Exploring Differences in Amorphous Layer Formation during FIB Sample Preparation between Metals and Non Metals

Abstract

Over the last decade dual beam FIB has proved increasingly useful in a laboratory setting. Researchers have used dual beam FIB's in a wide variety of experiments, including but in no way limited to fabrication of nano pillars for micromechanical testing, modification of circuits during use, and the creation of tooling for nanoscale machining. The field of high resolution microscopy in particular has seen huge benefits from FIB processing in the form of high quality orientation specific lamella. One drawback of the technique is the formation of damage caused as a result of the impact of fast gallium ions on samples. This damage takes the form of the formation of lattice defects (e.g., dislocation loops) and/or amorphous layers on the surfaces of the FIB'd samples. The latter form of damage can degrade the quality of high resolution images and so the mechanism of its formation has been the subject of considerable study. However, these various studies of amorphization during FIB processing have largely focused on the effect of gallium ions on electronic materials. This concentration of research is likely due to several primary factors, the extensive use of and investment by the semiconductor industry in FIB research and the geometrical factors complicating measurement of amorphous damage. In our research, we have found that there are considerable differences between the amounts of amorphous damage between, on the one hand, semiconductors and, on the other hand, metallic materials. The present paper focuses on attempts to identify the origin for these differences.The main problem in studying the formation of amorphous damage is the consistency in results as a function of the specifics of sample geometry. Most often, thin FIM lamellae are used, and the problem with the use of such samples is the difficulty in distinguishing between "primary" damage and redeposition of material that has been initially milled from the sample surface [1]. In the present study, needle shaped samples have been employed as these allow for consistent measurements of damage formation (as a function of sample thickness, which can be readily determined from the local diameter of the needle), and by positioning of a needle in close proximity to one being milled, the phenomenon of re-deposition can also be examined. Multiple needles may be made using a single FIB trenching operation. In this study one 28 µm * 1.5 µm * 4.5 µm foil was sufficient to create six posts suitable for needle shaping and one slab suitable for creation of a TEM lamella. Redeposition was noted to be a major concern when making closely spaced needles, in Figure 1 . Needles made at 30 kV using NiAl and commercially pure metals such as Ti, Ni, Al, and Fe all contained disordered layers with average thicknesses ranging from a ≈ 2 nm to ≈ 8 nm.

Comparing Xe+pFIB and Ga+FIB for TEM sample preparation of Al alloys: Minimising FIB‐induced artefacts

“…Amorphised layers induced by ion beam milling have been reported in several studies using both TEM and EBSD (electron backscatter diffraction) 24,40,41 . For example, Presley et al reported an amorphised layer with a similar 4 nm thickness for a 30 kV Ga + FIB milling of Al7075 29 . Kelley et al 42 compared amorphisation for Si between Ga + and Xe + and found Xe + ions created a thinner amorphous layer, consistent with our results although the layers were much larger (21.5 nm with Ga + 42 ) due to the nature of covalent bonds of Si and faster migration of ion‐induced lattice defects in Al 43 …”

“…pillars and found that Xe + ‐prepared pillars had higher yield strengths than Ga + ‐prepared pillars because it avoided Ga segregation on the grain boundary that reduced the strength of polycrystalline aluminium pillars. However, questions remain as to the suitability of Xe + pFIB for the preparation of Al alloy TEM specimens, such as compositional contamination, structural changes (eg new phases), 26,27 ion‐induced damage (eg dislocation loops), 28 and amorphisation, 2,29,30 Xe ‘bubble’ or precipitate formation, 27,31,32 as well as levels of redeposition 5 , 14 …”