edfas.org 15 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 2 in an antiphase manner across scales from J = 1 to J = 5. More precisely, the µSHD curve of the underfocus image is concave upward with the minimum µSHD in orientation L = 5, while the overfocus image is concave downward with the maximum µSHD in the same orientation L = 5. This result agrees well with Vanderlinde’s description that the under- and overfocus images exhibited sharp edges only in the vertical and horizontal direction, respectively. Moreover, it is reasonable to locate the µSHD curve of the true focus image between the under- and over-focus ones. Note that the under, over, and true focus µSHD curves converge at larger scales at J = 7 and J = 8. The µSHDs of the sharper image after astigmatism on smaller scales, that is, J = 1 to J = 4, are found to increase substantially, but vary less significantly between different orientations at each scale. On larger scales from J = 5 to J = 8, all µSHD curves exhibit a large peak at L = 5, and the effect of astigmatism only shows a small increase in µSHD. With the above calibration, Vanderlinde’s experience on adjusting focus and astigmatism correction can be quantitatively described and passed onto novice electron microscopists. The µSHDs of a true focus image lie right in between these of the under- and over-focus images on the smaller scales, but maintain the largest on the larger scales. An astigmatism correction makes the µSHDs of the image the largest across all scales and also varies less significantly with orientation on smaller scales. Such quantitative information can also be used to realize autonomous focus of SEM. ENERGY DISPERSIVE X-RAY SPECTROSCOPY Figures 5a to c show the EDS mapping of O, Si, and Al on a sample consisting of 1.25 µm width and 0.5 µm thickness Al metal lines in a SiO2 dielectric with a 5 keV electron beam energy, respectively. Considering the energy of the characteristic x-rays of the elements and the penetration depth of the electron beam, the choice of 5 keV electrons is adequate, and the three elemental maps are well defined compared with the SEM image provided in the original reference. Because the bright dots matter in the elemental maps, we concentrate on smaller scales for this case. Examining the µSHD curves a to c in Fig. 5g on scales from Fig. 4 Illustration of the underfocus (a), overfocus (b), true focus (c), and true focus after astigmatism (d) SEM images. Images are adapted from Vanderlinde.[16] (e) The µSHD curves of the images. (a) (b) (e) (c) (d)
RkJQdWJsaXNoZXIy MTYyMzk3NQ==