Automatic <scp>FIB‐SEM</scp> Preparation of Straight Pillars for In‐Situ Nanoindentation

In‐situ indentation tests in FIB‐SEMs are a powerful tool to characterize the mechanical deformation properties of matter at the micron scale [1,2]. FIB milling is used to produce micrometer sized – usually cylindrical – pillars from the bulk, while SEM imaging allows to determine the geometry of the pillars prior, during and after the load‐displacement data acquisition. In this work, different automatic workflows were tested for the preparation of high aspect‐ratio pillars with well‐defined geometries, in particular with perfectly perpendicular side walls. A state‐of‐the‐art FIB‐SEM instrument was used to fabricate the pillars. They were machined by milling a series of concentric rings with decreasing FIB currents into the sample. Hereby, the sample was at 54° tilt to ensure normal incidence of the FIB. The last and smallest ring was milled with a 3 nA probe, which yielded a slightly material dependent pillar wall angle of around 2° to the sample normal. After this pre‐preparation step, the geometry of the pillars was refined further to achieve perfectly perpendicular pillar side walls using lathe milling [3]. The ideal cylindrical geometry is highly desirable, because it is easier to model for a reliable analysis of the load‐displacement measurement. Two different lathe milling techniques were implemented in this work and compared. They both involve a number of FIB milling steps each performed at different sample rotations to shape the pillar wall along its whole circumference. After each sample rotation the pillar needs to be repositioned accurately by means of SEM and FIB image recognition of fiducial marks on the sample. The first approach, #1, is similar to the one described in [3]. The walls of the pillar are shaped from the side by FIB milling at zero degree stage tilt as shown in Figure 1(a). For sample repositioning a single fiducial is used which is placed – for symmetry reasons – exactly in the center of the pillar (see Figs. 1(b) and (c)). Approach #1 was automated using the application programming interface (API) of the FIB‐SEM instrument. Including lathe milling the total preparation time per typical pillar adds up to about an hour. Because of the space needed for the fiducial mark only pillars with diameters, d>5 µm, can be fabricated automatically in this way. The need to fabricate smaller pillars with d<5 µm motivated an alternative and new lathe milling workflow, #2 (see Figure 2). Here, the walls of the pillar are shaped from the pillar top (sample at 54° tilt), as it was done in the pre‐preparation step, too. By slightly under‐tilting the sample a few degrees an edge of the pillar was exposed to the FIB for machining (Fig 2(a)). The sample was then rotated and repositioned for the next milling step. This process was iterated to cover the full circumference of the pillar. In order to reduce the number of iterations the milling was done following the green boomerang type of shape depicted in Figure 2(b). Only eight iterations – as compared to at least 18 with approach #1 – were needed to obtain an almost perfectly circular pillar cross section (see Fig. 2(c)). In summary, the new lathe milling process can be used to machine very small pillars. It can be combined easily with the pillar pre‐preparation step for a fully automatic pillar preparation. Further, because it gets along with less iterations, it is faster than previous approaches.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.