Written in the Stars: the Moment our Galaxy Began Boiling

Title: Birth of the Galactic Disk Revealed by the H3 Survey

Authors: Charlie Conroy, David H. Weinberg, Rohan P. Naidu, Tobias Buck, James W. Johnson, Phillip Cargile, Ana Bonaca, Nelson Caldwell, Vedant Chandra, Jiwon Jesse Han, Benjamin D. Johnson, Joshua S. Speagle, Yuan-Sen Ting, Turner Woody, Dennis Zaritsky

First Author’s Institutions: Center for Astrophysics | Harvard & Smithsonian, Cambridge, MA, 02138, USA

Status:  Submitted to ApJ

The Milky Way has a messy history.  In its 13+ billion years of life, it has experienced varying stages of evolution–star formation, dwarf galaxy accretion, structural changes–and things will only get messier moving forward (for example, the Milky Way is projected to collide with Andromeda within the next 4.5 billion years).  Understanding the Milky Way’s interesting past is important not only for understanding the structural and chemical evolution of the Milky Way but also for informing the evolution of spiral galaxies beyond our own.

One way to study the Milky Way’s evolutionary history is by observing its current state and searching for clues about its past.  Stars are great tools for this purpose.  Though stars make up just ~4% of the mass in the Milky Way (the rest of it lies in dark matter, gas, and dust), they are unique in that they carry the imprint of the Milky Way’s past in their present chemical and orbital characteristics.  Some astronomers like to call stars the “time capsules” of the cosmos for this reason.

Today’s authors use the chemistry and orbital characteristics of stars in the Milky Way to identify a transitional period in our Galaxy’s history during which the Milky Way went from ‘simmering’ to ‘boiling’ and cooked up the Milky Way disk.

Step One: Sieve for In Situ Stars

The principal aim of today’s work is to find hints about the Milky Way’s past by studying the behavior of stars born within the Milky Way (in situ). The authors pull from the H3 spectroscopic and Gaia astrometric surveys to create their sample of Milky Way stars with chemical and kinematic data. Because the authors want only in situ stars, they must filter out all stars that were likely accreted by the Milky Way and thus carry the imprint of their own extra-Galactic history.  To do this, the authors apply cuts to the survey data, selecting only stars that are on relatively circular (eccentricity < 0.8), prograde orbits (as opposed to retrograde orbits, which go against the bulk rotational trend of the Milky Way).  The idea here is that accreted stars tend to have orbital characteristics that are physically chaotic and do not obey the general trend of Milky Way stars.

Figure 1: A plot of the ages of in situ stars in the Milky Way as a function of their iron content ([Fe/H]).  The authors recreate the expected trend between these two quantities: as the Galaxy ages (downward on the y-axis), its overall iron content increases (rightward on the x-axis).  This is because iron is synthesized in dying stars and released into the interstellar medium with each new generation. Figure 4 in the paper.

Once the authors extract their in situ stellar population, they set off to analyze their sample and see what they can learn about the Milky Way’s evolutionary history.  They first determine the ages of their stars using isochronal fitting.  They recover the expected trend between iron content ([Fe/H], also sometimes called metallicity) and age (Figure 1), where younger stars have higher [Fe/H] and older stars have lower [Fe/H].  This trend is expected because as the Milky Way ages, it becomes more metal-rich.  This is an effect of stellar nucleosynthesis: stars generally create metals by fusing lighter elements into heavier ones, and when stars die (be it as supernovae or planetary nebulae), they release their nucleosynthetic output into the interstellar medium.  New stars are born of this metal-rich material, and the cycle continues.  For this reason, when we look at Figure 1, it makes sense that the younger stars have higher iron content.

Step Two: Sprinkle Stars onto Tinsley-Wallerstein Diagram

Next, the authors plot their sample of in situ stars on the Tinsley-Wallerstein diagram, which compares the abundance of various alpha-elements in a star relative to iron ([alpha/Fe]) to its overall iron content, where the alpha-elements include O, Ne, Mg, Si, S, Ar, Ca, and Ti.  The Tinsley-Wallerstein diagram is a popular tool for studying the star formation history of a select stellar population.  It works by leveraging the different timescales at which iron and the alpha-elements are created and released into the interstellar medium.  Iron has two main creation sources: the deaths of massive stars (called a core-collapse or Type II supernovae) and the explosions of white dwarfs in binary systems (called a Type 1a supernovae).  In other words, each time one of these events occurs in our Galaxy, the Galactic content of iron increases.  This is why we see the correlation between stellar age and iron content in Figure 1.  Unlike iron, the alpha-elements are created exclusively in dying massive stars.  The alpha-elements are released into the interstellar medium on rapid time scales because high-mass stars are short-lived.  Iron, on the other hand, has a more steady release into the interstellar medium because it comes from both dying massive stars and exploding white dwarfs, which are low-mass and thus long-lived.

Figure 2: The alpha-to-iron abundance ([alpha/Fe]) of in situ stars in the Milky Way as a function of their iron content ([Fe/H]).  We observe a negative slope at [Fe/H] < -1.3, and a peculiar hump at [Fe/H] > -1.3.  This can best be explained by a period of slow star formation at early Galactic times (when the overall iron content was low) followed by a sudden explosion of star formation that produced most of the stars we observe today in the disk.  Figure 5 in the paper.

If we put this all together, we can conclude that early in the Milky Way’s life, when stars were only just being formed, the overall iron content was low.  When the most massive stars began to die off, the Galactic content of both iron and the alpha elements began to increase.  Eventually, when lower-mass stars caught up and began to die as well, their iron output diluted the alpha content of the interstellar medium and thus reduced the alpha to iron ratio of our Galaxy.

This is exactly what today’s authors see when they plot their stars on the Tinsley-Wallerstein diagram (Figure 2). Their stars fall into two major populations: the high-alpha, low-iron population and the low-alpha, high-iron population, as expected.  However, what’s particularly interesting is the behavior of their stars at the intersection (also called the ‘knee’) of the two populations.  The ‘knee’ of their diagram is not as simple as expected.  The authors observe a strong negative slope at [Fe/H] < -1.3, followed by a peculiar hump around [Fe/H] ~ -1.3.

A Boost of Star Formation when the Galaxy was ~1 Billion Years Old

What does the behavior of their stars in the Tinsley-Wallerstein diagram tell us?  The strong negative slope where [Fe/H] < -1.3 suggests that Galactic star formation rate was slow during this time period, with the alpha-elements initially produced by dying massive stars being steadily offset by the delayed enhancement of iron from dying low-mass stars.  At [Fe/H] ~ -1.3, there is a sudden enhancement in alpha-elements in the Galaxy relative to iron.  Because the alpha-elements trace massive star death, this means that the Galaxy experienced a sudden burst of star formation at this epoch.  Interestingly, the authors find that the stars created during and after this starburst (i.e. [Fe/H] >= 1.3) have less chaotic kinematics than stars created before it (i.e. [Fe/H] < 1.3).

Figure 3: Galactic evolutionary models of the alpha-to-iron abundance ([alpha/Fe]) of Milky Way stars as a function of their iron content ([Fe/H]). Note how varying the overall Galactic star formation efficiency (SFE) in the model changes its overall shape in the diagram.  The best fit model to the data (bold red line above) suggests that the Milky Way underwent a period of slow star formation followed by a sudden burst of star formation that created most of the disk stars we observe today.  Figure 6 in the paper.

If we put this all together, we can infer that our Milky Way was lethargic at early times, forming stars at a slow rate.  Suddenly, once the Galactic iron content reached [Fe/H] ~ -1.3, the Milky Way perked up, forming stars at a rapid rate.  Stars born beyond this burst of star formation display calmer kinematics that are expected of stars in the disk.  We can thus deduce that the bulk of the disk stars we observe today were created during and beyond this star burst event.  The authors denote these two phases (pre- and post-starburst) as the ‘simmering’ and ‘boiling’ phases of the Milky Way, and they estimate that this transition likely occurred approximately 13 Gyrs ago.  The authors confirm this by comparing their data to Galactic chemical evolution models (Figure 3).

What happened at [Fe/H] ~ -1.3 to trigger this intense star formation in our Galaxy and thus the birth of the Galactic disk?  No one knows for sure.  It is possible that a major merger with another proto-galaxy could have increased the Galactic gas supply and triggered the formation of new stars.  It is also possible that the Galaxy simply ingested enough gas from the intergalactic medium to finally begin forming stars efficiently.  What we do know for sure, though, is that the Milky Way has had a messy evolutionary history, and its stars continue to reveal new secrets about our Galaxy’s interesting past.

Astrobites Edited By: Graham Doskoch

Featured Image Credit: Background galaxy–Stargate series, a work belonging to MGM Television Entertainment; Pot–Shutterstock

About Catherine Manea

Catherine is a 3rd year PhD student at the University of Texas at Austin. Her research is in Galactic Archaeology, the practice of using the dynamical and chemical information of individual stars to study the evolution of our Milky Way. She is particularly interested in pushing chemical tagging, the practice of tracing stars back to their birth sites, to new limits.

Discover more from astrobites

Subscribe to get the latest posts to your email.


  1. Absolutely loved reading this!


Leave a Reply