Stellar nuclear astrophysics environments often involve rapid nuclear burning in fast mixing, convective environments. Depending on the specific conditions, the result can be nuclear flash burning, or a deflagration, with strongly altered turbulence morphology. These effects are observable, and in this way details of the hydrodynamic simulations can be validated.
Of specific interest is the situation shown in a snapshot (right) of run sbm04_conc2a3 which shows a layer of stellar material with different composition: red in the top layer contains mainly hydrogen, blue is mainly carbon. Most of the carbon-rich (blue) layer is mixing rapidly (convection) after a short while because it is heated from a nuclear source near the bottom of the simulations box. The top layer containing H is initially not mixed (stable layer). Eventually convection will entrain some hydrogen into the carbon-rich layer where it burns very efficiently, releasing heat and driving mixing itself. The simulation requires high-accuracy reconstruction of material interfaces. The effect of this burn and mix can be compared to observations.
In order to study hydrodynamics code and algorithm properties relevant for this simulation in isolation we look at an overdense blob of material with fuel 1 (hydrogen) in a gravitationally stratified environment of fuel 2 (carbon), as shown in the snapshot (left) of run tnbe_conc2a3. In that movie the left panel shows again fuel1 as red in the blob, and fuel 2 as blue. The right panel shows the ash appearing at the mix interface of fuel_1 and fuel_2. As the blob sinks down, mixing at the interface of the blob allows the reaction fuel_1 + fuel_2 -> ash to release energy, and thereby add buoyancy. This eventually reverses the decent of the mix of ashes and fuels into a rise. The exact simulation results depend on modeling mixing at material interfaces, because the mix of the two fuels is directly proportional to the energy released, which in turn leads to the reversal of flow direction. The fact that the two fuels involved are on opposite sides of the material interface makes this problem particularly sensitive to the accurate treatment of material interfaces. We investigate and test different strategies for this problem.