To improve the durability of water splitting reaction, we have developed a NiO/AlGaN photoanode in which highly crystalline AlGaN grown on a GaN substrate is covered with a protective layer of NiO [1]. InGaN, a mixed crystal of GaN and InN, has a narrower band gap than AlGaN or GaN, and it has been reported that InGaN photoanodes have higher solar-to-hydrogen conversion efficiencies compared with ones made from those materials [2]. To improve their efficiency further, it is necessary to increase the photocurrent by increasing the thickness of InGaN with the aim of maximally absorbing the light transmitted into it. However, increasing the thickness may increase dislocations that occur when growing the InGaN layer. In this study, to investigate the effect of the thickness of the InGaN layer on photoelectrochemical properties, we measured the photocurrent in a water splitting reaction using NiO/InGaN photoanodes with different InGaN thicknesses. In addition, we measured the CO2 reduction products after 350 h of continuous light irradiation using the best-performing photoanode (NiO/InGaN on GaN substance) in gas-phase CO2 reduction reaction system. In0.02Ga0.98N (thickness: 100, 200, 300, 400, 500, 600, and 700 nm)/n-GaN heterostructures were grown by metal organic chemical vapor deposition on a sapphire substrate. The InGaN layer had a composition gradient in which the In composition increased from the InGaN/n-GaN interface toward the InGaN surface [2]. The thickness of In0.02Ga0.98N was evaluated by secondary ion mass spectroscopy. The light transmittance was measured with a spectrophotometer, and the arithmetic average roughness (R a) of the In0.02Ga0.98N surface was evaluated with atomic force microscopy. Ni thin film was fabricated on the In0.02Ga0.98N surface by vacuum evaporation. The sample was then heat-treated at 563 K for 60 min in air to form the NiO layer [1]. NiO/In0.02Ga0.98N/n-GaN as the photoanode and Pt wire as the cathode were immersed in 1 mol/L NaOH and the water splitting reaction proceeded. The photocurrent was measured with a potentio-galvanostat under application of 1.7 V and irradiation with a Xe lamp adjusted to the same intensity as AM1.5G (λ ≤ 374 nm). Next, to quantify the CO2 reduction products, NiO/In0.02Ga0.98N (thickness: 500 nm)/n-GaN on GaN substrate was prepared as the photoanode. Thermocompression was used to prepare a composite sheet using gold fiber as a cathode and NafionTM as an electrolyte membrane [3]. The sheet was set so as to separate the oxidation and reduction chambers. 0.5 mol/L CsOH was used in the oxidation chamber and CO2 was directly supplied at 45 sccm to the cathode in the reduction chamber. The photocurrent was measured under similar conditions as the water splitting reaction. The produced gas was determined every 10 h with gas chromatography, and the produced liquid after 350 h was determined with high-performance liquid chromatography. The photocurrent density measured 1 min after the light irradiation ceased increased with thickness from 100 to 500 nm, while it decreased for thicknesses more than 500 nm. The maximum photocurrent density was 0.81 mA/cm2. It was considered that the light absorption ability of the In0.02Ga0.98N layer increased with its thickness. Moreover, since R a increased, it was thought that the crystallinity of the In0.02Ga0.98N decreased as its thickness increased. These results indicate that increasing the In0.02Ga0.98N thickness increases the number of generated electron-hole pairs because of the increased light absorption. On the other hand, these also indicate that the decrease in crystallinity increases the recombination probability of electron-hole pairs. Here, it was assumed that the photocurrent density reached its maximum when the thickness was 500 nm. The results of measuring the CO2 reduction products using the NiO/In0.02Ga0.98N (thickness: 500 nm)/n-GaN on the GaN substrate photoanode and the composite sheet are shown in Fig. 1. CO and HCOOH were detected, with the cumulative amounts of product per photoanode area being 3237 and 263 µmol/cm2, respectively, after 350 h. The average Faradaic efficiencies of CO and HCOOH during 350 h were 83 and 6%, respectively. The cumulative amount of O2 was 1865 µmol/cm2, and the average Faradaic efficiency during 350 h was 96%. Therefore, the NiO/In0.02Ga0.98N (thickness: 500 nm)/n-GaN on GaN substrate was found to be an effective photoanode for a gas-phase CO2 reduction reaction system with high efficiency and durability.[1] Y. Uzumaki et al., PRiME 2020, L04-3054 (2020).[2] K. Ohkawa, ECS Transactions 66 (1), 135-138 (2015).[3] S. Sato et al., 2023 ECSJ Fall Meeting, S2-1-09 (2023) (in Japanese).Fig. 1. Amount of CO, HCOOH, H2, and O2 measured using NiO/In0.02Ga0.98N (thickness: 500 nm)/n-GaN on GaN substrate photoanode versus irradiation time. Figure 1
Read full abstract