Edison Sung
Application of Machine Learning
Application of Machine Learning
Edison Sung
Dr. Ang-Yu Lu
Building a universal categorizing method to efficiently graphene and their resulting Raman spectroscopy
Abstract
Graphene, a single layer of carbon atoms arranged in a hexagonal lattice, exhibits many unique optical, electrical, and mechanical properties. To fully take advantage of graphene’s properties, a reliable large-scale synthesis method is required. Currently, researchers have developed many synthesis methods for monolayer graphene growth. Therefore, establishing a universal metric to evaluate the quality of graphene monolayers is crucial for graphene-based applications. All of the graphene samples analyzed were produced through chemical vapor deposition (CVD) and were monolayered. Raman spectroscopy was used as a tool to analyze the graphene samples. Non-linear curve fitting and data visualization were used to determine the relationships between the various features of a Raman spectrum and how they were affected by the external perturbations. It was discovered that features from all three bands: G band, 2D band, and D band are necessary for effective analysis of the quality of graphene produced. It was also discovered that the 2D band is sensitive to strain, the G band is sensitive to doping, and the D band is highly sensitive to both. With combining 2D width and the intensity ratio of D/G, we can disentangle the external perturbations, such as strain and doping, to establish the metrics of Raman measurement for CVD-graphene samples. Future work must be done on expanding the number of samples with a greater variety of synthesis methods in order to get more comprehensive results.