Characterization and Modeling of Ferromagnetic Particulate Nanocomposites for Strain and Fracture Sensing

This dissertation investigates the multiphysical behavior of multi-walled carbon nanotubes (MWCNTs)/polydimenthylsiloxane (PDMS) composites containing chain-structured nickel particles for strain sensing. Compared with traditional strain gauges, this novel strain sensor exhibits high flexibility, large elongation, and high strain sensitivity and therefore has a wide application in structural health monitoring and fracture detection with minimal surface preparation. The scope of this study covers the material fabrication, numerical simulation of microstructure evolution, micromechanics-based characterization and modeling for the multi-physical properties, and experimental investigation of the strain sensitivity in sensing applications. MWCNT/PDMS composites with chain-structured ferromagnetic particles were fabricated using a solution casting method under an external magnetic field. Different concentrations of MWCNTs, as well as ferromagnetic particles, were well mixed in the pre-polymer matrix. An external magnetic field was applied during the curing process to align the particles into a chain structure. The morphology of MWCNTs and chain-structured nickel particles in the PDMS were investigated using an optical microscope and a scanning electron microscope. The electrical properties such as a percolation threshold and electrical conductivity of MWCNT/PDMS composites with different concentrations of chain-structured ferromagnetic particles were investigated for strain sensing application. For MWCNT/PDMS composites, a simplified model has been developed to predict their effective electrical conductivity. MWCNTs are well dispersed in a PDMS matrix, and the mixture is then cured and cast into thin films for electrical characterization. The MWCNTs are assumed to be statistically uniformly distributed in the PDMS matrix with the three-dimensional (3D) waviness. As the proportion of MWCNTs increases to a certain level, namely the percolation threshold, the discrete MWCNTs start to connect with each other, forming a 3D network which exhibits a significant increase in the effective electrical conductivity. The eight-chain model has been used to predict the effective electrical conductivity of the composite, in which the contact resistance between MWCNTs has been considered through the Simmons' equation. The eight-chain network features can be significantly changed to adjust to modifications to the mixing process, MWCNT length and diameter, and clustering and curling of MWCNTs. A Gaussian statistics-based formulation is used to calculated the effective length of a single MWCNT which is well dispersed in the matrix. The modeling results for the effective electrical conductivity agree with the experiments very well, they are highly dependent on the contact resistance between MWCNTs and the waviness of the MWCNTs. The effect of inter-nanotube distance and diameter of MWCNTs on the effective electrical conductivity of the MWCNT/PDMS composite is also discussed.Micromechanics-based modeling method of the microstructure evolution of ferromagnetic particles moving in the PDMS pre-polymer has been developed to understand the alignment mechanism and to optimize the fabrication procedure. Under a uniformly applied magnetic field, in the neighborhood of ferromagnetic particles, the magnetic field will be significantly distorted and the magnetic force induced will align particles into short chains, which will further merge into long chains. The experiments have been simulated with the equivalent inclusion method. This study has led to the development of a novel strain sensor. Both MWCNTs and ferromagnetic particles enhanced the electrical conductivity of the nanocomposites, but they exhibited different effects on the strain sensitivity of the sensor. When the proportion of MWCNTs that are well dispersed in PDMS is higher than the percolation threshold, the strain sensitivity reduces with the increase of MWCNTs in general; whereas a higher volume fraction of FPs produces a higher strain sensitivity when the chain-structure of FPs is sustained. The mechanisms causing this interesting phenomenon have been demonstrated through the microstructural evolution and micromechanics-based modeling method. These findings indicate that an optimal design of the volume fraction of FPs and MWCNTs exists to achieve the best strain sensitivity of this type of sensors. It is demonstrated that the nanocomposites containing 20 vol.% of nickel particles and 0.35 wt.% MWCNTs exhibits a high strain sensitivity of ~80