Hydrocode Modeling of Detonation Failure in Vapor-Deposited Explosive Films with Explicit 3D Interfacial Roughness

Recent experiments conducted at Sandia National Laboratories (SNL) have investigated the role of a variable substrate interfacial roughness on detonation propagation in vapor-deposited high explosive (HE) films. From these experiments, it was shown that non-ideal interfaces can affect both the detonation velocity (i.e., at a given HE layer thickness) as well as the critical thickness for self-propagating detonations. Preliminary simulations using SNL’s hydrocode, CTH, initially assumed idealized 3D smooth interfaces to investigate the detonation velocity as a function of HE film thickness. Unfortunately, such oversimplistic consideration, which neglected the role of interface roughness, resulted in a poor match between the experimental results and model predictions.

Therefore, a new 3D computational study is presented that leverages a novel capability in CTH for using voxel-map import of an image stack, whereby 3D renderings of non-ideal interfaces may be ingested from either CT or profilometer scan data. As such, the current simulations are better suited to explore new features; specifically, finite film width and the roughness intensity between the vapor-deposited film and inert substrate. The current modeling results show that non-ideal 3D interfaces can increase the failure thickness of an explosive film. Additional analysis of the shock front curvature for each configuration is compared to Detonation Shock Dynamics (DSD) theory. Finally, the transverse kinetic energy is studied as a potential loss mechanism contributing to the velocity deficit. Statistical regression models are presented to identify trends between the various observables and detonation failure. Overall, this work analyzes the role that complex interfaces play on affecting the detonation velocity versus film thickness curves for vapor-deposited HEs, and a new hypothesis is presented as to the mechanism(s) driving the observed behaviors, which have so far remained elusive to model.