Research projects

VESTRAL: galaxies star by star in the first billion years

PIs: Dylan Nelson and Matthew Smith; co-PIs: Annalisa Pillepich and Ralf Klessen

What happens when galaxies are simulated finely enough to follow individual stars and single supernova explosions? VESTRAL is a new project of very high-resolution cosmological “zoom-in” simulations of galaxy formation. It models gas down to temperatures of 10 K, resolves the interstellar medium below one parsec, and forms and evolves stars one by one, so that each supernova and its feedback on the surrounding gas is captured explicitly. With these simulations, we study how star clusters and nuclear star clusters form, how the first galaxy disks emerge, how supernova-driven bubbles shape young galaxies, and how the first black holes are seeded and grow. The first VESTRAL papers will appear soon.

TNG-Cluster

PIs: Annalisa Pillepich and Dylan Nelson

TNG-Cluster is a new addition to the IllustrisTNG suite of cosmological magnetohydrodynamical simulations of galaxy formation. It focuses on >350 massive galaxy clusters with all the physics of IllustrisTNG and with a numerical resolution that allows to robustly model simultaneously the large-scale structure out of which clusters form but also the galaxies therein. See here for info, projects and results.

Cosmic-Key (ERC 101087822)

PI: Annalisa Pillepich

The key to precise and accurate cosmology? Simulating the physics that shapes gaseous haloes. We aim at developing, providing, and using the simulations, mocks, tools, and methods to maximize the scientific return of cosmological surveys. In particular, we focus on galaxy clusters and their potential to constrain the parameters that describe the functioning and evolution of the Universe as a whole.

ERGO-ML

PI: Annalisa Pillepich

Is it possible to infer the past merger and assembly history of galaxies based on their photometric images? Or can we learn anything about the supermassive black holes at the center of massive galaxy clusters from the maps of their core regions from X-ray spectroscopy? With ERGO-ML (Extracting Reality from Galaxy Observables with Machine Learning), we use machine learning tools to infer unobservable physical properties of galaxies, and galaxies groups and clusters, from their observable manifestations, especially at the so-called map or field level.

IllustrisTNG

PI: Volker Springel; PIs of TNG50: Dylan Nelson and Annalisa Pillepich

I am one of the developers of the IllustrisTNG simulations and co-PI of the TNG50 simulation. These are gravity + magnetohydrodynamics simulations of large cosmological volumes that simultaenously resolve thousands of galaxies therein. Their outcome has been used by now in more than one thousand papers! Among them are those from my team, whereby we have extracted novel and useful sicentific notions across a wide range of topics: see below and see more at the IllustrisTNG webpage

Cosmology with eROSITA and the physics of galaxy clusters

eROSITA has been launched in 2019 and is delivering an unprecedented all-sky map of the X-ray universe, including about hundred of thousands galaxy clusters. These can be used to place constraints on our cosmological models, including the level and evolution of Dark Energy. In parallel, we can and need to use simulations to quantify and predict the properties of the intra-cluster plasma, including its observational signatures as well as the effects of magnetic fields, shocks and feedback from the central super massive black holes.

Cluster galaxies and evolution of galaxies in different environments

We use simulations to quantify the effects of environment into the evolution of galaxies, their star formation activity, morphologies, and stellar and gaseous content. For example, satellite galaxies orbiting within the potential well of massive galaxy clusters undergo stripping of their materials, including the gas which sometimes produce extended wakes, or Jellyfish Galaxies. See our Citizen Scientist Project on Zooniverse

Assembly and properties of Milky Way and Andromeda-like galaxies

One of the most intriguing challenges of our field is to understand our Galaxy, the Milky Way and our nieghbor Andromeda. We aim to model all aspects of it. We develop and use simulations (TNG0, IllustrisTNG, Illustris and Eris) to understand how special our Galaxy and Andromeda are, how they formed and how to interpret observational data.
See our special data release of TNG50 MW/M31-like galaxies here!

Galaxy assembly, accreted stars and stellar haloes

Simulations have shown that galaxies assemble their stars by both making them (star formation) and by accreting and merging with other galaxies (accretion). Many of these accreted or ex-situ stars can be found at large distances around galaxies, forming the so called stellar haloes. We use simulations to quantify the assembly of galaxies and their stellar haloes across these different assembly channels, make predictions for future surveys and do comparisons to observations.

Effects of baryons on (cold) dark matter

Simulations including only the effects of gravity (and hence only dark matter) have provided an exquisitely precise picture of the properties and distribution of dark matter, in haloes and across the large scale structure. However, recently, simulations including star formation, gas physics and feedback seem to suggest that such picture is more complex than anticipated…

Our Projects