When a shock wave propagates across the interface between fluids of different acoustic impedance two fundamental events follow: the refraction of the shock wave and the baroclinic generation of vorticity on the interface. This latter, in turn, causes the development of distinct vortical structures which lead to the distortion and eventual breakup of the original interface shape, and ultimately the mixing of the two fluids.
This type of events occur at microscopic scales (implosion of microtargets for the achievement of nuclear fusion by inertial confinement) and astronomical scales as well (overturn of supernova cores, possibly the formation of large jets). Similarly, experiments are conducted over a very broad range of length, time and energy scales: laser-driven experiments depositing kJ-MJ of energy on targets a few mm in diameter; and shock tube experiments where gas interfaces several cm across are impulsively accelerated.
In this talk I will describe the results of some laboratory and computational experiments we have performed on spherical and 2-D interfaces. The experiments are performed in a vertical shock tube of large, square internal cross section (25 $\times$ 25 cm$^2$) with structural capability to whitstand the pressure load consequent to an $M=5$ shock wave in atmospheric air reflecting off the end wall.
The spherical interface is a soap bubble filled with argon, falling freely in nitrogen. The bubble is imaged continuously during its free fall and single-exposure images are recorded at several delay times after the bubble has been accelerated by a planar shock wave. The experiment is numerically simulated using the {\it Raptor} code, made available by our colleagues at LLNL. $\it Raptor$ is a finite volume code that solves the Euler equations using a Riemann solver with Colella's ``Piecewise Linear Method" for data reconstruction at cell interfaces. The code incorporates dynamic adaptive mesh refinement for concentrating computational resources on locally fine grids in regions of large thermodynamic gradients. A comparison of the laboratory and computational measurement of the size of the bubble's large geometrical features is presented, together with estimates of the flow circulation.
The 2-D interface is prepared by flowing different gases from opposite ends of the shock tube to form a stagnation plane and by oscillating two rectangular pistons on opposite sides of this plane: depending on the amplitude and frequency of the oscillations, different 2-D shapes can be imposed on the interface. Preliminary results show good qualitative comparison to previous similar experiments by J. Jacobs at the University of Arizona and suggest great potential for this interface-generation technique.
