Gas-phase synthesis of plasmonic nanoparticles with robust high bandgap shell materials: A study of Cu@CaF2 with AI supported transmission electron microscopy analysis

Cu@CaF2 plasmonic nanoparticle (NP) films are physically synthesized, and their morphology, and electronic and optical properties are thoroughly investigated. Cu NPs are generated using a gas aggregation source assisted by magnetron sputtering, while CaF2 coatings are deposited by thermal evaporation. Scanning Electron Microscopy, Atomic Force Microscopy (AFM), and Transmission Electron Microscopy (TEM) provide a clear picture of the film morphology. The Cu NP shapes remain substantially unaffected by the deposition of CaF2, acting as nucleation centers for fluoride growth, which effectively forms a protective shell. With increasing CaF2 thickness, the shells extend to form islands, which eventually coalesce into a complex film morphology. Computer vision methods based on Mask Regional convolutional neural network, one of the leading deep learning architectures for object detection, are employed to fully automate particle analysis, exploiting its capabilities to perform detailed statistical evaluation of NP size and shape from a large number of images. AFM and TEM reveal that the NPs have a lateral size of 〈d〉 = 13.8 ± 0.9 nm and an oblate spheroid shape with aspect ratio  = d/h ≈ 1.2. In situ XPS data show that the chemical state of the NPs is unaffected by the presence of CaF2. Finally, optical data obtained with a UV–vis–near IR spectrometer and simulated using the Maxwell–Garnett approximated extinction cross section demonstrate that the Cu localized surface plasmon resonance remains robust under prolonged atmospheric exposure, a fundamental property that is crucial for applications in photovoltaics and optoelectronics.