A Foundation for the Atomistic Simulation of Solid Rocket Propellants

The rocket boosters on some of the most powerful space launch vehicles in the world use solid propellants. Solid propellant rocket boosters provide prodigious thrust for launching satellites and other important payloads without the need for cryogenic refrigeration and at lower design cost than launch systems employing liquid propellants. However, solid fuels present several unique challenges due to the complex nature of the materials. In particular, these propellants are composed of a particulate oxidizer that is held together by a polymeric binder which can age and degrade over time. The rate of this degradation process depends on many factors such as the presence of fuel contaminants, such as oxygen and water, environmental conditions (e.g., temperature variation and ultraviolet radiation), and the different oxidizer–polymer interfaces present in the system.

Atomistic simulation provides a versatile framework for examining the effects of each of the degradation factors above both individually and in combination. In this webinar, we describe the development of classical atomistic models and their use for evaluating the physical properties of the polymeric binder hydroxyl-terminated polybutadiene (HTPB) and the particulate oxidizer ammonium perchlorate (AP). The HTPB polymer is computationally cross-linked using a directed diffusion approach and then quenched to temperatures between -40 and 40 °C. We then computed the temperature-dependent elastic properties of HTPB using uniaxial compressive and tensile deformation simulations. Classical forcefields for AP were optimized against experimental physical properties and the AP–HTPB interface was simulated to examine the interaction of hydroxyl groups with the AP surface. Our simulations of the AP–HTPB interfaces explain and corroborate the experimental observation of facet-dependent interfacial energy. Furthermore, we discovered that the hydroxyl groups in the HTPB polymer melt exhibit associative behavior that could impact the spatial distribution of cross-links. However, as one might expect for polymer systems, our atomistic simulations of elastic behavior do not extrapolate accurately to engineering behavior because they do not capture all the configurational changes that can occur at the relevant timescales of minutes and longer. This limitation of atomistics underscores the need to augment the insights gained from atomistic approaches by recourse to mesoscale models, such as those provided in MedeA.