Title:
Analysis of fibrillar structures for the engineering of polymeric transistors
Analysis of fibrillar structures for the engineering of polymeric transistors
Author(s)
Persson, Nils Erland
Advisor(s)
Grover, Martha A.
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
Image processing software was developed and applied to the analysis of polymer nanofiber microstructures in poly(3-hexylthiophene) (P3HT)-based organic field effect transistors. A large database of processing, structural, and electrical property data was collected from the fabrication of bottom gate, bottom contact thin film transistors, including over 100 Atomic Force Microscopy images of densely packed, semi-flexible P3HT nanofibers. Fibers were extracted from the images through a combination of anisotropic diffusion filtering, skeletonization, and vectorization via active contours. Image processing parameters were trained through a machine learning approach, yielding structural measurement error of less than 10% when compared to manual fiber tracing. Numerical methods for the calculation of quantitative structural metrics were developed, including fiber length and width distributions, fiber packing density, fiber alignment, and the decay length of orientational order. Results indicated that fiber alignment was linearly correlated with charge carrier mobility in P3HT-based thin film transistors when fibers were oriented perpendicular to the direction of charge transport. Fiber length was the strongest predictor of alignment, with short fibers causing defects in alignment in blade-coated thin films. These findings suggest that P3HT nanofibers act not as conduits for charge transport, but as hubs through which charges can access longer polymer chains that carry them across the film. Process-structure relationships for the nucleation, growth and alignment of P3HT nanofibers are quantified and discussed for a variety of solution processing techniques, including sonication, UV irradiation, poor solvent addition, aging, and microfluidic flow processing. Finally, techniques from data science and computer vision are applied to the analysis of a large materials microscopy database, demonstrating a new system for the discovery and analysis of microstructural data. The image processing software, GTFiber, is packaged for Windows and Mac and available at [gtfiber.github.io].
Sponsor
Date Issued
2017-08-03
Extent
Resource Type
Text
Resource Subtype
Dissertation