Hyperspectral Imaging and Machine Learning for Nanomaterials Characterization

Nanoscale material systems are of great importance for answering open questions in funda- mental science and enabling advancements in next-generation technologies. This dissertation explores the structural, electronic, and optical properties of several nanomaterial systems at the atomic-, nano-, and micro- length scales. A collection of analytical characterization tools, non-trivial imaging methods, and advanced data analysis techniques are presented. This methodology is essential for finding and understanding new physics in nanomaterial systems; it is also an important tool for characterizing new nanomaterial systems and determining the relationship between synthetic parameters and relevant material properties. Applications of these techniques to a wide array of scientific questions are given throughout the thesis.A newly discovered source of highly localized light emission in hexagonal boron nitride (h-BN) is presented. This luminescent color center is closely related to unique impurities and native defects that occur exclusively in high-quality samples grown using a special high-pressure, high-temperature synthetic routine. An electron beam is used as a probe to characterize the luminescence, as well as activate and deactivate the light emission. The electron beam can also be used to deterministically create arbitrary nanoscale light emission patterns in h-BN synthesized using more conventional methods. A variety of h-BN polytypes are discussed and the effects of growth parameters, geometry, dimensionality, impurity concentration, and sample preparation on light emission are discussed.Transition-metal dichalcogenides (TMDs) are an exciting class of Van der Waals (VDW) materials that display a variety of unique phenomena in their 2D (monolayers) and 1D (nanotubes and nanoribbons) forms that are absent in the bulk material. A special synthesis method which allows for targeted growth of specific nanostructures is presented along with characterization of the TMD nanomaterials. A combination of special transmission electron microscopy (TEM) techniques and advanced data analysis methodology allows for rapid determination of the grain structure of continuous, nanocrystalline, monolayer TMD films.Gallium chalcogenides (GCs) are another class of VDW materials, closely related to TMDs, that are of significant interest to the nanoscience community. The type and concentration of chalcogen atoms in these materials can be varied, resulting in a alloys with tunable structural, electronic, and optical properties. A combination of TEM, angle-resolved photoemission spectroscopy (ARPES), and photoluminescence (PL) is used to fully characterize the quasiparticle band structure, optical bandgap, and crystalline phase of the alloys across their compositional range. Monolayers of these materials exhibit a distinct change in band structure, resulting in the material transitioning from direct bandgap in the bulk to indirect bandgap in the 2D limit.The characterization and analysis methods discussed here are broadly applicable to a wide variety of nanomaterial systems. Widespread adoption of machine learning and automation in the nanoscience community is a promising route toward high-throughput scientific inquiry