was changed from −8 to +8 V. The ZnO-nanowire-based device shows a current rise at about 4 V in forward bias, and the I(forward)/I(reverse) ratio at 8 V is about 11.3. However, UV illumination (365 nm) induces a significant increase in the current of the reverse region, as indicated by a red line. The increased current at −8 V between dark and UV illumination is about one order of magnitude. The enhanced photoconductive gain of the nanowire-based device is attributed to the large surface-to-volume ratio and the high density of deep-level surface traps. In the case of ZnO-based UV sensors, it has been generally acceptedReference Soci, Zhang, Xiang, Dayeh, Aplin, Park, Bao, Lo and Wang24 that oxygen molecules are adsorbed onto ZnO surfaces by capturing free electrons from the n-ZnO semiconductor 
${\rm{[}}{{\rm{O}}_{\rm{2}}}{\rm{(g)}} + {{\rm{e}}^ - } \to {\rm{O}}_{\rm{2}}^ - {\rm{(ad)]}}$, which has a low conductive depletion layer that forms near the surface. The electron–hole pairs are photogenerated [hν → e− + h+], when the photon energy illumination exceeds band gap energy E 0 (hν > E g). The holes migrating to the surface along the potential gradient produced by band-bending discharge the negatively charged adsorbed ions 
${\rm{[}}{{\rm{h}}^{\rm{ + }}}{\rm{ + O}}_{\rm{2}}^ - {\rm{(ad)}} \to {{\rm{O}}_{\rm{2}}}{\rm{(g)]}}$, and consequently, oxygen is photodesorbed from the surfaces, thereby decreasing the free carrier concentration and decreasing the width of the depletion layer. The high density of trap states usually found in nanowires due to the dangling bonds at the surface readily act as oxygen-related hole-trap states in the reaction 
${\rm{[}}{{\rm{h}}^{\rm{ + }}}{\rm{ + O}}_{\rm{2}}^ - {\rm{(ad)}} \to {{\rm{O}}_{\rm{2}}}{\rm{(g)]}}$ under UV illumination. The photogenerated electron–hole pairs are separated by an applied bias leading to the steady-state persistent photocurrent in the nanowire-based device structures. For the ZnO-film-based UV sensor, a very little increase in current was observed in the reverse bias after UV illumination.
The current–voltage (I–V) characteristics of the ITO/p-NiO/n-ZnO/AZO heterostructures: (a) ZnO-nanowire-based sample (b) ZnO-film-based sample.
In the UV sensors operating with a p–n junction, when the depletion layer is exposed to UV illumination (365 nm), electron–hole pairs are generated, which move into the n-type and p-type regions respectively in reverse bias. We confirmed that there is no distinguishable change in the current after UV illumination, ignoring the photoresponse for the NiO with the band gap (3.7 eV). Therefore, the current change by the 365-nm photoillumination is attributed to the ZnO nanowire arrays with a band gap of 3.3 eV. The responsivities of the p-NiO/n-ZnO heterostructure UV sensors measured at a reverse bias of −5 V are shown in Fig. 5. This result shows how much the ZnO nanowires generate charge carriers by photons with specific wave lengths ranging from 350 to 700 nm. The UV photoresponsivity of the p-NiO/n-ZnO nanowire heterostructure is ∼0.57 A/W. The responsivity shows the highest value at 370 nm corresponding to the ZnO band gap. Interestingly, this device exhibits a weak response in the visible region, unlike the ZnO/Si heterojunction. The ZnO-film-based UV sensor shows a similar trend but with a much lower photoresponse than the ZnO-nanowire-based sample. This is attributed to the high surface-to-volume ratio of the ZnO nanowires. Such a photoresponse is strongly related to the oxygen adsorption/desorption process in the presence of trap states at the nanowire surface. In contrast, the oxygen adsorption/desorption process at the grain boundaries and at the surfaces contributed to the photoresponse in the thin-film-based photosensors.Reference Ji, Peng, Su, Young, Wu and Cheng25 The enhanced surface area of the active ZnO nanowires leads to an increase of the photocurrent during the UV illumination. The maximum photocurrent of the ZnO-nanowire-based UV sensor is 3.3 × 10−7 A for 300 s. Nevertheless, the photoresponse in the UV region is relatively low compared to a Schottky-contact-based photosensor.Reference Kim, Jung, Kwon and Cho15 This may result from the lower crystalline quality of the NiO film, which induces leakage current and charge scattering.Reference Kwon, Chun, Han and Cho26
Spectral responsivities of the ZnO-based UV sensors measured at −5 V applied bias.
Figure 6 shows the transient response of the current measured by periodical UV illumination under the bias of −3 V. More distinguishable on/off characteristics were observed in the nanowire-based sensor. However, these p–n diodes using p-NiO show delayed relaxation behavior under the dark conditions. The slow recovery is due to the charge trapping at the high-density grain boundaries of the polycrystalline NiO film. One of the most noticeable characteristics of our UV sensor is its high transparency (or semitransparency), despite p–n junction devices. The layers used in this study are all-oxide, including Corning glass, AZO bottom electrode, ZnO nanowires, NiO film, and ITO top electrode. Thus, if we optimize the deposition conditions of p-type NiO, better UV sensors can be fabricated with high transparency.
The transient response of the current measured by periodical 365-nm UV irradiation under a bias of −3 V.
IV. CONCLUSIONS
All-oxide heterostructure UV photosensors made with the structure of ITO/p-NiO/n-ZnO nanowires/AZO/glass substrate have been fabricated. The ZnO nanowires were grown by MOCVD, and the ITO and p-NiO layers were deposited by RF magnetron sputtering. The p-NiO/n-ZnO heterostructure resulted in a p–n diode with good rectifying properties. The ZnO nanowires with high aspect ratio were synthesized as a sensing layer on the AZO electrode. UV photosensors were successfully fabricated with an operating wave length of 370 nm in the reverse bias region. The transmittance of this device was about 60% in the visible region, resulting in semitransparent UV photosensors. Moreover, the ZnO-nanowire-based UV sensor showed higher photoresponse compared to the ZnO-film-based UV sensor. This was attributed to the single-crystal ZnO nanowire arrays with high surface-to-volume ratio. Therefore, NiO is a good candidate as p-type oxide in optoelectronic applications.
ACKNOWLEDGMENTS
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by Ministry of Education, Science and Technology (Grant No. 2012-0001447), the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (Grant No. 2012-0038) and the Energy International Collaboration Research & Development Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) funded by the Ministry of Knowledge Economy (MKE) (Grant No. 2011-8520010050).
