Etch-Pit Density Investigation on Both Polar Faces of ZnO Substrates

Abstract

It is well known that the presence of dislocations in electronic and optoelectronic devices based on semiconductor p-n junctions can lead to short circuits of these junctions and thus to malfunction. The main cause of dislocations in homoepitaxial layers is threading dislocations present in the initial substrate. Thus, it is extremely important to investigate defect structure, particularly dislocations appearing on the surface of commercial wafers, which are employed for epitaxial growth. Zinc oxide is a very promising material for applications in thermally stable optoelectronic devices, such as light-emitting diodes and laser diodes emitting in the ultraviolet spectral region, mostly because of its high exciton-binding energy 60 meV. Much work has been done on the heteroepitaxy of ZnO, for example, on sapphire, in order to get good-quality epitaxial layers. However, due to problems associated with heteroepitaxy of ZnO, such as high dislocations density as a result of lattice mismatch and diffusion of Al atoms into epitaxial layers in the case of growth on sapphire, obtaining p-type ZnO has rather proven to be challenging. Great attention is now being focused on ZnO substrates with the intention to realize homoepitaxially grown p-doped layers. This technology has not yet been substantially developed and it is still not clear on which polar face of ZnO substrates p-doping can be efficiently and reproducibly realized. In this situation, both zinc- and oxygen-polar faces of substrates have to be characterized. In the case of compound semiconductors with polar faces, revealing etch pits by chemical etching on both faces is typically not possible. 1 It was shown that zinc- and oxygen-terminated sides of ZnO bulk wafers are influenced by acids in drastically different ways. 2 This behavior was explained in terms of the surface-bonding model. The etching rate of oxygen-terminated ZnO is much higher than that of zinc-terminated ZnO. The morphologies are also qualitatively different after chemical treatment. The etching of oxygenterminated ZnO results in the formation of so-called hillocks, which do not provide any useful information about the structure of defects. In contrast to these hillocks, etch pits appear on the opposite polar side. Revealing these etch pits by wet-chemical etching is normally used to define the density of threading dislocations on zinc face. As an alternative to wet-chemical etching, it is possible to employ thermal treatment. Gu et al. showed the appearance of atomic terraces after annealing in air for 3 h at 1050°C. 3 Graubner et al. state that annealing in an oxygen environment causes the formation of atomic terraces only at temperatures higher than 1100°C. 4 In 1971, the appearance of hexagonally formed etch pits on an oxygen face of ZnO bulk wafer after thermal treatment in helium atmosphere at 1200°C was reported by Wolf et al. 5 After considering these reports, questions about the dynamics of the thermal treatment process arise. The influence of the gas environment and temperature effect are not clear. In this work, we present the results on chemical and thermal etching of ZnO substrates supplied by two companies: TokyoDenpa and Crystec. The first approach was used to reveal etch pits on a zinc-polar face and the latter one on an oxygen-polar face. The thermal treatment under special conditions was also used to improve the morphology of ZnO substrates. Experimental For chemical etching, the following acids and their solutions were used: 10% and 1% solutions of HCl, 15% and 1% solutions of H3PO4, 20% and 1% solutions of HNO3, and a 1% solution of CH3COOH. In most cases, the samples from both suppliers were etched at room temperature. The etching process with acetic acid was conducted at room temperature and at 5°C. The etching durations were specially optimized in each case. The samples were treated in 10% and 1% HCl for 6 s, in a 15% solution of H3PO4 for 40 s, and in 1% H3PO4 for 5 s. The etching duration in 20% HNO3 was 90 s and in 1% HNO3, 30 s. The smallest etching duration 2s was for CH3COOH. Thermal treatment was done at 1050, 1100, and 1150°C, for durations from 1 to 5 h. The samples from Crystec were annealed at 1050°C in quartz‐glass oven in excess air and nitrogen for 3 and 4 h. Crystec substrates were also thermally etched at 1100°C for 3, 4, and 5 h and at 1150°C for 1, 2, and 3 h. Atomic force microscopy AFM measurements were done with a DME microscope in alternating current ac mode before and after chemical and thermal treatment.