Al2O3 heterostructure photoelectrodes with different Al layer thicknesses

Author’s Accepted Manuscript Efficient photoelectrochemical water splitting using CuO nanorod/Al2O3 heterostructure photoelectrodes with different Al layer thicknesses Jin-wook Ha, Hyukhyun Ryu, Won-Jae Lee, JongSeong Bae www.elsevier.com/locate/physb

PII: DOI: Reference:

S0921-4526(17)30285-5 http://dx.doi.org/10.1016/j.physb.2017.05.052 PHYSB309976

To appear in: Physica B: Physics of Condensed Matter Received date: 7 March 2017 Revised date: 28 May 2017 Accepted date: 29 May 2017 Cite this article as: Jin-wook Ha, Hyukhyun Ryu, Won-Jae Lee and Jong-Seong Bae, Efficient photoelectrochemical water splitting using CuO nanorod/Al2O heterostructure photoelectrodes with different Al layer thicknesses, Physica B: Physics of Condensed Matter, http://dx.doi.org/10.1016/j.physb.2017.05.052 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Efficient photoelectrochemical water splitting using CuO nanorod/Al2O3 heterostructure photoelectrodes with different Al layer thicknesses

Jin-wook Haa, Hyukhyun Ryua,*, Won-Jae Leeb , Jong-Seong Baec a

Department of Nano Science & Engineering, High Safety Vehicle Core Technology Research Center, Inje University, Gimhae-si, Gyeongnam 621-749, Republic of Korea

b

Department of Materials and Components Engineering, Dong-Eui University, Busan 614-714, Republic of Korea c

Division of Analysis & Research, Korea Basic Science Institute, Busan 618-230, Republic of Korea

*Corresponding author. Tel.: +82-55-320-3874. [email protected] (Hyukhyun Ryu)

Abstract In this study, an efficient water splitting technique was investigated with CuO nanorod/Al2O3 heterostructure photoelectrodes. Cupric oxide (CuO) nanorods were grown on a fluorine-doped tin oxide (FTO) glass substrate by using modified-chemical bath deposition. In addition, Al thin films were deposited on the CuO nanorods using thermal evaporation, and then, aluminum oxide (Al2O3) layers were formed in open air to create the CuO nanorod/Al2O3 structure. In this study, the morphological, optical and structural properties of CuO nanorod/Al2O3 photoelectrodes were analyzed according to the various thicknesses of the Al layers, and the effects of the thickness on the photoelectrochemical (PEC) properties were mainly discussed. We obtained a maximum photocurrent value of -2.26 mA/cm2 (-0.55 V vs. SCE) and a theoretical solar-to-hydrogen (STH) conversion efficiency of 1.61% using the Al 30-nm thick sample, which had the largest amount of the Al2O3 layer, as confirmed by X-ray photoelectron spectroscopy (XPS).

Keywords Photoelectrochemical (PEC), cupric oxide (CuO), aluminium oxide (Al2O3), photoelectrode

Introduction The photoelectrochemical (PEC) water splitting method is one of the cleanest and most advantageous technologies in terms of using unlimited solar energy. [1] Research into water splitting technology was initiated by Honda and Fujishima in 1972, and many studies have been carried out to date. The photoelectrode materials used for water splitting technique research can be roughly divided into oxide and non-oxide semiconductors. TiO2, ZnO, Cu2O, CuO, WO3 and BiVO4 have been representatively studied as oxide type semiconductor materials. [2-5] Among the photoelectrode material candidates, CuO, a p-type semiconductor, is the most suitable material for absorbing solar energy because it has the narrowest band gap among the oxide semiconductor candidates. [6] In addition, CuO is suitable for alternative energy technology research because of its abundant reserves, low manufacturing price, and non-toxic properties. [7] However, in practice, it is difficult to satisfy the various conditions required for the commercialization of water splitting technology using only a single material. [8] Ultimately, heterostructure photoelectrode research based on CuO materials is required for the preparation of photoelectrodes with high practicality. Recently, a number of research studies have been published using heterostructures of CuO with various materials to increase the efficiency of photoelectrodes. [9-13] In this regard, the Ho-Kimura research team reported a very interesting result. In their study, the growth of ZnO, TiO2, and Al2O3 layers on Cu2O/CuO layers was investigated, and the photoelectrode efficiency was studied. When the Al2O3 layer was utilized as a heterostructure layer with CuO, it acted as an electron acceptor layer, and an improved photocurrent density value was obtained compared to other materials such as ZnO and TiO2. [13] In addition, Al2O3 layers have been used as photocatalytic layers for p-type semiconductor photoelectrodes, such as NiO or SiO, to increase their conductivity as well as to decrease the charge carrier recombination phenomenon on the surface, and enhanced photocurrent density values have been reported. [14, 15] Therefore, Al2O3 materials have demonstrated good potential for forming heterostructure materials with CuO, which is a p-type semiconductor, but a more detailed study is highly required. During studies dealing with heterogeneous materials, it is very important to consider their compatibility while maintaining the inherent properties of each material. Particularly, when a one-dimensional nanostructure is exploited as an active layer for preparing highly efficient heterostructure photoelectrodes, the preservation of the morphological properties is important to maintain the advantages of the active layer, such as

a large surface area and excellent charge transfer ability. However, in the previous studies, the Al2O3 preparation methods were either anodic oxidation of the sample in a precursor solution or chemical oxidation of the precursor solution by drop casting. [13, 14] During these processes for the fabrication of CuO/Al2O3 heterostructures, it is relatively difficult to maintain the morphological properties of the CuO nanorod structure. Here, we deposited the Al layer on the CuO nanorods using a thermal evaporator. In addition, the Al layer was oxidized in air to form an Al2O3 layer of approximately 2 to 10 nm. As a result, the Al2O3 layer was prepared relatively easily. [16] Therefore, the morphological properties of the CuO nanorods were maintained. In this study, we investigated the photoelectrochemical properties of the CuO nanorod/Al2O3 heterostructures according to various Al layer thicknesses. For the analysis, the morphological, structural and optical properties of the CuO nanorod/Al2O3 heterostructures with various Al layer thicknesses were analyzed using field emission scanning electron microscopy (FE-SEM), X-ray diffraction (XRD), UV-vis spectroscopy and X-ray photoelectron spectroscopy (XPS). The photoelectrochemical performance was measured using a three-electrode system under 1-sun illumination conditions (1.5 AM filter, 100 mA/cm2) in a 1 M KOH aqueous solution. In addition, a MottSchottky plot was used to investigate the electrochemical impedance spectroscopy (EIS) properties of the prepared photoelectrode.

Experimental The CuO nanorod structures were grown on a CuO seed layer using the modified-chemical bath deposition (M-CBD) method. The solute and solvent of the solution used in the M-CBD method were prepared by stirring an 0.02 M copper acetate monohydrate solution and deionized water at 300 RPM for 30 minutes, respectively, and ammonia 25% aqueous solution was added at a volume ratio of 50: 1. The solution was uniformly injected at a feed rate of 0.6 ml/min for 10 minutes to grow the CuO nanorod structures. The M-CBD method required the solution to be at a constant concentration and volume in the reactor. The solution was supplied to the prepared seed layer using a peristaltic pump (Masterflex L/S, EW-7523-90). During this time, the FTO glass substrate was directly in contact with the hot plate and heated to 175 °C. To reduce the temperature difference between the solution outside the reactor and the solution inside the reactor, the solution was supplied to the reactor after passing through a 60 °C bath. After growth by the M-CBD method, the grown

sample was rinsed and dried using deionized water and purified air, respectively. Thereafter, an annealing process was conducted using a furnace at 400 °C for 5 minutes in air. The Al layer was deposited by a thermal evaporator onto the grown CuO nanorods. A 3 × 3 mm pellet (purity: 99.999%) was used as the Al source, and a tungsten boat was used. The Al layer was deposited at a rate of 3 ~ 5 Å/s under a high vacuum pressure of

Torr. To study the efficiency of the photoelectrode

with different Al thicknesses, the Al thickness was adjusted to 0, 10, 30, 50, and 70 nm by using a thickness monitor sensor. After the deposition process, the CuO nanorod/Al structure was exposed to air to oxidize the Al layer. In detail, for sufficient oxidization of the Al layer to the Al2O3 layer, the samples were stored for one day at room temperature in air. The morphological properties of the grown CuO nanorod/Al2O3 structures were analyzed using an FESEM (Quanta 200 FEG), and the optical properties were analyzed using UV-vis spectroscopy (SCINCO, S3100). The structural properties were analyzed by X-ray diffraction (Cu-Kα = 1.54060 Å), and an X-ray photoelectron spectroscopy analysis was carried out to identify the chemical bonding state of the Cu, Al and O atoms and to determine the atomic ratio of each element in the CuO nanorod/Al2O3 heterostructure. The photoelectrochemical (PEC) performance of the FTO glass/CuO nanorod/Al2O3 structure was measured using a potentiostat (Digy-Ivy, DY2111). The reference electrode, counter electrode and working electrode were a saturated calomel electrode (SCE), graphite rod, and FTO glass/CuO nanorod/Al2O3 structure, respectively. The electrolyte aqueous solution used was 1 M KOH (pH 14), and I-V (Current-Potential) curves were analyzed by applying a potential from 0 to -0.65 V under 1-sun illumination (1.5 AM filter, 100 mW/cm²) using a 300 W xenon (Xe) lamp. In addition, to analyze the EIS properties of the photoelectrode, a Mott-Schottky measurement was carried out under the following conditions: a 1 M KOH electrolyte solution, 1 kHz frequency, 10 mV amplitude potential, and an applied potential from 0.1 to 0.7 V in dark conditions.

Results and discussion Figure 1 shows the FE-SEM images of the FTO glass/CuO nanorod/Al2O3 structure with different Al thicknesses. The Al 0-nm sample, that is, with no Al deposition, shows the surface of the CuO nanorods. As the Al layer was deposited to 10 nm, the gaps between the CuO nanorods started to fill, and the surface was gradually covered. After reaching an Al thickness of 50 nm, aggregation on the surface was observed, and the CuO nanorod surface was fully covered. Figure 1 (f) shows the cross-section view images of the prepared

samples, and the thickness of the CuO layer was found to be approximately 1.2 μm. Figure 2 shows the results of XRD analysis to confirm the structural properties of the FTO glass/CuO nanorods/Al2O3 with various Al layer thicknesses. Figure 2 (a) shows the XRD spectra, which show the ( (

) and (

̅ ),

) planes of the monoclinic cupric oxide (ICSD card no. 98-009-2364) at 35.57°, 38.85° and

53.58°, respectively. Other peaks were found at 26.53°, 33.71°, 37.76°, 51.51° 61.59°, 65.5°, and 78.24°,which were attributed to the tin oxide (SnO2) (ICSD card no. 98-003-9178) from the FTO substrate. To prevent the deformation of the CuO nanorod structure by thermal energy during the deposition of the Al layer, substrate heating was not applied and a post heat treatment process was not conducted. Generally, amorphous structures are grown when the thermal evaporation process is performed at low substrate temperature conditions, [17] and the crystallization temperature of Al2O3 is known to be higher than 1000 °C. [18] Therefore, it was considered that a very thin amorphous Al2O3 layer was formed because it was oxidized naturally at room temperature in air. No Al2O3-related peaks appeared in the XRD spectrum. Figure 2 (b) shows the XRD peak intensities for the (

̅ ), (

) and (

) planes according to the various thicknesses of the Al layers. In all samples, the (

peak showed a relatively stronger peak intensity than the ( that the CuO crystals preferentially grew in the ( nanorods were mainly grown along the (

̅ ) and (

)

) peaks; therefore, it was considered

) direction. In our previous studies, we reported that CuO

) direction by using the M-CBD method. [19-21] The (

) plane

had the lowest density of Cu2+ ions in the CuO unit cell, [22] and therefore, the passivation by NH3 occurred less along the (

) plane. Consequently, the growth of the CuO nanorods along the (020) plane was preferentially

achieved, as shown in Fig. 2 (b). As the thickness of the Al layer increased, the intensity of the (020) peak gradually decreased, which is thought to be influenced by the Al and Al2O3 layers. Generally, the X-ray diffraction method detects the diffracted secondary electrons from a sample. Secondary electrons are greatly influenced by the thickness and composition of a sample. Therefore, the amorphous Al and Al2O3 layers can interfere with the movement of the secondary electrons generated by diffraction from the CuO lattice. As the Al and Al2O3 layers became thicker, the amount of interrupted secondary electrons was increased, which resulted in the decrease of the (020) peak intensity. Figure 3 shows the optical properties of the FTO glass/CuO nanorod/Al2O3 structures according to the Al layer thickness using UV-vis spectroscopy. Figure 3 (a) shows a plot of (αhν)2 and hν. Figure 3 (b) shows the optical energy bandgap of each sample obtained by extrapolation of the linear part of the plot in Figure 3 (a). All samples had an absorption coefficient that obeyed the following photon energy formula: (αhν)2 = B(hν−Eg),

where α is the absorption coefficient of the material, hν is the photon energy, B is a constant related to the material, and Eg is the optical bandgap of the material. The optical energy bandgap of all the samples was represented within the range of 1.6 ~ 1.9 eV, which is similar to the previously reported optical energy bandgap of the CuO material (1.4 ~ 1.7 eV), [19-21] and the change in the optical energy bandgap trend is considered to be affected by the Al and Al2O3 layers. The Al 10-nm and Al 30-nm samples indicated that the optical energy bandgap slightly increased compared to the Al 0-nm samples, and then, the optical energy bandgap value decreased as the Al thickness increased to 70 nm. In general, the amorphous Al2O3 layer is known to have a wide optical bandgap of approximately 5.2 to 6.4 eV. [23,24] Therefore, in the case of the samples with Al layer thicknesses of 10 and 30 nm, the Al layer conformed to the shape of the CuO nanorods, and it was expected to form an Al2O3 layer having a relatively large amount and large surface area that could influence the increase of the optical energy bandgap. However, in the case of the samples with Al layer thicknesses of 50 and 70 nm, the surface morphology was similar to that of a thin film due to the gaps being filled between the CuO nanorods by the Al layer, and the Al2O3 layer on the surface could be reduced while the thicker Al layer existed beneath the formed Al2O3 surface layer. Therefore, the optical energy bandgap values of the Al 50- and 70- nm samples were decreased because of less light transmittance. Figure 4 shows the XPS results from the qualitative analysis of the FTO glass/CuO nanorod/Al2O3 structures with various Al layer thicknesses. Figure 4 (a) shows the total XPS spectra of the samples. Cu 2p3, O 1s and C 1s peaks were confirmed in all of the samples, and Al 2p peaks were confirmed in the Al 10-nm sample. Basically, the XPS spectrum expresses the kinetic energy of the photoelectrons and the number of electrons escaping from the incident the X-ray beam. However, since the inelastic scattering of electrons occurs with a high probability in the solid sample, the peak spectrum of the material in the sample does not appear. [25] Thus, only surfaces a few nanometers thick on the grown structure can be examined. Therefore, it is considered that the XPS method is very effective for the analysis of the bonding states and atomic content ratios of heterostructure surfaces. Figure 4 (b) is the magnified graph for the peak of Cu2p3. As shown in Fig. 4 (b), the binding energy of Cu2p3/2 was found at 933.27 eV, and the satellite peak of Cu2p3/2 peak was observed from the CuO sample. The peak spectrum and peak position at 933.27 eV were attributed to the CuO material. [26, 27] The peak intensity of Cu2p3/2 and the satellite of Cu2p3/2 decreased as the Al layer became thicker, which indicates that the CuO nanorods exposed on the surface gradually decreased. The binding energy values of the Cu2p3/2 peak were confirmed at 933.27, 932.65, 932.48, 932.5 and 932.7 eV according to the various Al layer

thicknesses of 0, 10, 30, 50 and 70 nm, respectively. We found that the samples with Al layers of 30 nm had the lowest binding energy values. In general, it is known that the binding energy value can change depending on the bonding state of the atoms and surrounding atoms. That is, as more atoms form a metallic bond, the binding energy is lower, and as the degree of oxidation increases, the binding energy is higher. [11, 27, 28] When Al 30nm sample was compared to the Al 10-nm sample, it was found that more Cu-Al bonds existed, and the Al layer was more uniformly deposited on the CuO nanorods. As the Al layer thickness increased to 50 nm and 70 nm, the distance from the CuO nanorods to the Al surface increased, which resulted in the decrease of the detected Cu-Al bonds and a smaller shift of the Cu peak. Figure 4 (c) shows the magnified graph of the Al 2p peak, and the Al 2p peaks were observed at a binding energy of 74.1 eV for all of the samples except the Al 0-nm sample. The peak for the binding energy at 74.1 eV was attributed to the Al 2p1/2 (metal) and Al 2p (Al oxide), which proves the formation of the Al2O3 phase. Therefore, it is considered that the largest amount of the Al2O3 layer was formed on the Al 30-nm sample because of the strongest peak. Figure 4 (d) shows the magnified graph for the O 1s peak. From the CuO sample, the typical metal oxide peak (right) and the hydrated metal oxide surface peak (left) were observed. [29] The O 1s peaks were observed for the Al deposited samples, and the strongest peak intensity was observed for the Al layer thickness of 30 nm. This means that the Al2O3 layer was formed by the natural oxidation reaction between Al atoms with O atoms existing in the air, and the oxidation reaction proceeded the most in the Al 30-nm sample. In summary, Figure 4 (b), (c) and (d) show that the sample with an Al layer thickness of 30 nm was the most uniformly deposited on the CuO nanorods and had the largest amount of the Al2O3 layer. Figure 5 shows the photoelectrochemical properties of the FTO glass/CuO nanorod/Al 2O3 photoelectrodes according to various Al layer thicknesses. Figure 5 (a) represents the current density according to the applied potentials of 0 to -0.65 V (vs. SCE) under 1-sun illumination (AM 1.5, 100 mW/cm2). As the potential increased in the negative direction, the current increased, which means all samples exhibited the properties of a p-type semiconductor. Figure 5 (b) represents the maximum photocurrent density obtained at 0.55 V (vs. SCE) according to the Al thickness. As the thickness of the Al layer increased to 30 nm, the photocurrent density value tended to increase gradually, and the maximum photocurrent density of -2.26 mA/cm2 was observed with the Al layer thickness of 30 nm. However, the photocurrent density value decreased when the thickness of the Al layer was greater than 30 nm. The photocurrent density value is considered to be highly related to the heterostructure of the CuO nanorod/Al2O3 layers and the formation of Al2O3 layer as a

function of the Al thickness. As discussed above, the existence of the Al2O3 layer was confirmed by XPS measurements, and we found that the 30-nm thick Al layer FTO glass/CuO nanorod/Al2O3 structure photoelectrode was the most successfully grown in terms of the photocurrent density. This result was due to the Al 30-nm sample having a uniform Al2O3 layer along with the one-dimensional CuO nanorods as discussed in the SEM and XPS results. In other words, the Al 30-nm sample maintained the advantages of one-dimensional photoelectrodes and had the largest amount of Al2O3, which is known to have very positive effects on the photoelectrochemical properties, as mentioned above. Generally, the charge separation ability and life time of the electron hole pairs (EHP) are important factors for the improvement of photoelectrochemical properties. In addition, the structural and morphological properties of photoelectrode surfaces have been known to have a great influence on the generation and separation of electron-hole pairs. [4] A similar study was reported by Gao et al., which reported that Al2O3 layers can be used as barrier layers on TiO2 layers to minimize the surface charge recombination, improving the photocurrent density value. [30] On the other hand, the 10-nm thick Al layer sample was too thin to cover the entire surface of the CuO nanorods, so we could not fully expect the positive Al2O3 effect. In the case of the Al layer thicknesses of 50 nm and 70 nm, the photocurrent density values were very low because the surface area of the Al2O3 was reduced by the filled CuO nanorods and the small thickness of the existing Al2O3 layer. This caused these samples to lose the advantages of a onedimensional photoelectrode and the contribution of the Al2O3 layer. A reduced surface area means a decreased contact area with the electrolyte and a decreased photocurrent density by the degraded charge transfer ability. In addition, the Al layer predominantly existed on the CuO nanorod layer rather than the Al2O3 layer, and consequently, the Al layer contributed to the occurrence of a dark current and alleviated the merits of the Al2O3 layer, which could increase the surface charge recombination. Figure 5 (c) represents the theoretical solar-tohydrogen (STH) conversion efficiency calculated by the following equation according to the applied potential: η = I (1.23 – V)/Jlight ×100%

(1)

where I is the photocurrent (mA), 1.23 is the theoretical potential (V) required for water splitting, V is the applied external potential (V), and Jlight is the irradiance of the solar simulator (mW/cm2). The results showed that the theoretical STH conversion efficiency was similar to the photocurrent density result, and the maximum STH conversion efficiency was 1.61% at 0.51 V (vs. RHE) for the Al 30-nm sample. Figure 6 shows the impedance analysis results for FTO glass/CuO nanorod/Al2O3 photoelectrode

samples with different Al thicknesses using a 1 M KOH electrolyte without light irradiation. The flat band potential and acceptor density were obtained from the following equation: 2 1/𝐶𝑠𝑐 = (2/q𝜀𝑠 𝜀0 𝑁𝐴 ) [ 𝑉𝑎𝑝𝑝 – (𝑉𝑓𝑏 – 𝐾𝑏 𝑇/𝑞)]

(2)

where 𝜀0 is the permittivity in vacuum, 𝜀𝑠 is the relative permittivity of the CuO electrode, q is the electronic charge of the carrier, 𝑁𝐴 is the acceptor density concentration, T is the absolute temperature, 𝐾𝑏 is the Boltzmann constant, 𝐶𝑆𝐶 is the capacitance of the space charge, 𝑉𝑓𝑏 is the flat-band potential of the semiconductor/electrolyte interface, and 𝑉𝑎𝑝𝑝 is the applied potential to the electrode. [5, 13] The flat band potential values obtained in this study were in the range of 0.142 ~ 0.17 V, which were in the previously reported value range of 0 ~ 0.3 V for p-type CuO. [6, 31, 32] It was found that the flat band potential results had very similar trends to the photocurrent density results as well as the acceptor density results. As the Al layer thickness increased to 30 nm, the values for the flat band potential and the acceptor density increased, and then, these values were decreased in the Al layer thicknesses of 50 nm and 70 nm. We obtained the highest flat band potential of 0.17 V and the highest acceptor density of 10.3×1023 cm3 using the Al 30-nm sample, which had the maximum photocurrent density value, as shown in Fig. 5 (b). Chiang et al. reported that porous nanostructured CuO photoelectrodes exhibit a high charge carrier density and a high photocurrent density because the charge carriers have a short travel distance when the electrolyte penetrates into nanostructures with a high surface area at the electrode/electrolyte interface. [6] In this study, the Al 30-nm sample was considered to have the largest Al2O3 surface area; therefore, the maximum charge carriers had a short travel distance. Consequently, we found that the Al 30-nm sample had the highest photocurrent density, flat band potential and maximum acceptor density due to the positive effects of the Al2O3 layer and the one-dimensional shape of the CuO nanorod/Al2O3 heterostructure photoelectrodes.

Conclusion A study of the photoelectrochemical properties of CuO nanorod/Al2O3 heterostructure photoelectrodes was conducted with different Al layer thicknesses. Based on the investigation of the morphological, optical, and structural properties and the qualitative analysis, the Al 30-nm sample had the most uniform Al2O3 layer over the CuO nanorods. The Al 30-nm photoelectrode maintained the merits of the one-dimensional CuO nanorod

structures and had the largest amount of the Al2O3 layer, which had a positive effect by decreasing the surface charge recombination. By analyzing the impedance results of the prepared photoelectrodes, the highest flat band potential of 0.170 V and the highest charge carrier concentration of 10.3×1023 cm3 were obtained from the Al 30-nm sample. As a result, the Al 30-nm sample generated the highest photocurrent density of -2.26 mA/cm2 (0.55 V vs. SCE) and the highest theoretical solar-to-hydrogen (STH) conversion efficiency of 1.61%.

Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2016R1D1A3B01008959).

References

[1] M. Gratzel, Photoelectrochemical cells, Nature 414 (2001) 338-344. [2] A. Fujishima, K. Honda, Photolysis-decomposition of water at the surface of an irradiated semiconductor, Nature 238 (1972) 37-38. [3] J. R. McKone, N. S. Lewis, H. B. Gray, Will Solar driven Water-splitting Devices See the Light of Day?, Chem. Mater. 26 (2014) 407-414. [4] H. M. Chen, C. K. Chen, R. S. Liu, L. Zhang, J. Zhang, D. P. Wilkinson, Nano-architecture and material designs for water splitting photoeletrodes, Chem. Soc. Rev. 41

(2012) 5654-5671.

[5] S. Masudy-Panah, R. S. Moakhar, C. S. Chua, H. R. Tan, T. I. Wong, D. Chi, G. K. Dalapati, Nanocrystal engineering of sputter-grown CuO Photoeletrode for Visible-Light-Driven Electrochemical Water Spliting, ACS appl. Mater. Interfaces 8 (2016) 1206-1213. [6] C. Chiang, Y. Shin, K. Aroh, S. Ehrman Copper oxide photocathode prepared by a solution based process, Int. J. Hydrogen Energy 37 (2012) 8232-8239. [7] Y. S. Chaudhary, A. Agrawal, R. Shrivastav, V. R. Satsangi, S. Dass, A study on the photoeletrochemical properties of copper oxide thin films, Int. J. Hydrogen Energy 29 (2004) 131-134. [8] K. Sivula, F. L. Formal, M. Gratzel, Solar Water Splitting: Progress Using Hematite (α-Fe2O3) Photoeletrodes, ChemSusChem. 4 (2011) 432-449.

[9] A. A. Dubale, C. Pan, A. G. Tamirat, H. Chen, W. Su, C. Chen, J. Rick, D. W. Ayele, B. A. Aragaw, J. Lee, Y. Yang, B. Hwang, Heterostructured Cu2O/CuO decorated with nickel as a highly efficient photocathode for photoeletrochemical water reduction, J.

Mater. Chem. A 3 (2015) 12482-12499.

[10] X. Guo, P. Diao, D. Xu, S. Huang, Y. Yang, T. Jin, Q. Wu, M. Xiang, M. Zhang, CuO/Pd composite photocathode for photoeletrochemical hydrogen evolution reaction, Int. J. Hydrogen Energy 39 (2014) 76867696. [11] Y. Lim, C. S. Chua, C. J. J. Lee, D. Chi, Sol-gel deposited Cu2O and CuO thin films for photocatalytic water splitting, Phys. Chem. Chem. Phys. 16 (2014) 25928-25934. [12] P. Wang, X. Wen, R. Amal, Y. H. Ng Introducing a protective interlayer of TiO 2 in Cu2O-CuO heterojunction thin film as a highly stable visible light photocathode, RSC adv. 5 (2015) 5231-5236. [13] S. Ho-kimura, S. J. A. moniz, J. Tang, I. P. Parkin, A Method for Synthesis of Renewable Cu 2O Junction Composite Electrodes and Their Photoelectrochemical Properties, ACS sustainable Chem. Eng. 3 (2015) 710717. [14] C. Hu, K. Chu, Y. Zhao, W. Y. Teoh, Efficient Photoelectrochemical Water Splitting over Anodized pType NiO Porous Films, ACS appld. Mater. Interfaces 6 (2014) 18558-18568. [15] M. J. Choi, J. Jung, M. Park, J. Song, J. Lee, J. H. Bang, Long-term durable silicon photocathode protected by a thin film Al2O3/SiOx layer for photoelectrochemical hydrogen evolution, J. Mater. Chem. A. 2 (2014) 2928-2933. [16] H. Oh, k. Jang, C. Chi, impedance Characteristics of Oxide layers on Aluminum, Bull. Korean Chem. Soc. 20 (1999) 1340-1344. [17] J. M. Lackner, W. Waldhauser, R. Berghauser, R. Ebner, B. Major, A. Flan, G. Jakopic, NEW TRENDS IN COATING: ROOM TEMPERATURE DEPOSITION OF TITANIUM-BASED FILMS, Materials Engineering 140, (2004) 611-615. [18] L. Zhang, H. C. Jiang, C. Liu, J. W. Dong, P. Chow, Annealing of Al2O3 thin films prepared by atomic layer deposition, J. Phys. D: Appl. Phys. 40 (2007) 3707-3713. [19] H. B. Oh, H. Ryu, W. J. Lee, Vertical Growth of Cupric Oxide Nanorods for photoelectrode Using a Modified Chemical Bath Deposition Method, J. Electrochem. Soc. 161 (2014) H578-H582. [20] H. B. Oh, H. Ryu, W. J. Lee, Effects of Growth Temperature on Cupric Oxide Nanorod Photoelectrodes Using a Modified Chemical Bath Deposition, J. Electrochem. Soc. 161 (2014) H633-H636. [21] H. B. Oh, H. Ryu, W. J. Lee, Effects of copper precursor concentration on the growth of cupric oxide nanorods for photoelectrode using a modified chemical bath deposition method, J. Alloy. Compd. 620 (2015) 55-59.

[22] Z. Zhang, H. Sun, X. Shao, D. Li, H. Yu, M. han, Three-dimensionally Oriented Aggregation of a Few Hundred Nanoparticles into Monocrystalline Architectures, Adv. Mater. 17 (2005) 42-47. [23] V. F. Korzo, V. N. Chernyaev, Band structure of localized electron sites in non-crystalline Al2O3 films, Thin Solid Films 34 (1976) 381-385. [24] W. M. M. Kessels, J. A. V. Delft, G. Dingemans, M. M. Mandoc, Review on the prospects for the use of Al2O3 for high-efficiency solar cells, Workshop (Dept. of Applied Physics, Eindhoven University of Technology (TU/e), The Netherlands) [25] C. C. Chusuei, D. W. Goodman, X-Ray Photoelectron Spectroscopy, in Encyclopedia of Physical Science and Technology (Ed: R. A. Meyers), Elsevier B. V., ScienceDirect, 17 (2004) 921-938. [26] G. Avgouropoulos, T. Ioannides, Selective CO oxidation over CuO-CeO2 catalysts prepared via the ureanitrate combustion method, Appl. Catal., A 244 (2003) 155-167. [27] S. Poulston, P. M. Parlett, P. Stone, M. Bowker, Surface Oxidation and Reduction of CuO and Cu2O studied Using XPS and XAES, Surf. Interface Anal. 24 (1996) 811-820. [28] S. Stankovich, D. A. Dikin, R. D. Piner, Ke. A. Kohlhaas, A. Kleinhames, Y. Jia, Y. Wu, S. T. Nguyen, R. S. Ruoff, Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide, Carbon 45 (2007) 1558-1565. [29] J. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Systematic XPS studies of metal oxides, hydroxides and peroxides, Phys. Chem. Chem. Phys. 2 (2000) 1319-1324. [30] X. Gao, D. Guan, J. Huo, J. Chen, C. Yuan, Free standing TiO2 nanotube array electrodes with an ultra-thin Al2O3 barrier layer and TiCl4 surface modificationfor highly efficient dye sensitized solar cells, Nanoscale 5 (2013) 10438-10446. [31] C. Chiang, Y. Shin, S. Ehrman, Li Doped CuO Film Electrodes for Photoelectrochemical cells, J. Electrochem. Soc. 159 (2012) B227-B231. [32] K. L. Hardee, A. J. Bard, Semiconductor Electrodes X. Photoelectrochemical Behavior of Several polycrystalline Metal Oxide Electrodes in Aqueous Solutions, J. Electrochem. Soc. 124 (1977) 215-224.

Figure captions Fig. 1. Top view SEM images of the CuO nanorod/Al2O3 structures with various Al layer thicknesses. (a) 0 nm (b) 10 nm (c) 30 nm (d) 50 nm (e) 70 nm and (f) a cross-section image of the FTO glass/CuO nanorod/Al2O3 structure. Fig. 2. (a) XRD spectra and (b) peak intensities of the (111), (11 ̅ ) and (020) planes with various Al layer thicknesses.

2

Fig. 3. UV-vis spectra of CuO nanorod/Al2O3 structures. (a) (αhν) versus hν plot and (b) optical energy bandgap values with various Al layer thicknesses.

Fig. 4. The XPS data of the CuO nanorod/Al2O3 structures with various Al layer thicknesses from 0 to 70 nm. (a) XPS spectra, (b) Cu 2p3 peak, (c) Al 2p peak, and (d) O 1s peak.

Fig. 5. The PEC properties of CuO nanorod/Al2O3 photoelectrodes with various Al layer thicknesses.

(a) Current densities at applied potentials ranging from 0 to -0.65 V (vs. SCE), (b) photocurrent densities measured at -0.55 V (vs. SCE), and (c) theoretical solar to hydrogen efficiency values.

Fig. 6. The EIS data of the CuO nanorod/Al2O3 photoelectrodes with various Al layer thicknesses from 0 nm to 70 nm in dark measurement conditions. (a) Mott-Schottky plot and (b) flat-band potential and acceptor density plot.