Abstracts Category : Other

The impact of supernova feedback on the evolution of the ISM

by Anabele-Linda Pardi

The aim of this thesis is to investigate the formation and destruction of molecular gas as a component of the turbulent, multiphase, magnetised interstellar medium (ISM). In the pursuit of a better understanding of the impact of stellar feedback has on the evolution of the ISM, we use simulations of (256 pc)3 periodic boxes and 0.5x0.5x10 kpc3 stratified disk slabs modelling local parts of the Galactic disk with the Eulerian, adaptive mesh refinement (AMR), three dimensional, magnetohydrodynamic code FLASH 4. The simulations include self-gravity, magnetic fields, supernova (SN) explosions and a chemical network following H+, H, H2, C+ and CO.The first part of the thesis focuses on the H2 molecule formation and density-magnetic field scaling for various initial magnetic field strengths (0, 6x10-3, 0.3, 3 microG) in an environment stirred by randomly placed SNe with a fixed rate. We find that with increasing initial field strengths the simulations develop an ISM with a more homogeneous density/temperature structure and with a decreasing H2 mass fraction. We extend this study to include SNe spatial correlated with their parental molecular clouds (MCs). With the intension to create a more realistic model, we vary the fraction of SNe exploding in peak density regions and the ones traveling away from the MCs. We include self-gravity, magnetic fields and also vary the initial gas density and the SN rate. Our results show that the fraction of SNe exploding in their natal MC have a significant impact on the structure of the ISM by reducing the H2 mass fraction and the volume filling fraction of the hot gas up to an order of magnitude. Small variations of certain percentages of SN going off in MCs (depending on the magnetisation of the environment, on self-gravity and the SN rate) create dramatic transitions in the temperature and chemical structure of ISM. The last part of the thesis is dedicated to the study of the self-consistent SN-driven ISM evolution (a more realistic implementation than the fixed SN rate) with an extended simulation volume, including magnetic fields, self-gravity and the Galactic potential. We test the Schmidt-type star formation recipe extensively used in cosmological simulations and we find that the current implementation does not meet the criterion that the cooling time of the supernova remnants (SNRs) should be at least eight times longer than the sound-crossing time. This is an important issue as some of the simulations produce star formation surface densities comparable to the Kennicutt-Schmidt law but fail to produce a hot phase. We also derive recommended densities for the injection regions that ensure that for a given resolution and star formation efficiency, the SNRs expand before they radiate all their energy away. For practical reasons, we also put forward two potential solutions for the over-cooling problem like to model a more realistic ISM taking into account that a certain fraction of the massive OB stars travel away from their natal MCs and that they have differentAdvisors/Committee Members: White, Simon D. M. (advisor).