Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Effects of feedback and resolution on galaxy formation at z = 3

Åhman, Milles LU (2026) FYSK04 20261
Department of Physics
Astrophysics
Abstract
Astrophysicists have long studied the early Universe in an attempt to explain where galaxies, and by proxy us, come from. In this thesis, the global properties of galaxies at z = 3, where star formation plays a large part , are studied when changing the stellar feedback strength and resolution in simulations. The simulation code used is based on RAMSES, using a similar setup to VINTERGATAN. Four simulations were run: one with fiducial supernova energies, one with half the fiducial energy, one with twice the energy, and lastly a fiducial simulation with lower spatial and mass resolution. These simulations have shown that when increasing feedback, the gas outflows from galaxies increase, enriching the circumgalactic medium (CGM) with metals... (More)
Astrophysicists have long studied the early Universe in an attempt to explain where galaxies, and by proxy us, come from. In this thesis, the global properties of galaxies at z = 3, where star formation plays a large part , are studied when changing the stellar feedback strength and resolution in simulations. The simulation code used is based on RAMSES, using a similar setup to VINTERGATAN. Four simulations were run: one with fiducial supernova energies, one with half the fiducial energy, one with twice the energy, and lastly a fiducial simulation with lower spatial and mass resolution. These simulations have shown that when increasing feedback, the gas outflows from galaxies increase, enriching the circumgalactic medium (CGM) with metals and higher temperatures. Furthermore, higher feedback spreads gas more evenly across the galactic
disk, broadening the stellar and dark matter distribution. High-resolution simulations using this feedback model seem to overestimate the Kennicutt-Schmidt slope, while the low-resolution simulations seem to follow it precisely. This leads to the low-resolution run corresponding to doubling the feedback in high-density regions in most global properties and distributions. Lowering the resolution does, however, incorrectly overestimate the gas temperature in both the galactic disk and CGM compared to all high-resolution simulations. (Less)
Popular Abstract
What is the oldest thing we can see from the Earth? At first glance, this might seem like a nonsensical question, but to astronomers, it is very much not. As you probably know, light has a finite speed, roughly 3 * 10^8 m/s. This is incredibly fast on Earth but quite slow in space. Therefore, if we see the light from a galaxy, odds are that the image we see is really several million, if not billions, of years old. This gives astronomers a sort of "time machine" where they can see the Universe as it was much earlier in its life.

The farthest galaxies we have managed to observe are galaxies actively being formed. They form from huge clouds of dark matter collapsing under gravity. Dark matter is invisible and doesn't interact with normal... (More)
What is the oldest thing we can see from the Earth? At first glance, this might seem like a nonsensical question, but to astronomers, it is very much not. As you probably know, light has a finite speed, roughly 3 * 10^8 m/s. This is incredibly fast on Earth but quite slow in space. Therefore, if we see the light from a galaxy, odds are that the image we see is really several million, if not billions, of years old. This gives astronomers a sort of "time machine" where they can see the Universe as it was much earlier in its life.

The farthest galaxies we have managed to observe are galaxies actively being formed. They form from huge clouds of dark matter collapsing under gravity. Dark matter is invisible and doesn't interact with normal matter in any other way than gravitationally. As dark matter collapses, it forms "halos" which act like a sink, attracting gas to their centres. This gas can then collapse to form stars, which is what we can then observe with our telescopes. This is also a very violent process, with most young galaxies eventually being eaten by others while they grow. During these collisions, the star formation rate can go up from the colliding gas. The process of star formation is therefore essential for understanding galaxy formation.

Most galaxies can regulate their own star formation using supernovae. Supernovae are giant explosions that happen at the end of the lives of the most massive stars. These supernovae release so much energy that the gas in whole galaxies are affected through higher temperatures and more chaotic distributions. THis lowers the star formation rate in the host galaxy. The more stars that form, the more supernovae will be triggered, which in turn suppresses the formation of stars. This becomes a negative feedback loop, which can suppress star formation by a large factor. This process lies, therefore, at the heart of galaxy formation, yet it is not well understood.

As these galaxies are so far away, it is very difficult for us to understand how these processes actually work with telescopes alone. To improve our understanding, we simulate the earliest galaxies on large computers, changing parameters until the result of our simulations matches our observations. We can then study the simulations and learn much more about the processes than our limited observations.

In this thesis, I am analysing four simulations, three where I vary how energetic the supernovae are and one where I vary the resolution of the simulation. Through this analysis, I should be able to determine how simulations are affected by supernovae and their resolution. We might then compare the results with observations and evaluate what supernova energy matches observations. With better simulations, we can learn much more about this very turbulent era of the universe's history. (Less)
Popular Abstract (Swedish)
Vad är den äldsta saken vi kan se från jorden? Detta kan verka som udda fråga, men för astronomer är den helt rimlig. Du vet antagligen att ljuset har ändlig hastighet, ungefär 3 * 10^8 m/s. Detta är otroligt snabbt på jorden, men inte lika så i rymden. Alltså, om vi med ljus observerar en galax är antagligen den bild vi ser miljoner, om inte miljarder, år gammal. Detta ger astronomer en sorts "tidsmaskin" som låter dem observera universum som det var mycket tidigare i sitt liv.

De mest avlägsna galaxerna vi har lyckats observera är galaxer som precis håller på att skapas. Galaxer bildas i enorma moln av mörk materia som långsamt kollapsar under sin egen gravitation. Mörk materia är en osynlig typ av materia som inte interagerar med... (More)
Vad är den äldsta saken vi kan se från jorden? Detta kan verka som udda fråga, men för astronomer är den helt rimlig. Du vet antagligen att ljuset har ändlig hastighet, ungefär 3 * 10^8 m/s. Detta är otroligt snabbt på jorden, men inte lika så i rymden. Alltså, om vi med ljus observerar en galax är antagligen den bild vi ser miljoner, om inte miljarder, år gammal. Detta ger astronomer en sorts "tidsmaskin" som låter dem observera universum som det var mycket tidigare i sitt liv.

De mest avlägsna galaxerna vi har lyckats observera är galaxer som precis håller på att skapas. Galaxer bildas i enorma moln av mörk materia som långsamt kollapsar under sin egen gravitation. Mörk materia är en osynlig typ av materia som inte interagerar med synlig materia på något sätt förutom just tyngdkraften. När den mörka materien kollapsar skapar den stora moln som i sin tur attraherar gas. Denna gas kollapsar sedan för att skapa stjärnor, vilket är det våra teleskop observerar. Detta är en mycket turbulent process i vilken de flesta unga galaxer blir uppätna av större galaxer i stora kollisioner. I dessa kollisioner kan stjärnbildning öka på grund av kollisioner i gasen. Hur stjärnor bildas är därför ett otroligt viktigt ämne för att förstå galaxer.

De flesta galaxer kan reglera sin egen stjärnbildning med hjälp av supernovor. Supernovor är stora explosioner som sker i slutet av massiva stjärnor liv. Supernovorna släpper ut så mycket energi att gasen i hela galaxerna blir påverkad med höjda temperaturer och en mer kaotisk fördelning. Detta sänker hur snabbt stjärnbildning sker. Ju fler stjärnor som skapas, desto fler supernovor kommer att skapas, vilket i sin tur sänker stjärnbildningen i galaxen. Detta blir en feedback loop som sänker hastigheten i vilka galaxer bildar stjärnor med en stor faktor. Denna process tros på så sätt ligga i självaste hjärtat av galaxbildning. Trots detta är den rätt dåligt förstådd av forskningen.

Eftersom dessa galaxer är så långt borta är det väldigt svårt för oss att förstå dessa processer med enbart teleskop. För att förstå detta bättre har vi därför vänt oss till datorsimuleringar. Vi simulerar galaxernas födelse samtidigt som vi ändrar på underliggande processerna tills vi får något som ser ut som våra observationer. Vi kan sedan studera simuleringarna i mycket högre detalj och därför lära oss mer om hur verkligheten fungerar.

I detta projektet studerar jag fyra simuleringar. I tre av dem varierar jag mängden energi som varje supernova utlöser och i den sista minskar jag upplösningen. Genom denna analys borde jag kunna dra slutsatser om hur galaxer påverkas av just supernovor och hur simuleringarna påverkas av deras upplösning. Vi borde sedan kunna jämföra de simulerade galaxerna med riktiga galaxer i observationer och från det kunna komma fram vilka supernovor som är mest sanningsenliga. Med bättre simuleringar kan vi lära oss mer om hur universum såg ut i sina tidigaste stadier. (Less)
Please use this url to cite or link to this publication:
author
Åhman, Milles LU
supervisor
organization
course
FYSK04 20261
year
type
M2 - Bachelor Degree
subject
keywords
Galaxies, Galaxy formation, Simulations, Stellar Feedback
report number
2026-EXA260
other publication id
2026-EXA260
language
English
id
9241888
date added to LUP
2026-06-19 17:49:36
date last changed
2026-06-19 17:49:36
@misc{9241888,
  abstract     = {{Astrophysicists have long studied the early Universe in an attempt to explain where galaxies, and by proxy us, come from. In this thesis, the global properties of galaxies at z = 3, where star formation plays a large part , are studied when changing the stellar feedback strength and resolution in simulations. The simulation code used is based on RAMSES, using a similar setup to VINTERGATAN. Four simulations were run: one with fiducial supernova energies, one with half the fiducial energy, one with twice the energy, and lastly a fiducial simulation with lower spatial and mass resolution. These simulations have shown that when increasing feedback, the gas outflows from galaxies increase, enriching the circumgalactic medium (CGM) with metals and higher temperatures. Furthermore, higher feedback spreads gas more evenly across the galactic
disk, broadening the stellar and dark matter distribution. High-resolution simulations using this feedback model seem to overestimate the Kennicutt-Schmidt slope, while the low-resolution simulations seem to follow it precisely. This leads to the low-resolution run corresponding to doubling the feedback in high-density regions in most global properties and distributions. Lowering the resolution does, however, incorrectly overestimate the gas temperature in both the galactic disk and CGM compared to all high-resolution simulations.}},
  author       = {{Åhman, Milles}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Effects of feedback and resolution on galaxy formation at z = 3}},
  year         = {{2026}},
}