edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 25 NO. 4 24 test to deliver a multiscale 3D analysis platform that does not compromise high-resolution imaging with accurate and automatic material removal capability. Future work will address the sensitivity of our RR control in the presence of multilayer materials with minimum hardness variations, for example in case of Si/SiGe nanolaminates representing a technologically relevant case. ARTIFACT-FREE VOLUMETRIC MAPPING Modes such as scanning spreading resistance microscopy (SSRM) and conductive atomic force microscopy (C-AFM) have proven their value in many areas of materials research and nanoelectronics FA. For example, using C-AFM as a secondary mode of operation, substantial scan-to-scan variation in the electrical profiles have been observed and are intrinsically related to the variability found in the tip-sample junction of high-pressure scans. Figure 5 shows the comparison between two independent probes, alternating their operation on the same area of the sample surface after removing the TiN top-electrode. The electrical profiles are acquired in a location with a high density of dot cells (C-AFM image in Fig. 5a), after the top layer (TiN) is removed and the oxide layer is exposed. Both profiles are collected at the same location at the depth of 40 nm. This ensures it is positioned well below the topelectrode in the oxide layer. By alternating the probes while maintaining the sample location, one can compare the data acquired for two different conductive probes, namely a diamond probe (CDT-NCHR, k = 92 N/m) and a Pt/Ir coated Si probe (ATEC-CONTPt, k = 7 N/m). Note that here the diamond probe is deliberately used at the same high-pressure conditions applied during the process of material removal, while the Pt/Ir probe is scanned at low force, i.e., the minimum force needed to obtain a good electrical contrast. The results indicate striking differences for the same structures when measured with the two probes. For the case of the high-pressure scan with the diamond probe, multiple shape and conductive artifacts can be identified. For example, the nominal diameter of the dots (i.e., 100 nm) is often magnified and shown with a distorted round shape (inset Fig. 5b left). Similarly, the electrical contrast presented blurred features with the presence of double-tip artifacts. In contrast, Fig. 5c (right inset) shows the measurement obtained with the Pt/Ir coated Si probe, measuring at a low pressure in the exact same location of the sample. Here, the read-out provides a uniform leakage current distribution across the oxide surface, with a dimension and shape that is in line with the nominal values. This demonstrates the critical role of probe switching for the acquisition and three-dimensional segmentation of the (artifact-free) dataset in high-resolution tomographic imaging with scalpel SPM. Analysis of a 3D NAND dummy structure tested the volumetric mapping Fig. 5 (a) Schematic of the MIM structures used to test the suppression of artifacts by alternating probes in C-AFM experiment. (b) High-pressure electrical scan of C-AFM where the current profile presents blurred features with the presence of double-tip artifacts. (c) Artifact-free measurement obtained with the Pt/Ir coated probe, measuring at a low-pressure in the exact same location of the sample, thanks to alternating the removal and sensing scans. (d) Combined data set of conductive mapping then merged to build 3D conductive tomogram of the 3D NAND dummy sample. (d) (c) (a) (b)
RkJQdWJsaXNoZXIy MTMyMzg5NA==