Development And Application Of The Reaxff Reactive Force Field Method To Ferroelectrics And Crosslinked Polyethylene

In order to design and optimize advanced materials such as ferroelectrics and polymers it is essential to obtain detailed, atomistic-scale insight of the materials characteristics. Quantum chemical (QC) methods are considered the most accurate technique among atomistic simulation methods; however, these methods are computationally highly expensive, making them impractical for larger systems. The ReaxFF reactive force field method can model chemical reactions based on the bond order concept without the full expense of QC methods, and it has been successfully applied to a wide range of systems, including covalent and metal oxide/hydride/carbide materials. In this dissertation, ReaxFF reactive force fields are developed and verified against density functional theory (DFT) data for barium titanate (BaTiO3) and crosslinked polyethylene (XLPE) materials to improve atomistic-scale understanding of these materials. Subsequently, these ReaxFF force fields were used in molecular dynamics simulations used to elucidate the reactive dynamical response of these materials. Ferroelectric perovskites such as BaTiO3 have seen numerous applications in nonvolatile memories, transducers, micro sensors and capacitors because of their unique properties such as spontaneous polarization, piezoelectric and pyroelectric effects, as well as large dielectric constants. In this dissertation, we developed an easily extendable atomistic ReaxFF reactive force field for BaTiO3 that can capture both its field- as well as temperature-induced ferroelectric hysteresis and corresponding changes due to surface chemistry and bulk defects. Using our force field, we were able to reproduce and explain a number of experimental observations: (1) existence of a critical thickness of 4.8 nm below which ferroelectricity vanishes in BaTiO3; (2) migration and clustering of oxygen vacancies (OVs) in BaTiO3 and reduction in the polarization and the curie temperature due to the OVs; (3) domain wall interaction with surface chemistry to influence ferroelectric switching and polarization magnitude. This new computational tool opens up a wide range of possibilities for making predictions for realistic ferroelectric interfaces in energy-conversion, electronic and neuromorphic systems. XLPE has emerged as an outstanding insulator for high-voltage direct current power transmission cables due to its favorable dielectric properties, low water permeability, structural integrity at high temperature and chemical resistance. Dicumyl peroxide (DCP) is the most commonly used peroxide crosslinking agent for polyethylene (PE) in high voltage power cables. The DCP reactions in the PE matrix lead to the formation of a range of byproducts, some of which remain in the final XLPE product and may have adverse effects on the cable function and its long-term properties. Currently, our knowledge of the effects of the byproducts and chemical reactions involved during the crosslinking procedure on the properties of the final XLPE product is limited. By understanding how the crosslinking byproducts change the XLPE properties, improvements may be formulated relative to conventional XLPE cables. In order to design and optimize XLPE cables, it is crucial to obtain atomistic-scale insight of XLPE chemistry since each and every byproduct in XLPE affects differently the electrical properties of the polymer and thus the effect of each of the byproducts should be investigated. In this dissertation, we developed a ReaxFF force field validated against DFT data obtained for XLPE chemistry. Using this force field, we studied the effects of different parameters such as temperature, density, type of peroxide, and the ratio of peroxides to PE on the formation of byproducts, distribution of functional groups, and crosslinking. Our results indicate that a moderate curing temperature rise to 500 K leads to an increased crosslinking extent, however, temperature rise above 500 K may have adverse effects on the PE crosslinking. Additionally, our results indicate that elevating the density improves the PE crosslinking. Our study showed that a high ratio of DCP to PE can increase the amount of generated byproducts but may not necessarily lead to an increased amount of XLPE. Our MD results also indicate that the presence of an external electric field had almost no effect on crosslinking and that di-(1-decyl-1-phenylundecyl) peroxide, may not be as efficient as DCP in XLPE production. These results indicate that ReaxFF based molecular dynamics (MD), validated by experiments, is an efficient tool for analyzing -- and improving -- the conditions of polymerization chemistry. As the final part of this dissertation, we present an eReaxFF-based MD simulation framework, which includes an explicit electron description verified against DFT data, to investigate the roles of XLPE byproducts and processing variables such as density and voids on the time to dielectric breakdown (TDDB) of PE. The eReaxFF method is an extension of the ReaxFF method with description of an explicit electron-like or hole-like particle. Our simulation results indicate that an increase in density of PE increases the TDDB, however, adding a byproduct with positive electron affinity (EA) such as acetophenone can reduce the TDDB. Furthermore, during the electrical breakdown in PE, electrons tend to migrate through voids when transferring from anode to cathode. In comparison to neutral acetophenone, we find that the acetophenone radical anion can significantly reduce the energy barrier and the reaction energy of secondary chemical reactions.