Bias-dependent Photoabsorption Properties of GaN Porous Structures under Back-side Illumination

The high-density formation of semiconductor nano- and micro-structures—targeting applications such as optoelectronic devices and high-efficiency energy-conversion devices—has been intensely researched. Among these structures, porous structures formed by an electrochemical process are one of the most promising due to their unique features. The purpose of this study is to clarify the photoabsorption properties of GaN porous structures formed by the photo-assisted electrochemical process. GaN is one of the most attractive materials because of its chemical stability and its potential to achieve direct photoelectrolysis by solar power without consuming electric power. Fundamental photocurrent on porous structures was first measured and compared with that on GaN planar substrates. The mechanism of the observed photoabsorption properties of GaN porous structures was then investigated from both experimental and theoretical aspects.

A custom-made cell equipped with a crystal window and indium-tin-oxide (ITO) plate was used for both the porous formation and spectroscopic measurements. The electrochemical cell has three electrodes: an n-type GaN electrode as a working electrode (W.E.), a platinum (Pt) counter electrode (C.E.), and Ag/AgCl reference electrode (R.E.). GaN-porous structures were formed using a photo-assisted electrochemical process in BSI mode [1]. As a UV light source for the photo-assisted electrochemical process, a monochromatic light with a wavelength of 370 nm was irradiated using a xenon (Xe) lamp through an optical band-path filter. The pore diameter and density of the GaN-porous samples were estimated to be about 25 nm and 6.5×10¹⁰ cm^-2, respectively, which remain almost the same throughout the porous layer. It was previously demonstrated that deeper pores were obtained in BSI mode, as compared with those obtained in the FSI mode, because the pores are only etched at the pore tips due to the supply of holes.

Photocurrents on the planar and porous electrodes measured in the NaCl electrolyte under the light irradiation with an intensity of 100 µW/cm². For the planar electrode, large photocurrents were observed under light with wavelength of 360 nm. As the wavelength of the monochromatic light increased, the photocurrents drastically decreased. This result indicates that the band-edge absorption around 3.4 eV, corresponding to photon energy with a wavelength of 365 nm, plays an important role in the generation of photocarriers on the planar electrode. As expectedly, larger photocurrents were observed on the porous electrode, reflecting the large surface area of the electrode and its low photo-reflectance properties due to the high-density array of pores [2]. It is also noted that the photocurrents were even observed under illumination with wavelength of 380 nm, corresponding to a photon energy of 3.26 eV, which is smaller than the bandgap energy of the GaN bulk (i.e., 3.4 eV). The photocurrents obtained on the porous electrode are 11 times larger under the light with a wavelength of 370 nm and 380 nm. These results clearly indicate that light with photon energy, hν, below the bandgap contributes to the generation of photocarriers on GaN-porous structures.

The Franz-Keldysh effect is one possible phenomenon to explain the present results, in which high electric field causes a redshift of the photoabsorption edge, leading to the presence of absorption below the bulk bandgap. Generally, a high electric field is applied to the GaN surface or interface due to Fermi-level, E_F, pinning at about 1.0 eV below the conduction band minimum, E_C. For example, when doping density is as high as 10¹⁸ cm^-3, the internal electric field in a thin depletion layer with width of several dozen nanometers reaches 5×10⁵V/cm. In such a situation, the wave function of an electron in the conduction band and a hole in the valence band are characterized by tails in the forbidden bandgap, resulting in band-to-band transition at an energy lower than the bandgap. To clarify the possibility that the Franz-Keldysh effect can explain that observation, the potential distribution of the GaN porous structures was calculated by a computer program. In this calculation, no external voltage was applied, but the potential distribution near the pore tips was strongly modified, where the contour intervals of the potential became narrow. These results suggest that the observed redshift of the photoabsorption edge arises from the high-electric field concentrated at the pore tips and that it is consistently explained by the Franz-Keldysh effect.

[1] A. Watanabe, Y. Kumazaki, Z. Yatabe and T. Sato, ECS Electrochem. Lett. 4 (2015) H11.

[2] Y. Kumazaki, A. Watanabe, Z. Yatabe and T. Sato, J. Electrochem. Soc. 161 (2014) H705.