Star formation efficiency of giant molecular clouds in spiral galaxies

Abstract

Context. Star formation (SF) remains an incompletely understood concept essential to many fields of astrophysics. Observations show that star-forming giant molecular clouds (GMCs) form around 1 per cent or less stars per free-fall time. This trend is observed over all nearby galaxies observed to date by the Physics at High Angular resolution in Nearby GalaxieS ALMA Survey (PHANGS-ALMA) collaboration. This makes SF a highly inefficient process. It is not known whether this is an effect of stellar feedback processes that regulate SF, or if clouds are intrinsically inefficient at converting gas into stars. Aims. This thesis aims to analyse a suite of simulations implementing different SF models to determine which model types can reproduce observed star formation efficiencies (SFEs). We further investigate the effects of stellar feedback, specifically the extent to which it can regulate SFE. Methods. Isolated Milky Way mass galaxy simulations are compared to data from the PHANGS-ALMA collaboration. The simulations are mock observed to enable a direct comparison with the analysis of Leroy et al. (2025). Results. The results show that intrinsically inefficient models are better at reproducing observational data. Feedback has a small effect on local properties with a larger influence on global parameters, and even low complexity models, such as fixed input efficiencies, can match the low observed efficiencies with the correct input efficiency. In contrast, models with high intrinsic SFEs and strong feedback, which are often employed in cosmological simulations, are shown to fail at reproducing observed cloud-scale trends and SFEs. Conclusions. The findings give an indication of which models are reasonable at matching observed data and what parameters are important to determine SFEs. These findings indicate which models best reflect observational data and how feedback regulates properties. The results provide guidance in selecting and calibrating SF models for future galaxy-scale simulations.

Popular Abstract

The stars you see in the night sky were once part of a cold cloud of gas, but exactly how these clouds collapse to form stars is an open question in astrophysics. Stars are one of the key players in galactic evolution. Their formation and evolution are what drive the development and structure of galaxies. By converting interstellar gas into stellar cores, they ignite nuclear burning at their cores, creating new elements and producing immense amounts of power. They disperse the new material and energy into their surroundings through supernovae, one of the most energetic events in our universe. This changes their surroundings by blowing gas away, heating dust grains and ionising atoms around them. Because stars influence their surroundings so strongly, it is a key interest to understand how stars form and how efficiently they do so. Current theories and observations agree that stars form in clouds of interstellar molecular gas. However, the underlying physics of how these stars form and the influence of cloud properties are not well understood. Most models assume that stars are born in the collapse of particularly dense regions of the cloud. This happens in both the centre of the cloud and in shock fronts caused by turbulent gas that travels faster than the local speed of sound. These shock fronts are similar to what happens when a plane breaches the sound barrier. Instead of waiting millions of years to observe how stars form, this work investigates simulations of virtual galaxies. By starting all simulations off the same way and letting them develop over a span of several hundred million years, we can look at the differences caused by different star formation models. To be able to compare the simulations with large-scale surveys of galaxies outside our own, we observe our simulations as if we were real astronomers looking through a telescope. In this way, we can calculate several key properties of the clouds and their efficiency at forming stars. This then allows us to examine the different models and determine which families of models best reproduce the observed data.

The main interest of the thesis is to figure out which models match the observed fraction of gas that gets turned into stars during the collapse of a cloud. We also investigate how the effects of newly formed stars influence the birth of future stars. We find that different models can produce very different outcomes that completely change how galaxies look. Some produce clumpy gas while others create cloudless, dispersed galaxies. Compared to observational data, the results show that the observations are best matched with models that are internally converting only very little of their gas into stars. We also observe that while the effects of newly formed stars have a large impact on the general structure of galaxies, they do not influence individual clouds as much.

This work indicates which models accurately describe what is happening in gas clouds in our universe and how they form stars. The results allow us to choose better models when simulating galaxies, putting us one step closer to having fully accurate simulations of the universe around us. Helping us to understand both how our galaxy formed and how it will develop in the future.

• author: Kamleitner, Elias
• supervisor: Oscar Agertz
• organization: Department of Physics; Astrophysics
• course: FYSK04 20261
• year: 2026
• type: M2 - Bachelor Degree
• subject: Physics and Astronomy
• keywords: Star formation, star formation models, molecular clouds, galaxy simulation, feedback
• report number: 2026-EXA262
• other publication id: 2026-EXA262
• language: English
• id: 9242174
• date added to LUP: 2026-08-31 09:40:09
• date last changed: 2026-08-31 09:40:09