Dissertation/Thesis Abstract

Efficient Computational Methods for the Study of Charge Transport in Disordered Organic Semiconductors
by Brown, J. S., Ph.D., University of Colorado at Boulder, 2019, 153; 27663249
Abstract (Summary)

High fidelity computational methods are key for identifying fundamental properties of organic semiconductors. This dissertation focuses on both improving computational bottle necks in performance and accuracy and identifying new mechanisms governing charge transport.

Kinetic Monte Carlo charge transport simulations are a powerful tool for understanding the electrical properties of disordered organic semiconductors. However, the simulations are computationally intensive, particularly when sites of low energy are strongly coupled. A hybrid algorithm that combines Monte Carlo method with the Chemical Master equation is proposed. By dynamically applying the Chemical Master equation, the state space of problematic regions can be coarsened leading to performant algorithms. Testing of the hybrid method shows that it is able to accurately reproduce the dynamics with less than 5 % error while achieving large performance gains. Speedup, particularly in disordered systems, has been shown to increase by several orders of magnitude.

Application of kinetic Monte Carlo charge transport simulations to organic semiconductors has helped clarify how energetic correlations effect the charge carrier mobility. A seeded nucleation mechanism was proposed to help explain charge transport in non-polar organic semiconductors exhibiting an exponential density of states (DOS). A heuristic implementation of the mechanism is able to both account for Poole-Frenkel behavior at low fields and the exponential tails observed in the DOS.

Finally, a multi-scale approach has been leveraged to correlate the molecular structure of PTB7 and PCE10 with their electronic DOS. Improvements in the methodology are proposed through a fragment basis method and tested on PTB7 and PCE10. Performance improvements are achieved by reducing the self consistent calculations to the linear scaling regime. Comparison with full ab initio calculations show the fragment basis approach is able to accurately reproduce the molecular orbital energies and wave functions. The DOS of PTB7 show excellent qualitative agreement with experimental results found in the literature.

Indexing (document details)
Advisor: Shaheen, Sean
Commitee: McLeod, Robert, Park, Wounjhang, Larsen, Ross, Glaser, Matthew, Baumeier, Björn
School: University of Colorado at Boulder
Department: Electrical Engineering
School Location: United States -- Colorado
Source: DAI-B 81/7(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computational physics, Computational chemistry, Electrical engineering
Keywords: Charge transport, Correlations, Master equation, Monte Carlo, Tight binding
Publication Number: 27663249
ISBN: 9781392534762
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest