A Sneak Peek into Early-Universe Star Formation with Boötes I

This image from the Hubble Space Telescope shows the Large Magellanic Cloud nebula N44, which is home to a large star cluster. Credit: ESA/Hubble & NASA, D. Gouliermis; CC BY 4.0

Title: Probing the IMF in the Early Universe — Direct Measurements in the Boötes I UFD with JWST/NIRCam
Authors: Keyi Ding et al.
First Author’s Institution: University of Maryland
Status: Accepted ApJ
The Stellar Initial Mass Function

Figure 1: Several different formulations for the IMF. Astronomers have found that the Milky Way IMF seems to follow either a broken power law (Kroupa), or a log-normal (Chabrier) distribution. Credit:JohannesBuchner; CC BY-SA 4.0

A fundamental concept in astronomy is the stellar initial mass function (IMF). The IMF describes the number of stars of each mass that form from a single birth cloud. If the IMF has a negative slope, it means most stars are low in mass; if it has a positive slope, it means most stars are high in mass. Figure 1 shows some examples of commonly used IMFs, all of which have a negative slope, since we see far more low-mass stars than high-mass stars in the Milky Way. You may notice that several models deviate from a straight line at about 0.5 solar mass. This point, known as the turnover or critical mass, is crucial in determining the exact shape of the IMF.

For a relatively simple concept, the IMF is shaped by incredibly complicated underlying physical mechanisms. Things like turbulence, magnetic fields, and chemical enrichment all play a role in shaping the observed IMF. Additionally, the IMF is a pretty fundamental quantity. An enormous amount of astronomy research relies on assumptions made about the IMF. For example, since most stars are low in mass, and low-mass stars are dimmer, astronomers use the IMF to convert the amount of light in a galaxy to the number of stars; if there are more or fewer low-mass stars than we expect, our measurements will be wrong.

A big question surrounding the IMF is whether or not it’s universal. In the Milky Way, astronomers have been able to measure the IMF accurately, and they have found that it seems to be the same regardless of which bunch of stars we use to measure it.¹ However, we know that galaxies in the early universe were very different compared to today. Things get tricky when you acknowledge that most measurements made of early-universe galaxies rely on modeling tools that are entirely reliant on assuming an IMF. This gnarly little detail makes measuring the IMF in the early universe especially valuable to astronomers.

What Makes Ultra-Faint Dwarfs So Special?

Today’s authors attempt to measure the early-universe IMF using a local relic, an ultra-faint dwarf galaxy (UFD). You might describe UFDs as “incredibly funky little galaxies.” They’re much less massive than the Milky Way, with about 10,000 times less stellar mass. What stars they do have tend to be very old and metal poor. The nature of UFDs has led many astronomers to think of them as fossils: relatively untouched galaxies formed in the early universe. Since we think UFDs are fossils of earlier galaxies, measuring the IMF in a UFD tells us whether the IMF was the same in the early universe as it is today. The authors focus on Boötes I, a relatively luminous UFD orbiting the Milky Way. Figure 2 shows Boötes I as seen by the Sloan Digital Sky Survey.

Figure 2: Boötes I as seen by the Sloan Digital Sky Survey. Because they’re so diffuse, UFDs look less like galaxies and more like groups of stars. Credit: Vasily Belokurov – SDSS-II Collaboration

The IMF in Boötes I

Measuring the IMF can get tricky — it’s typically pretty difficult to measure the mass of each individual star in a galaxy. Thankfully, Boötes I is close enough that we can do exactly that! Using JWST’s NIRCam instrument, today’s authors obtain imaging of Boötes I that is sensitive enough to extract roughly 10,000 stars belonging to the galaxy.

To measure the IMF from the observed population of stars, the authors use a modified version of Starwave, a Bayesian inference tool. In short, the tool takes in some assumptions about the population of stars in the galaxy, then generates many potential color–magnitude diagrams for various parameter selections. You can then assess how well each simulated color–magnitude diagram fits the observed data, thus determining likely parameters for the stellar population. The authors apply their tool for three different IMF models, testing how close the IMF in Boötes I is to that of the Milky Way. This allows them to determine how well the IMF compares to that of the Milky Way, which is typically thought of as a broken power law or log-normal distribution. If the IMF in Boötes I aligns with the Milky Way’s IMF, we’ll have a solid piece of evidence for a truly invariant IMF across cosmic time, allowing astronomers to rest easy knowing our modeling efforts haven’t been bunk this whole time.

So… What Did We Learn?

The authors find that a single power law can be ruled out to a good degree of confidence. This is good, as Milky Way studies show undeniable evidence of a turnover in the distribution. The broken power law and log-normal models both fit the observed data relatively well, aligning well with Milky Way–derived IMFs. All in all, they find solid evidence for an invariant IMF in the early universe.

However, they are unable to say with absolute certainty that the IMF is invariant. Given that their data are nearly perfect (that is, it’s pretty much impossible to get better data for Boötes I), the authors emphasize that a larger sample of UFD IMFs is needed to truly rule out an invariant IMF, but we’re certainly taking steps in the right direction! As the age-old saying goes, “more data are needed!”

Original astrobite edited by Maggie Verrico.

Source: American Astronomical Society/AAS-Nova

---

Editor’s Note: The Milky Way IMF may vary from star cluster to star cluster; as described in this AAS Nova highlight from 8 July 2026, recent research using data from the Gaia spacecraft has found evidence for IMF variation in our galaxy. ↩︎

---

About the author, Drew Lapeer:

Drew is a first-year PhD student at the University of Massachusetts Amherst. They are broadly interested in the evolution of galaxies, with a focus on the impact of cosmic feedback on the galactic ecosystem. In their free time, they enjoy reading, rock climbing, hiking, and baking!

---

Record-breaking Faint Satellite Galaxy of the Milky Way Discovered

An international team led by researchers from Tohoku University has found an extremely faint dwarf satellite galaxy of the Milky Way. The team's discovery is part of the ongoing Subaru Strategic Survey using Hyper Suprime-Cam. The satellite, named Virgo I, lies in the direction of the constellation Virgo. At the absolute magnitude of -0.8 in the optical waveband (Note), it may well be the faintest satellite galaxy yet found. Its discovery suggests the presence of a large number of yet-undetected dwarf satellites in the halo of the Milky Way and provides important insights into galaxy formation through hierarchical assembly of dark matter.

Figure 1: The position of Virgo I in the constellation of Virgo (left). The right panel shows a density map of Virgo I's member stars in a 0.1 deg x 0.1 deg area, based on the stars located inside the green zone in the color-magnitude diagram of Virgo I shown in Figure 4. The color range from blue -> white -> yellow -> red indicates increasing density. (Credit: Tohoku University/National Astronomical Observation of Japan)

Movie: An animation showing locations of Milky Way Galaxy's satellite galaxies, featuring the newly discovered Virgo I. An image captured from the animation is shown here. The computer graphics was created using Mitaka, a four-dimensional digital universe viewer. In the image from the Subaru Telescope, Green circles denote the member candidate stars that might belong to Virgo I. (Credit: NAOJ)

Currently, some 50 satellite galaxies to the Milky Way have been identified. About 40 of them are faint and diffuse and belong to the category of so-called "dwarf spheroidal galaxies" (Figure 2). Many recently discovered dwarf galaxies, especially those seen in systematic photometric surveys such as the Sloan Digital Sky Survey (SDSS) and the Dark Energy Survey (DES) are very faint with absolute luminosity in the optical waveband below -8 magnitude. These are so-called "ultra-faint dwarf galaxies". However, previous searches made use of telescopes with a diameter of 2.5 to 4 meters, so only satellites relatively close to the Sun or those with higher magnitudes were identified. Those that are more distant or faint ones in the halo of the Milky Way are yet to be detected (Figure 3).

Figure 2: Satellite galaxies associated with the Milky Way Galaxy. Squares are Large and Small Magellanic Clouds and circles are dwarf spheroidal galaxies.

Figure 3: False-color dwarf galaxy images taken with Subaru Telescope. Left: Leo II (V-band absolute magnitude MV = -11.9 mag). Middle: Boötes I (MV = -6.3 mag), where both images are taken with Suprime Cam. Right: HSC image of Virgo I (MV = -0.8 mag). Ultra-faint dwarf galaxies (Boötes I and Virgo I) are hard to detect from these images.

The combination of the large aperture of 8.2-meter Subaru Telescope and the large field-of-view Hyper Suprime-Cam (HSC) instrument is very powerful in this study. It enables an efficient search for very faint dwarf satellites over large areas of the sky. The first step in searching out a new dwarf galaxy is to identify an over density of stars in the sky, using photometric data. Next is to assess that the over dense appearance is not due to line-of-sight or accidental juxtapositions of unrelated dense fields, but is really a stellar system. The standard method for doing this is to look for a characteristic distribution of stars in the color-magnitude diagram (comparable to the Hertzsprung-Russell diagram (middle and left panels of Figure 4)). Stars in a general field shows no particular patterns in this diagram (right panel of Figure 4).

Figure 4: Stars in the color-magnitude diagram. Old stellar populations show a characteristic distribution along the curve seen in the diagram. From left to right: Boötes I, Segue I, Virgo I, and a general field outside Virgo I. The spatial distribution of the stars, which are located inside the green band for Virgo I, is shown in the right panel of Figure 1. Note that stars in a general field outside Virgo I (right panel) show no characteristic feature.

Finding Virgo I

Daisuke Homma, a graduate student at Tohoku University, found Virgo I under the guidance of his advisor, Masashi Chiba, and their international collaborators. "We have carefully examined the early data of the Subaru Strategic Survey with HSC and found an apparent over density of stars in Virgo with very high statistical significance, showing a characteristic pattern of an ancient stellar system in the color-magnitude diagram," he said. "Surprisingly, this is one of the faintest satellites, with absolute magnitude of -0.8 in the optical waveband. This is indeed a galaxy, because it is spatially extended with a radius of 124 light years – systematically larger than a globular cluster with comparable luminosity."

The faintest dwarf satellites identified so far was Segue I, discovered by SDSS (-1.5 mag) and Cetus II in DES (0.0 mag). Cetus II is yet to be confirmed, as it is too compact as a galaxy. Virgo I may ultimately turn out to be the faintest one ever discovered. It lies at a distance of 280,000 light years from the Sun, and such a remote galaxy with faint brightness has not been identified in previous surveys. It is beyond the reach of SDSS, which has previously surveyed the same area in the direction of the constellation Virgo (Figure 5).

Figure 5: The relation between the distance from the Sun and absolute magnitude in optical waveband for Milky Way satellites discovered so far. Virgo I is extremely faint and distant from the Sun and is beyond the reach of SDSS. Except for Virgo I, DES mostly discovers those outside SDSS's limit.

According to Chiba, the leader of this search project, the discovery has profound implications. "This discovery implies hundreds of faint dwarf satellites waiting to be discovered in the halo of the Milky Way," he said. "How many satellites are indeed there and what properties they have, will give us an important clue of understanding how the Milky Way formed and how dark matter contributed to it."

Using HSC to Trace Galaxy Formation 

Formation of galaxies like the Milky Way is thought to proceed through the hierarchical assembly of dark matter, forming dark halos, and through the subsequent infall of gas and star formation affected by gravity. Standard models of galaxy formation in the context of the so-called cold dark matter (CDM) theory predict the presence of hundreds of small dark halos orbiting in a Milky Way-sized dark halo and a comparable number of luminous satellite companions. However, only tens of satellites have ever been identified. This falls well short of a theoretical predicted number, which is part of the so-called "missing satellite problem". Astronomers may need to consider other types of dark matter than CDM or to invoke baryonic physics suppressing galaxy formation to explain the shortfall in the number of satellites. Another possibility is that they have seen only a fraction of all the satellites associated with the Milky Way due to various observational biases. The issue remains unsolved.

One of the motivations for the Subaru Strategic Survey using HSC is to do increase observations in the search for Milky Way satellites. The early data from this survey is what led to the discovery of Virgo I. This program will continue to explore much wider areas of the sky and is expected to find more satellites like Virgo I. These tiny companions to be discovered in the near future may tell us much more about history of the Milky Way's formation.

The team's finding is published in the Astrophysical Journal in its November 14, 2016 on-line version and November 20, 2016 in the printed version, Volume 832, Number 1. The title of the paper is "A New Milky Way Satellite Discovered in the Subaru/Hyper Suprime-Cam Survey" by D. Homma et al., which is also available in preprint from arXiv:1609.04346v2. This work is supported by a JSPS Grant-in-Aid for Scientific Research (B) (JP 25287062) and a MEXT Grant-in-Aid for Scientific Research on Innovative Areas (JP15H05889, JP16H01086).

Research Team: 

Daisuke Homma (Tohoku University, Japan), Masashi Chiba (Tohoku University, Japan), Sakurako Okamoto (Shanghai Astronomical Observatory, China), Yutaka Komiyama (National Astronomical Observatory of Japan (NAOJ), Japan), Masayuki Tanaka (NAOJ, Japan), Mikito Tanaka (Tohoku University, Japan), Miho N. Ishigaki (Kavli Institute for the Physics and Mathematics of the Universe (Kavli IPMU, WPI), University of Tokyo, Japan), Masayuki Akiyama (Tohoku University, Japan), Nobuo Arimoto (Subaru Telescope, NAOJ, USA), Jose A, Garmilla (Princeton University, USA), Robert H. Lupton (Princeton University, USA), Michael A. Strauss (Princeton University, USA), Hisanori Furusawa (NAOJ, Japan), Satoshi Miyazaki (NAOJ, Japan), Hitoshi Murayama (Kavli IPMU, WPI, University of Tokyo, Japan), Atsushi J. Nishizawa (Nagoya University, Japan), Masahiro Takada (Kavli IPMU, WPI, University of Tokyo, Japan), Tomonori Usuda (NAOJ, Japan), Shiang-Yu Wang (Institute of Astronomy and Astrophysics, Academia Sinica, Taiwan)

Note: 

For Comparison, the abolute magnitude in visible waveband for M31 (Andromeda Galaxy), Large Magellanic Cloud, and Small Magellanic Cloud are -21.77, -18.35 and -17.02, respectively, according to NASA/IPAC Extragalactic Database (http://ned.ipac.caltech.edu). 

Related Articles:

• Leo II: An Old Dwarf Galaxy with Juvenescent Heart

Source: Subaru Telescope