top of page


My research interests evolve around astrophysics, focusing on modelling the evolution of galaxies and their central supermassive black holes (BHs) within a cold dark matter cosmogony, and on establishing a close connection between theoretical predictions and observed galaxy spectra. Specifically, my research aims at unveiling the differential impact of internal processes such as energetic phenomena from stars and accreting BHs (i.e. active galactic nuclei, AGN) as well as external processes due to environment and mergers on galaxies and BHs from the earliest cosmic epochs until present-day. Therefor, a combination of different theoretical, numerical techniques is employed and developed: semi-analytic models (SAMs), cosmological large-scale and zoom-in hydrodynamic simulations as well as spectral evolution models based on detailed photoionisation calculations. Below, I briefly summarise some important achievements and on-going projects.

The cosmic evolution of supermassive black holes and AGN


To investigate the evolution of supermassive BHs and AGN in the centres of galaxies over cosmic time, we work on improving models for BH seeding, growth and feedback in both SAMs (Hirschmann+12a, Fontanot, DeLucia, Hirschmann+20) and large-scale cosmological hydrodynamic simulations (“Magneticum” suite, see background image; Hirschmann+14a, Steinborn, Dolag, Hirschmann+15, Steinborn, Hirschmann+18). The latter includes a simulation run with a volume of 500^3 cMpc, large enough to capture rare, luminous AGN. The figure on the left illustrates the distribution of gas, stars and BHs of a 70Mpc slice of this simulation at z=3,2,1,0 (different panels, Hirschmann+14). Further information on the Magneticum simulations can be found and corresponding data sets can be downloaded here.

As an example, our modelling advances allowed us to demonstrate that the observed “anti-hierarchical trend” in BH growth (massive BHs predominantly form before that of less massive BHs) is a natural outcome of a hierarchically growing Universe, if feedback processes due to stars and AGN are strongly regulating the gas content available for accretion onto the BH, and if nuclear activity is not only driven by merger events – resolving any initial tension with the currently favoured LambdaCDM cosmology.

The impact of energetic phenomena on massive galaxies

Over the last decade, feedback processes from both different evolutionary stages of massive stars and AGN turned out to be crucial ingredients for forming realistic galaxies. To explore the differential impact of various stellar and AGN feedback processes on integrated and spatially resolved properties of massive galaxies (gas content, composition, metallicity and kinematics, as well as stellar populations) at different cosmic epochs, we are performing and analysing sets of cosmological zoom-in simulations, by developing and implementing different feedback models (Hirschmann+12/13/15; Choi,..,Hirschmann+17/18/19). The movie below visualises the evolution of the stellar (left panel) and gas content (middle and right panels) of a massive, present-day galaxy in one of our cosmological zoom-in simulations. In these simulations, we model the momentum and energy release from AGN, motivated by observed nuclear accretion disk winds, and we account for Compton/Photoionisation-heating and radiation pressure due to X-ray radiation (Ostriker+10, Choi,..., Hirschmann+17).

  • In our classic view, stellar feedback was thought to mainly affect the evolution low-mass galaxies by regulating their star formation. Our zoom-in simulations, instead, showed that stellar population properties, stellar and gas metallicities, as well as stellar and gas kinematics of massive present-day galaxies can be also strongly influenced by stellar feedback, and not only by AGN feedback as initially thought (Hirschmann+13/15; Serra, ..., Hirschmann+14). These intriguing results have been calling for a rethinking of the conventional picture, pointing towards a complex superposition of both stellar and AGN feedback governing the assembly of massive galaxies.

  • AGN feedback still remains a key player for forming realistic massive present-day galaxies. Our cosmological zoom-in simulations predict massive, present-day galaxies with a realistically low stellar content at a given halo mass due to efficient heating and removal of gas in the ISM, as a consequence of our kinetic AGN feedback model. This process is crucial for re-distributing gas and metals not only in, but also around massive galaxies (e.g. regulating the hot, X-ray luminous halo gas) as well as for shaping other properties such as stellar populations, stellar and gas kinematics, morphologies, sizes, etc., largely consistent with observational data at different cosmic epochs (Choi, ..., Hirschmann+2017/18/20, Hirschmann+17; Brennan,..,Hirschmann+18; Frigo, Naab, Hirschmann+18; Florian, Ziegler, Hirschmann+20; Schimek, Hirschmann+in prep.; Schoenegger, Hirschmann+in prep.). 

Nature versus nurture: the role of environment for galaxy evolution


Observed properties of galaxies have long been known to depend on the large-scale environment. Despite much effort though, it is not yet clear which physical, i.e. internal or external, processes are mainly responsible for the observed trends (“nature” vs. “nurture” hypothesis). To explore the relative role of environment for galaxy evolution (among other science questions), we have been employing and developing the GAEA (GAlaxy Evolution and Assembly) semi-analytic model (DeLucia+14, Hirschmann+16, Fontanot,..,Hirschmann+20). The figure on the left shows the projected distribution of galaxies between z=0.08 and z=0.13, colour-coded by the cold (HI) gas content (Original by A. Zoldan). More information on our GAEA model can be found hereResults from the GAEA model published in Hirschmann+16 are fully available here.

By quantifying the environmental history of (satellite) galaxies (Hirschmann+13, Hirschmann+14), we found time-scales for suppressing star formation in galaxies at high densities to be largely determined by the self-regulation between star formation and stellar feedback (Hirschmann+16; Fontanot, Hirschmann+17) rather than by environmental processes. This suggested that the regulation of star formation of galaxies in dense environments is primarily governed by internal processes, with environment playing a secondary role, favouring the “nature” hypothesis (De Lucia, Hirschmann+19). Nevertheless, further improvements of models for environment, such as ram pressure stripping of cold and hot gas, were necessary in GEA, to accurately reproduce observed long quenching time-scales and quiescent fractions of low-mass galaxies at a given galaxy stellar mass (Xie, De Lucia, Hirschmann+20).

Interpretative framework for observed emission line galaxies

A fair comparison between observed and simulated galaxies is vital for interpreting observations and putting strong constraints on empirically motivated models adopted in simulations. However, such a confrontation can often be only approximate because of uncertain assumptions and degeneracies, when deriving physical quantities from observed galaxy spectra and their emission lines. To establish a more accurate interface between simulations and spectral observables, we have been developing a novel, numerical methodology to provide mock spectra with nebular emission lines of galaxies via coupling spectral evolution models to cosmological simulations (e.g. IllustrisTNG) and SAMs (GAEA, SantaCruz SAM, Hirschmann+17/19/21 in prep.), see figure below for an illustration. We trace emission lines emerging from gas elements in the ISM ionised by young stars, AGN, post-AGB stellar populations and shocks. 


With our methodology, the predicted mock emission lines of cosmological simulations and SAMs are largely consistent with the that of SDSS galaxies and their location in the “BPT diagram”. Our mock predictions further agree with the observed increase in [OIII]/Hb from low to high redshifts (Hirschmann+17), and they are roughly consistent with the observed trends between strong line-ratios and Oxygen abundances in local galaxies (Hirschmann+in prep.). This unique link between physical galaxy formation properties/processes and direct spectral observables is also expected to be crucial for the interpretation of observed spectra of distant galaxies (e.g. to identify their ionising source(s), Hirschmann+19) – measurable with excellent statistics and quality with future spectrographs, such as NIRSpec/JWST. At moment, we are working on improving spectral evolution models by various means, and on enhancing the connected between simulated ISM properties and photo-ionisation calculations. Furthermore, employing the BEAGLE (BayEsian Analysis of GaLaxy sEds) tool, we are developing a pipeline, where physical galaxy properties are derived from simulated spectra in the same way as from observed spectra.

"BlackDawn": Galaxy and black hole formation at cosmic dawn


With novel, large sets of high-resolution cosmological zoom-in simulations, dedicated to the first few billion years after the Big Bang, we aim to robustly answer fundamental, yet unsolved questions in the field of galaxy formation and evolution at cosmic dawn:

  • How did the first  galaxies and the first supermassive black holes form and evolve just a few hundred million years after the BB?

  • Which fraction of first galaxies do already contain a supermassive BH in their centres?

  • What sets the structure, dynamics and energetics of the ISM of galaxies at cosmic dawn, and which is the impact of released energy from accreting BHs?

  • How can emission lines of galaxy spectra be used to formulate a consistent theory for the evolution of extreme galactic environments at earliest cosmic epochs?

  • What is the origin of the increasingly large amount of massive quiescent galaxies just 1-2 Gyr after the BB? What causes the SF quenching in these objects?

With our work, we plan to provide the necessary theoretical background for interpreting existing emission line observations, and for designing future high-redshift surveys with revolutionary observatories such as the James Webb Space Telescope.

The figure on the left exemplarily shows the stellar and gas content 0.8 and 2.5 Gyr after the BB for one of our high-resolution, high-redshift cosmological zoom-in simulations. 

bottom of page