Artificial Intelligence design and synthesis of nanoparticles and nanoMaterials – AIM

Silicon (Si) cores coated with a thin shell (=10 nm) of a metallic coating (Au or Ag) are interesting for optical applications, particularly if they can be organized into metasurfaces. The strong magnetic dipole resonance of the silicon particle can be coupled to the electric dipole resonance of the shell for phase plates, near-perfect absorbers, and antireflection coatings. The nanostructured building blocks must present high homogeneity in the thin metallic shells. To find the synthetic conditions producing such nanomaterials is a daunting task, when restricted to conventional synthetic trial-and-error small-scale batch approaches. To economize time and resources, machine learning can be applied to a continuous flow synthetic process. Continuous shells of less than 10 nm in thickness have never been made with gold and silver coatings around silica or silicon. With appropriate placement of these core-shell particles on a surface, we will be able to spatially control the wavefront of light. To determine the optimal arrangement of nanostructured particles on a surface, libraries of metasurfaces can be simulated, and then used to train an artificial intelligence. The artificial intelligence can thus predict the desired spacing and organization of core-shell particles. These metasurfaces will then be experimentally generated using nanoimprint lithography to produce wells in a 2D substrate, followed by convective assembly to place the Si@metallic core-shell particles within the wells. The novelty of this project results from scientific and methodological innovations: the synthetic process, the nanomaterials generated, the expected optical properties and the machine learning techniques applied to the development of these metasurfaces. This project couples a team of chemists within the ICMCB with experimental and theoretical physicists located within the CRPP and the LAAS.