Periodic Reporting for period 2 - MULTIDIMSPEC (Multi-dimensional analysis of the metal-poor Galaxy)
Période du rapport: 2021-07-01 au 2022-12-31
The Milky Way is a puzzle made of hundreds of billions of individual pieces: a spectacular mixture of stars of all ages, some newly born and some as old as the Universe itself. With the combined data from telescopes on the ground and in space we determine the positions, motions, and chemical composition of millions of stars to reconstruct our Galaxy’s formation history. How many smaller galaxies have been cannibalized by the beautiful large spiral galaxy we live in today? The starlight also reveals the origin of elements; how much the stars synthesize and how much they destroy. Was the Li in your battery created in the Big Bang itself? Which violent stellar explosions created the Cu in your saucepan?
To find the answers to these and other questions requires development of instruments, analysis methods, and theory. Central to this ERC project are three optical multi-object spectrographs, each connected to a 4m-telescope: GALAH@AAT (Siding Spring Observatory, Australia), WEAVE@WHT (Roque de los Muchachos Observatory, La Palma, Spain) and 4MOST@VISTA (Paranal Observatory, Chile). Together, the surveys will collect tens of millions of stellar spectra, a regime change that can only be mastered by introducing elements of machine learning in the analysis and scientific exploitation.
Of fundamental importance is also the quality of synthetic stellar spectra, which we generate by computing a model of the stellar atmosphere from where light is emitted. If the model is based on erroneous physical assumptions or uncertain atomic data, we will infer inaccurate chemical abundances and fundamental stellar parameters. In this ERC project, we abandon the traditional modelling assumptions that stellar atmospheres are static and one-dimensional and that all matter and radiation are in equilibrium (LTE), which leads to improvements that are particularly significant for the oldest and most chemically pristine (metal-poor) stars in our Galaxy.
To find the answers to these and other questions requires development of instruments, analysis methods, and theory. Central to this ERC project are three optical multi-object spectrographs, each connected to a 4m-telescope: GALAH@AAT (Siding Spring Observatory, Australia), WEAVE@WHT (Roque de los Muchachos Observatory, La Palma, Spain) and 4MOST@VISTA (Paranal Observatory, Chile). Together, the surveys will collect tens of millions of stellar spectra, a regime change that can only be mastered by introducing elements of machine learning in the analysis and scientific exploitation.
Of fundamental importance is also the quality of synthetic stellar spectra, which we generate by computing a model of the stellar atmosphere from where light is emitted. If the model is based on erroneous physical assumptions or uncertain atomic data, we will infer inaccurate chemical abundances and fundamental stellar parameters. In this ERC project, we abandon the traditional modelling assumptions that stellar atmospheres are static and one-dimensional and that all matter and radiation are in equilibrium (LTE), which leads to improvements that are particularly significant for the oldest and most chemically pristine (metal-poor) stars in our Galaxy.
The surfaces of stars like our Sun are boiling; surface convection cause enormous hot bubbles of gas to rise and overturn, releasing their surplus energy before the cooled gas trickles back to the deeper layers of the atmosphere. This phenomenon must be modelled by 3D radiation-hydrodynamic simulations running on supercomputers. We have produced new models of metal-poor dwarf and giant stars, with upper surface layers that are significantly cooler than the traditional 1D hydrostatic models. Our re-analysis of the most metal-poor star found in the Milky Way points to a much higher content of C and Fe than previously thought, but still less than a thousandth of the amounts found in the Sun. This has consequences for our understanding of how star formation could take place in the newborn universe.
In collaboration with atomic physicists, we have assembled new atomic models for neutral and singly ionized Ti and Cu. The models contain information about their complex term structure and their radiative and collisional transition probabilities. Only a few years ago, such models contained very large uncertainties, due to the essentially unknown collisional cross-sections with neutral hydrogen, which is the main constituent of stellar atmospheres. Equipped with our new atomic models, we can now predict the strength of Ti and Cu lines in stellar spectra without resorting to the inaccurate LTE assumption. We show that metal-poor star abundances are significantly affected, but more work, in particular using 3D models, remains before the abundances of these astrophysically important elements can be placed on a firm footing.
Stars have excellent memories; their chemical compositions always mirror their birth clouds and their orbital properties, like total energy and angular momentum, are preserved despite numerous laps around the centre of the Galaxy. We have taken advantage of this and used unsupervised clustering algorithms and wavelet transforms to search for over-densities in the multi-dimensional data sets that combine chemistry, motions, and age. We identify remnants of very old clusters and dwarf galaxies scattered across the halo of the Milky Way, which helps us narrow down the sizes and timings of the giant engulfment events that produced them.
In collaboration with atomic physicists, we have assembled new atomic models for neutral and singly ionized Ti and Cu. The models contain information about their complex term structure and their radiative and collisional transition probabilities. Only a few years ago, such models contained very large uncertainties, due to the essentially unknown collisional cross-sections with neutral hydrogen, which is the main constituent of stellar atmospheres. Equipped with our new atomic models, we can now predict the strength of Ti and Cu lines in stellar spectra without resorting to the inaccurate LTE assumption. We show that metal-poor star abundances are significantly affected, but more work, in particular using 3D models, remains before the abundances of these astrophysically important elements can be placed on a firm footing.
Stars have excellent memories; their chemical compositions always mirror their birth clouds and their orbital properties, like total energy and angular momentum, are preserved despite numerous laps around the centre of the Galaxy. We have taken advantage of this and used unsupervised clustering algorithms and wavelet transforms to search for over-densities in the multi-dimensional data sets that combine chemistry, motions, and age. We identify remnants of very old clusters and dwarf galaxies scattered across the halo of the Milky Way, which helps us narrow down the sizes and timings of the giant engulfment events that produced them.
The main goal of this ERC project is to advance the state-of-art of stellar spectroscopy. In collaboration with researchers at Århus University, Uppsala University, and Australian National University, we work on the construction and public release of a grid of 3D model atmospheres, spanning a large range of stellar parameters. This will enable research teams worldwide to begin the transition from 1D to 3D atmospheres. The development goes hand-in-hand with the evolution of 3D non-LTE spectrum synthesis, a field in which we are pioneers.
The GALAH survey is the leading optical high-resolution survey in the world. We contributed to its third data release, which took place in Sep. 2021, presenting fundamental parameters and chemical abundances for up to 30 elements for more than half a million stars. In the two years to come,we expect GALAH and WEAVE to more than double this number, and in particular increase the number of metal-poor stars. We have developed a machinery to analyse such large stellar samples fully automatically with python, utilizing non-LTE radiative transfer for more than a dozen astrophysically important elements, thereby removing one of the largest error sources.
Equipped with the data analysis tools we have developed, the future data sets will help us identify more fine-grained structures in chemo-dynamic space. How many large accretion events did the Milky Way experience in its youth? How much of the stellar halo is made up of disrupted globular clusters? Can rare stars with peculiar compositions reveal unusual stellar explosions?
In summary, this ERC project is placed at the forefront of theoretical and observational developments in stellar astrophysics, advancing the frontier of the research field to address questions related to the origin and evolution of elements, stars, and entire galaxies.
The GALAH survey is the leading optical high-resolution survey in the world. We contributed to its third data release, which took place in Sep. 2021, presenting fundamental parameters and chemical abundances for up to 30 elements for more than half a million stars. In the two years to come,we expect GALAH and WEAVE to more than double this number, and in particular increase the number of metal-poor stars. We have developed a machinery to analyse such large stellar samples fully automatically with python, utilizing non-LTE radiative transfer for more than a dozen astrophysically important elements, thereby removing one of the largest error sources.
Equipped with the data analysis tools we have developed, the future data sets will help us identify more fine-grained structures in chemo-dynamic space. How many large accretion events did the Milky Way experience in its youth? How much of the stellar halo is made up of disrupted globular clusters? Can rare stars with peculiar compositions reveal unusual stellar explosions?
In summary, this ERC project is placed at the forefront of theoretical and observational developments in stellar astrophysics, advancing the frontier of the research field to address questions related to the origin and evolution of elements, stars, and entire galaxies.