Logo Logo
Help
Contact
Switch language to German
High-redshift star-forming galaxies
High-redshift star-forming galaxies
Ten billion years ago, the Universe was much more active than today. Irregular gas-rich and turbulent disc galaxies dominated the cosmic star formation history with star formation rates 10-100 times higher than found today. These galaxies often contain giant, kpc-sized clumps that are the launching sites of powerful outflows caused by stellar feedback. These massive objects are thought to play an important role in the bulge formation process, often found in the centre of a galaxy. The origin of the giant clumps can be explained by gravitational disc instability, which is supported by many simulations and observations. The observed high gas fractions, densities and high-velocity dispersions of these galaxies imply gravitationally unstable or marginally stable discs over cosmic time. These conditions allow large perturbations (Toomre-length) to grow that can lead to massive gravitationally bound clumps. We investigate the nature of these giant clumps in more detail by revisiting the linear perturbation theory for the thick disc approximation and employ very high- and low-resolution hydrodynamic simulations (~3-100 pc) with the adaptive mesh refinement code RAMSES. We find that the Toomre-length is around two times larger for a gas disc with a vertical profile (thick disc) than for the thin disc approximation, which is typically used to estimate clump properties. For the first time, we confirm that simulations show the growth of the predicted perturbations that establish as Toomre-rings. Clumps do not form directly on the predicted Toomre scale, which is commonly assumed. Instead, the initial growing structures (rings) collapse further, and clumps can fragment after reaching higher densities on a sub-Toomre-scale (2nd phase). These conditions lead to the formation of many and much smaller clumps than predicted by fragmentation on the Toomre scale. To avoid spurious fragmentation at maximum resolution, simulations typically use an artificial pressure floor (APF) that ensures a minimum Jeans-length, resolved by a few resolution elements. We show that in lower resolution simulations (like cosmological zoom-in simulations) this APF acts already at low densities, preventing the collapse phase and leading to massive and inflated clumps that are roughly — but by chance — on the initial Toomre scale. These results disprove the common hypotheses that giant clumps need to form directly on the Toomre scale and sub-fragment to a gravitationally bound substructure (top-down scenario) if the resolution is high enough. In the high-resolution simulations instead, we find that the many smaller clumps group to clump clusters (CCs) that appear as a giant and massive objects (bottom-up scenario) if we consider the resolution limit of observations at z~2. They reflect the masses, sizes and internal kinematic properties of giant clumps in high-redshift observations. In Genzel et al. (2011) they concluded from the inferred kinematics that their clumps are either pressure supported by high-velocity dispersion or are still undergoing collapse because of the small velocity gradients. In our simulations, the gradient is explained by a clump cluster rotating on kpc-scales around their centre of mass. Furthermore, the CCs predict a hierarchy of properties on several scales, which is dependent on the observational resolution. This can be tested by future telescopes. For high observational resolution, gravitationally bound clusters (closed clusters) are visible, and with a decreasing resolution, fewer objects can be identified, leading to massive and larger clumps that are represented by open clusters. The derived Toomre-length for the thick disc approximation, the finding of the clump clusters and the bottom-up scenario open the door for a new understanding of the nature of the observed giant clumps. The CC scenario could have strong implications for the internal evolution, lifetimes and the migration timescales of the observed giant clumps, bulge growth and AGN activity, stellar feedback and the chemical enrichment history of galactic discs.
Galaxies, High-Redshift, Gravitational Instability, Hydrodynamics
Behrendt, Manuel
2019
English
Universitätsbibliothek der Ludwig-Maximilians-Universität München
Behrendt, Manuel (2019): High-redshift star-forming galaxies. Dissertation, LMU München: Faculty of Physics
[thumbnail of Behrendt_Manuel.pdf]
Preview
PDF
Behrendt_Manuel.pdf

57MB

Abstract

Ten billion years ago, the Universe was much more active than today. Irregular gas-rich and turbulent disc galaxies dominated the cosmic star formation history with star formation rates 10-100 times higher than found today. These galaxies often contain giant, kpc-sized clumps that are the launching sites of powerful outflows caused by stellar feedback. These massive objects are thought to play an important role in the bulge formation process, often found in the centre of a galaxy. The origin of the giant clumps can be explained by gravitational disc instability, which is supported by many simulations and observations. The observed high gas fractions, densities and high-velocity dispersions of these galaxies imply gravitationally unstable or marginally stable discs over cosmic time. These conditions allow large perturbations (Toomre-length) to grow that can lead to massive gravitationally bound clumps. We investigate the nature of these giant clumps in more detail by revisiting the linear perturbation theory for the thick disc approximation and employ very high- and low-resolution hydrodynamic simulations (~3-100 pc) with the adaptive mesh refinement code RAMSES. We find that the Toomre-length is around two times larger for a gas disc with a vertical profile (thick disc) than for the thin disc approximation, which is typically used to estimate clump properties. For the first time, we confirm that simulations show the growth of the predicted perturbations that establish as Toomre-rings. Clumps do not form directly on the predicted Toomre scale, which is commonly assumed. Instead, the initial growing structures (rings) collapse further, and clumps can fragment after reaching higher densities on a sub-Toomre-scale (2nd phase). These conditions lead to the formation of many and much smaller clumps than predicted by fragmentation on the Toomre scale. To avoid spurious fragmentation at maximum resolution, simulations typically use an artificial pressure floor (APF) that ensures a minimum Jeans-length, resolved by a few resolution elements. We show that in lower resolution simulations (like cosmological zoom-in simulations) this APF acts already at low densities, preventing the collapse phase and leading to massive and inflated clumps that are roughly — but by chance — on the initial Toomre scale. These results disprove the common hypotheses that giant clumps need to form directly on the Toomre scale and sub-fragment to a gravitationally bound substructure (top-down scenario) if the resolution is high enough. In the high-resolution simulations instead, we find that the many smaller clumps group to clump clusters (CCs) that appear as a giant and massive objects (bottom-up scenario) if we consider the resolution limit of observations at z~2. They reflect the masses, sizes and internal kinematic properties of giant clumps in high-redshift observations. In Genzel et al. (2011) they concluded from the inferred kinematics that their clumps are either pressure supported by high-velocity dispersion or are still undergoing collapse because of the small velocity gradients. In our simulations, the gradient is explained by a clump cluster rotating on kpc-scales around their centre of mass. Furthermore, the CCs predict a hierarchy of properties on several scales, which is dependent on the observational resolution. This can be tested by future telescopes. For high observational resolution, gravitationally bound clusters (closed clusters) are visible, and with a decreasing resolution, fewer objects can be identified, leading to massive and larger clumps that are represented by open clusters. The derived Toomre-length for the thick disc approximation, the finding of the clump clusters and the bottom-up scenario open the door for a new understanding of the nature of the observed giant clumps. The CC scenario could have strong implications for the internal evolution, lifetimes and the migration timescales of the observed giant clumps, bulge growth and AGN activity, stellar feedback and the chemical enrichment history of galactic discs.