edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 25 NO. 3 20 wave by the unit cell of the crystal structure, meaning it is sensitive to the types of atoms present and their arrangements within the unit cell. Refinement, i.e., a comparison of simulated and experimental data to quantify the structure factor that best reproduces the experimental results, of energy filtered diffraction patterns (energy filtering removes inelastic scattering which improves the fidelity of the refinement) was used to quantify the concentration of nitrogen in dilute-nitride GaNxAs1-x epilayers as a function of position.[48] It also provided insight into whether N incorporated interstitially or substitutionally, as well as in identifying the presence of As vacancies. Quantitative refinement of structure factors can also be used to understand bonding through the generation of charge density maps.[49,50] Yet to date, most structure factor refinement using CBED has been carried out on bulk crystals intentionally positioning the electron probe away from defects, however, a spatially resolved method to understand changes to bonding in proximity to defects has been proposed.[51] Other new 4D-STEM techniques can provide high spatial resolution complements to measurements without site specificity. For example, dilatometry or positron annihilation spectroscopy have traditionally been used to measure bulk averages of vacancy concentrations, it has recently been shown that 4D-STEM can be used to monitor more local changes in point defect concentrations.[52] This was done by directly correlating independent measures of changes in lattice parameter and the volumetric expansion as external stimuli are applied (Fig. 4 panel i). Changes in defect concentrations were monitored in regions of approximately 100 nm2 during in situ heating of gold or as a function of electron radiation in aluminum films.[52] Because defects perturb the long-range order of a crystal and therefore the electron scattering process, several approaches look to analyze the symmetry of diffraction patterns to identify defects. One such method is coined as symmetry-STEM (S-STEM) and is based on the cross correlation of diffraction patterns with themselves after a symmetry operation (e.g., rotation or mirror) has been applied. In the analysis, intensity maxima occur when the symmetry operation is satisfied and intensity decreases at symmetry-breaking events. It was demonstrated that this approach can track changes in symmetry with atomic resolution (Fig. 4 panel ii). Because the image contrast of this correlation map can be interpreted as deviations from a specific symmetry operation, this approach can be used to identify regions of the crystal Fig. 4 Panel i reproduced from Mills.[52] Vacancy concentration mapping of Au as it is heated, independent measurements of lattice parameter and length quantified the increase in vacancies (ΔN/N). Panel ii reproduced from Krajnaka.[53] An example of symmetry-STEM analysis on [100] CeB6. S-STEM images for different applied-symmetry operations are compared with the traditional imaging modes of BF-, ABF-, and HAADF-STEM. FOUR-DIMENSIONAL SCANNING TRANSMISSION ELECTRON MICROSCOPY (continued from page 17)
RkJQdWJsaXNoZXIy MTMyMzg5NA==