Dynamic behavior of metal nanoparticles in MOF materials: analysis with electron microscopy and deep learning

Abstract

Electron microscopy is a key characterization technique for nanoscale systems, and electron microscopy images are typically recorded and analyzed in terms of the morphology of the objects under study in static mode. The emerging current trend is to analyze the dynamic behavior at the nanoscale observed during electron microscopy measurements. In this work, the study of the stability of MOF structures with different compositions and topologies under conditions of an electron microscope experiment revealed an unusual dynamic behavior of M NPs formed due to the electron-beam-induced transformation of specific frameworks. The transition to the liquid phase led to spatial movement, rapid sintering, and an increase in the M NPs size within seconds. In the case of copper nanoparticles, instantaneous sublimation was observed. The dynamic behavior of Co NPs was analyzed with a computational framework combining deep learning and classic computer vision techniques. The present study for the first time revealed unique information about the stability of a variety of MOFs under an electron beam and the dynamic behavior of the formed M NPs. The formation of Fe, Ni, Cu, and Co NPs was observed from a molecular framework with a specific subsequent behavior – a stable form for Fe, excessive dynamics for Co, and sublimation/condensation for Cu. Two important outcomes of the present study should be mentioned: (i) electron microscopy investigations of MOF samples should be made with care, as decomposition under an electron beam may lead to incorrect results and the appearance of “phantom” nanoparticles; and (ii) MOFs represent an excellent model for fundamental studies of molecular-to-nano transitions in situ in video mode, including a number of dynamic transformations.