UC Riverside

The ability to characterize nanoscale particle size and shape has gained increasing importance as the development of nanomaterial synthesis and application. Despite being the most convenient approach to characterize small features, the resolution of optical microscopy is severely limited by the optical diffraction limit. Scanning probe microscopy, including scanning electron microscopy (SEM), atomic force microscopy (SEM), and scanning tunneling microscopy (TEM), is able to resolve features even at sub-nanometer scale, and has become a vital tool in the development of materials science. However, these high-resolution microscopies usually require expensive apparatus and complex operating procedures. In this dissertation, we present a novel computational imaging platform to characterize nanoparticle morphology on a conventional optical microscope, being able to predict the size of nanoparticles with a mean error below 5%. The capability of the method is validated by experiments. This dissertation covers four parts:1. Deep learning-assisted optical microscopy for nanoparticle morphology characterization 2. Quantitative analysis of the mode evolution on a metal-coated optical fiber tip 3. Low-loss bending of single-mode optical fiber for tip design of near-field scanning optical microscopy (NSOM) 4. Magnetic field sensing using a silver nanowire (AgNW) nanofocusing probe tip