Seminars and Colloquia at ESO Garching and on the campus

Intermediate-mass forming stars (Herbig stars, 1.5 to 15 solar-mass young stellar objects) were first characterized as a population by George Herbig in 1960. Today, they are of great importance for planet formation studies and for linking the low- and high-mass star formation regimes. Historically, the study of the general properties of Herbig stars has been limited by the lack of a well-defined sample. In addition, only a few sources were known, most of them discovered serendipitously. However, this has recently changed thanks to the homogeneous discovery of many new stars of the class. In this talk, I provide a brief review of the history of Herbig stars and describe our current knowledge of this population of stars and their protoplanetary disks, which host key observational features for understanding planet formation and the interaction of young stars with their environments. I present the recent results of Stapper et al. (2026), who surveyed all known Herbig stars within 1 kpc, and I describe ongoing efforts to understand the puzzling nature of the more massive Herbig stars (over 8 solar masses). Forming stars that massive are not expected to be optically visible, and thus their study opens a new window into massive star formation.

The past decade of observations of directly imaged companions has signaled a revolution in exoplanet spectroscopy. The objects observed to date reveal an extraordinary diversity of atmospheres, spanning a wide range of ages, temperatures, metallicities, and cloud properties. In this talk, I will present various results obtained with the numerical frameworks I am developing to model these complex exoplanetary spectra: ForMoSA and Optex. Using ForMoSA, I will discuss key atmospheric retrieval findings on benchmark targets (e.g., AF Lep b, beta Pic b, VHS 1256 b) alongside statistical studies of field brown dwarf spectral libraries. Specifically, I will address how these results help us piece together the broader picture of planet formation and evolution, particularly in contrast to stellar formation mechanisms and how formation pathways impact observable spectral features. Finally, I will conclude by outlining future prospects for the ELT era, highlighting end-to-end simulation work for MICADO's Long-Slit Spectroscopy (LSS) mode and its potential to further uncover the atmospheric properties of directly imaged exoplanets.

Spatially resolved observations of nearby H II regions have long shown that the ionised interstellar medium (ISM) is far from uniform, exhibiting significant variations in density, temperature, and, in some cases, chemical composition. However, classical emission-line diagnostics often rely on simplified descriptions of the emitting gas. When applied to integrated spectra of high-redshift galaxies, where this internal complexity remains unresolved, such assumptions can bias the inferred physical conditions and chemical abundances. 

In this talk, I will present recent results obtained with HOMERUN, a multi-cloud photoionisation-modelling framework designed to account for the complex structure of the ISM. I will first show how the detailed modelling of individual well-characterised high-redshift galaxies reveals hidden density and ionisation structures that can bias classical metallicity measurements by 0.2–1 dex. I will also discuss how these structures complicate the interpretation of unusual abundance patterns and high-ionisation emission lines. Finally, I will then extend the analysis to ~2500 star-forming galaxies at 2 < z < 12, showing that high-redshift systems broadly follow the metallicity–ionisation sequence observed locally, while exhibiting significantly higher gas densities and substantial density stratification.

The distribution of galaxies throughout the Universe is mapped to ever better precision by an ambitious survey program, and it holds important clues about the nature of Dark Energy, Gravity, and Dark Matter. Yet, unlocking this information is challenging. State-of-the-art analyses compare theory and data in terms of summary statistics, foremost the galaxy power spectrum. In this compression, information is lost. Field-level analysis, in contrast, considers the three-dimensional galaxy density and can optimally constrain cosmological parameters. These gains can be significant, and on mock data, field-level analysis doubles the precision over the state-of-the-art for some parameters. As an additional benefit, it constrains the primordial density fluctuations that give rise to the observed galaxy distribution. A central ingredient to facilitate field-level analysis is the forward model that connects an assumption on cosmology and initial conditions to the observations. By describing galaxy bias in the forward model with the Effective Field Theory of Large Scale Structure (EFT of LSS), uncertainties in galaxy formation physics can be marginalized over. Nevertheless, the high precision in field-level analysis puts stringent requirements on the accurate representation of cosmological, observational, and systematic effects, and the high number of variables demands sophisticated algorithms for numerical inference. In this talk, I explain how field-level analysis can be implemented, why it should, and the challenges ahead in realizing the promise of robust and precise cosmological insights on the initial conditions and their cosmological evolution from observations.

The discovery of dusty star-forming galaxies (DSFGs) as major contributors to the Cosmic Infrared Background has transformed our understanding of galaxy evolution. With star formation rates reaching hundreds to thousands of solar masses per year, these galaxies account for a substantial fraction of cosmic star formation, yet the molecular gas reservoirs that fuel their activity and the physical conditions of their interstellar medium (ISM) remain challenging to constrain. In this talk, I will present results from VLA Vz-GAL, a large survey of CO(1-0) emission in 106 DSFGs at redshifts 1 to 6, exploring what these observations reveal about their molecular gas reservoirs. I will then introduce TUNER, a state-of-the-art turbulent, non-LTE radiative transfer framework that jointly models dust continuum and molecular-line emission to infer the physical conditions of the ISM. Finally, I will discuss how extending this approach to incorporate ISM heating and chemistry can help connect the molecular gas physics of high-redshift DSFGs and quasar hosts with that of local ultraluminous infrared galaxies, offering a broader perspective on the processes that regulate galaxy growth across cosmic time

Galaxies in the low-redshift Universe trace an intricate “cosmic web” pattern: dense clusters are linked by more tenuous filaments, which in turn form walls around giant underdense voids. The observed properties of galaxies vary systematically between these environments, pointing to a fundamental connection between galaxies and the cosmic web around them. I will summarize the empirical evidence for this link, and then illustrate how large hydrodynamical simulations help to understand its physical nature. I discuss the predicted structure of filaments, how well galaxies trace their underlying dark matter skeletons, and to what extent filaments influence the baryon content and star formation activity within their galaxies. For clusters, I show how like-with-like comparisons between simulations and observations can be enabled by simulation-based halo mass measurements of real systems. I discuss successes and failures of the recent COLIBRE simulation model emerging from these comparisons, and what they teach us about cluster galaxies in the real Universe. I end by outlining how deep spectroscopic surveys and further advanced simulations will soon enable us to probe the co-evolution of galaxies and their cosmic web environment over the majority of cosmic time.