edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 1 18 (the area averaged along one dimension) and 3D chemical maps of samples containing multiple layers, thereby resolving the composition and location of different elements segregated to interfaces or grain boundaries due to processing. Figure 4 shows an example of a high-entropy Cr0.2Mn0.2Fe0.2Co0.2Ni0.2 alloy in which ductility measurements recorded as a function of temperature suggested that grain boundary embrittlement was the reason for crack initiation.[13] To gain further insight into the embrittlement, an APT sample was extracted from a grain boundary in the alloy (Fig. 4a). The 3D tip reconstruction results revealed Cr-, Mn-, and Ni-rich regions at the grain boundary (Fig. 4b-d), and 1D composition profiles along different lines showed concentrations of these elements was up to 10 at.% greater than in the bulk. This fast nanosegregation during tensile testing at elevated temperatures reduces grain boundary cohesion and leads to embrittlement and grain boundary cracking. In the APT reconstruction, complex grain boundaries and the distribution of elements or clusters can be visualized. Site-specific specimen prep- aration has been used to access discrete nanoscale devices including p- and n-type MOSFETs, gate stacks and FinFET structures for APT analysis.[14-16] In one example,[14] alloying elements and dopants, such as Pt or As, are commonly present in the region near silicide contacts in an n-type MOSFET, but the exact role of stress and defects on the redistribution of those species upon annealing is still a topic of debate. In Fig. 5a, the 3D APT reconstruction shows the separate layers of the NiSi contact, poly-Si gate, and the SiOx spacers. The authors find notable clusters of platinum arsenide in the NiSi phase and inhomogeneous segregation to interfaces (not shown). However, 1D top-down composition profiles through a central 10 nm diameter cylindrical region either through a patterned device (Fig. 5b) or on an unpatterned blanket wafer (Fig. 5c) reveal little difference between the Pt and As distributions. Such results, according to the authors, indicate the total silicide process, rather than spatial confinement in the patterned device, controls element distributions upon annealing in these structures.[14] ADVANCED CHARACTERIZATION OF MATERIALS USING ATOM PROBE TOMOGRAPHY(continued from page 16) (a) Fig. 4 Example of segregation to grain boundaries and cluster identification. (a) TEM image of the alloy sample, (b) APT reconstruction “atom map” of the sample tip showing grain boundary segregation of Cr and Ni, (c) side (cross-section) view of the grain boundary with isoconcentration surfaces delineating areas of 23 at.% Ni and 24 at.% Cr enrichment, (d) plan view (looking perpendicular to the grain boundary), (e) 1D composition along arrow “E,” and (f) 1D composition along arrow “F” in section C. Re- produced from Ref 13 under the terms of the Creative Commons Attribution License. (b) (c) (d) (e) (f)
RkJQdWJsaXNoZXIy MTYyMzk3NQ==