CHEMSUSCHEM FULL PAPERS DOI: 10.1002/cssc.201402025
Cobalt-Phosphate-Assisted Photoelectrochemical Water Oxidation by Arrays of Molybdenum-Doped Zinc Oxide Nanorods Yan-Gu Lin,*[a] Yu-Kuei Hsu,[b] Ying-Chu Chen,[c] Bing-Wei Lee,[d] Jih-Shang Hwang,[d] LiChyong Chen,*[c] and Kuei-Hsien Chen*[c, e] We report the first demonstration of cobalt phosphate (Co-Pi)assisted molybdenum-doped zinc oxide nanorods (Zn1xMoxO NRs) as visible-light-sensitive photofunctional electrodes to fundamentally improve the performance of ZnO NRs for photoelectrochemical (PEC) water splitting. A maximum photoconversion efficiency as high as 1.05 % was achieved, at a photocurrent density of 1.4 mA cm2. More importantly, in addition to achieve the maximum incident photon to current conversion efficiency (IPCE) value of 86 %, it could be noted that the IPCE of Zn1xMoxO photoanodes under monochromatic illumination (450 nm) is up to 12 %. Our PEC performances are com-
parable to those of many oxide-based photoanodes in recent reports. The improvement in photoactivity of PEC water splitting may be attributed to the enhanced visible-light absorption, increased charge-carrier densities, and improved interfacial charge-transfer kinetics due to the combined effect of molybdenum incorporation and Co-Pi modification, contributing to photocatalysis. The new design of constructing highly photoactive Co-Pi-assisted Zn1xMoxO photoanodes enriches knowledge on doping and advances the development of highefficiency photoelectrodes in the solar-hydrogen field.
Introduction The efficient photoelectrolysis of water into hydrogen poses a long-standing challenge, with promise for solar energy conversion and storage in the form of synthetic fuels.[1, 2] Desirable for photoelectrolysis are strong visible-light absorption, efficient charge-carrier separation and transport, and facile interfacial charge-transfer kinetics, especially for the water oxidation reaction.[3] Despite important advances and intensive research and development efforts worldwide, no photoelectrochemical (PEC) system for solar hydrogen production has met the required performance benchmarks. In addition to the search for [a] Dr. Y.-G. Lin National Synchrotron Radiation Research Center Hsinchu 30076 (Taiwan) E-mail: [email protected] [b] Prof. Y.-K. Hsu Department of Opto-Electronic Engineering National Dong Hwa University Hualien 97401 (Taiwan) [c] Y.-C. Chen, Dr. L.-C. Chen, Dr. K.-H. Chen Center for Condensed Matter Sciences National Taiwan University Taipei 10617 (Taiwan) [d] B.-W. Lee, Prof. J.-S. Hwang Institute of Optoelectronic Sciences National Taiwan Ocean University Keelung 202 (Taiwan) [e] Dr. K.-H. Chen Institute of Atomic and Molecular Sciences Academia Sinica Taipei 10617 (Taiwan) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201402025.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
new photoactive materials, another fruitful approach is to modify the electronic properties of known semiconductors to further improve their photoactivity for water-splitting reactions. Abundant and inexpensive oxide semiconductors, such as ZnO or TiO2, have been recognized as promising photoelectrode materials, but their photoconversion efficiency is substantially limited owing to their large band-gaps and rapid charge recombination.[4, 5] Many works have been devoted to the mitigation of these intrinsic limitations, by using strategies involving doping and nanostructuring.[6–10] One-dimensional (1D) nanorod/nanowire-array photoelectrodes, offering large surface areas and short diffusion distances for photogenerated minority carriers, are expected to facilitate charge separation.[6, 11] Meanwhile, heteroatom doping into oxide semiconductors can not only increase the electronic conductivity of the semiconductor but can also narrow the band-gap, allowing to utilize more visible light.[6, 8, 9, 12–14] For example, nitrogendoped ZnO nanowires and nanotetrapods were found to ensure rapid electronic transport, efficient charge separation, and enhanced visible-light absorption.[6, 9] Incorporating carbon has also been shown to enhance the photoconversion efficiency of porous ZnO nanostructures for PEC water splitting.[13] Recently, doping with 4d transition metals, such as molybdenum, was reported to remarkably enhance the PEC performance of many photoanodes, including TiO2, BiVO4, and Fe2O3.[15–18] Nevertheless, 1-D molybdenum-doped ZnO nanostructures have not been reported for PEC water-splitting. In particular, while most investigations of Mo-doped ZnO have focused on the use of transparent conducting oxides for flat-panel display de-
ChemSusChem 0000, 00, 1 – 8
&1&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
vices,[19] thorough attempts to exploit these interesting doping effects for solar-energy-driven applications are lacking. On the other hand, it is equally important to accelerate the rate of oxygen evolution at photoanodes by using catalysts. Water oxidation generally occurs at significant overpotentials because it requires the removal of a total of four electrons and four protons from two molecules of water to form one molecule of oxygen. Therefore, surface modification of the photoanodes with Figure 1. SEM images of a) pristine, and b) Mo-doped ZnO NRs on FTO substrate. Inset: Photo of an as-synthean electrocatalyst is needed to sized sample. c) HRTEM image of Mo-doped ZnO NRs. d–g) Elemental mapping images of Zn, O, and Mo. achieve high water oxidation activity at low overpotentials. In order to lower the required overpotential, various catalysts have been intensively studied, such as IrO2, cobalt ions, and the cobalt phosphate catalyst, “Co-Pi”.[20, 21] Co-Pi in particular has received a lot of recent attention because it uses earthabundant elements, shows effective water oxidation characteristics, and is stable over time due to its “self-healing” mechanism.[22, 23] Consequently, it has also been applied to many potential photoanodes, including ZnO, BiVO4, Si, and Fe2O3, and has shown improvements in both current onset potential and photocurrent density.[18, 24–26] Herein, we show that the uniform deposition of Co-Pi-based oxygen evolution catalysts onto Mo-doped ZnO nanorods (NRs) is key for efficient photoelectrolysis, allowing for the preparation of next-generation PEC devices that could benefit from the light-scattering ability of high-aspect-ratio NRs, the unique advantages of molybdenum incorporation, and the excellent electrocatalysis characteristics of Co-Pi catalysts (Supporting Information, Figure S1). For ZnO-based photoelectrodes to be used at a scale commensurate with the global energy demand, methods to fabricate both doped and 1-D nanostructured ZnO using rapid and inexpensive preparation techniques are necessary. Electrodeposition is a simple and cost-effective technique that is very promising from this point of view. In this work, we demonstrate a facile and reliable procedure for fabrication of Co-Pi/ Zn1xMoxO hybrid nanoelectrodes in the absence of any surfactant agent. The Zn1xMoxO NR arrays were first grown on fluorine-doped tin oxide (FTO) glass substrates by using an in situ electrodeposition method. In the case of Co-Pi catalysts, a modified electrochemical deposition approach at room temperature reported earlier led to rapid formation of cable-like nanocomposites.[18] Herein, the PEC performance of Co-Pi/ Zn1xMoxO hybrid nanosystems in solar water-splitting without Figure 2. a) X-ray diffraction patterns of undoped and Mo-doped ZnO NRs. sacrificial reagents is systematically evaluated. Furthermore, we Inset: Magnified view of the region 368–378. b) Micro-Raman spectra of unalso perform microstructural characterization, elemental analydoped and Mo-doped ZnO NRs. sis, and electrochemical impedance spectroscopy (EIS) on the Mo-doped ZnO photoelectrodes to confirm the molybdenum doping and understand the correlation with improved PEC 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&2&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS performance and the surface photochemistry of ZnO. To the best of our knowledge, this represents the first case of realizing Co-Pi assisted Zn1xMoxO NR-array photoanodes for PEC water oxidation in a nonsacrificial electrolyte.
Results and Discussion
www.chemsuschem.org sults, which were performed to understand the electronic structure of the molybdenum-containing ZnO NRs (Figure 3). The O 1s core level spectra of Mo-doped ZnO NRs in Figure 3 b show a main peak at 531.2 eV, which can be deconvoluted into two peaks. The low binding energy peak at 530.6 eV, denoted by O1, corresponds to ZnO bonding, while the higher one at 532.5 eV (O2) is often attributed to oxygen vacancies in ZnO.[13] The XPS spectra for Mo 3d5/2 are also presented in Figure 3 a, showing a slight down-shift of the position of the main peak as compared to hexavalent Mo (Mo6 + , 232.5 eV).[17, 18] This is likely due to partial reduction of Mo6 + , which is consistent with the increase in density of oxygen vacancies (as electron donor) in ZnO through Mo incorporation, which involves transfer of electrons to Mo. It has also been reported that the Zn 2p binding energy shifts to a slightly lower value if Mo substitutes for O, owing to strong Mo–Zn bonding ( 10 kcal mol1).[28] However, the absence of any significant changes in the Zn 2p binding energies as determined by XPS (Figure 3 a) seems to imply that it is Zn, instead of O, which is substituted by Mo. Meanwhile, the Mo concentration in the Zn1xMoxO NRs was also estimated by XPS measurement to be about 1.7 at %. The Mo concentration can be tuned by changing the Na2MoO4 concentration in experimental conditions. Extended X-ray absorption fine structure (EXAFS) spectra taken around the Zn K-edge of Mo-containing ZnO NRs are shown in Figure 3 c. An apparent decrease in relative intensity and interatomic distance of the first shell and second shell (ZnO and ZnZn bonding) over ternary Zn1xMoxO NRs
A scanning electron microscopy (SEM) image of pristine ZnO NRs reveals a high surface density (ca. 109 rods cm2) with diameters of about 150 50 nm (Figure 1 a). Clearly, the in situ doping process does not alter the morphology of the nanostructures (Figure 1 b). A high-resolution transmission electron microscopy (TEM) image of doped ZnO is shown in Figure 1 c, and enables to determine the spacing of the (101) lattice plane of hexagonal ZnO crystal as 0.24 nm. More detailed TEM images with energy dispersive X-ray (EDX) elemental mapping of Zn, O, and Mo are shown in Figure 1 d–g, and confirms the presence of uniformly dispersed Mo atoms inside the ZnO NRs. X-ray diffraction (XRD) patterns of undoped and Mo-doped ZnO NRs are compared in Figure 2 a, showing that both ZnO NRs are of wurtzite structure and that there is no significant phase change upon Mo incorporation. In addition, an effect of Mo substitution in the ZnO matrix was a significant up-shift in the (101) peak of ternary Zn1xMoxO NRs relative to pure ZnO, indicating a decrease in the lattice parameter of ZnO (Figure 2 a, inset). This slight contraction of the ZnO lattice upon Mo incorporation suggests that the smaller-sized Mo ion (Mo6 + , 0.62 ) has most likely substituted for the larger Zn ion (Zn2 + , 0.74 ). To gain further insight into the evolution of defects with Mo incorporation, micro-Raman analyses were performed (Figure 2 b). The presence of a sharp and dominant E2 (high) mode at 438 cm1, which is associated with the vibration of oxygen atoms in ZnO, indicates the wurtzite nature of ZnO.[27] Significant variations in the position as well as the full width at half maximum (FWHM) of the peak corresponding to the E2 (high) mode occur, owing to the formation of compressive strain resulting from Mo doping. Furthermore, the peak at 575 cm1 corresponding to the A1 (LO) mode originates from oxygen vacancies in ZnO,[27] implying Mo incorporation may induce oxygen vacancies in the ZnO lattice. This could be further confirmed by X-ray photoelectron spectroscopy (XPS) and X-ray ab- Figure 3. XPS spectra of a) Mo 3d, Zn 2p3/2, and b) O 1s transitions for Mo-doped ZnO NRs. c) Zn K-edge EXAFS sorption spectroscopy (XAS) re- spectra, and d) O K-edge XAS spectra of Mo-doped ZnO NRs and pure ZnO samples. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&3&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS would be observed compared to pure ZnO. This can be mainly attributed to the strong distortion of the ZnO lattice resulting from the positioning of Mo in the ZnO framework. The slight decrease of the ZnO lattice parameter upon Mo incorporation suggests that the smaller-sized Mo atom has most likely substituted for the larger Zn atom, which is consistent with XRD and XPS results. We further fitted the results of pure ZnO and Mocontaining ZnO samples at the Zn K-edge (Supporting Information, Table S2). Meanwhile, Figure 3 d shows the O K-edge XAS spectra of ternary Zn1xMoxO NRs. The broadening of the absorption peak at ca. 537 eV for Mo-doped ZnO NRs, compared to pure ZnO structure, can be assigned to the presence of oxygen vacancies, suggesting that Mo incorporation induces oxygen vacancies in the ZnO lattice.[29] In addition, the overall intensities of the features of Mo-doped ZnO NRs are higher than those of pure ZnO structures, reflecting the increase in number of unoccupied O 2p–Zn 4s hybrid states. The reason is Mo incorporation, which may create a high density of oxygen vacancies that serve as electron donors, so that more electrons are transferred from O 2p states to Mo. To verify that Mo incorporation enhances light harvesting, we measured UV-vis absorption spectra of Mo-containing ZnO nanoelectrodes (Figure 4 a). In comparison to pristine ZnO, it is apparent that Mo-doped ZnO exhibits a significant enhancement in the absorption of visible light. This finding is commensurate with the red-shift of the band gap in the ternary Zn1xMoxO NRs. Besides, to quantify the PEC performance, the incident photon to current conversion efficiency (IPCE) was measured to study the photoresponse of pristine and Mo-containing ZnO NRs as a function of incident light wavelength (Figure 4 b). Pristine ZnO has a strong photoresponse only in the near-UV region. Surprisingly, the doped sample displays substantial photoactivity in the visible-light region, from 400 to 550 nm, in addition to strong photoresponse in the near-UV region. Particularly, the IPCE of Mo-doped ZnO photoelectrodes at the monochromatic wavelength of 450 nm is up to 12 %. Mo modification substantially improves the light utilization and conversion efficiency in the visible region of interest, as compared to the pure ZnO structure. The increase of the IPCE in the visible range may be probably attributed to the bandgap narrowing of ZnO and/or amorphous MoO2 phase in the ZnO. These results are in good agreement with the UV-visible absorption data (Figure 4 a). To further determine the exact efficiency of the water-splitting reaction under simulated AM 1.5 illumination, systematic PEC measurements without sacrificial reagents were carried out in a 0.5 m aqueous solution of Na2SO4 with a pH of 6.8. Figure 5 a depicts a set of linear-sweep voltammograms recorded on pristine ZnO, Mo-doped ZnO, and Co-Pi/Mo-doped ZnO photoanodes in the dark and under light illumination of 100 mW cm2. Linear-sweep voltammagrams in the dark from 0.2 to + 1 V show the current density range of 107 A cm2. The Mo-containing ZnO NRs manifest a pronounced photocurrent, which increases to 0.9 mA cm2 under illumination. Clearly, the photocurrent density of Mo-containing ZnO nanoelectrodes is about 2.3 times greater than that of pristine ZnO, suggesting that ZnO incorporated with Mo can effectively improve 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
Figure 4. a) Absorption spectra of undoped and Mo-doped ZnO NRs. b) Measured IPCE spectra of pristine ZnO NRs and Mo-doped ZnO NRs in the region of 300–650 nm at a potential of 0.4 V (vs. Pt) in two-electrode system.
PEC water-splitting. The maximum photocurrent density (1.4 mA cm2) generated from Co-Pi/Mo-doped ZnO hybrid nanoelectrodes is about 3.5 times greater than that of pristine ZnO. These results are comparable to those of many oxidebased photoanodes in recent reports.[6, 8, 18, 30–32] We also measured the amount of hydrogen and oxygen evolved, determined by the gaseous product in the PEC cell. Generally, in PEC water splitting, two reaction processes happen: Photoanode reaction : 2 H2 O ! O2 þ 4 Hþ þ 4 e
ð1Þ
Photocathode reaction : 2 H2 O þ 2 e ! H2 þ 2 OH
ð2Þ
The rate of H2 and O2 evolution measured for our Co-Pi/Modoped ZnO hybrid nanocomposites are about 21 and 11 mmol h1, respectively. The yield of H2 after illumination is much lower than the calculated theoretical value, based on the photocurrent data and Faraday’s law (the theoretical value ChemSusChem 0000, 00, 1 – 8
&4&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
Figure 5. a) Linear sweep voltammagrams, collected at a scan rate of 10 mV s1 at applied potentials from 0.2 to + 1 V (vs. Ag/AgCl) for pristine ZnO, Mo-doped ZnO, and Co-Pi/Mo-doped ZnO photoanodes in the dark and under light illumination of 100 mW cm2. b) Photoconversion efficiency of the PEC cells for Mo-doped ZnO and Co Pi/Mo-doped ZnO photoanodes as a function of applied potential.
is 26 mmol h1, in the ideal case). This is possibly due to the formation of water from H2 and O2 on the photoelectrodes. Furthermore, in order to quantitatively explore the photoactivity of Mo-doped ZnO and Co-Pi/Mo-doped ZnO NRs, the efficiency (h) of applied bias photon-to-current generation was calculated using an equation in which the contribution due to applied potential is subtracted from the total efficiency.[33] Significantly, it is worth mentioning that the plot of efficiency versus applied potential (Figure 5 b) shows that the maximum value of efficiency is 1.05 % at an applied potential of + 0.3 V, which is also higher than recent reported values for other oxide-based photoanodes.[6, 9, 34] To understand the intrinsic electronic properties of ternary Zn1xMoxO NRs in the electrolyte solution, we measured electrochemical impedance spectroscopy (EIS) in the dark coupled to Mott–Schottky analysis, commonly used to determine both donor density and flatband potential at semiconductor/liquid contacts. Capacitances at the semiconductor/electrolyte interface with the use of an equivalent circuit are described by a Mott–Schottky plot.[35] All 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ZnO NR samples show a positive slope in the Mott– Schottky plots, as expected for n-type semiconductors (Figure 6). The estimated flat-band potentials of the pure ZnO and Mo-doped ZnO are 0.28 V and 0.34 V, respectively. Apparently, Mo-containing ZnO has a more negative flat-band potential than pure ZnO on account of the Mo-doping effect. In addition, the calculated donor densities of the pure ZnO and Mo-doped ZnO are 6 1018 and 3 1020 cm3, respectively. The enhanced donor density is due to the created oxygen vacancies, which are known to be an electron donor for oxide-based semiconductors.[7] The increased donor density improves the charge transport in ZnO, as well as the electron transfer at the interface between the ZnO and FTO substrate. Moreover, the increased electron density is expected to shift the Fermi level of ZnO toward the conduction band. The upward shift of the Fermi level facilitates charge separation at the semiconductor/electrolyte interface, by increasing the degree of band bending at the ZnO surface. Therefore, the reduced band gap, increased carrier concentration, and improved interfacial catalysis are believed to be the major reasons for the improvement in photoactivity of PEC water-splitting based on Co-Pi modified Mo-doped ZnO NR photoelectrodes. Finally, the structural stability during PEC water splitting is another important factor to evaluate for the potential PEC materials. Time-dependent measurement carried out on a representative Co-Pi modified Mo-doped ZnO NRs displays a minimal degradation rate after 11 h of continuous running in the water-oxidation process at 1 V (Supporting Information, Figure S3). Meanwhile, after 11 h of continuous PEC reaction, ICP-MS was applied to monitor the concentration of Zn2 + ions in the electrolyte. Only a trace amount of Zn2 + ion could be detected in the photoreaction solution, showing that the photocorrosion factor would be minor for our developed Co-Pi modified Mo-doped ZnO NRs.
Figure 6. Mott–Schottky plots for pure ZnO and Mo-doped ZnO NRs in 0.5 m Na2SO4.
ChemSusChem 0000, 00, 1 – 8
&5&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS Conclusions An efficient solar hydrogen system is developed based on CoPi assisted ternary Zn1xMoxO nanorod (NR) photoelectrodes. This is the first example of in situ doping of zinc oxide nanorods with molybdenum for solar water splitting in a nonsacrificial electrolyte. Relative to pristine ZnO NR photoanodes, the photocurrent of Co-Pi/Zn1xMoxO NRs increases substantially (up to ca. 250 %). The highest photoconversion efficiency obtained from our Mo-containing ZnO NRs is ca. 0.47 %, much better than undoped ZnO NRs, and subsequent modification with Co-Pi further enhances the photoconversion efficiency of the Mo-containing ZnO NRs to ca. 1.05 %. Furthermore, concerning the maximum incident photon to current conversion efficiency (IPCE) value of 86 %, it could be noted that the IPCE of Co-Pi-assisted Zn1xMoxO photoanodes under monochromatic illumination (450 nm) is up to 12 %. The improvement in photoactivity of PEC water-splitting may be attributed to the enhanced visible-light absorption, increased charge-carrier densities, and improved interfacial charge-transfer kinetics due to the combined effects of molybdenum incorporation and Co-Pi modification contributing to photocatalysis. Because the Co-Pi assisted Zn1xMoxO NR photoanodes are convenient to fabricate and highly active, they can serve as a useful substitution for ZnO in a variety of solar-energy-driven applications including PEC water splitting, photocatalysis, and solar cells.
Experimental Section The Zn1xMoxO NR arrays were grown on an FTO glass, serving as substrate and working electrode, by typical electro-deposition method. A piece of Pt foil and a standard Ag/AgCl electrode were used as the counter and reference electrodes, respectively. Aqueous 0.1–0.3 m KCl and 5–10 mm ZnCl2 solutions with 25–75 mm Na2MoO4 were used as electrolytes. H2O2 (5–10 mm) was used as source of oxygen. The applied potential was operated at 0.8 V versus Ag/AgCl for 30 min and the solution temperature was maintained at 80 8C. Then, the resulting Mo-doped ZnO NR arrays were heated in vacuum to 450 8C for 3–6 h. Finally, Co-Pi catalysts were deposited potentiostatically at 1.1 V versus Ag/AgCl in a solution of 0.1–0.2 mm cobalt nitride in 0.1 m potassium phosphate buffer at pH 7 for several minutes. For material characterization, SEM measurements were made on a JEOL 6700 field-emission SEM. For obtaining TEM images, the Mo-doped ZnO NR products on the FTO substrate were scratched and dispersed onto a carbon-coated Cu grid, and analyzed using a JEOL JEM-2100 TEM system. The XRD analyses were performed on a Bruker D8 Advance diffractometer with a Cu target emitting X-ray radiation at 1.54 (40 kV, 40 mA). Micro-Raman analyses were performed on a Jobin Yvon Labram HR800 spectrometer. XPS spectra were obtained using a Microlab 350 system. XAS analyses were performed in National Synchrotron Radiation Research Center. UV-vis absorption spectra were carried out with a JASCO V670 instrument. In PEC water-splitting reaction, a Co-Pi/Mo-doped ZnO photoanode was used as the working electrode with surface area of 0.5–1 cm2, a platinum plate as counter electrode, and Ag/ AgCl as reference electrode. All PEC studies were operated in a 0.5 m Na2SO4 solution (pH 6.8) as supporting electrolyte medium by using an Electrochemical Multichannel Solartron Analytical System. The water-splitting photoelectrode was illuminated with 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org the AM 1.5G simulated solar light at 100 mW cm2. This illumination source matches the shape and intensity of the Air Mass 1.5 Global (AM 1.5 G) G173 standard set forth by the American Society of Testing and Materials. The IPCE was calculated using the following equation: IPCE [%] = (1240 photocurrent density [mA cm2] 100)/(wavelength [nm] photon flux [mW cm2]). IPCE measurements were performed by 2-electrode system under monochromatic irradiation, emitted from a 150 W Xe lamp equipped with bandpass filters and a variable neutral density filter. Monochromatic photon fluxes at each wavelength were measured using an optical photodiode power meter and were adjusted to around 0.1 mW cm2 using the neutral density filter. Finally, the PEC cell was constructed to allow the collection of gases generated at the electrodes. For products analysis, the amounts of evolved gases after illumination were determined with gas chromatography (Shimadzu, MS-5A column, Ar carrier).
Acknowledgements This work was supported by the Ministry of Science and Technology and National Synchrotron Radiation Research Center. We gratefully thank MOST, NSRRC, IAMS, and NTU for financial support in this project. Keywords: doping · molybdenum · photocatalysis · water splitting · zinc oxides [1] N. S. Lewis, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2006, 103, 15729 – 15735. [2] Y. G. Lin, Y. K. Hsu, A. M. Basilio, Y. T. Chen, K. H. Chen, L. C. Chen, Opt. Express 2014, 22, A21 – A27. [3] Z. Li, W. Luo, M. Zhang, J. Feng, Z. Zou, Energy Environ. Sci. 2013, 6, 347 – 370. [4] J. Shi, X. Wang, Energy Environ. Sci. 2012, 5, 7918 – 7922. [5] Y. G. Lin, Y. K. Hsu, Y. C. Chen, S. B. Wang, J. T. Miller, L. C. Chen, K. H. Chen, Energy Environ. Sci. 2012, 5, 8917 – 8922. [6] X. Yang, A. Wolcott, G. Waang, A. Sobo, R. C. Fitzmorris, F. Qian, J. Z. Zhang, Y. Li, Nano Lett. 2009, 9, 2331 – 2336. [7] G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang, Y. Li, Nano Lett. 2011, 11, 3026 – 3033. [8] S. Hoang, S. Guo, N. T. Hahn, A. J. Bard, C. B. Mullins, Nano Lett. 2012, 12, 26 – 32. [9] Y. Qiu, K. Yan, H. Deng, S. Yang, Nano Lett. 2012, 12, 407 – 413. [10] M. Xu, P. Da, H. Wu, D. Zhao, G. Zheng, Nano Lett. 2012, 12, 1503 – 1508. [11] Y. G. Lin, Y. C. Chen, C. Feldmann, J. T. Miller, L. C. Chen, K. H. Chen, Y. K. Hsu, ChemCatChem. 2014, 1684 – 1690. [12] G. Liu, L. C. Ying, J. Wang, P. Niu, C. Zhen, Y. Xie, H. M. Cheng, Energy Environ. Sci. 2012, 5, 9603 – 9610. [13] Y. G. Lin, Y. K. Hsu, Y. C. Chen, L. C. Chen, S. Y. Chen, K. H. Chen, Nanoscale 2012, 4, 6515 – 6519. [14] I. S. Cho, C. H. Lee, Y. Feng, M. Logar, P. M. Rao, L. Cai, D. R. Kim, R. Sinclair, X. Zheng, Nat. Commun. 2013, 4, 1723. [15] A. Kleiman-Shwarsctein, Y. S. Hu, A. J. Forman, G. D. Stucky, E. W. McFarland, J. Phys. Chem. C 2008, 112, 15900 – 15907. [16] X. Yu, C. Li, Y. Ling, T. A. Tang, Q. Wu, J. Kong, J. Alloys Compd. 2010, 507, 33 – 37. [17] W. Luo, Z. Yang, Z. Li, J. Zhang, J. Liu, Z. Zhao, Z. Wang, S. Yan, T. Yu, Z. Zou, Energy Environ. Sci. 2011, 4, 4046 – 4051. [18] S. K. Pilli, T. E. Furtak, L. D. Brown, T. G. Deutsch, J. A. Turner, A. M. Herring, Energy Environ. Sci. 2011, 4, 5028 – 5034. [19] C. Wu, J. Shen, J. Ma, S. Wang, Z. Zhang, X. Yang, Semicond. Sci. Technol. 2009, 24, 125012. [20] S. D. Tilley, M. Cornuz, K. Sivula, M. Grtzel, Angew. Chem. Int. Ed. 2010, 49, 6405 – 6408; Angew. Chem. 2010, 122, 6549 – 6552.
ChemSusChem 0000, 00, 1 – 8
&6&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS [21] J. Sun, D. K. Zhong, D. R. Gamelin, Energy Environ. Sci. 2010, 3, 1252 – 1261. [22] M. W. Kanan, D. G. Nocera, Science 2008, 321, 1072 – 1075. [23] Y. Surendranath, M. W. Kanan, D. G. Nocera, J. Am. Chem. Soc. 2010, 132, 16501 – 16509. [24] D. K. Zhong, J. Sun, H. Inumaru, D. R. Gamelin, J. Am. Chem. Soc. 2009, 131, 6086 – 6087. [25] E. M. P. Steinmiller, K.-S. Choi, Proc. Natl. Acad. Sci. USA 2009, 106, 20633 – 20636. [26] J. J. H. Pijpers, M. T. Winkler, Y. Surendranath, T. Buonassisi, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2011, 108, 10056 – 10061. [27] P. K. Giri, S. Bhattacharyya, D. K. Singh, R. Kesavamoorthy, B. K. Panigrahi, K. G. M. Nair, J. Appl. Phys. 2007, 102, 093515. [28] M. Kuhn, J. A. Rodriguez, Surf. Sci. 1995, 336, 1 – 12. [29] S. Krishnamurthy, C. McGuinness, L. S. Dorneles, M. Venkatesan, J. M. D. Coey, J. G. Lunney, C. H. Patterson, K. E. Smith, T. Learmonth, P. A. Glans, T. Schmitt, J. H. Guo, J. Appl. Phys. 2006, 99, 08M111. [30] Q. Peng, B. Kalanyan, P. G. Hoertz, A. Miller, D. H. Kim, K. Hanson, L. Alibabaei, J. Liu, T. J. Meyer, G. N. Parsons, J. T. Glass, Nano Lett. 2013, 13, 1481 – 1488.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org [31] R. Franking, L. Li, M. A. Lukowski, F. Meng, Y. Tan, R. J. Hamers, S. Jin, Energy Environ. Sci. 2013, 6, 500 – 512. [32] O. Zandi, B. M. Klahr, T. W. Hamann, Energy Environ. Sci. 2013, 6, 634 – 642. [33] Z. Chen, T. F. Jaramillo, T. G. Deutsch, A. K. Shwarsctein, A. J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E. W. McFarland, K. Domen, E. L. Miller, J. A. Turner, H. N. Dinh, J. Mater. Res. 2010, 25, 3 – 16. [34] Y. Hou, F. Zuo, A. Dagg, P. Feng, Angew. Chem. Int. Ed. 2013, 52, 1248 – 1252; Angew. Chem. 2013, 125, 1286 – 1290. [35] A. Wolcott, W. A. Smith, T. R. Kuykendall, Y. Zhao, J. Z. Zhang, Adv. Funct. Mater. 2009, 19, 1849 – 1856.
Received: February 6, 2014 Revised: March 13, 2014 Published online on && &&, 0000
ChemSusChem 0000, 00, 1 – 8
&7&
These are not the final page numbers! ÞÞ
FULL PAPERS Y.-G. Lin,* Y.-K. Hsu, Y.-C. Chen, B.-W. Lee, J.-S. Hwang, L.-C. Chen,* K.-H. Chen* && – && Cobalt-Phosphate-Assisted Photoelectrochemical Water Oxidation by Arrays of Molybdenum-Doped Zinc Oxide Nanorods
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Molybdenum and Co: A photocatalytic system comprising an array of zinc oxide nanorods (NRs) doped with molybdenum is reported. The addition of cobalt phosphate (Co-Pi) further enhances the performance of the Zn1xMoxO NR-array photoanodes. The results demonstrate that the system can serve as visible-light-sensitive photofunctional electrodes to fundamentally improve the performance of ZnO-based photoanodes for photoelectrochemical water oxidation.
ChemSusChem 0000, 00, 1 – 8
&8&
These are not the final page numbers! ÞÞ
Cobalt-Phosphate-Assisted Photoelectrochemical Water Oxidation by Arrays of Molybdenum-Doped Zinc Oxide Nanorods Yan-Gu Lin,*[a] Yu-Kuei Hsu,[b] Ying-Chu Chen,[c] Bing-Wei Lee,[d] Jih-Shang Hwang,[d] LiChyong Chen,*[c] and Kuei-Hsien Chen*[c, e] We report the first demonstration of cobalt phosphate (Co-Pi)assisted molybdenum-doped zinc oxide nanorods (Zn1xMoxO NRs) as visible-light-sensitive photofunctional electrodes to fundamentally improve the performance of ZnO NRs for photoelectrochemical (PEC) water splitting. A maximum photoconversion efficiency as high as 1.05 % was achieved, at a photocurrent density of 1.4 mA cm2. More importantly, in addition to achieve the maximum incident photon to current conversion efficiency (IPCE) value of 86 %, it could be noted that the IPCE of Zn1xMoxO photoanodes under monochromatic illumination (450 nm) is up to 12 %. Our PEC performances are com-
parable to those of many oxide-based photoanodes in recent reports. The improvement in photoactivity of PEC water splitting may be attributed to the enhanced visible-light absorption, increased charge-carrier densities, and improved interfacial charge-transfer kinetics due to the combined effect of molybdenum incorporation and Co-Pi modification, contributing to photocatalysis. The new design of constructing highly photoactive Co-Pi-assisted Zn1xMoxO photoanodes enriches knowledge on doping and advances the development of highefficiency photoelectrodes in the solar-hydrogen field.
Introduction The efficient photoelectrolysis of water into hydrogen poses a long-standing challenge, with promise for solar energy conversion and storage in the form of synthetic fuels.[1, 2] Desirable for photoelectrolysis are strong visible-light absorption, efficient charge-carrier separation and transport, and facile interfacial charge-transfer kinetics, especially for the water oxidation reaction.[3] Despite important advances and intensive research and development efforts worldwide, no photoelectrochemical (PEC) system for solar hydrogen production has met the required performance benchmarks. In addition to the search for [a] Dr. Y.-G. Lin National Synchrotron Radiation Research Center Hsinchu 30076 (Taiwan) E-mail: [email protected] [b] Prof. Y.-K. Hsu Department of Opto-Electronic Engineering National Dong Hwa University Hualien 97401 (Taiwan) [c] Y.-C. Chen, Dr. L.-C. Chen, Dr. K.-H. Chen Center for Condensed Matter Sciences National Taiwan University Taipei 10617 (Taiwan) [d] B.-W. Lee, Prof. J.-S. Hwang Institute of Optoelectronic Sciences National Taiwan Ocean University Keelung 202 (Taiwan) [e] Dr. K.-H. Chen Institute of Atomic and Molecular Sciences Academia Sinica Taipei 10617 (Taiwan) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201402025.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
new photoactive materials, another fruitful approach is to modify the electronic properties of known semiconductors to further improve their photoactivity for water-splitting reactions. Abundant and inexpensive oxide semiconductors, such as ZnO or TiO2, have been recognized as promising photoelectrode materials, but their photoconversion efficiency is substantially limited owing to their large band-gaps and rapid charge recombination.[4, 5] Many works have been devoted to the mitigation of these intrinsic limitations, by using strategies involving doping and nanostructuring.[6–10] One-dimensional (1D) nanorod/nanowire-array photoelectrodes, offering large surface areas and short diffusion distances for photogenerated minority carriers, are expected to facilitate charge separation.[6, 11] Meanwhile, heteroatom doping into oxide semiconductors can not only increase the electronic conductivity of the semiconductor but can also narrow the band-gap, allowing to utilize more visible light.[6, 8, 9, 12–14] For example, nitrogendoped ZnO nanowires and nanotetrapods were found to ensure rapid electronic transport, efficient charge separation, and enhanced visible-light absorption.[6, 9] Incorporating carbon has also been shown to enhance the photoconversion efficiency of porous ZnO nanostructures for PEC water splitting.[13] Recently, doping with 4d transition metals, such as molybdenum, was reported to remarkably enhance the PEC performance of many photoanodes, including TiO2, BiVO4, and Fe2O3.[15–18] Nevertheless, 1-D molybdenum-doped ZnO nanostructures have not been reported for PEC water-splitting. In particular, while most investigations of Mo-doped ZnO have focused on the use of transparent conducting oxides for flat-panel display de-
ChemSusChem 0000, 00, 1 – 8
&1&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
vices,[19] thorough attempts to exploit these interesting doping effects for solar-energy-driven applications are lacking. On the other hand, it is equally important to accelerate the rate of oxygen evolution at photoanodes by using catalysts. Water oxidation generally occurs at significant overpotentials because it requires the removal of a total of four electrons and four protons from two molecules of water to form one molecule of oxygen. Therefore, surface modification of the photoanodes with Figure 1. SEM images of a) pristine, and b) Mo-doped ZnO NRs on FTO substrate. Inset: Photo of an as-synthean electrocatalyst is needed to sized sample. c) HRTEM image of Mo-doped ZnO NRs. d–g) Elemental mapping images of Zn, O, and Mo. achieve high water oxidation activity at low overpotentials. In order to lower the required overpotential, various catalysts have been intensively studied, such as IrO2, cobalt ions, and the cobalt phosphate catalyst, “Co-Pi”.[20, 21] Co-Pi in particular has received a lot of recent attention because it uses earthabundant elements, shows effective water oxidation characteristics, and is stable over time due to its “self-healing” mechanism.[22, 23] Consequently, it has also been applied to many potential photoanodes, including ZnO, BiVO4, Si, and Fe2O3, and has shown improvements in both current onset potential and photocurrent density.[18, 24–26] Herein, we show that the uniform deposition of Co-Pi-based oxygen evolution catalysts onto Mo-doped ZnO nanorods (NRs) is key for efficient photoelectrolysis, allowing for the preparation of next-generation PEC devices that could benefit from the light-scattering ability of high-aspect-ratio NRs, the unique advantages of molybdenum incorporation, and the excellent electrocatalysis characteristics of Co-Pi catalysts (Supporting Information, Figure S1). For ZnO-based photoelectrodes to be used at a scale commensurate with the global energy demand, methods to fabricate both doped and 1-D nanostructured ZnO using rapid and inexpensive preparation techniques are necessary. Electrodeposition is a simple and cost-effective technique that is very promising from this point of view. In this work, we demonstrate a facile and reliable procedure for fabrication of Co-Pi/ Zn1xMoxO hybrid nanoelectrodes in the absence of any surfactant agent. The Zn1xMoxO NR arrays were first grown on fluorine-doped tin oxide (FTO) glass substrates by using an in situ electrodeposition method. In the case of Co-Pi catalysts, a modified electrochemical deposition approach at room temperature reported earlier led to rapid formation of cable-like nanocomposites.[18] Herein, the PEC performance of Co-Pi/ Zn1xMoxO hybrid nanosystems in solar water-splitting without Figure 2. a) X-ray diffraction patterns of undoped and Mo-doped ZnO NRs. sacrificial reagents is systematically evaluated. Furthermore, we Inset: Magnified view of the region 368–378. b) Micro-Raman spectra of unalso perform microstructural characterization, elemental analydoped and Mo-doped ZnO NRs. sis, and electrochemical impedance spectroscopy (EIS) on the Mo-doped ZnO photoelectrodes to confirm the molybdenum doping and understand the correlation with improved PEC 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&2&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS performance and the surface photochemistry of ZnO. To the best of our knowledge, this represents the first case of realizing Co-Pi assisted Zn1xMoxO NR-array photoanodes for PEC water oxidation in a nonsacrificial electrolyte.
Results and Discussion
www.chemsuschem.org sults, which were performed to understand the electronic structure of the molybdenum-containing ZnO NRs (Figure 3). The O 1s core level spectra of Mo-doped ZnO NRs in Figure 3 b show a main peak at 531.2 eV, which can be deconvoluted into two peaks. The low binding energy peak at 530.6 eV, denoted by O1, corresponds to ZnO bonding, while the higher one at 532.5 eV (O2) is often attributed to oxygen vacancies in ZnO.[13] The XPS spectra for Mo 3d5/2 are also presented in Figure 3 a, showing a slight down-shift of the position of the main peak as compared to hexavalent Mo (Mo6 + , 232.5 eV).[17, 18] This is likely due to partial reduction of Mo6 + , which is consistent with the increase in density of oxygen vacancies (as electron donor) in ZnO through Mo incorporation, which involves transfer of electrons to Mo. It has also been reported that the Zn 2p binding energy shifts to a slightly lower value if Mo substitutes for O, owing to strong Mo–Zn bonding ( 10 kcal mol1).[28] However, the absence of any significant changes in the Zn 2p binding energies as determined by XPS (Figure 3 a) seems to imply that it is Zn, instead of O, which is substituted by Mo. Meanwhile, the Mo concentration in the Zn1xMoxO NRs was also estimated by XPS measurement to be about 1.7 at %. The Mo concentration can be tuned by changing the Na2MoO4 concentration in experimental conditions. Extended X-ray absorption fine structure (EXAFS) spectra taken around the Zn K-edge of Mo-containing ZnO NRs are shown in Figure 3 c. An apparent decrease in relative intensity and interatomic distance of the first shell and second shell (ZnO and ZnZn bonding) over ternary Zn1xMoxO NRs
A scanning electron microscopy (SEM) image of pristine ZnO NRs reveals a high surface density (ca. 109 rods cm2) with diameters of about 150 50 nm (Figure 1 a). Clearly, the in situ doping process does not alter the morphology of the nanostructures (Figure 1 b). A high-resolution transmission electron microscopy (TEM) image of doped ZnO is shown in Figure 1 c, and enables to determine the spacing of the (101) lattice plane of hexagonal ZnO crystal as 0.24 nm. More detailed TEM images with energy dispersive X-ray (EDX) elemental mapping of Zn, O, and Mo are shown in Figure 1 d–g, and confirms the presence of uniformly dispersed Mo atoms inside the ZnO NRs. X-ray diffraction (XRD) patterns of undoped and Mo-doped ZnO NRs are compared in Figure 2 a, showing that both ZnO NRs are of wurtzite structure and that there is no significant phase change upon Mo incorporation. In addition, an effect of Mo substitution in the ZnO matrix was a significant up-shift in the (101) peak of ternary Zn1xMoxO NRs relative to pure ZnO, indicating a decrease in the lattice parameter of ZnO (Figure 2 a, inset). This slight contraction of the ZnO lattice upon Mo incorporation suggests that the smaller-sized Mo ion (Mo6 + , 0.62 ) has most likely substituted for the larger Zn ion (Zn2 + , 0.74 ). To gain further insight into the evolution of defects with Mo incorporation, micro-Raman analyses were performed (Figure 2 b). The presence of a sharp and dominant E2 (high) mode at 438 cm1, which is associated with the vibration of oxygen atoms in ZnO, indicates the wurtzite nature of ZnO.[27] Significant variations in the position as well as the full width at half maximum (FWHM) of the peak corresponding to the E2 (high) mode occur, owing to the formation of compressive strain resulting from Mo doping. Furthermore, the peak at 575 cm1 corresponding to the A1 (LO) mode originates from oxygen vacancies in ZnO,[27] implying Mo incorporation may induce oxygen vacancies in the ZnO lattice. This could be further confirmed by X-ray photoelectron spectroscopy (XPS) and X-ray ab- Figure 3. XPS spectra of a) Mo 3d, Zn 2p3/2, and b) O 1s transitions for Mo-doped ZnO NRs. c) Zn K-edge EXAFS sorption spectroscopy (XAS) re- spectra, and d) O K-edge XAS spectra of Mo-doped ZnO NRs and pure ZnO samples. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&3&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS would be observed compared to pure ZnO. This can be mainly attributed to the strong distortion of the ZnO lattice resulting from the positioning of Mo in the ZnO framework. The slight decrease of the ZnO lattice parameter upon Mo incorporation suggests that the smaller-sized Mo atom has most likely substituted for the larger Zn atom, which is consistent with XRD and XPS results. We further fitted the results of pure ZnO and Mocontaining ZnO samples at the Zn K-edge (Supporting Information, Table S2). Meanwhile, Figure 3 d shows the O K-edge XAS spectra of ternary Zn1xMoxO NRs. The broadening of the absorption peak at ca. 537 eV for Mo-doped ZnO NRs, compared to pure ZnO structure, can be assigned to the presence of oxygen vacancies, suggesting that Mo incorporation induces oxygen vacancies in the ZnO lattice.[29] In addition, the overall intensities of the features of Mo-doped ZnO NRs are higher than those of pure ZnO structures, reflecting the increase in number of unoccupied O 2p–Zn 4s hybrid states. The reason is Mo incorporation, which may create a high density of oxygen vacancies that serve as electron donors, so that more electrons are transferred from O 2p states to Mo. To verify that Mo incorporation enhances light harvesting, we measured UV-vis absorption spectra of Mo-containing ZnO nanoelectrodes (Figure 4 a). In comparison to pristine ZnO, it is apparent that Mo-doped ZnO exhibits a significant enhancement in the absorption of visible light. This finding is commensurate with the red-shift of the band gap in the ternary Zn1xMoxO NRs. Besides, to quantify the PEC performance, the incident photon to current conversion efficiency (IPCE) was measured to study the photoresponse of pristine and Mo-containing ZnO NRs as a function of incident light wavelength (Figure 4 b). Pristine ZnO has a strong photoresponse only in the near-UV region. Surprisingly, the doped sample displays substantial photoactivity in the visible-light region, from 400 to 550 nm, in addition to strong photoresponse in the near-UV region. Particularly, the IPCE of Mo-doped ZnO photoelectrodes at the monochromatic wavelength of 450 nm is up to 12 %. Mo modification substantially improves the light utilization and conversion efficiency in the visible region of interest, as compared to the pure ZnO structure. The increase of the IPCE in the visible range may be probably attributed to the bandgap narrowing of ZnO and/or amorphous MoO2 phase in the ZnO. These results are in good agreement with the UV-visible absorption data (Figure 4 a). To further determine the exact efficiency of the water-splitting reaction under simulated AM 1.5 illumination, systematic PEC measurements without sacrificial reagents were carried out in a 0.5 m aqueous solution of Na2SO4 with a pH of 6.8. Figure 5 a depicts a set of linear-sweep voltammograms recorded on pristine ZnO, Mo-doped ZnO, and Co-Pi/Mo-doped ZnO photoanodes in the dark and under light illumination of 100 mW cm2. Linear-sweep voltammagrams in the dark from 0.2 to + 1 V show the current density range of 107 A cm2. The Mo-containing ZnO NRs manifest a pronounced photocurrent, which increases to 0.9 mA cm2 under illumination. Clearly, the photocurrent density of Mo-containing ZnO nanoelectrodes is about 2.3 times greater than that of pristine ZnO, suggesting that ZnO incorporated with Mo can effectively improve 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
Figure 4. a) Absorption spectra of undoped and Mo-doped ZnO NRs. b) Measured IPCE spectra of pristine ZnO NRs and Mo-doped ZnO NRs in the region of 300–650 nm at a potential of 0.4 V (vs. Pt) in two-electrode system.
PEC water-splitting. The maximum photocurrent density (1.4 mA cm2) generated from Co-Pi/Mo-doped ZnO hybrid nanoelectrodes is about 3.5 times greater than that of pristine ZnO. These results are comparable to those of many oxidebased photoanodes in recent reports.[6, 8, 18, 30–32] We also measured the amount of hydrogen and oxygen evolved, determined by the gaseous product in the PEC cell. Generally, in PEC water splitting, two reaction processes happen: Photoanode reaction : 2 H2 O ! O2 þ 4 Hþ þ 4 e
ð1Þ
Photocathode reaction : 2 H2 O þ 2 e ! H2 þ 2 OH
ð2Þ
The rate of H2 and O2 evolution measured for our Co-Pi/Modoped ZnO hybrid nanocomposites are about 21 and 11 mmol h1, respectively. The yield of H2 after illumination is much lower than the calculated theoretical value, based on the photocurrent data and Faraday’s law (the theoretical value ChemSusChem 0000, 00, 1 – 8
&4&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
Figure 5. a) Linear sweep voltammagrams, collected at a scan rate of 10 mV s1 at applied potentials from 0.2 to + 1 V (vs. Ag/AgCl) for pristine ZnO, Mo-doped ZnO, and Co-Pi/Mo-doped ZnO photoanodes in the dark and under light illumination of 100 mW cm2. b) Photoconversion efficiency of the PEC cells for Mo-doped ZnO and Co Pi/Mo-doped ZnO photoanodes as a function of applied potential.
is 26 mmol h1, in the ideal case). This is possibly due to the formation of water from H2 and O2 on the photoelectrodes. Furthermore, in order to quantitatively explore the photoactivity of Mo-doped ZnO and Co-Pi/Mo-doped ZnO NRs, the efficiency (h) of applied bias photon-to-current generation was calculated using an equation in which the contribution due to applied potential is subtracted from the total efficiency.[33] Significantly, it is worth mentioning that the plot of efficiency versus applied potential (Figure 5 b) shows that the maximum value of efficiency is 1.05 % at an applied potential of + 0.3 V, which is also higher than recent reported values for other oxide-based photoanodes.[6, 9, 34] To understand the intrinsic electronic properties of ternary Zn1xMoxO NRs in the electrolyte solution, we measured electrochemical impedance spectroscopy (EIS) in the dark coupled to Mott–Schottky analysis, commonly used to determine both donor density and flatband potential at semiconductor/liquid contacts. Capacitances at the semiconductor/electrolyte interface with the use of an equivalent circuit are described by a Mott–Schottky plot.[35] All 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ZnO NR samples show a positive slope in the Mott– Schottky plots, as expected for n-type semiconductors (Figure 6). The estimated flat-band potentials of the pure ZnO and Mo-doped ZnO are 0.28 V and 0.34 V, respectively. Apparently, Mo-containing ZnO has a more negative flat-band potential than pure ZnO on account of the Mo-doping effect. In addition, the calculated donor densities of the pure ZnO and Mo-doped ZnO are 6 1018 and 3 1020 cm3, respectively. The enhanced donor density is due to the created oxygen vacancies, which are known to be an electron donor for oxide-based semiconductors.[7] The increased donor density improves the charge transport in ZnO, as well as the electron transfer at the interface between the ZnO and FTO substrate. Moreover, the increased electron density is expected to shift the Fermi level of ZnO toward the conduction band. The upward shift of the Fermi level facilitates charge separation at the semiconductor/electrolyte interface, by increasing the degree of band bending at the ZnO surface. Therefore, the reduced band gap, increased carrier concentration, and improved interfacial catalysis are believed to be the major reasons for the improvement in photoactivity of PEC water-splitting based on Co-Pi modified Mo-doped ZnO NR photoelectrodes. Finally, the structural stability during PEC water splitting is another important factor to evaluate for the potential PEC materials. Time-dependent measurement carried out on a representative Co-Pi modified Mo-doped ZnO NRs displays a minimal degradation rate after 11 h of continuous running in the water-oxidation process at 1 V (Supporting Information, Figure S3). Meanwhile, after 11 h of continuous PEC reaction, ICP-MS was applied to monitor the concentration of Zn2 + ions in the electrolyte. Only a trace amount of Zn2 + ion could be detected in the photoreaction solution, showing that the photocorrosion factor would be minor for our developed Co-Pi modified Mo-doped ZnO NRs.
Figure 6. Mott–Schottky plots for pure ZnO and Mo-doped ZnO NRs in 0.5 m Na2SO4.
ChemSusChem 0000, 00, 1 – 8
&5&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS Conclusions An efficient solar hydrogen system is developed based on CoPi assisted ternary Zn1xMoxO nanorod (NR) photoelectrodes. This is the first example of in situ doping of zinc oxide nanorods with molybdenum for solar water splitting in a nonsacrificial electrolyte. Relative to pristine ZnO NR photoanodes, the photocurrent of Co-Pi/Zn1xMoxO NRs increases substantially (up to ca. 250 %). The highest photoconversion efficiency obtained from our Mo-containing ZnO NRs is ca. 0.47 %, much better than undoped ZnO NRs, and subsequent modification with Co-Pi further enhances the photoconversion efficiency of the Mo-containing ZnO NRs to ca. 1.05 %. Furthermore, concerning the maximum incident photon to current conversion efficiency (IPCE) value of 86 %, it could be noted that the IPCE of Co-Pi-assisted Zn1xMoxO photoanodes under monochromatic illumination (450 nm) is up to 12 %. The improvement in photoactivity of PEC water-splitting may be attributed to the enhanced visible-light absorption, increased charge-carrier densities, and improved interfacial charge-transfer kinetics due to the combined effects of molybdenum incorporation and Co-Pi modification contributing to photocatalysis. Because the Co-Pi assisted Zn1xMoxO NR photoanodes are convenient to fabricate and highly active, they can serve as a useful substitution for ZnO in a variety of solar-energy-driven applications including PEC water splitting, photocatalysis, and solar cells.
Experimental Section The Zn1xMoxO NR arrays were grown on an FTO glass, serving as substrate and working electrode, by typical electro-deposition method. A piece of Pt foil and a standard Ag/AgCl electrode were used as the counter and reference electrodes, respectively. Aqueous 0.1–0.3 m KCl and 5–10 mm ZnCl2 solutions with 25–75 mm Na2MoO4 were used as electrolytes. H2O2 (5–10 mm) was used as source of oxygen. The applied potential was operated at 0.8 V versus Ag/AgCl for 30 min and the solution temperature was maintained at 80 8C. Then, the resulting Mo-doped ZnO NR arrays were heated in vacuum to 450 8C for 3–6 h. Finally, Co-Pi catalysts were deposited potentiostatically at 1.1 V versus Ag/AgCl in a solution of 0.1–0.2 mm cobalt nitride in 0.1 m potassium phosphate buffer at pH 7 for several minutes. For material characterization, SEM measurements were made on a JEOL 6700 field-emission SEM. For obtaining TEM images, the Mo-doped ZnO NR products on the FTO substrate were scratched and dispersed onto a carbon-coated Cu grid, and analyzed using a JEOL JEM-2100 TEM system. The XRD analyses were performed on a Bruker D8 Advance diffractometer with a Cu target emitting X-ray radiation at 1.54 (40 kV, 40 mA). Micro-Raman analyses were performed on a Jobin Yvon Labram HR800 spectrometer. XPS spectra were obtained using a Microlab 350 system. XAS analyses were performed in National Synchrotron Radiation Research Center. UV-vis absorption spectra were carried out with a JASCO V670 instrument. In PEC water-splitting reaction, a Co-Pi/Mo-doped ZnO photoanode was used as the working electrode with surface area of 0.5–1 cm2, a platinum plate as counter electrode, and Ag/ AgCl as reference electrode. All PEC studies were operated in a 0.5 m Na2SO4 solution (pH 6.8) as supporting electrolyte medium by using an Electrochemical Multichannel Solartron Analytical System. The water-splitting photoelectrode was illuminated with 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org the AM 1.5G simulated solar light at 100 mW cm2. This illumination source matches the shape and intensity of the Air Mass 1.5 Global (AM 1.5 G) G173 standard set forth by the American Society of Testing and Materials. The IPCE was calculated using the following equation: IPCE [%] = (1240 photocurrent density [mA cm2] 100)/(wavelength [nm] photon flux [mW cm2]). IPCE measurements were performed by 2-electrode system under monochromatic irradiation, emitted from a 150 W Xe lamp equipped with bandpass filters and a variable neutral density filter. Monochromatic photon fluxes at each wavelength were measured using an optical photodiode power meter and were adjusted to around 0.1 mW cm2 using the neutral density filter. Finally, the PEC cell was constructed to allow the collection of gases generated at the electrodes. For products analysis, the amounts of evolved gases after illumination were determined with gas chromatography (Shimadzu, MS-5A column, Ar carrier).
Acknowledgements This work was supported by the Ministry of Science and Technology and National Synchrotron Radiation Research Center. We gratefully thank MOST, NSRRC, IAMS, and NTU for financial support in this project. Keywords: doping · molybdenum · photocatalysis · water splitting · zinc oxides [1] N. S. Lewis, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2006, 103, 15729 – 15735. [2] Y. G. Lin, Y. K. Hsu, A. M. Basilio, Y. T. Chen, K. H. Chen, L. C. Chen, Opt. Express 2014, 22, A21 – A27. [3] Z. Li, W. Luo, M. Zhang, J. Feng, Z. Zou, Energy Environ. Sci. 2013, 6, 347 – 370. [4] J. Shi, X. Wang, Energy Environ. Sci. 2012, 5, 7918 – 7922. [5] Y. G. Lin, Y. K. Hsu, Y. C. Chen, S. B. Wang, J. T. Miller, L. C. Chen, K. H. Chen, Energy Environ. Sci. 2012, 5, 8917 – 8922. [6] X. Yang, A. Wolcott, G. Waang, A. Sobo, R. C. Fitzmorris, F. Qian, J. Z. Zhang, Y. Li, Nano Lett. 2009, 9, 2331 – 2336. [7] G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang, Y. Li, Nano Lett. 2011, 11, 3026 – 3033. [8] S. Hoang, S. Guo, N. T. Hahn, A. J. Bard, C. B. Mullins, Nano Lett. 2012, 12, 26 – 32. [9] Y. Qiu, K. Yan, H. Deng, S. Yang, Nano Lett. 2012, 12, 407 – 413. [10] M. Xu, P. Da, H. Wu, D. Zhao, G. Zheng, Nano Lett. 2012, 12, 1503 – 1508. [11] Y. G. Lin, Y. C. Chen, C. Feldmann, J. T. Miller, L. C. Chen, K. H. Chen, Y. K. Hsu, ChemCatChem. 2014, 1684 – 1690. [12] G. Liu, L. C. Ying, J. Wang, P. Niu, C. Zhen, Y. Xie, H. M. Cheng, Energy Environ. Sci. 2012, 5, 9603 – 9610. [13] Y. G. Lin, Y. K. Hsu, Y. C. Chen, L. C. Chen, S. Y. Chen, K. H. Chen, Nanoscale 2012, 4, 6515 – 6519. [14] I. S. Cho, C. H. Lee, Y. Feng, M. Logar, P. M. Rao, L. Cai, D. R. Kim, R. Sinclair, X. Zheng, Nat. Commun. 2013, 4, 1723. [15] A. Kleiman-Shwarsctein, Y. S. Hu, A. J. Forman, G. D. Stucky, E. W. McFarland, J. Phys. Chem. C 2008, 112, 15900 – 15907. [16] X. Yu, C. Li, Y. Ling, T. A. Tang, Q. Wu, J. Kong, J. Alloys Compd. 2010, 507, 33 – 37. [17] W. Luo, Z. Yang, Z. Li, J. Zhang, J. Liu, Z. Zhao, Z. Wang, S. Yan, T. Yu, Z. Zou, Energy Environ. Sci. 2011, 4, 4046 – 4051. [18] S. K. Pilli, T. E. Furtak, L. D. Brown, T. G. Deutsch, J. A. Turner, A. M. Herring, Energy Environ. Sci. 2011, 4, 5028 – 5034. [19] C. Wu, J. Shen, J. Ma, S. Wang, Z. Zhang, X. Yang, Semicond. Sci. Technol. 2009, 24, 125012. [20] S. D. Tilley, M. Cornuz, K. Sivula, M. Grtzel, Angew. Chem. Int. Ed. 2010, 49, 6405 – 6408; Angew. Chem. 2010, 122, 6549 – 6552.
ChemSusChem 0000, 00, 1 – 8
&6&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS [21] J. Sun, D. K. Zhong, D. R. Gamelin, Energy Environ. Sci. 2010, 3, 1252 – 1261. [22] M. W. Kanan, D. G. Nocera, Science 2008, 321, 1072 – 1075. [23] Y. Surendranath, M. W. Kanan, D. G. Nocera, J. Am. Chem. Soc. 2010, 132, 16501 – 16509. [24] D. K. Zhong, J. Sun, H. Inumaru, D. R. Gamelin, J. Am. Chem. Soc. 2009, 131, 6086 – 6087. [25] E. M. P. Steinmiller, K.-S. Choi, Proc. Natl. Acad. Sci. USA 2009, 106, 20633 – 20636. [26] J. J. H. Pijpers, M. T. Winkler, Y. Surendranath, T. Buonassisi, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2011, 108, 10056 – 10061. [27] P. K. Giri, S. Bhattacharyya, D. K. Singh, R. Kesavamoorthy, B. K. Panigrahi, K. G. M. Nair, J. Appl. Phys. 2007, 102, 093515. [28] M. Kuhn, J. A. Rodriguez, Surf. Sci. 1995, 336, 1 – 12. [29] S. Krishnamurthy, C. McGuinness, L. S. Dorneles, M. Venkatesan, J. M. D. Coey, J. G. Lunney, C. H. Patterson, K. E. Smith, T. Learmonth, P. A. Glans, T. Schmitt, J. H. Guo, J. Appl. Phys. 2006, 99, 08M111. [30] Q. Peng, B. Kalanyan, P. G. Hoertz, A. Miller, D. H. Kim, K. Hanson, L. Alibabaei, J. Liu, T. J. Meyer, G. N. Parsons, J. T. Glass, Nano Lett. 2013, 13, 1481 – 1488.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org [31] R. Franking, L. Li, M. A. Lukowski, F. Meng, Y. Tan, R. J. Hamers, S. Jin, Energy Environ. Sci. 2013, 6, 500 – 512. [32] O. Zandi, B. M. Klahr, T. W. Hamann, Energy Environ. Sci. 2013, 6, 634 – 642. [33] Z. Chen, T. F. Jaramillo, T. G. Deutsch, A. K. Shwarsctein, A. J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E. W. McFarland, K. Domen, E. L. Miller, J. A. Turner, H. N. Dinh, J. Mater. Res. 2010, 25, 3 – 16. [34] Y. Hou, F. Zuo, A. Dagg, P. Feng, Angew. Chem. Int. Ed. 2013, 52, 1248 – 1252; Angew. Chem. 2013, 125, 1286 – 1290. [35] A. Wolcott, W. A. Smith, T. R. Kuykendall, Y. Zhao, J. Z. Zhang, Adv. Funct. Mater. 2009, 19, 1849 – 1856.
Received: February 6, 2014 Revised: March 13, 2014 Published online on && &&, 0000
ChemSusChem 0000, 00, 1 – 8
&7&
These are not the final page numbers! ÞÞ
FULL PAPERS Y.-G. Lin,* Y.-K. Hsu, Y.-C. Chen, B.-W. Lee, J.-S. Hwang, L.-C. Chen,* K.-H. Chen* && – && Cobalt-Phosphate-Assisted Photoelectrochemical Water Oxidation by Arrays of Molybdenum-Doped Zinc Oxide Nanorods
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Molybdenum and Co: A photocatalytic system comprising an array of zinc oxide nanorods (NRs) doped with molybdenum is reported. The addition of cobalt phosphate (Co-Pi) further enhances the performance of the Zn1xMoxO NR-array photoanodes. The results demonstrate that the system can serve as visible-light-sensitive photofunctional electrodes to fundamentally improve the performance of ZnO-based photoanodes for photoelectrochemical water oxidation.
ChemSusChem 0000, 00, 1 – 8
&8&
These are not the final page numbers! ÞÞ
