Photoelectrochemical hydrogen generation using graded In-content InGaN photoelectrode structures

Photoelectrochemical hydrogen generation using graded In-content InGaN photoelectrode structures

Item Type Article

Authors Ohkawa, Kazuhiro;Uetake, Yusuke;Velazquez-Rizo, Martin;Iida, Daisuke

Citation Ohkawa, K. et al., 2019. Photoelectrochemical hydrogen generation using graded In-content InGaN photoelectrode structures. Nano Energy, 59, pp.569–573. Available at: http://

dx.doi.org/10.1016/j.nanoen.2019.03.011.

Eprint version Publisher's Version/PDF

DOI 10.1016/j.nanoen.2019.03.011

Publisher Elsevier BV

Journal Nano Energy

Rights This is an open access article under the CC BY license (http://

creativecommons.org/licenses/BY/4.0/).

Download date 2024-01-22 11:17:53

Item License http://creativecommons.org/licenses/BY/4.0/

Link to Item http://hdl.handle.net/10754/631563

Contents lists available atScienceDirect

Nano Energy

journal homepage:www.elsevier.com/locate/nanoen

Full paper

Photoelectrochemical hydrogen generation using graded In-content InGaN photoelectrode structures

Kazuhiro Ohkawa

^a,∗

, Yusuke Uetake

^b

, Martin Velazquez-Rizo

^a

, Daisuke Iida

^a

aElectrical Engineering, King Abdullah University of Science and Technology, Thuwal, 23955-6900, Saudi Arabia

bDepartment of Applied Physics, Tokyo University of Science, Katsushika, Tokyo, 125-8585, Japan

A R T I C L E I N F O

Keywords:

InGaN photoelectrode NiO cocatalyst Hydrogen generation Nitrides

Metalorganic vapor phase epitaxy Graded structure

A B S T R A C T

We have improved the InGaN/GaN heterointerface to achieve higher energy conversion efficiency by replacing a uniform InGaN layer with a graded In-content InGaN layer. Even In0.08Ga0.92N/GaN heterostructure has a large conduction band offset, which is large enough to suppress the photocurrent in the photocatalytic system. The graded In-content InGaN structures were grown by metalorganic vapor-phase epitaxy by changing the TMInflow rate gradually. X-ray reciprocal space mapping confirmed the graded structures. The graded InGaN/GaN structure significantly increased photocurrent and H2generation by 50% and more compared with the con- ventional uniform InGaN/GaN structures.

1. Introduction

The energy demand of modern societies can be only satisfied using fossil fuels, which generate dangerous pollutants for the environment.

To diminish the consumption of fossil fuels is necessary to use alter- native energy sources, such as hydrogen, which has the potential to overcome the disadvantages of pollutant fuels. However, it is needed to find an efficient and environmental-friendly method to take advantage of its ecological benefits. One of the most promising options to generate hydrogen is through the photocatalytic water splitting. The method of a semiconductor photoelectrode can generate spatially separated hy- drogen and oxygen gases from water [1]. The photo-absorption phe- nomenon creates electron-hole pairs at the semiconductor surface re- gion. The surface band bending of an n-type semiconductor can separate the pairs into electrons and holes spatially. The electrons and holes reduce and oxidize water resulting in hydrogen and oxygen gas generation, respectively. Also, the semiconductor conduction and va- lence band edges must straddle the reduction and oxidation levels of water. The band edge positions can play critical roles in photocurrent and H2generation by water splitting [1–8].

We have developed nitride photocatalyst to generate H2from aqu- eous solution [9]. The H2generation under zero bias has realized by optimizing the surface band structure [10,11]. The potential and cap- abilities of thinfilm GaN-based photoelectrodes have been reported so far [12–15]. We found out NiO cocatalyst for nitride photocatalysts [16]. The NiO cocatalyst can prevent the photocorrosion of nitride

layers and improve the energy conversion efficiency. The durability of GaN photocatalysts is more than 500 h [17]. Nitride photocatalyst with NiO cocatalyst also can reduce CO2molecules, that is so-called artificial photosynthesis [18,19]. As far as the material for nitride is GaN, effi- ciency is low because its bandgap of 3.42 eV [20] can absorb only the light shorter than 363 nm in wavelength. To enhance the solar-to-hy- drogen energy conversion efficiency, the introduction of InGaN is mandatory since the bandgap of InGaN changes from 3.42 down to 0.67 eV [21] with the In content. The narrower bandgap can absorb more solar light. Recent nitride photoelectrode structures have uniform InGaN alloys grown on GaN [22–26]. They reported improved effi- ciency. However, the InGaN/GaN photocatalysts have large barriers for electrons as conduction band offsets (ΔEc) at the heterointerface. Since these barrier heights are much larger than the thermal energy at room temperature, the photocurrents for the conventional InGaN/GaN pho- tocatalysts are presumably suppressed.

In this paper, we report the growth of the graded In-content InGaN/

GaN layers and investigate the effect of the graded InGaN on the pho- tochemical phenomena.Fig. 1shows the schematic band diagrams of InGaN/GaN heterostructures. We have evaluated the ΔEc of In0.08Ga0.92N/GaN heterointerface which is 0.32 eV [6]. In general,ΔEc

can estimate 0.23–0.36 eV by some literature so that our value is rea- sonable [27–30]. SuchΔEcvalue is too large for the electron transfer from the InGaN photoabsorption region to the highly-conductive n-type GaN at room temperature. In the case of the graded InGaN layer as shown inFig. 1(b), the In-content changes from 0% at the InGaN/GaN

https://doi.org/10.1016/j.nanoen.2019.03.011

Received 20 July 2018; Received in revised form 7 February 2019; Accepted 2 March 2019

∗Corresponding author.

E-mail address:kazuhiro.ohkawa@kaust.edu.sa(K. Ohkawa).

Nano Energy 59 (2019) 569–573

Available online 07 March 2019

2211-2855/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

T

interface to a certain value at the surface. Therefore, there is no step- like energy barrier in the conduction band. We can expect more smooth electron transfer compared to that in the uniform InGaN structure, which would increase the overall energy conversion efficiency.

2. Material and methods

The samples used in our experiments were grown by metalorganic vapor-phase epitaxy (MOVPE) in a single-wafer horizontal reactor. The material precursors were trimethylgallium (TMGa), trimethylindium (TMIn), and NH3for Ga, In, and N, respectively. H2-diluted 10-ppm silane was used for n-type doping.Fig. 2shows the schematic cross- sectional views of the samples structures. A 3-μm-thick Si-doped n-type GaN was grown on thec-plane of a sapphire substrate covered with a low-temperature GaN buffer layer [31]. We fixed the carrier con- centration of Si-doped n-type GaN at 1 × 10¹⁸cm⁻³to realize high electrical conductivity for all of our samples. Afterward, we grew an unintentionally doped 110-nm-thick uniform and graded InGaN pho- toabsorption layers with the maximum In-contents of 8 and 16%. The graded InGaN layer was grown by linearly changing theflow rate of TMIn.Table 1indicates the sample details about InGaN structures and their growth conditions. The X-ray reciprocal-space mapping evaluates the In-content and lattice relaxation of InGaN layers. After the epitaxial growth, NiO islands as cocatalysts were deposited on InGaN surfaces [16]. NiO cocatalyst seems to improve water oxidation. We observed the enhanced photocurrent both for water splitting and artificial pho- tosynthesis by introducing NiO cocatalysts [16,17,32]. In the photo- electrochemical experiment setup, an InGaN photoelectrode and a Pt counterelectrode were immersed in 1.0 mol/L NaOH solution in a quartz container. We used a Xe-lamp (Eagle Engineering, Japan) as an irradiation light source with a power densityfixed at 100 mW/cm²at the front of the quartz container. The area of InGaN photoelectrode irradiated was 1.0 cm², and the remaining surface was covered with epoxy resin. The electrical current was measured between the elec- trodes using an ammeter. We have not applied the extra bias between the InGaN photoelectrodes and Pt counterelectrodes. The gas generated from the Pt counterelectrode was analyzed by gas chromatography to determine the volume of H2gas in the produced gas. We evaluated the crystal quality of the samples by X-ray diffraction and transmission electron microscopy (TEM). The X-ray (λ= 1.540598 Å) is given by the Cu-Kα1 line using a (220) Ge symmetric double crystal X-ray dif- fractometer (PANalytical X'Pert). The TEM observation (Titan, FEI) is under the 300 kV.

3. Results and discussion

We evaluated the structures and their crystallinities of the uniform and graded InxGa1-xN (x = 0.08 and 0.16) layers grown on n-GaN by the X-ray diffraction reciprocal-space mapping.Fig. 3 shows that the uniform samples have clear two spots related with GaN and InGaN, and the graded InGaN layers exhibit appreciable tails from the GaN spots.

Those tails are the evidence that the In content of InGaN changes gra- dually. It is also found that InGaN layers (sample A, B, and D) were completely coherently grown on the GaN layer because their in-plane lattice constant has the same value, indicating that these layers had a highly crystalline quality. The sample C also seems to be coherently grown on the GaN layer, however, sample C has a broadened InGaN spot, meaning that the InGaN layer was slightly lattice-relaxed because the critical thickness of In0.16Ga0.84N on GaN is only 90 nm [33]. As a result, the InGaN layers thicknesses were smaller than their critical thicknesses, otherwise, misfit dislocations would have introduced in the layers [33]. Since In incorporation causes strain, the net In amount in Fig. 1.Concept of (a) uniform and (b) graded InGaN photoelectrode band

structures under light irradiation. Electrons created by light absorption go into the inside, and these move to the Pt counterelectrode through the Ohmic contact to the n-type GaN.

Fig. 2.Schematic cross-section views of (a) uniform and (b) graded InGaN photoelectrode structures.

Table 1

InGaN structures and their growth conditions.

Sample A Sample B Sample C Sample D

Structure Uniform Graded Uniform Graded

In-content 0.08 0–0.08 0.16 0–0.16

Tg(^oC) 774 774 772 772

TMGaflow (μmol/min) 5.7 5.7 5.7 5.7 TMInflow (μmol/min) 4.1 0.4–4.1 24.6 0.4–24.6

NH3flow (slm) 4.5 4.5 18 18

V/III ratio 21000 33000–21000 27000 132000–27000

K. Ohkawa, et al. Nano Energy 59 (2019) 569–573

570

the graded In0.16Ga0.84N layer is smaller than that in the uniform layer.

Therefore, the graded In0.16Ga0.84N layer keeps quality, but the uniform one is not. We also observed the InGaN layers by cross-section TEM as shown inFig. 4. The graded InGaN structure only introduces the V-pits via the threading dislocations. However, the uniform InGaN structure indicates the misfit dislocations (MDs) from the InGaN/GaN hetero- interface caused by strain relaxation [33]. It is supporting the X-ray diffraction results. We confirmed that the graded In0.16Ga0.84N layers are high-quality even through high In content.

Time dependences of the photocurrent and H2generation using the samples A-D are shown inFig. 5. No bias among the electrodes was applied for the time-dependent experiments. The photocurrents mea- sured were very stable. At the end of the photoelectrolysis experiments, the surface of the InGaN photoelectrodes did not show any etching due to photocorrosion.

The uniform In0.08Ga0.92N layer exhibited the constant photocurrent of 1.23 mA, but in the case of the graded structure, the photocurrent was 1.89 mA. The improvement in photocurrent was more than 50%. It indicates that the graded structure enhances the photocurrent by re- moving the step-like potential barrier caused by the conduction band offsetΔEcamong GaN and InGaN. H2evolution is also improved by introducing the graded InGaN structure. The energy conversion effi- ciency from Xe-lamp light energy to hydrogen energy defined as the following equation

Fig. 3.X-ray diffraction reciprocal-space maps of (a,c) uniform InGaN layers (i.e., sample A and C) and (b,d) graded InGaN layers (i.e., sample B and D), respectively. It can be obtained the lattice-spacing of (0004) and(112¯0)in GaN from the coordinate of the position by(112¯4)reflection.

Fig. 4.Bright-field cross-section TEM images of the (a) graded In_0.16Ga_0.84N/n- GaN and (b) uniform In0.16Ga0.84N/n-GaN interfaces. The graded InGaN structure only exhibited V-pits and threading dislocations. In contrast, the uniform InGaN structure displayed the introduction of MDs from the InGaN/

GaN heterointerface. The iridium and carbon on top of the samples were de- posited as protective layers to minimize the surface damage during the lamella preparation.

Fig. 5.Time dependences of photocurrent and H₂gas generation from uniform and graded InGaN photoelectrodes without extra bias. Samples A and B are uniform and graded In_0.08Ga_0.92N, samples C and D are uniform and graded In_0.16Ga_0.84N, respectively.

= × η G× n

P A Δ ^o

(1) whereΔG^ois the Gibbs energy of H2combustion (237.13 kJ/mol),nis the amount of generation rate H2molecules,Pis the power density of irradiated light,Ais the irradiation area, andtis the measurement time of the photoelectrolysis experiment. The efficiencies of the uniform and graded In0.08Ga0.92N layers were 1.4 and 2.3%, respectively. Their Faradic efficiencies were 95 and 99%, respectively. The energy con- version efficiency has been improved as well with the photocurrent by the graded InGaN structure.

In the case of In0.16Ga0.84N, the uniform layer was not high quality as can been seen from the broadened spot of InGaN inFig. 3(c). The photocurrent and the energy conversion efficiency of the sample were as low as 0.27 mA and 0.1%, respectively. In0.16Ga0.84N can absorb more light compared with In0.08Ga0.92N and creates more electron-hole pairs around the surface region. However, the degraded In0.16Ga0.84N layer contains a lot of defects, and defects recombine the generated electrons and holes, resulting in low efficiency.Fig. 3(d) indicates that the graded In0.16Ga0.84N layer is high-quality. Therefore, we observed significant improvement both in the photocurrent and the H2genera- tion. A fourfold improvement was seen in the photocurrent by the graded structure. On the other hand, H2 generation by the graded structure was one order more than that by the uniform one. The one digit difference is originated from very low efficiency in the uniform In0.16Ga0.84N layer since we need photocurrent larger than 0.2 mA to observe H2 generation. We must point out the apparent higher effi- ciency from the graded In0.08Ga0.92N layer rather than the graded In0.16Ga0.84N layer. We assume one of this reason that the charge po- larization can be also present because of the high internal stresses due to the lattice mismatch with increasing In content [34,35]. The direc- tion of the piezoelectricfield of InGaN layers in the (0001) polar or- ientation is from the top surface to bottom [36,37]. According to sev- eral reports, the photoabsorption devices (i.e., solar cells) are much influenced by the carrier transports by the piezoelectricfield [38,39].

Such charge polarization would occur a band bending in the InGaN layer even though the graded structures, affecting the separation of electrons and holes. We think that the graded structures can remove the step-like potential barrier, however, it still has influenced by the pie- zoelectric field. Further study will be necessary to understand this point.

4. Conclusions

We investigated the improvement of the InGaN photoelectrode to generate H2gas from aqueous solutions. The graded In-content InGaN photoelectrode structures were grown by MOVPE changing the TMIn flow rate gradually. The significant increment in the photocurrent was observed by introducing the graded InGaN layer instead of the uniform one since the graded structure makes flat of the step-like potential barrier in the conduction band at the InGaN/GaN heterointerface. The energy conversion efficiency of the graded InGaN photoelectrode was improved more than 50% compared to the uniform one. We conclude that the graded InGaN structure is effective to improve the energy conversion efficiency of InGaN/GaN photocatalysts.

References

[1] A. Fujishima, K. Honda, Nature 238 (1972) 37–38.

[2] K. Domen, A. Kudo, T. Onishi, J. Catal. 102 (1986) 92–98.

[3] T. Takata, Y. Furumi, K. Shinohara, A. Tanaka, M. Hara, J.N. Kondo, K. Domen, Chem. Mater. 9 (1997) 1063–1064.

[4] H. Kato, A. Kudo, J. Phys. Chem. B 105 (2001) 4285–4292.

[5] K. Maeda, T. Takata, M. Hara, N. Saito, Y. Inoue, H. Kobayashi, K. Domen, J. Am.

Chem. Soc. 127 (2005) 8286–8287.

[6] K. Fujii, M. Ono, T. Ito, Y. Iwaki, A. Hirako, K. Ohkawa, J. Electrochem. Soc. 154 (2007) B175−B179.

[7] G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R.C. Fitzmorris, C. Wang, J.Z. Zhang, Y. Li, Nano Lett. 11 (2011) 3026–3033.

[8] M.G. Kibria, S. Zhao, F.A. Chowdhury, Q. Wang, H.P.T. Nguyen, M.L. Trudeau, H. Guo, Z. Mi, Nat. Commun. 5 (2014) 3825.

[9] K. Fujii, T. Karasawa, K. Ohkawa, Jpn. J. Appl. Phys. 44 (2005) L543−L545.

[10] M. Ono, K. Fujii, T. Ito, Y. Iwaki, T. Yao, K. Ohkawa, J. Chem. Phys. 126 (2007) 054708.

[11] Y. Iwaki, M. Ono, K. Yamaguchi, K. Kusakabe, K. Fujii, K. Ohkawa, Phys. Status Solidi C 5 (2008) 2349–2351.

[12] K. Fujii, K. Ohkawa, Jpn. J. Appl. Phys. 44 (2005) L909−L911.

[13] K. Fujii, K. Ohkawa, J. Electrochem. Soc. 153 (2006) A468−A471.

[14] K. Fujii, M. Ono, Y. Iwaki, K. Sato, K. Ohkawa, T. Yao, J. Phys. Chem. C 114 (2010) 22727–22735.

[15] K. Fujii, Y. Iwaki, H. Masui, T.J. Baker, M. Iza, H. Sato, J. Kaeding, T. Yao, J.S. Speck, S.P. DenBaars, S. Nakamura, K. Ohkawa, Jpn. J. Appl. Phys. 46 (2007) 6573–6578.

[16] T. Hayashi, M. Deura, K. Ohkawa, Jpn. J. Appl. Phys. 51 (2012) 112601.

[17] K. Ohkawa, W. Ohara, D. Uchida, M. Deura, Jpn. J. Appl. Phys. 52 (2013) 08JH04.

[18] S. Yotsuhashi, M. Deguchi, Y. Zenitani, R. Hinogami, H. Hashiba, Y. Yamada, K. Ohkawa, APEX 4 (2011) 117101.

[19] T. Sekimoto, H. Hashiba, S. Shinagawa, Y. Uetake, M. Deguchi, S. Yotsuhashi, K. Ohkawa, J. Phys. Chem. C 120 (2016) 13970–13975.

[20] D. Brunner, H. Angerer, E. Bustarret, F. Freudenberg, R. Höpler, R. Dimitrov, O. Ambacher, M. Stutzmann, J. Appl. Phys. 82 (1997) 5090–5096.

[21] T. Araki, Y. Saito, T. Yamaguchi, M. Kurouchi, Y. Nanishi, H. Naoi, J. Vac. Sci.

Technol., B 22 (2004) 2139–2143.

[22] K. Fujii, K. Kusakabe, K. Ohkawa, Jpn. J. Appl. Phys. 44 (2005) 7433–7435.

[23] J. Li, J.Y. Lin, H.X. Jiang, Appl. Phys. Lett. 93 (2008) 162107.

[24] T. Ogita, Y. Uetake, Y. Yamashita, A. Kuramata, S. Yamakoshi, K. Ohkawa, Phys.

Status Solidi 212 (2015) 1029–1032.

[25] T. Sekimoto, H. Hashiba, S. Shinagawa, Y. Uetake, M. Deguchi, S. Yotsuhashi, K. Ohkawa, Jpn. J. Appl. Phys. 55 (2016) 088004.

[26] M. Ebaid, D. Priante, G. Liu, C. Zhao, M.S. Alias, U. Buttner, T.K. Ng, T.T. Isimjan, H. Idriss, B.S. Ooi, Nano Energy 37 (2017) 158–167.

[27] K.A. Wang, C. Lian, N. Su, D. Jena, J. Timler, Appl. Phys. Lett. 91 (2007) 232117.

[28] P.D.C. King, T.D. Veal, C.E. Kendrick, L.R. Bailey, S.M. Durbin, C.F. McConville, Phys. Rev. B 78 (2008) 033308.

[29] P.G. Moses, C.G. Van de Walle, Appl. Phys. Lett. 96 (2010) 021908.

[30] L. Dong, J.V. Mantese, V. Avrutin, Ü. Özgür, H. Morkoç, S.P. Alpay, J. Appl. Phys.

114 (2013) 043715.

[31] S. Nakamura, Jpn. J. Appl. Phys. 30 (1991) L1705–L1707.

[32] S. Yotsuhashi, M. Deguchi, Y. Zenitani, R. Hinogami, H. Hashiba, Y. Yamada, K. Ohkawa, Jpn. J. Appl. Phys. 51 (2012) 02BP07.

[33] D. Iida, Y. Kondo, M. Sowa, T. Sugiyama, M. Iwaya, T. Takeuchi, S. Kamiyama, I. Akasaki, Phys. Status Solidi RRL 7 (2013) 211–214.

[34] T. Takeuchi, S. Sota, M. Katsuragawa, M. Komori, H. Takeuchi, H. Amano, I. Akasaki, Jpn. J. Appl. Phys. 36 (1997) L382−L385.

[35] T. Takeuchi, C. Wetzel, S. Yamaguchi, H. Sakai, H. Amano, I. Akasaki, Y. Kaneko, S. Nakagawa, Y. Yamaoka, N. Yamada, Appl. Phys. Lett. 73 (1998) 1691–1693.

[36] S.F. Chichibu, T. Azuhata, T. Sota, T. Mukai, S. Nakamura, J. Appl. Phys. 88 (2000) 5153–5157.

[37] H. Masui, H. Asamizu, T. Melo, H. Yamada, K. Iso, S.C. Cruz, S. Nakamura, S.P. DenBaars, J. Phys. D Appl. Phys. 42 (2009) 135106.

[38] C.J. Neufeld, S.C. Cruz, R.M. Farrell, M. Iza, J.R. Lang, S. Keller, S. Nakamura, S.P. DenBaars, J.S. Speck, U.K. Mishra, Appl. Phys. Lett. 98 (2011) 243507.

[39] M. Sumiya, T. Honda, L. Sang, Y. Nakano, K. Watanabe, F. Hasegawa, Phys. Status Solidi 212 (2015) 1033–1038.

Kazuhiro Ohkawa is a professor at King Abdullah University of Science and Technology (KAUST). After re- ceiving his M.Sc. degree from the University of Tokyo, he worked at Panasonic, Central Research Laboratories, and received his Ph.D. degree from the University of Tokyo in 1991 while he was associated with Panasonic. He moved to the University of Bremen, Germany, where he was a pro- fessor from 1996 to 1998. He worked at Tokyo University of Science from 1998 to 2016. His researchfield is optical- electrical devices such as LEDs, laser diodes, and photo- catalysts based on nitride semiconductors.

K. Ohkawa, et al. Nano Energy 59 (2019) 569–573

572

Yusuke Uetakereceived his M.Sc. degree in Applied Physic from Tokyo University of Science in 2016.

Martin Velazquez-Rizois currently a Ph. D. student at Energy Conversion Devices and Materials Laboratory at KAUST. He received his M. Sc. in Computer, Electrical and Mathematical Sciences and Engineering from KAUST in 2017.

Daisuke Iidais currently a Research Scientist at Energy Conversion Devices and Materials Laboratory at KAUST. He received his Ph.D. in Materials Science and Engineering from Meijo University in 2011. He worked as an assistant professor at Tokyo University of Science from 2014 to 2017. His researchfield is III-nitride semiconductors based optoelectronic devices.