Fascination with Technology Milky Way significantly older than previously thought

Related Vendors

Leibniz-Gemeinschaft e.V.

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

In our "Fascination with Technology" section, we present impressive projects from research and development to designers every week. Today: how machine learning is revising the history of the Milky Way's formation.

Machine learning sheds new light on the formation history of our Milky Way: A surprising discovery about the development of our galaxy, based on data from the Gaia mission, revealed a large number of old stars on similar orbits as our Sun. They formed the thin disk of the Milky Way less than one billion years after the Big Bang, several billion years earlier than previously thought.

The Milky Way consists of a large halo, a central bulge and a bar, a thick disk, and a thin disk. Most stars are located in the so-called thin disk of our Milky Way and follow an organized rotation around the galactic center. Middle-aged stars, like our 4.6 billion-year-old Sun, belong to the thin disk, which is thought to have formed about 8 to 10 billion years ago.

What lies behind galactic archaeology

Understanding how the Milky Way was formed is an important goal of galactic archaeology. For this purpose, detailed maps of the galaxy are needed that show the age, chemical composition, and movements of the stars. These maps, known as chrono-chemo-kinematic maps, help to understand the history of our galaxy. Creating these detailed maps is a challenge because it requires large datasets of stars with precisely known ages.

A common approach to tackling this challenge is the study of very metal-poor, old stars that provide a window into the early days of the Milky Way. Very metal-poor stars are known to be old because they were among the first stars to form when the universe was still largely composed of hydrogen and helium, before many of the heavier elements were generated and distributed by subsequent generations of stars.

The old stars in the disk suggest that the formation of the Milky Way's thin disk began much earlier than previously thought, about four to five billion years ago.

Samir Nepal, AIP

Data set from the Gaia mission used

Using a dataset from the Gaia mission of the European Space Agency (ESA), an international team led by astronomers from the Leibniz Institute for Astrophysics Potsdam (AIP) studied stars in the vicinity of the Sun, about 3200 light-years around the Sun. They discovered a surprising number of very old stars in thin disk orbits; most of them are older than 10 billion years, some even older than 13 billion years. These old stars show a wide range of metal compositions: some are very metal-poor (as expected), while others have twice the metal content of our much younger Sun, indicating that there was rapid metal enrichment in the early phase of the Milky Way's development.

"These old stars in the disk suggest that the formation of the Milky Way's thin disk began much earlier than previously thought, about four to five billion years ago," explains Samir Nepal from the AIP and lead author of the study. "This study also shows that our galaxy had intense star formation in early epochs, leading to very rapid metal enrichment in the inner regions and the formation of the disk. This discovery aligns the timeline of disk formation in the Milky Way with the timeline of high-redshift galaxies observed by the James Webb Space Telescope (JWST) and the Atacama Large Millimeter Array (ALMA) radio telescope. The results indicate that cold disks were able to form and stabilize very early in the history of the universe, providing new insights into galaxy evolution."

Research uses advanced machine learning technologies

"Our study suggests that the Milky Way's thin disk may have formed much earlier than we thought, and that its formation is closely linked to the early chemical enrichment in the innermost regions of our galaxy," explains Cristina Chiappini. "The combination of data from different sources and the application of advanced machine learning technologies have enabled us to increase the number of stars with high-quality stellar parameters —a crucial step that has led our team to these new insights."

More than 800,000 stars analyzed

The findings were made possible by the third data release of the Gaia mission. The team analyzed the stellar parameters of more than 800,000 stars using a novel machine learning method that combines information from different data types to obtain improved stellar parameters with high precision. These precise measurements include gravity, temperature, metal content, distances, kinematics, and the ages of the stars.

In the future, a similar machine learning procedure will be used to analyze millions of spectra collected during the 4MIDABLE-LR survey with the 4-Meter Multi-Object Spectroscopic Telescope (4MOST), which will commence operations in 2025.