Defect formation during chlorine-based dry etching and their effects on the electronic and structural properties of InP/InAsP quantum wells

International audienceThe general objective is the investigation of the defects formed by dry etching tools such as those involved in the fabrication of photonic devices with III-V semiconductors. Emphasis is put on plasma exposures with chlorine-based chemistries. In addition to identifying these defects and describing their effects on the electro-optic and structural properties, the long-term target would be to predict the impact on the parameters of importance for photonic devices, and possibly include these predictions in their design. The work is first centered on explaining the experimental methodology. This methodology starts with the design and growth of a quantum well structure on indium phosphide, including ternary indium arsenide/phosphide quantum wells with graded arsenic/phosphor composition. These samples have then been characterized by luminescence methods (photo- and cathodoluminescence), high-resolution transmission electron microscopy, and secondary ion mass spectrometry. As one of the parameters of importance in this study, the authors have also included the doping level. The samples have been exposed to the etching plasmas for "short" durations that do not remove completely the quantum wells, but change their optical signature. No masking layer with lithographic features was involved as this work is purely oriented to study the interaction between the plasma and the samples. A significant difference in the luminescence spectra of the as-grown undoped and doped samples is observed. A mechanism describing the effect of the built-in electric field appearing as a consequence of the doping profile is proposed. This mechanism involves quantum confined Stark effect and electric-field induced carrier escape from the quantum wells. In the following part, the effects of exposure to various chlorine-based plasmas were explored. Differences are again observed between the undoped and doped samples, especially for chemistries containing silicon tetrachloride. Secondary ion mass spectrometry indicates penetration of chlorine in the structures. Transmission electron microscopy is used to characterize the quantum well structure before and after plasma bombardment. By examining carefully the luminescence spectral properties, the authors could demonstrate the influence of the etching plasmas on the built-in electric field (in the case of doped samples), and relate it to some ionic species penetrating the structures. Etching plasmas involving both chlorine and nitrogen have also been studied. The etching rate for these chemistries is much slower than for some of the silicon tetrachloride based chemistries. Their effects on the samples are also very different, showing much reduced effect on the built-in electric field (for the doped samples), but significant blue-shifts of the luminescence peaks that the authors attributed to the penetration of nitrogen in the structures. Nitrogen, in interstitial locations, induces mechanical compressive stress that accounts for the blue-shifts. Finally, from the comparison between secondary ion mass spectrometry and luminescence spectra, the authors suggest some elements for a general mechanism involved in the etching by chloride-chemistries, in which a competition takes place between the species at the surface, active for the etching mechanism, and the species that penetrate the structure, lost for the etching process, but relevant in terms of impact on the electro-optic and structural features of the exposed materials. © 2016 American Vacuum Society