Nanomaterials are nanostructured materials in which one of its dimensions is at least smaller than 100 nm. The unique size, shape, structure, and high surface area-to-volume ratio of nanomaterials result in outstanding optical, electrical, mechanical, and magnetic properties versus their bulk counterparts. In particular, carbon nanotubes (CNTs) possess outstanding electrical, mechanical, and electromechanical properties due to their unique structure and carbon-carbon covalent bonds. To take advantage of these properties for use in real-world applications, one approach is to disperse CNTs in a polymer matrix to form nanocomposites. However, the underlying mechanisms of how CNT-based nanocomposites derive their electromechanical properties remains poorly understood, despite the vast amount of experimental and numerical modeling work done in this area.

In this thesis, a CNT-based thin film model was derived based on percolation theory. One dimensional CNT elements were randomly distributed in a predefined two-dimensional area, and their electrical and electromechanical properties were simulated. The objective was to evaluate the main parameters that influence the model’s electrical and electromechanical properties. Numerical simulation results showed that the percolation threshold and electrical characteristics of the thin film are affected by CNT lengths, concentrations, and intrinsic piezoresistivity. It was also found that the electromechanical behavior of the model was characterized by linear piezoresistivity. The CNT-based thin film model became less sensitive to applied strains as CNT concentration increased. Furthermore, an important finding was that, near the percolation threshold, inconsistent strain sensing response was observed.

In order to improve the accuracy of predicting the bulk-scale electromechanical behavior of CNT-based thin film, the model was updated with multi-scale experimental measurements. Atomic force microscope images of CNT-based thin films were acquired, and image analysis was conducted to measure the physical characteristics of as-dispersed CNTs, which were then incorporated in the model. A key finding was that the morphology of the CNT network is an important parameter that governs bulk nanocomposite electrical and electromechanical behavior. The models were validated by conducting electromechanical experiments that characterized thin film behavior fabricated using different parameters and subjected to different loading excitations. The model was able to accurately characterize the electromechanical properties of these films.

This dissertation also explores the use of numerical models to guide the design of electrical time-domain reflectometry (ETDR) sensors that featured CNT-based thin film sensing elements. An advantage offered by the proposed sensor is that, unlike currently available distributed strain sensing systems, implementation of ETDR sensors is easier due to their simple system architecture and low manufacturing and installation costs. The previous simulation results showed that the CNT-based thin film was more sensitive to strains as CNTs were more aligned. Therefore, CNT sensing elements with aligned CNTs were integrated in the ETDR setup, and the sensor showed better strain sensing performance. The two most important research contributions of this dissertation were: (1) the systematic investigation conducted to uncover the fundamental material mechanisms that governed CNT-based thin film electromechanical behavior; and (2) the development of a numerical model for sensor design and optimization.