May_EDFA_Digital

edfas.org ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 23 NO . 2 16 step, a region of interest was imaged at high resolution. In the first demonstration, this scan covered an area 40 µm in diameter and demonstrated a 3D resolution of 18.9 nm. The volume of this measurement is 15 times larger compared to what was demonstrated in tomography mode, and was only limited by the available measurement time, which was two days in this case. Figure 5 shows a 3D rendering of the resulting 3D dataset. An animated video of this volume rendering is available at https:// youtu.be/GvyTiK9CNO0. A virtual slice of a small region of this dataset, taken from the second lowest layer in the chip, is shown in Fig. 6a. Figure 6b shows the result of copying the same chip and performing imaging by mechanical delayering to this layer and scan- ning electron microscopy (SEM). CONCLUSION AND OUTLOOK It has been shown how3D x-ray ptychog- raphy can bridge the imaging gap between conventional x-ray microscopy and electron microscopy. In tomography geometry a resolution of currently 14.6 nm was reached in integrated circuits, and the method has proven useful to investigate solder media having features on the nanoscale. For chips, the requirement of producing isolated samples for tomography is removed in laminog- raphy geometry, where entire ICs can be mounted and any region of interest can be selected and imaged at selectable resolution levels. The obtained results show Fig. 4 (a) Illustration of the LamNI measurement geometry. The coherent x-rays propagate from right to left. The sample is mounted on a 2D scanner and moves in a plane perpendicular to a rotation stage. In contrast to tomography the axis of rotation is not perpendicular to the x-ray propagation direction. (b) Rendering of the design of the LamNI microscope. Accurate sample scanning is achieved via dedicated laser interferometric measurements and closed-loop positioning. Figure reproduced from Holler et al. [29] Fig. 5 3D rendering of the laminography measurement dataset. (a) A combined rendering of the low resolution and high resolution dataset. (b) A rendering of only the volume measured at high resolution. (c) A rendering of a sub-volume of (b). In front of this region a functional unit of the circuit is color coded. Figure adapted from Holler et al. [28] (a) (b) (a) (b) (c) (continued on page 18)