Abstract (eng)
The implementation of star formation in numerical simulations on galaxy scales and cosmological structures is still a matter of much debate and multiple indispensable experiments. The reason for this are the irresolvable physical processes on sub-grid scales. A large variety of different methods have been developed and applied over the last decades. Nevertheless, we are still in the test phase in which way the choice of parameter sets are affecting the simulation results.
In order to examine this issue further, we present N-body/SPH simulations of isolated dwarf galaxies for two different star-formation recipes, one with the commonly used star-formation threshold parameters and one derived under the assumption of self-regulation equations. Further emphasis has been put on possible choices of free parameters and the smoothing length of SPH particles. For these studies we use the public version of Gadget-2 with self-implemented sub-resolution physics as a single-phase gas description. Moreover, we directly compare these results with a multi-phase Nbody/SPH code ''cdSPH'', developed in our group, where the same sub-resolution physics as e.g. gas cooling and stellar feedback by energy and chemical yields, have been implemented, while the interstellar medium is divided into the two dominating gas phases. These are allowed to evolve dynamically and energetically separately, but are coupled by not freely parameterizable drag and mass exchange. Here we will discuss differences in star-formation rates, gas dynamics, and chemical evolution of the different components. Our results suggest that artificial parameters, like e.g. the cooling shutoff time, strongly influence the different star formation recipes with varying strengths. The system reacts very sensitively on the choice of the gravitational softening length, which can also influence star-formation recipes in different ways, depending on how the density and/or the star-formation volume is implemented. If the supernova efficiency is chosen too high, a dwarf galaxy system can easily get torn apart. Surprisingly, sudden starbursts are found in this isolated dwarf galaxy simulation, where large regions become unstable, collapse and produce 10^5 to 10^6 solar masses, within just a few Myr. This occurs only for short cooling shutoff times. Comparisons between the single- and multi-phase descriptions show
a strong impact on the star-formation rate. Additionally, in the single-phase model, the amount of ejected gas from the central region is clearly lower due to the inherent mixing compared to the multi-phase simulations, which also affects the chemical enrichment of the system.