Optimization of the preparation of GaN-based specimens with low-energy ion milling for (S)TEM

We report on optimization of electron transparent GaN based specimens for transmission electron microscopy (TEM) and scanning TEM (STEM) studies by combining focused ion beam thinning and low-energy (≤500 eV) Ar-ion milling. Energy dependent ion milling effects on GaN based structures are investigated and the quality of ion milled samples is compared with that of specimens prepared by wet chemical etching. Defects formed during ion milling lead to amorphization of the specimen. The experimental results are compared with Monte-Carlo simulations using the SRIM (stopping and range of ions in matter) software. Specimen thickness was deduced from high-angle annular dark field STEM images by normalization of measured intensities with respect to the intensity of the scanning electron probe and comparison with multislice simulations in the frozen lattice approach. The results show that the thickness of the amorphous surface layer can be successfully reduced below 1 nm by low energy ion milling, leading to a homogeneous image contrast in TEM and STEM, so that good conditions for quantitative analysis can be achieved. For an ion energy of 400 eV the thickness measurements resulted in an etching rate of about 6-8 nm/min.