Tech News
← Back to articles

Limit of atomic-resolution-tomography reconstruction of amorphous nanoparticles

read original related products more articles

(a) A simulated projection image (300 × 300 pixels) at the sampling of 0.33 Å/pixel and D = 5.6 × 104 e/Å2, using a 200 kV electron beam of 25 mrad convergence angle and a background noise of 1 e. The simulated images were centered using atoms at the surface extremities (Supplementary Note 1). (b) The preprocessed image of (a) with up-sampling of 3 times and noise smoothing, and (c) the corresponding projection obtained from the reconstructed tomogram using 55 projections from −70° to 70°. (b) The input atomic model with Pt, Pd, and Ni for the heavy, medium, and light atoms and plots showing three types of atoms separately. (c) The determined atomic structure from the simulated AET experiment as shown in the same way as (b). (d) The measured integrated atomic peak intensity histograms for three electron fluences. The atomic peaks were identified using a background threshold (1% of the maximum tomogram intensity). (e) The assignment of atomic types using the k-means clustering algorithm in ref. 3 and the CT obtained at D = 5.6 × 104 e/Å2. The low success rates for atomic identification are caused by the overlap of atomic peak intensities, which is also seen in the uploaded experimental data in ref. 3.