Composite materials are used as structural materials in many modern applications, ranging from sports equipment to aircraft, in order to improve mechanical properties while reducing the overall weight of the structure. Composite materials, consisting of fibers, particles or other inclusions embedded in a polymer matrix, can be tailored for proper strength and stiffness in chosen directions. However, the long term performance of such composites under extreme conditions is still unknown for most composite materials. For example, aircraft composites experience large cyclic variations in temperature and pressure and are subjected to constant vibrational stresses. Such variations in external conditions, along with the presence of humidity and other reactive gases like ozone, oxygen, etc., could result in long term reduction in strength and stiffness. Unfortunately, experimental characterization of such structural changes in the time scales appropriate for the active life time of composite structures is expensive, time consuming and may not be accurate. Computational methods, on the other hand, offer a cost sensitive alternate approach, in order to understand the mechanisms involved in the structural changes at various length scales.
Failure that occurs at the structural level results from accumulated damage at lower length scales. At the molecular level, bond breakage and molecular sliding contribute to regional weakness realized at the meso-scale. Integration of such localized damage will ultimately result in structural failure. Molecular dynamics (MD) simulations are used to understand damage at the lower length scales, but these must be coupled to models at the meso and continuum levels. The above figure illustrates a MD simulation of bond breakage in an epoxy composite resulting from oxidation and hydrothermal effects. From such molecular level events, a continuum level damage parameter can be calculated using the accrued bond breakage with its directionality, which can serve as a measure of the amount of damage in the system and can indicate the remaining time for reliable operations. This parameter can be experimentally obtained using a combination of x-ray analysis, neutron scattering, PALS, etc., for validating the computational procedure.
Composites, since they consist of two or more components, allow us to vary the components and their relative amounts, in order to obtain desired properties. In this way, computational approaches to materials design are very well suited for composite materials. We propose to apply these materials design approaches to tailor composites with longer lifespan (or time to failure), under a given set of external conditions. Starting with MD simulation data, meso-scale models will be developed. These models will account for the degree of bond breakage, the net directionality of breakage, the likely location of breakage, and incorporate the evolution of accumulated damage as a function of time. The models will be implemented and validated, and then used to characterize failure across large families of composite materials. The goal is both to understand the science behind failure caused by different conditions, and the design of new composite materials. Students will develop skills in materials science, mathematical modeling, development of computer implementations for large-scale calculations, and statistical data analysis algorithms.