Investigating the influence of particle dispersion on the electromechanical properties of nanoparticle-based conductive polymer composites

Abstract

Nanoparticle-based conductive polymer composites (CPCs) are an attractive solution to many engineering problems due to their advantageous electromechanical and structural properties. CPC-based strain and fatigue sensing transducers have generated tremendous interest in recent decades, especially in the fields of smart materials and structural health monitoring. The viability of deploying CPC transducers for strain and fatigue sensing has been demonstrated repeatedly, yet a lack of fundamental understanding of underlying electromechanical phenomena, such as percolation, piezoresistivity, and the influence of preparation methods on such phenomena, has limited the widespread adoption of CPC-based transducers in real-world applications. An attempt to reduce necessary experimental characterization efforts, the task of predicting electromechanical properties of CPC-based transducers for engineering applications presents a unique challenge considering the multi-scale nature of nanoparticle-based CPCs. Traditional bulk property prediction models can effectively describe basic CPC behaviors, but the stochastic nature of CPC materials requires more complex models to elucidate the relationship between nanoparticle dispersion and CPC electromechanical characteristics at various length scales. In this dissertation, the influence of stochastic particle dispersion on the electromechanical properties of spherical nanoparticle-based CPCs is investigated through a combination of multi-scale stochastic modeling and experimental validation. The development of a high-fidelity, numerical, stochastic modeling algorithm, capable of generating representative volume elements (RVEs) for electromechanical interrogation, with nanoscale feature resolution and primary particle agglomeration control is discussed in great detail. Investigations of various algorithms for efficiently populating RVEs with primary particles are presented. The need for and efficacy of spanning multiple length scales (nanometer to millimeter) via the aforementioned stochastic model placement algorithms is explored. CPC thin film samples, composed of epoxy and carbon black, are experimentally manufactured via spin coating. The translucent properties of the spun thin films are favorable for optical microscopy imaging, which is then conducted to validate and improve stochastic model fidelity. Quantification of experimental CPC mixture dispersion is achieved via post-processing of optical micrographs. Statistical analysis of agglomeration shape and size descriptors (average agglomerate area, roundness, circularity, etc.) from stochastically generated RVE images are directly compared to experimental datasets, and model parameter tuning methods are presented. Bulk CPC samples are electromechanically characterized via uniaxial tensile and fatigue testing. Stochastically generated RVEs are subjected to simulated uniaxial strain, and changes in the resulting simulated random resistor networks are interrogated by applying Kirchhoff’s Laws and the theory of tunneling conductivity. Analytical models for predicting the piezoresistive response of spherical particle CPC volumes under uniaxial compression are adapted to agree with experimentally realized behaviors. Results from simulated, analytically modeled, and experimentally derived datasets are compared. The implications of this research, avenues for improvement, and recommendations for future work are discussed.