The semiconductor and memory industries have experienced a major transition from conventional planar devices to complex 3D architectures, such as 3D NAND and FinFETs. New materials and designs have raised metrology complexity and greatly increased the number of critical dimension measurements required for process control. 3D NAND presents many metrology challenges including low contrast features, high aspect ratio trenches, and holes in multiple stacks. Many technologies, such as STEM/TEM imaging, critical dimension small angle X-ray spectroscopy (CD-SAXS) and through focus scanning optical microscopy, have been explored to address these metrology challenges. The techniques in their present forms haven't adequately achieved the sensitivity and resolution to measure the required critical dimensions of these 3D NAND structures. Automated STEM/EDS metrology capabilities were evaluated using a commercially available 3D NAND memory device to compare critical dimensions to STEM images. Site navigation, feature alignment and tracking, layer alignment, focus, EDS mapping, on-line elemental quantification, and CD metrology were done entirely in automation using an iFast recipe. Active drift correction by beam shift and passive drift correction by frame integration were used to eliminate the effects of sample drift. Quantification of the 512 x 512 maps was done automatically at speeds more that 10x faster than manual routines in Esprit. Figures 1(b) and 1(c) are examples of a quantified EDS color map and individual grey scale elemental maps used for metrology measurements. Light elements (N and O), along with heavier elements, were detected with high contrast for robust metrology. The green arrows in Figure 1 (c) indicate regions of interest placed by an automated measurement recipe. The thickness of individual layers was obtained using an Edgefinder activity, which collapses a two dimensional spatial array of pixel intensities into a 1-dimensional plot of intensities as a function of distance.
Automated S/TEM (auto-S/TEM) and by extension automated energy dispersive x-ray spectroscopy (auto-EDS) enable the collection of large volumes of data that historically was not feasible on a sustained basis using manual S/TEM operation. Automation removes variabilities attributed to operator training and bias while enabling repeatable, precise, and consistent data. This paper examines general considerations for high volume and auto-EDS metrology. Maps are evaluated relative to analysis time and the requisite signal-to-noise (SNR) necessary for adequate measurement precision. Auto-EDS parameters (dwell time and mapping time) were investigated to maximize EDS net counts and SNR for throughput. A methodology was presented to monitor the SNR ratio of critical dimensions which can be extended to further understand the minimum value needed to maintain an acceptable measurement precision. It is important to further study other means of increasing the SNR, such as beam current, accelerating voltage, and larger solid-angle detectors.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.