It is a miraculous quirk of nature that any material can survive application of the extreme electric fields necessary to extract atoms from a specimen surface one-at-a-time without bulk rupture of the material itself. The field of atom probe tomography (APT) relies upon this natural quirk and continues to find an ever widening variety of materials can be successfully investigated [1], yet the issue of premature specimen failure remains critical for continued field growth and adoption [2].Since the primary goal of an ordinary APT analysis is to gather compositional information related to a materials science problem, improving the success rate of the investigation is a secondary luxury when a single, high quality dataset is sufficient for the analysis need. In the current study, analysis yield improvement is the primary purpose of the investigation. We report on our results to date relating how analysis conditions affect yield for our standard specimen.For this study we chose a material that has many material characteristics common to the microelectronics industry but also with a history of low (but non-zero) yield in our laboratory. As illustrated in Figure 1, the material consists of a 12 nm oxide grown on a Si <100> substrate with an additional 100 nm of phosphorous doped poly-silicon was grown and implanted with boron [3,4]. Historically, the analysis yield of the doped and implanted poly-silicon region was relative high while that of the oxide was extremely low. Our goal was to statistically analyze the analysis yield for this specimen type and then consider how various variables, both specimen preparation and analysis conditions, effect yield. The initial considerations included specimen size, as measured by the tip diameter at the oxide, cap size, detection rate (DR), and laser pulse energy.For the initial investigation, some 66 specimens were manufactured with a similar geometry and a variety of oxide dimensions to look for correlation between analysis yield and specimen size. In Figure 1, the protective nickel cap is visible at the specimen apex with the bright 12 nm oxide observable 100 nm below. A narrow shank angle geometry was chosen due to its reproducibility. This allowed for predictable analysis evolution near and through the primary region of interest (the 12 nm SiO2). Based on these 66 analysis attempts, a number of statistically significant conclusions were drawn for this sample type: first, yield through the poly-silicon is very high. In fact, additional data not reported here puts the overall yield for this region near 95%. Specimens rarely fail here regardless of analysis conditions. Second, DR does affect yield through the oxide. Lowering the DR from 0.3% to 0.1% increases yield from 24% to 82% which is statistically significant at better than a 95% confidence level. Third, specimens fail much more often through the low-to-high field Si/SiO2 interface than the high-to-low field SiO2/Si interface. In fact, every specimen that survived through the top high-to-low interface continued analysis in...
Specimen survivability is a primary concern to those who utilize atom probe tomography (APT) for materials analysis. The state-of-the-art in understanding survivability might best be described as common-sense application of basic physics principles to describe failure mechanisms. For example, APT samples are placed under near-failure mechanical-stress conditions, so reduction in the force required to initiate field evaporation must provide for higher survivability—a common sense explanation of survivability. However, the interplay of various analytical conditions (or instrumentation) and how they influence survivability (e.g., decreasing the applied evaporation field improves survivability), and which factors have more impact than others has not been studied. In this paper, we report on the systematic analysis of a material composed of a silicon-dioxide layer surrounded on two sides by silicon. In total, 261 specimens were fabricated and analyzed under a variety of conditions to correlate statistically significant survivability trends with analysis conditions and other specimen characteristics. The primary result suggests that, while applied field/force plays an obvious role in survivability for this material, the applied field alone does not predict survivability trends for silicon/silicon-dioxide interfaces. The rate at which ions are extracted from the specimen (both in terms of ions-per-pulse and pulse-frequency) has similar importance.
Continuing advances in Atom Probe Tomography and Focused Ion Beam Scanning Electron Microscope technologies along with the development of new specimen preparation approaches have resulted in reliable methods for acquiring 3D subnanometer compositional data from device structures. The routine procedure is demonstrated here by the analysis of the silicon-germanium source-drain region of a field effect transistor from a de-packaged off-the-shelf 28 nm design rule graphics chip. The center of the silicon-germanium sourcedrain region was found to have approximately 180 ppm of boron and the silicide contact was found to contain both titanium and platinum.
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.