Astrobites RSS

NGC 1277 as seen by Hubble

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: RUBIES: Evolved Stellar Populations with Extended Formation Histories at z ∼ 7 − 8 in Candidate Massive Galaxies Identified with JWST/NIRSpec
Authors: Bingjie Wang et al.
First Author’s Institution: Pennsylvania State University
Status: Published in ApJL

When JWST was launched, scientists were hopeful that the telescope would be able to answer hundreds of questions about the early universe. Luckily, it has been able to answer many of these questions. And even more luckily, observations of the early universe with JWST have been able to generate even more questions we don’t know the answers to!

One new conundrum scientists are running into is that it looks like some galaxies in the early universe have what we call evolved stellar populations. This means that even though these galaxies are only a couple hundred million years old (super young, right?), they already have ~100-million-year-old stars, meaning that these galaxies had to make a large amount of stars in only a short span of time. We’re not used to seeing this many stars form this fast — in fact, we expect that star formation should be slow this early in the universe. So how do we reconcile this?

The authors of today’s article look in more depth at the spectra of three such galaxies to figure out if these galaxies have really been forming stars for millions of years already. These three galaxies are located at redshifts of z = 6.7–8.4, which places them just 600 million years after the Big Bang.

plot of spectra of three galaxies

Figure 1: Balmer break at rest wavelength 3645 Angstroms (Å) observed in the galaxy spectra. The strength of the break (the relative flux difference before and after the break) is 1.71. [Wang et al. 2024]

Balmer Breaks Create Questions

First, the authors looked to confirm that evolved stellar populations do in fact exist in these galaxies. One foolproof way to do this is to look for a Balmer break in the galaxies’ spectra. The Balmer break comes from the atmospheres of older stars, indicating that the galaxy contains a population of older stars and has had little star formation for at least the past 100 million years. Excitingly, the authors confirm that there is in fact a Balmer break in the spectra of these three galaxies, as shown in Figure 1. Unfortunately, the Balmer break only tells you that there are older stars, not necessarily how old they are or how many of them there are.

Stellar Populations or Active Galactic Nuclei?

The authors then turn to understanding the finer details of the stellar population, specifically its formation history. The authors found evidence of broad lines in their emission-line spectra, signaling that there might be active galactic nucleus activity in at least two of these galaxies. Broad lines come from Doppler broadening, where photons emitted by an active galactic nucleus have a wide distribution of velocities, resulting in a broad emission line. However, the authors aren’t entirely ready to say conclusively that there are active galactic nuclei in the spectra.

In the local, low-redshift universe, we’re pretty good at understanding how these broad lines relate to active galactic nuclei because we have hundreds of spectra to test. But at high redshift, we don’t have a lot of data. We aren’t sure why supermassive black holes act the way they do early in the universe, and we don’t want to assume they act the same as the ones we observe in the local universe. So researchers have to rely on testing theoretical models against high-redshift galaxy observations.

To test how much of the light in these galaxies is coming from active galactic nuclei or evolved stellar populations, the authors test three different models: 1) The light is mostly from stars. 2) It’s from a mixture of stars and active galactic nuclei. 3) It’s mostly from active galactic nuclei. The authors create the model spectra using Prospector and compare the modeled spectra to the observed spectra to see how well they match.

Model Comparisons

The results of the modeling are shown in Figure 2. The most important thing the authors are testing here is what these different levels of contribution from the active galactic nucleus mean for the formation history of the stellar population. For instance, the model with the maximal stellar contribution means that the stellar population had to form earlier, because there need to be enough stars to create all the light seen. The minimal stellar model can form stars a bit later because the active galactic nucleus contributes more to the light seen, so you don’t need as many stars producing a lot of light. However, there is no way to verify which model is more accurate at this stage, so the authors turn to theorizing how each model could fit into the evolutionary stages of later galaxies.

comparison of modeled spectra to observed spectra

Figure 2: Left: Each model compared to the spectrum for one of the observed galaxies. Right: Predicted star formation rate as a function of time since Big Bang for each model. As seen, models with higher active galactic nucleus contribution predict later star formation. [Wang et al. 2024]

Progenitors of Massive Quiescent Galaxies?

To find out how their models fit into the evolution of galaxies, the authors then compare the different suggested star formation histories of the medium and maximal stellar contribution models with the projected histories of massive quiescent galaxies that exist at z ~ 3-5. These massive quiescent galaxies lack an explanation for how they got so massive yet have such slow star formation now. As can be seen in Figure 3, the star formation rate traced across the age of the universe for the massive galaxies (black, blue, green, and yellow distributions) matches much better in the maximal stellar contribution model. The authors posit that these galaxies could create a lot of stellar mass in a short time, then have their star formation slowed greatly to become these massive quiescent galaxies. With this explanation, the authors note that these galaxies would be a relatively small fraction of all the galaxies in the early universe.

comparison of star formation histories of massive quiescent galaxies and the RUBIES sample galaxies

Figure 3: Comparison of medium (purple) and maximal (red) stellar models with the projected formation histories of a sample of massive quiescent galaxies at z ~ 3-5. [Wang et al. 2024]

Of course, there is still the complication that there’s convincing evidence that there are active galactic nuclei in this sample of galaxies. Also, having this many stars form this early is still an uncomfortable thought for many astronomers. The jury is still out on these galaxies and how much star formation they had early on, but future studies looking into the far infrared can hopefully distinguish between light from stars and active galactic nuclei and lend more insight into what makes up these galaxies. And hopefully with JWST we can find some more answers while exploring even more questions.

Original astrobite edited by Amaya Sinha.

About the author, Caroline von Raesfeld:

I’m a second-year PhD student at Northwestern University. My research explores how we can better understand high-redshift galaxy spectra using observations and modeling. In my free time, I love to read, write, and learn about history.

Illustration of gravitational waves emanating from a pair of black holes approaching a merger

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: The Metallicity Dependence and Evolutionary Times of Merging Binary Black Holes: Combined Constraints from Individual Gravitational-Wave Detections and the Stochastic Background
Authors: Kevin Turbang et al.
First Author’s Institution: Free University of Brussels and University of Antwerp
Status: Published in ApJ

The Gravitational Wave Ocean

Before the final few seconds of the gravitational wave signal from a merger washes up on LIGO–Virgo–KAGRA detectors on Earth, the black holes or neutron stars that caused the powerful waves were actually orbiting around each other for much, much longer — possibly for hundreds of millions of years. The gravitational waves emitted prior to the pair’s final hurrah are too quiet to be picked up via ground-based interferometry with current detectors, and they instead wash together in a background hum like waves out at sea.

The length of time between two black holes forming and their eventual merging — otherwise known as delay time — is a useful signature for telling us where the black holes came from and the lives of the stars from which they were born. Fortunately, we don’t have to sit around for a billion years watching the cosmos play at spinning black holes around each other (though it would be fun to help if you were an interstellar cat). The distribution of time delays can be calculated using the number of merging black holes per given time at a certain redshift (called the merger rate), something that we can measure with current detections of merging black holes with ground-based detectors; as long as we know the rate of star formation and how this varies with the metallicity of the forming stars.

Instead of waiting for one binary system to merge somewhere in the cosmos, we can understand the time-delay distribution by detecting many black holes that have formed and merged at different points in time. The different information that multiple detections give us allows us to paint a picture of this time-delay distribution without waiting the lifetime of a black hole binary. There’s one crux, however: with the limited sensitivity of current ground-based gravitational wave detectors, we only really resolve merging black holes out to a redshift of around 1, about 8 billion years ago, when the universe had already started slowing its pace of forming stars. The black holes merging closer than this distance make loud enough gravitational waves that they stand out from the noise, making them resolvable, while beyond this distance gravitational waves that are quieter and harder to detect are lost in detector noise, and hence are unresolved. This means we can only understand the merger rate at small redshifts with resolved detections, while minimising the information we can get about the time-delay distribution.

Heading for Choppy Waters

Fortunately, today’s authors have a trick up their sleeve. By combining the information from the current detections loud enough to individually identify and the information from the background hum of merging black holes too weak to detect, the so-called stochastic background, they can help bypass this problem. The stochastic background is made up of unresolved signals from many more inspiralling black holes — a random sea of gravitational waves — and provides vastly more information about the black hole population at higher redshifts. The authors show that by using information about the stochastic background, we can extract information about the shape of the time-delay distribution (i.e., whether most black holes take a long time to merge or whether most merge relatively quickly after forming), the minimum time a black hole binary takes to merge, and the maximum metallicity environment that can create merging black holes. However, this background has so far been too quiet to detect. Today’s authors show that predicting these parameters using only resolved signals doesn’t tell us much about the time-delay distribution, but things change significantly when including different models of the stochastic background.

Even though we haven’t detected this stochastic background, we can still make predictions about the time-delay distribution because we know that the stochastic signal needs to be quieter than we can currently hear in the detectors. Therefore, there should be fewer merging binaries making up the background, making it not loud enough to meet the detection threshold. This means we have an estimate of the energy per unit frequency of the stochastic background, without having even detected it.

The authors’ model assumes the time-delay probability distribution is a power law with slope κ. A steeply positively sloped time-delay distribution would mean that there is very little probability of any objects merging quickly after they have formed and it’s very likely that black holes take a long time to merge. A steeply negative slope would mean that it’s very likely that black holes merge quickly after forming. Using the resolved detections and this stochastic background estimate, the authors find κ is strongly constrained to be negative. Whether the slope would be positive or negative is determined by the low-redshift black hole population — a positive slope would mean that since every greater time delay is more probable than the smaller time delay before, every day there would be more and more black holes merging. Fortunately, we don’t need to explain an ever-increasing number of merging black holes, as measurements at low redshifts show that there are fewer merging black holes closer to us as far as our resolved detections reach. Meanwhile, they can’t place any constraints on the minimum delay time or the maximum sensitivity. This is a consequence of the estimated stochastic background being much weaker than what we can detect with the current observations.

Our Next Great Voyages

With the next big set of improvements to the LIGO detectors currently being implemented comes two possibilities: either this stochastic background is detected, or it remains too quiet to be picked up. The authors show that either of these cases will be more informative about the time-delay distribution than what the current observations can say. By simulating a stochastic background for each of these future cases and adding it to the current resolved detections, they can say more about the astrophysics of the time delay and metallicities of the black hole mergers in the universe.

For the case of a future detection of the stochastic background, the slope of the time-delay distribution is likely to be highly negative, while the metallicity parameter and the minimum time delay are both constrained to smaller values, as seen in Figure 1. This means that a universe with a loud stochastic background has shorter delay times with lots of black holes merging in the early universe and at low metallicities.

corner plot of the three astrophysical parameters of the time-delay distribution

Figure 1: A corner plot of the three astrophysical parameters of the time-delay distribution (from left to right: the time delay power-law slope κ, log of the maximum metallicity in solar metallicities, and log of the minimum time delay in billions of years) measured with current resolved gravitational wave observations, and a stochastic background detected with future LIGO–Virgo–KAGRA detectors. The three panels at the top of each column show the 1D probability distributions of the three parameters, where the dotted line is the current constraints on these parameters. The three panels in the bottom left represent the joint probability distributions between pairs of parameters, showing the correlations between the parameters. [Turbang et al. 2024]

On the other hand, a non-detection at higher sensitivities means that the time-delay slope is more likely to be weakly negative, with larger values of maximum metallicity and minimum time delay. In this case, there are fewer probable values of the time-delay power-law slope (we can understand this as a narrower probability distribution of κ in Figure 2, meaning a smaller uncertainty on this value), while there is a larger uncertainty on the maximum metallicity and minimum time-delay parameters. Essentially, this result is because a shallower slope of the time delay (a more even distribution of delay times across cosmic time) is the only way to get a stochastic background that is too quiet to be detected with future sensitivities. A quiet background means that there are fewer farther away black holes merging to contribute to the background, and so there can’t have been many more black holes merging at the start of the universe as there are now.

corner plot of the three astrophysical parameters of the time-delay distribution in the case of a non-detected stochastic background

Figure 2: The same as Figure 1 but considering a non-detected stochastic background at future detector sensitivity. In this case the value of κ is still negative but closer to zero, meaning a less steep distribution of black hole time delays. There is a wider range of possible maximum metallicities and minimum time delays. [Turbang et al. 2024]

So how do we tell how long it takes for two black holes in the universe to merge? No matter if we see it or not, the stochastic background will give us more information when the next sets of gravitational wave observations are gathered. With the currently ongoing fourth observing run of the LIGO–Virgo–KAGRA detectors, hopefully it won’t be long before we can sail the stochastic seas, so get to your seats for the great black hole race of the universe!

Original astrobite edited by Evan Lewis.

About the author, Storm Colloms:

Storm is a postgraduate researcher at the University of Glasgow, Scotland. They work on understanding populations of binary black holes and neutron stars from the gravitational wave signals emitted when they merge, and what that tells us about the lives and deaths of massive stars. Outwith astrophysics they spend their time taking digital and film photos and making fun doodles of their research.

emission shell surrounding the X-ray binary CI Camelopardalis

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Discovery of an [O III] Emission Shell Around the X-ray Binary CI Cam
Authors: Robert Fesen et al.
First Author’s Institution: Dartmouth College
Status: Published in RNAAS

From Be Star to X-ray Binary

Among the spectral classes of normal stars that we know and love, there is one particularly eccentric type called the Be stars. Unfortunately, no, the “e” there does not stand for eccentric. Much like the B-type stars, Be stars are incredibly hot and massive objects. However, what sets them apart is the notable presence of excess emission features in their spectral profiles, in contrast with the absorption lines typically seen in the spectra of regular B stars. Thus, the “e” stands for emission.

Our star today, CI Camelopardalis (CI Cam), was initially cataloged as a Be star, but when it was later found that the spectrum was particularly rich in forbidden emission lines, it was reclassified as a B[e]star. The brackets designate the forbidden lines that come from an improbable transition between metastable energy levels in atoms or ions, which can only be observed in low-density astrophysical environments. As an example of this notation, the doubly ionized oxygen forbidden line is written as [O III].

Artist’s illustration of an X-ray binary system

Figure 1: Artist’s illustration of an X-ray binary system that consists of a normal star and a compact object. The gravity of the compact object pulls matter off the normal star creating an accretion disk, shown in blue. [European Space Agency, NASA and Felix Mirabel (the French Atomic Energy Commission & the Institute for Astronomy and Space Physics/Conicet of Argentina)]

CI Cam underwent a dramatic X-ray outburst in 1998, which not only revealed the presence of an unseen compact companion, but also firmly established the system as an X-ray binary. An X-ray binary is a fascinating duo of a normal star and a compact object, which could be a white dwarf, a neutron star, or a black hole (see Figure 1). The two objects orbit closely around a common center of mass, emitting intense X-ray radiation. This high-energy emission is produced by the accretion process in which material from the star gets pulled onto the compact object by its immense gravitational field. As this material spirals inward, it heats up to temperatures in the millions of degrees, causing it to radiate very brightly in the X-ray part of the electromagnetic spectrum. Interestingly, the exact nature of CI Cam’s compact companion remains unclear to this day.

The Discovery

In a serendipitous finding, today’s authors uncovered a previously unknown shell of emitting gas surrounding the CI Cam system. This shell, spanning about 8 arcminutes by 12 arcminutes on the sky, is brightest in the forbidden [O III] emission line at 5007 Angstroms (within the optical or visible light regime). The discovery unfolded from the optical data of the large supernova remnant G150.3+4.5 (G150) that was observed as part of a limited wide-field imaging survey of optical emission associated with galactic remnants. This survey was led by Robert Fesen from Dartmouth College (the first author of today’s article). The emission shell was spotted adjacent to the remnant’s filaments along its northwestern edge (see Figure 2).

Zoomed out and detailed views of the newly discovered emission shell

Figure 2: Composite image showing the emission shell around CI Cam and the western portion of the galactic supernova remnant G150.3+4.5. Blue represents the [O III] emission and red represents the Hα emission. [Adapted from Fesen et al. 2024]

The presence of [O III] emission indicates the existence of extremely hot ionizing sources that can provide the high-energy photons needed to doubly ionize oxygen atoms. Typically, strong [O III] emissions are seen to have filamentary or arc-like structures tracing the interfaces between ionized and neutral gas regions. These emissions can also be enhanced by the presence of shocks, in addition to photoionization by hot sources. An emission shell featuring strong [O III] emission around CI Cam with faint Hα emissions (see Figure 3) implies the presence of high-velocity shocks of about 90 to 150 km/s, potentially driven by CI Cam’s outflows into the surrounding interstellar medium.

Two grayscale views of the shell in O III and H alpha filters

Figure 3: Monochromatic images of the emission shell around CI Cam as seen in [O III] (top) and in Hα (bottom). [Adapted from Fesen et al. 2024]

By adopting CI Cam’s newly refined Gaia distance of 4.1 kiloparsecs (about 13,000 light-years), the physical size of the shell, assuming it is indeed the result of CI Cam’s outflows, spans approximately 9.5 x 14.3 parsecs (31 x 47 light-years). To put this into perspective, the longer side of the nebula is roughly equivalent to a hundred billion times the distance between New York City and Los Angeles via the I-40. This dimension is comparable to the dimensions of nebulae observed around Wolf–Rayet stars, which are known to generate such large structures through their powerful stellar winds and outflows.

Due to the similarity in [O III] brightness between the shell and the supernova remnant G150’s filaments, their close proximity raised the question of a potential physical association between the two objects. However, the authors found no direct evidence in the morphology linking the shell to the remnant on the imaging data. They also took into account the available distance estimates for G150 derived from radio and γ-ray observations, which suggested a much closer distance compared to CI Cam. The authors argued that if G150 were actually at CI Cam’s distance, it would have to be a much larger and more ancient supernova remnant, which seems highly unlikely given its bright optical filaments that suggest a relatively young age. The authors then concluded that the apparent proximity of the shell to G150 is likely just a chance alignment with Cl Cam.

While X-ray binaries are primarily studied in the X-ray, optical observations offer a complementary perspective, helping us to understand the details of the binary’s interactions with its surrounding medium. That the size of the emission shell around CI Cam is comparable to those created by powerful Wolf–Rayet stars provides evidence for the strong stellar winds and the transient outburst that CI Cam has unleashed over an extended period. This, in turn, promises to shed light on the mass-loss mechanisms and accretion processes in this class of objects. This discovery not only highlights the significance of wide-field imaging surveys in unveiling previously unknown structures around peculiar objects, but also opens up a new observational window for advancing our understanding of how such systems enrich and shape their cosmic neighborhoods.

Original astrobite edited by Ivey Davis.

About the author, Janette Suherli:

Janette is a PhD student at University of Manitoba in Winnipeg, Canada. Her research focuses on the utilization of integral field spectroscopy for the studies of supernova remnants and their compact objects in the optical. She is also the current chair of Graduate Student Committee for the Canadian Astronomical Society (CASCA). She grew up in Indonesia where it is summer all year round! Before pursuing her PhD in astrophysics, Janette worked as a data analyst for a big Indonesian tech company, combating credit card fraud.

photograph of the Sun with plasma streaming from its surface

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: The Death of Vulcan: NEID Reveals That the Planet Candidate Orbiting HD 26965 Is Stellar Activity
Authors: Abigail Burrows et al.
First Author’s Institution: Dartmouth College and NASA’s Jet Propulsion Laboratory
Status: Published in AJ

It’s possible you may have heard of HD 26965, otherwise known as 40 Eridani A, the stellar host for the fictional planet Vulcan (homeworld of Spock, of Star Trek fame). Back in 2018, two teams of astronomers announced the likely presence of a super-Earth — Vulcan, perhaps — orbiting HD 26965, based on radial-velocity measurements from a variety of instruments. These astronomers measured the planet’s mass to be equivalent to more than 8 Earths and the orbital period to be roughly 42 days. But before Trekkies could break out the Romulan ale, researchers noted that more work was needed to fully separate the radial-velocity signals from those of the exoplaneteer’s eternal enemy: stellar activity, which can mimic planetary signals.

The Fate of Vulcan

Since then, study after study has highlighted stellar activity as the likely primary source of Vulcan’s radial-velocity signal, raising further uncertainty about the existence of the planet. What might bring us closer to a definite answer would be to combine detailed radial-velocity analysis techniques with what we already know about the effects of solar activity on radial-velocity measurements. This knowledge can be applied to high-resolution spectra of HD 26965 taken by NEID at the Kitt Peak National Observatory. NEID represents the latest and greatest in ground-based spectrometers, ensuring clearer, more frequent data than previous studies had access to.

Important to determining the status of HD 26965 b (the standard name for our Vulcan) is the concept of phase lag. Based on what we know from studying solar activity, the observational parameters used to study stellar activity can become offset in time from the effects of stellar activity on radial-velocity measurements. This would mean that stellar activity might not have been fully corrected for in early studies of this planet. If a decaying starspot or plage were present on HD 26965, correcting for it would lead to a reduction in the radial-velocity signal.

Combining Radial Velocity Analysis Techniques

To account for this phase lag, the authors first compute what are essentially smoothed radial-velocity signals by taking the sum of every radial velocity corresponding to every spectral line, weighted by their corresponding errors. These are corrected for NEID systematic quirks (such as by converting measurements into the stellar frame from the observatory frame NEID usually works in). This produces a “template” radial velocity that the authors compare each observed spectral line to, noting differences between observed radial velocity and “template” radial velocity. This comparison allows them to view how phase lag may affect the amplitude and location of the radial-velocity signal.

The Hits Just Keep On Coming

The observed NEID radial velocities have a similar 42-day period to the planet model proposed by an earlier study, but Figure 1 shows that they are out of phase by 30–40%! Straight off the bat, it’s not looking so good for Vulcan with strike number one. In addition, comparing these radial velocities to activity indicator lines Ca Ⅱ, H, and K reveals that they share a similar period but do not seem to be correlated. Strike number two!

observed radial velocities compared to the model including a planet in the system

Figure 1: The NEID-observed radial velocities of HD 26965 (marked in purple) compared to the planet model proposed by an earlier study (Ma et al. 2018, marked in yellow). The NEID radial velocities are out of phase with the planet model by 30–40%. [Burrows et al. 2024]

Going back to phase lag, the authors calculate phase offsets for all activity indicators using a Gaussian process, finding a consistent phase lag between 4.65 and 6.67 days, more than 10% of the star’s rotation period. Shifting data with this phase lag leads to significant reduction of the radial-velocity signal strength when modeled with a 42-day period, and overall over a variety of periods, seen in the periodograms in Figure 2. Strike number three!

radial-velocity measurements and periodograms per stellar activity

Figure 2: Left: A variety of corrected radial-velocity measurements per activity metric, marked with purple circles, accompanied by original radial velocities, marked with gray squares. Right: Periodograms for corrected radial velocities, marked in solid purple, and for original radial velocities, marked in gray dashes. The red diamonds denote periods with the highest remaining power. At the supposed 42-day period, the probability of that being the actual period (called the power) drops significantly for most activity indicators. [Burrows et al. 2024]

Stop, Stop, He’s Already Dead!

These are just a small sampling of demonstrations in the article — ultimately, they all seem to point towards stellar activity probably being the source of the radial-velocity signal from HD 26965, unfortunately for Vulcan.

Fortunately for us, this article aims to act as a Swiss army knife of sorts. It accomplished several things — first, the authors showed that phase lags may be important in determining the relationship between radial velocity and whatever may have affected it. Second, a bundling of analyses makes for a stronger case compared to a single analysis technique on its own, giving more solid evidence for the stellar activity hypothesis.

Last but not least, the methodologies used in this article can be used on radial velocities from other stars, including our own Sun! In addition to testing this multi-pronged analysis on other types of stars, the authors hope to use it on our Sun during especially turbulent activity periods. By refining our understanding of how our Sun and other stars fluctuate through time, we better our exoplanet detection techniques and our chances of finding habitable worlds.

Original astrobite edited by Dee Dunne.

About the author, Diana Solano-Oropeza:

I’m a first-year astronomy PhD student at Cornell University, where I study exoplanets, stars, and habitability using Gaia data. I earned my BS in physics at Drexel University before entering the Bridge to the PhD in STEM program at Columbia University. There, I researched TESS-detected exoplanets for two years. My hobbies include practicing Muay Thai, writing fiction, and playing video games. You can check out my website at https://dianasolano-oropeza.com/.

Illustration of the exoplanet WASP-12 b.

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Doomed Worlds I: No New Evidence for Orbital Decay in a Long-Term Survey of 43 Ultra-Hot Jupiters
Authors: Elisabeth R. Adams et al.
First Author’s Institution: Planetary Science Institute
Status: Published in PSJ

Ultra-hot Jupiters: Fleeting Beauties?

Ultra-hot Jupiters are gas giants orbiting close to their stars, with orbital periods less than roughly three days. Because these planets are large and close to their stars, they produce large signals, making them promising targets for detection and characterization. But, you know what they say: all good things must come to an end. These planets are expected to experience large tidal effects from their stars, resulting in a loss of angular momentum, orbital decay, and, eventually, the star engulfing the planet.

Several lines of evidence support the picture that ultra-hot Jupiters are subject to orbital decay over long timescales. For instance, stars hosting hot Jupiters tend to be younger than the average exoplanet host star, and ultra-hot Jupiters are rarer around older host stars. A recent research article even reports a direct detection of a planetary engulfment event from the sudden, short-lived increase in brightness of a faint star. While this evidence paints a compelling picture, it is difficult to estimate how quickly we expect ultra-hot Jupiters to experience orbital decay given theoretical uncertainty in stellar tidal effects.

Keeping Time: Working Hard or Hardly Working?

Because we expect orbital decay to occur and we know of thousands of transiting exoplanets, some of which have been observed for decades, several teams have searched for orbital decay and found two promising detections: WASP-12 b and Kepler-1658 b. Searching for orbital decay relies on the detection of transit-timing variations. This is when a planet passes in front of its star along our line of sight earlier or later than expected. There are many sources of transit-timing variations in addition to orbital decay, including precession, perturbations from companion planets or stars, or acceleration of the host star toward Earth.

Let’s say we observe the transit of a planet at time t = 0 and know its period, P. We expect to observe transits at time P, 2P, 3P, etc. In the case of orbital decay, the period of the planet is getting shorter as time goes on, meaning we need to factor in an additional quadratic term encoding the rate at which the period shrinks. Then, to detect a statistically significant signal of orbital decay, we need to show that the quadratic model fits the data better than the constant-period linear model. The authors of today’s article attempt to do exactly this but with an impressive level of care and attention to detail.

This science depends upon accurate and precise measurements of transit times for each planet in the authors’ sample, most of which have been observed by several teams with various instruments and methodologies over years or decades. Moreover, each transit time must be reported in one unified timing system (click here for more info on one of the most common timing systems). Not every transit observation properly identifies its timing system or accurately converts between timing systems, meaning any historical inaccuracies complicate such studies.

Statistical Methods: Comparing Models

The authors of today’s article compile transit times for 43 ultra-hot Jupiters and take new transit data for six of those planets to extend the temporal baseline of observations. To assess whether the linear (constant period) or quadratic (changing period) model fits the data better, the authors use the Bayesian information criterion (BIC), a model selection criterion that awards a good fit but penalizes additional parameters to avoid overfitting. The authors calculate the difference in the BIC (ΔBIC) between the linear and quadratic models, with a larger ΔBIC suggesting the quadratic model is preferred.

The authors additionally perform a variety of steps to ensure the quality of the data. They perform omit-one tests, where individual transit times are removed from the analysis and flagged if they alter the ΔBIC result by more than 25%. This step is essential since one transit time recorded inaccurately or in the wrong system could result in a spurious detection of orbital decay. The authors additionally perform a “rescaling test,” where the error bars are scaled up to account for unrealistically small error bars in reported transit times.

Results

As shown in Figure 1, four planets out of the sample of 43 had a ΔBIC above the detection threshold, including WASP-12 b, which had been found previously to show orbital decay. The authors measure WASP-12 b’s period to be shrinking by 30 milliseconds per year, matching previously reported values. The planets WASP-121 b and WASP-46 b show tentative period increases, but these results are highly dependent on one or a few data points, warranting further observations. The planet TrES-1 b has prior tentative claims of its period decreasing, and the authors find a tentative period decrease of 18 milliseconds per year. However, this rate of period shrinkage suggests stellar tidal effects that would differ greatly from theoretical predictions, perhaps suggesting a cause of period decrease other than orbital decay.

plot of delta BIC for each of the 43 planets studied

Figure 1: The value of each planet’s ΔBIC shown relative to a threshold of ΔBIC = 30 (top), zoomed in results (middle), and rescaled results (bottom) with scaled up uncertainties, indicating only WASP-12 b definitively shows signs of orbital decay. [Adams et al. 2024]

Only one planet, WASP-12 b, was found to have a clear period decrease after rescaling error bars, as shown in the bottom panel of Figure 1. The authors predict that if the orbits of the other planets in the sample were decaying as rapidly as the orbit of WASP-12 b, they could have found significant detections of period decrease in roughly half the sample. There is thus no evidence that orbital decay is common among ultra-hot Jupiters, which is possibly confounding considering the other lines of evidence that suggest ultra-hot Jupiters are subject to decay. Though patience is required, as time goes on, it will be possible to search for orbital decay around more planets at higher precision, helping us ascertain the ultimate fate of close-in planets.

Original astrobite edited by Ivey Davis.

About the author, Kylee Carden:

I am a first-year PhD student at The Ohio State University, where I am an observer of planets outside the solar system. I’m involved with the transiting exoplanet survey of the upcoming Roman Space Telescope and working with high-resolution spectroscopic observations of exoplanet atmospheres. I am a huge fan of my cat Piccadilly, cycling, and visiting underappreciated tourist sites.

Images of six "little red dot" galaxies from JWST

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: A Census of Photometrically Selected Little Red Dots at 4 < z < 9 in JWST Blank Fields
Authors: Vasily Kokorev et al.
First Author’s Institution: Kapteyn Astronomical Institute
Status: Published in ApJ

Imagine you peek into a kindergarten class and, to your shock, you see that all the children are well over 6 feet tall. That is precisely how astronomers felt when data from JWST showed galaxies with massive black holes just a few hundred million years after the birth of the universe. Some of these black holes have been measured to be a million times more massive than the mass of our Sun, and astronomers are puzzled as to how they could have gained so much mass in such a short time.

The earliest galaxies likely to host these black holes show up as little red dots in the images JWST took of the early universe (as seen in the long exposure images; see the bottom image in Figure 1). They are believed to be compact (with a small radius) galaxies with a Type I active galactic nucleus and obscured (covered by dust), which accounts for their red color and why they are easily observed in the infrared. The spectrum of the galaxy has a “V” shape, a blue continuum from the unobscured part of the active galactic nucleus in the galaxy, and a red continuum from the obscured part (see Figure 1). These little red dots have appeared in several images taken by JWST, hinting that plenty of massive black holes are lurking in the early universe.

spectrum and images of a little red dot galaxy

Figure 1: The characteristic spectrum of little red dots (on the left), with the compact source contributing to the dust-free blue color in the continuum (top right) and the dust-reddened part (bottom right). [Adapted from Kokorev et al. 2024]

Where You Look Matters!

It is vital to systematically look at these little red dot galaxies to understand how many massive black holes were in the early universe. Two factors can introduce biases in the counting of these galaxies. One is the phenomenon of cosmic variance: is JWST just preferentially looking in a direction with many little red dots, or should we expect the same number even when it looks at different parts of the universe? The other is how crowded it is in the direction in which you are observing: if you have a lot of stars in the direction you are looking, they could be misclassified as little red dots or vice versa. If you happen to have plenty of massive galaxy clusters in the direction you are looking, they may create an illusion that more of these little red dots exist than their true numbers (a phenomenon known as gravitational lensing).

To minimize the errors caused by these effects in determining the number of little red dots in the early universe, the authors of today’s research article specifically look at large areas (640 arcminutes2) of the sky by combining JWST data from various programs. This would minimize the effects of cosmic variance as you can measure the numbers over a bigger area of the sky. They also look specifically at data in blank fields (defined here as areas on the sky without galaxy clusters), which helps them determine the accurate number of objects per unit volume. All these galaxies are photometrically selected (i.e., chosen only from looking at images rather than spectra), meaning there is limited spectroscopic data to help confirm what kind of objects they are. Galaxies are determined to be little red dots based on their red colors and how compact they are in the images. Using the obtained fluxes, the authors then construct spectral energy distributions to determine the redshifts (z) of the sample. Limiting the sample to z > 4 (for the early universe), the authors end up with 260 little red dot galaxies

Do Not Judge a Galaxy by Its Size (in Your JWST Image!)

On calculating the total luminosity of the little red dots and comparing it to their redshifts, the authors find a large number of bright little red dots at redshift of z = 5 (around a billion years after the beginning of the universe). The number of little red dots is almost 100 times more than the number of ultraviolet-selected quasars, which are active galactic nuclei identified using another method. They also find that computer models are unable to reproduce the high fraction of the bright galaxies they uncover at high redshifts (left side of Figure 2). The authors derive the mass of the black holes at redshifts of z = 4.5–6.5 to be around 106–108 solar masses, indicating that these black holes were already massive a few hundred million years after the Big Bang. They find deviations from the predicted number densities of massive black holes at these redshifts from galaxy simulations. This is likely because the galaxies that host more gigantic black holes are very dusty, and thus, their spectra do not have any blue continuum. They may then be missed from selections of little red dots as one of the factors it depends on is the characteristic “V”-shaped spectrum, which would need a contribution from the blue continuum (right side of Figure 2).

plots showing the number density of little red dots as a function of luminosity and black hole mass

Figure 2: Left: The number density of the little red dots as a function of luminosity at 6.5 < z < 8.5 with the predicted values from simulations indicated by the blue solid line. Right: The number density as a function of black hole mass at 4.5 < z < 6.5. The observed number density of more massive black holes is lower than the values predicted by simulations. [Adapted from Kokorev et al. 2024]

While spectroscopy is a more reliable method to identify massive black holes, many galaxies that host black holes can still be picked out using near-infrared colors and photometry, which is a much less expensive technique. The challenge lies in ensuring that the photometrically selected sources are reliable, and the authors of today’s article made great use of this technique. Follow-up spectroscopic studies of these photometrically selected samples can help us further understand the exact nature of the black holes. Such studies are already underway, and we can look forward to finally making sense of how these black holes became so massive in such a short time after the formation of the universe!

Original astrobite edited by Delaney Dunne.

About the author, Archana Aravindan:

I am a PhD candidate at the University of California, Riverside, where I study black hole activity in small galaxies. When I am not looking through some incredible telescopes, you can usually find me reading, thinking about policy, or learning a cool language!

two images of Dracula's Chivito

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Dracula’s Chivito: Discovery of a Large Edge-On Protoplanetary Disk with Pan-STARRS
Authors: Ciprian T. Berghea et al.
First Author’s Institution: US Naval Observatory
Status: Published in ApJL

Where Planets Are Born

Studying protoplanetary disks helps us understand how planets, including those in our solar system, are born. These disks are vast and flared structures, consisting of dust and gas orbiting a young star. Protoplanetary disks contain the remnants of the stellar birth process, in which a collapsing molecular cloud gives rise to a central star surrounded by a swirling disk of material. Protoplanetary disks are vital to observe as they are the birth sites of planets. The tiny dust particles come together, sticking to each other and forming larger bodies. This process, influenced by gravity, gas, and radiation, leads to the birth of planets in developing planetary systems.

Meet the Vampire Sub

This research article features a large edge-on protoplanetary disk that was stumbled upon when going through images from the Pan-STARRS research project as a part of  a study of active galactic nucleus candidates. This disk is one of the largest known disks in the sky and is oriented edge on, completely obscuring its central star. Associated with a source of infrared light in the same region of the sky, IRAS 23077+6707, the disk spans approximately 11″ in apparent size, with a very faint structure in the disk’s northern part extending out to about 17″. It is possibly the largest protoplanetary disk (by angular extent) discovered to date. The structure of this disk is reminiscent of the popular Gomez’s Hamburger, which is not associated with any star-forming region, just like the subject of this article. The similarity to a sandwich, along with the fang-like structures in the northern part of the disk as seen in the images in Figure 1, earned the IRAS 23077+6707 protoplanetary disk the name “Dracula’s Chivito” (chivito is a type of sandwich and the national dish of Uruguay, where one of the co-authors is from).

Left: Pan-STARRS image of Dracule’s Chivito with a bright blue-purple light in the center, as a structure resembling a sandwich. It has faint filaments extending out like “fangs”. Right: A model of the disk on the left, which looks smoother and the fangs are shown to be a part of larger disk envelope, which is a hazy blue light emanating from the disk in the center which has a pink hue

Figure 1: Pan-STARRS image of Dracula’s Chivito (left) and the associated radiative transfer model (right). In the northern part of the disk, very faint filaments are seen extending 17″ out from the edges of the disk. These filaments have been nicknamed “fangs.” These are interpreted as the edges of the disk envelope, as reproduced in the model image. These fangs are likely obscured by a cloud in the southern part of the disk. [Adapted from Berghea et al. 2024]

Decoding DraChi

Analysis: Images of Dracula’s Chivito (henceforth referred to as DraChi) were obtained in the grizy filters of Pan-STARRS (Figure 1). These images and data from the Galaxy Evolution Explorer (GALEX) (ultraviolet), the Two Micron All-Sky Survey (2MASS) (infrared), the Infrared Astronomical Satellite (IRAS) (infrared), and AKARI (infrared) were used for photometric analysis, i.e., flux or brightness measurements. These data were used to construct the spectral energy distribution of the disk (Figure 2), together with a radiative transfer model generated using the code HOCHUNK3D. Radiative transfer models help us understand the disk geometry and how light is scattered by the dust grains in the disk. This light, which we see along our line of sight, is plotted as a spectral energy distribution. Spectral energy distributions give information about how much energy an object gives off at different wavelengths.

A plot of the spectral energy distribution of the disk as obtained from brightness measurements from different instruments and compared to the spectrum from the model. The different colored dots representing data are mostly fit well by the model except in near and mid infrared wavelengths (from about 10 to 50 microns)

Figure 2: The spectral energy distribution of the disk, using photometric data from the image, the model, and other sources as mentioned in the article. Brightness is plotted as a function of wavelength. The data (colored dots) match the model (the solid red line shows the model without extinction, and the dashed red line shows the model with extinction) in most places except in near- and mid-infrared wavelengths, which could be due to discrepancies between instruments or possible variability in the disk’s luminosity. [Berghea et al. 2024]

Distance: It is hard to estimate DraChi’s distance because it is not associated with a known star-forming region. Accurate distances to local molecular clouds are vital to locating protoplanetary disks and comprehending planet formation processes. Therefore, using the Gaia DR3 data for nearby stars, the extinction of the disk was estimated to find the distance to the nearest interstellar clouds, and DraChi is hence estimated to be 978 light-years away.

Basic Properties: The spectral energy distribution suggests that the host star of the disk is a pre-main-sequence star of type A with a temperature of about 6500–8500K. The images of the disk and the resultant radiative transfer model constrain the disk inclination to be between 80° and 84°. The scale height of the disk is about 25–50 au at a radius of 500 au (astronomical units), where 1 au is the distance from Earth to the Sun. Using the distance and the angular extent of the disk in the sky, the disk’s radius is estimated to be 1,650 au. The radiative transfer model, based on the scattering of the light, quantifies the mass of the disk to be about 0.2 times that of the Sun.

The Fangs: The authors noticed two “fang-like” features in the northern part of the disk, and these features were also reproduced in the model of the disk. The “fangs” closely resemble the “edge” of the shadow created by the disk in the bright surrounding envelope. They could be filaments due to a possible outflow from the central part of the disk, which is characteristic of a young disk at the end of the Class I phase (~0.5 million years old). It is possible that the fangs are present in the south, but this region is likely obscured in the images from Pan-STARRS and could be perhaps seen in infrared (longer-wavelength) imaging.

Is DraChi the Only One?

The short answer is no. DraChi is certainly different from most other protoplanetary disks, given its size and large distance from any known star-forming regions. But the existence of Gomez’s Hamburger proves there are more such disks awaiting discovery. A disk as young as DraChi is vital to understanding planet formation in its earlier stages, and its large size makes for interesting future observations using more sensitive instruments.

Original astrobite edited by Kylee Carden and Jessie Thwaites.

About the author, Maria Vincent:

Maria is a PhD candidate in astronomy at the Institute for Astronomy, University of Hawai’i at Manoa. Her research focuses on adaptive optics and high-contrast imaging science and instrumentation with ground-based telescopes. Driven by a fascination with planet formation and the intricate processes shaping our solar system, she uses the Subaru Coronagraphic Extreme Adaptive Optics suite to observe and study morphological features of protoplanetary disks in near-infrared wavelengths, aiming to understand disk structure and processes governing planet formation. On the instrumentation side, she is working on designing and constructing an optical testbed to test and characterize a new deformable mirror as part of the upcoming High-order Advanced Keck Adaptive Optics upgrade. Outside of work, she enjoys blogging, mystery, historical and science fiction literature and cinemedia, photography, hiking, and travel.

Illustration of a binary system undergoing mass transfer

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Expansion of Accreting Main-Sequence Stars During Rapid Mass Transfer
Authors: Mike Y. M. Lau et al.
First Author’s Institution: Heidelberg Institute for Theoretical Studies
Status: Published in ApJL

In our universe, most stars live in binaries. These stars live together, orbiting around each other, evolving and influencing each other’s lives. For many stars, this can be a tumultuous relationship with periods of mass transfer, altering the stars’ lives from the typical single-star evolution from a main-sequence star to a red giant to a stellar remnant. Mass transfer allows stars to grow and shrink in mass through their evolution, changing their identities and speeding up or slowing down their evolution. Mass transfer can also cause stars to go supernova.

diagram of the stages of Roche lobe filling

Figure 1: An Illustration of Roche lobe filling of binary stars in three different scenarios, A, B, and C. The left column shows the gravitational potential wells for each star, while the right column outlines the stars and their Roche lobes (the teardrop shapes around them). In scenario A, the matter of both stars is within each potential well and the stars are detached. In scenario B, the left star has expanded and filled its Roche lobe, causing matter to accrete onto the right star. In scenario C, both stars have filled their Roche lobes and orbit in a contact binary. [Storm Colloms, adapted from Pringle and Wade, Interacting Binary Stars (1985, Cambridge University Press)]

Mass transfer occurs when one of the stars in a binary system expands in radius during its evolution. If this expansion pushes the outside layers of the star beyond the gravitational influence of the star, the outer layers will be pulled from the expanded donor star onto the other star in the binary. The region of gravitational influence of each star is called its Roche lobe, illustrated in Figure 1.

As common as mass transfer is in the lives of stars, there are a lot of unknowns about mass transfer processes. Astronomers make simulations of stars transferring mass to try to understand these unknowns better. Simulating a fully accurate model of these processes requires many computational resources to model the hydrodynamic processes in all dimensions, and so models always make some simplifying assumptions.

One big simplifying assumption is how much of the donor star’s mass can be packed onto the accreting star. Previous models assumed that the accreting star does not change with time when accepting mass from the donor star. However, today’s authors are actually able to investigate how the properties of the accreting star change through the mass-transfer period using a stellar evolution code called Modules for Experiments in Stellar Astrophysics (MESA).

The authors show that how the accretor’s mass and radius respond to mass transfer depends on two timescales: 1) the thermal timescale of the accretor — this is the timescale on which the accretor can radiate away all of its thermal energy, and 2) the timescale of the accretion — this is the rate of mass transfer from the donor to the accretor. The thermal timescale of the accreting star is generally much longer than the mass-transfer timescale, meaning that during mass transfer the accreting star is fed mass much more quickly than it can “swallow” it, or have it thermally cool and settle onto the star. However, the accreting star can store this accreted matter in an envelope around it, stuffing it in its cheeks much like a hamster, as long as this does not spill over the edge of the Roche lobe — hence why the authors coin the name “hamstars” for these stars. This “cheek stuffing” makes the accreting star grow in radius and cool much like a typical main-sequence star becoming a red giant, as shown in Figure 2. As the star accretes more and more and its luminosity increases, its thermal timescale decreases, and it can radiate away the energy of accreting mass and “swallow” the matter faster. The star then shrinks from its expanded radius, having grown several times in mass.

HR diagrams showing the evolution of the accreting star during mass transfe

Figure 2: Hertzsprung–Russell diagrams showing the evolution of the accreting star during mass transfer. The three panels show the temperature versus luminosity for accretors of different initial masses, and each track in the panels is a different rate of mass transfer in the simulation. Most of the stars cool at some point in evolution, expanding in radius, before the accreted matter can cool onto the star and it shrinks and heats again, resuming its usual evolution having grown several times in mass. [Lau et al. 2024]

How much the radius expands during mass transfer changes the fraction of mass lost by the donor that is added to the accretor. This is called the accretion efficiency, and this quantity is often assumed to be an arbitrary constant. With the authors’ prescription for the change in radius of the accretor, they describe a new way of calculating the accretion efficiency that considers the accretor’s expansion. This depends on the initial mass of the accreting star, and it’s found to be higher than typical assumptions for low-mass (less than 15 solar masses) accreting stars, and lower for high-mass (more than 20 solar masses) stars.

The accretion efficiency determines how much of the mass accreted actually makes it onto the accreting star, versus how much might be lost during mass transfer. While low-mass stars can pack a lot of accreting matter in their hamster cheeks, high-mass stars have a lower limit for how much matter they can chew. These higher-mass accreting stars make up observed populations of X-ray binaries and the progenitors of pairs of black holes and/or neutron stars. Compared to current models, the authors’ prescription for accretion efficiency is in better agreement with the observed population of X-ray binaries and predicts fewer merging black hole binaries that could be observed with gravitational waves, something we could investigate further with more observations.

While today’s article considers the accretor’s response to mass transfer (pinning down the similarities between stars and rodents), it does not account for the accreting star’s spin or mass loss. If a star is spinning fast, it could limit how much the accreting star can expand. Some mass can also be swept away by stellar winds, meaning that not all of the mass lost by one star is accreted onto the other. Future work is needed to incorporate these additional factors before we can look up into the cosmos for a great hamstar hunt!

Original astrobite edited by Lucie Rowland.

About the author, Storm Colloms:

Storm is a postgraduate researcher at the University of Glasgow, Scotland. They work on understanding populations of binary black holes and neutron stars from the gravitational wave signals emitted when they merge, and what that tells us about the lives and deaths of massive stars. Outwith astrophysics they spend their time taking digital and film photos, and making fun doodles of their research.

Illustration of a Neptune-like exoplanet

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Scaling K2. VII. Evidence for a High Occurrence Rate of Hot Sub-Neptunes at Intermediate Ages
Authors: Jessie L. Christiansen et al.
First Author’s Institution: NASA Exoplanet Science Institute, California Institute of Technology
Status: Published in AJ

Exoplanets come in all shapes and sizes. Some are large and gaseous, like Jupiter or Neptune, while others are small and rocky, like Earth. Curiously, there seems to be a lack of a certain size of planet — those that have radii 1.5–2 times Earth’s radius  — and no one is sure why. Since we can detect exoplanets with sizes above and below this radius range, this dearth of planets is not likely due to a detection bias, but rather some physical mechanism at play. Astronomers believe that planets that form in this intermediate size range have a hard time holding onto their atmospheres and shrink due to some mass-loss mechanism. The authors of today’s research article investigate which mechanism is the most likely culprit for this observed phenomenon.

Mind the (Radius) Gap

Since the first detection of an exoplanet in 1992, the number of confirmed exoplanets has been increasing astronomically. Now astronomers have detected enough planets (5,612 and counting) to try to understand the underlying distributions between different populations of planets. When plotting the radius distribution of planets detected by the Kepler space telescope, Fulton et al. 2017 found that the radius distribution of small planets is bimodal — that is, there are two distinct peaks corresponding to two different planet populations, known as super-Earths and sub-Neptunes, and few planets in between them (see Figure 1). Super-Earths are rocky planets with sizes around 1.5 times that of Earth, while sub-Neptunes are gaseous planets, similar in composition to Neptune, but with radii between 2 and 3 times the radius of Earth. The scarcity of planets between these two populations, between 1.5 and 2 times the size of Earth, is often referred to as the “radius gap” or “radius valley,” where the occurrence rate of known exoplanets is much lower than expected.

plot of planet occurrence per star versus radius for planets with periods less than 100 days

Figure 1: The histogram of planet radii for a sample of Kepler planets with orbital periods shorter than 100 days. The super-Earth and sub-Neptune regimes are highlighted in light red and light cyan, respectively. The “radius gap” corresponds to the decreased occurrence rate between 1.5 and 2.0 Earth radii. [Adapted from Fulton et al. 2017]

The cause of this radius valley is still an unsolved mystery. Many astronomers believe that planets that form in this valley aren’t massive enough to hold on to their gaseous outer layers, causing them to shrink over time until they are left with only their rocky cores. The two leading theories for what causes this atmospheric loss are photoevaporation and core-powered mass loss. In the photoevaporation scenario, planets of this size have their gaseous atmospheres almost entirely stripped away by the high-energy radiation from their host stars. This process is predicted to occur in the first 100 million years following the planet’s formation.

On the other hand, the core-powered mass-loss scenario predicts that the energy generated from the planet’s own core as it cools propagates to its outer layers and blows away its atmosphere if the planet is not initially massive enough to hold onto it. This process is much slower, taking a whopping 0.5–2 billion years to complete. Both theories have a lot of compelling evidence, so which is more likely? In order to address this question, the authors of today’s article turned to NASA’s K2 mission.

Photoevaporation or Core-Powered Mass Loss?

While NASA’s Kepler space telescope was originally designed to observe just one patch of the sky, its high-precision pointing began to fail in 2013, causing it to stray away from its original field of view. Luckily, it was able to be repurposed into the K2 mission, where it instead began to observe stars along the ecliptic plane. Its new field of view contained the Praesepe and Hyades star clusters, two populations of stars between 600 million and 800 million years old. Because planets are roughly the same age as their host stars, astronomers were able to detect a large number of planets old enough to experience photoevaporation, but not old enough to have undergone core-powered mass loss. The authors looked for sub-Neptune-sized planets in these two clusters (i.e., the planet population just above the radius valley) and predicted that if they observed a high occurrence rate of sub-Neptunes, they can conclude that photoevaporation could not be the leading cause of this mysterious mass loss.

And that is exactly what they observed. As shown in Figure 2, the occurrence rate of sub-Neptunes is 79%–107% for stars in the Praesepe cluster. This occurrence rate is much higher than the occurrence rate of ~20% for sub-Neptunes around field stars (i.e., stars not associated with a cluster), which can be much older. This result is also consistent with a previous result from clusters observed by the Transiting Exoplanet Survey Satellite (TESS), where the authors found a fractional occurrence rate of ~100% for sub-Neptunes that were roughly 50 million years old. If photoevaporation were the driving mechanism of mass loss, the occurrence rate of sub-Neptunes should be closer to that of planets around field stars, since their atmospheres would have been completely stripped by photoevaporation by this age.

occurrence rates of sub-Neptunes over time

Figure 2: The occurrence rates of sub-Neptunes over time, from the young cluster stars detected with TESS (15–450 Myr) from Fernandes et al. 2023, to the intermediate-age (600–800 Myr) Praesepe stars detected with K2 from this paper, to the old field K2 stars (3–9 Gyr). The shaded pink region indicates the timescales where photoevaporation is predicted to occur (<100 Myr) and the orange shaded region shows the timescale where core-powered mass loss is predicted to occur (0.5–2 Gyr). The drop in occurrence rate occurs after the core-powered mass-loss timescale, making it the more favorable mass-loss mechanism. [Christiansen et al. 2023]

So, does that mean core-powered mass loss is the mechanism that is guilty of stealing these planet’s atmospheres and causing this radius valley? Perhaps! But the jury is still out. While this result does leave core-powered mass loss as the more favorable mass-loss mechanism, there may be many more possibilities we haven’t yet considered. The picture is still incomplete, but this research gives us a little more insight into what happens to small planets as they evolve.

Original astrobite edited by Junellie Gonzalez Quiles.

About the author, Tori Bonidie:

I am a 4th-year PhD candidate studying exoplanets at the University of Pittsburgh. Prior to this, I earned my BA in astrophysics at Franklin and Marshall College, where I worked on pulsar detection as a member of NANOGrav. In my free time you can find me cooking, napping with my cat, or reading STEMinist romcoms!

photograph of the ultra-faint dwarf galaxy Leo IV

Editor’s Note: Astrobites is a graduate-student-run organization that digests astrophysical literature for undergraduate students. As part of the partnership between the AAS and astrobites, we occasionally repost astrobites content here at AAS Nova. We hope you enjoy this post from astrobites; the original can be viewed at astrobites.org.

Title: Extended Stellar Populations in Ultra-Faint Dwarf Galaxies
Authors: Elisa A. Tau, A. Katherina Vivas, and Clara E. Martínez-Vázquez
First Author’s Institution: University of La Serena
Status: Published in AJ

The Runts of the Galaxy Litter

The first galaxies formed only a few hundred million years after the Big Bang. Galaxies start out small, then grow by pulling in the surrounding gas and merging with each other. Most galaxies today are about 100,000 light-years across (roughly the size of the Milky Way), but some of the largest galaxies can be millions of light-years across. At the other end of the scale, we have dwarf galaxies, which are closer to 10,000 light-years across. As a result of their small size, dwarf galaxies have few stars and aren’t very luminous. Dwarf galaxies often orbit larger galaxies as satellites, similar to how moons orbit planets and planets orbit stars.

Dozens of dwarf galaxies have been discovered orbiting the Milky Way. You’ve likely already heard of the most famous of these — the Large and Small Magellanic Clouds, which are visible from the Southern Hemisphere. However, most of the dwarf galaxies orbiting the Milky Way are much smaller and dimmer than the Magellanic Clouds, making them difficult to make out with the naked eye. Astronomers often identify dwarf galaxies by finding a collection of stars and gas that moves together with star-mapping surveys like Gaia. Ultra-faint dwarf galaxies are particularly difficult to detect because they have so few stars; where a normal dwarf galaxy may have millions or billions of stars, ultra-faint dwarfs usually only contain thousands or tens of thousands. Ultra-faint dwarf galaxies lack stars because most of the mass in these galaxies consists of dark matter.

A topic of discussion among astronomers studying dwarf galaxies is how large these galaxies can be. The size of a dwarf galaxy can be determined by measuring the distance to the farthest stars that belong to the galaxy. Stars located near the edges of dwarf galaxies would belong to a stellar halo, which contains older stars. Stellar halos around dwarf galaxies are incredibly difficult to detect because they contain so few stars, so it is not clear how often ultra-faint dwarfs host stellar halos or what the halo properties are. One way to trace stars that may belong to stellar halos is with RR Lyrae variable stars. In addition to being older stars, meaning we expect them to be more common in stellar halos, these stars are a type of “standard candle,” meaning they have a known luminosity or energy output. RR Lyrae stars pulsate, and the variation in their luminosities is related to the period of their pulsations, so astronomers can infer an RR Lyrae star’s luminosity by measuring changes in its brightness. Previous studies have attempted to look for stellar halos in ultra-faint dwarfs with RR Lyrae stars, but this method was challenging to carry out before the data and telescopes that became available in the last few years. The authors of today’s article are continuing the search for RR Lyrae stars in the outer reaches of ultra-faint dwarfs to probe the presence of extended stellar populations and stellar halos.

Finding Ultra-Faint Dwarfs

The authors compiled a list of known ultra-faint dwarfs around the Milky Way, many of which were discovered by the Dark Energy Survey (DES). They also compiled a sample of RR Lyrae stars that were detected by a series of stellar surveys, including Gaia, the Zwicky Transient Facility (ZTF) surveys, DES, and the first Panoramic Survey Telescope and Rapid Response System (Pan-STARRS) survey.

But how do astronomers actually determine which stars belong to an ultra-faint dwarf, especially when these galaxies contain so few stars? Astronomers use a couple of quantities to determine this. First, they are looking for a collection of stars that are in the same area of the sky and are all roughly the same distance from Earth. The authors measure the distance to each star by using the star’s luminosity, since it is related to the distance by the equation L = 4πFd2, where L is the luminosity (total power emitted), F is the flux (power per unit area detected at Earth), and d is the distance. Again, astronomers know the luminosities of these stars because the luminosity of an RR Lyrae star is related to how quickly its light output varies. By comparing this known luminosity to how much light they actually detect, astronomers can estimate the distance between us and the RR Lyrae star.

The next main variable the authors use is proper motion, which is a measurement of how quickly an object moves across the sky. A group of stars with similar proper motions are likely moving together, which is necessary for the collection to be called a dwarf galaxy. Astronomers consider the angular distance between each RR Lyrae star and the center of the dwarf galaxy. Then, they use the distances and proper motions they measured to determine if that RR Lyrae star belongs to the dwarf galaxy or the Milky Way.

Lastly, we need to discuss a quantity called the half-light radius, which the authors denote by Rh. The half-light radius is the radius out to which half the light emitted by a galaxy is contained. Since galaxies don’t have clear edges, the half-light radius provides a useful scale for measuring galaxy sizes. Note that the full radius of a galaxy isn’t 2 Rh since galaxies are brighter in the center and dimmer near the edges. The authors are looking for RR Lyrae stars that are several half-light radii from the centers of their ultra-faint dwarfs, which is where they believe the stellar halos would exist.

Living on the Edge

The authors identify more than 100 RR Lyrae stars in ultra-faint dwarfs from the surveys mentioned above and find that nearly half of the ultra-faint dwarfs contain at least one RR Lyrae star. Some examples of the RR Lyrae stars and their locations in the galaxies are shown in Figure 1.

Plots showing the positions of RR Lyrae stars in four of the ultra-faint dwarfs used in this study

Figure 1: A sample of ultra-faint dwarf galaxies and their associated RR Lyrae stars. Symbols correspond to the catalog that the star’s position comes from, and the contours represent different radii. Note that some of the stars for the Bootes dwarf galaxies may actually belong to the Sagittarius Stream, which is a part of the Milky Way. [Tau et al. 2024]

These ultra-faint dwarfs all contain at least one RR Lyrae star at a distance of more than 4 Rh, which is what the authors consider to be the boundary between the main body of the galaxy and the stellar halo surrounding it. Since RR Lyrae stars exist here, we suspect other stars do too, indicating that these ultra-faint dwarfs do contain stars that have a large spatial extension from the main body of the dwarf galaxy and may belong to a stellar halo.

A more comprehensive showcase of the RR Lyrae stars is highlighted in Figure 2. While most RR Lyrae stars (and, therefore, most stars) are expected to be within 2 Rh of the center, several have been found well beyond this distance. This again highlights that RR Lyrae stars can be located beyond the main body of the ultra-faint dwarfs in what are likely stellar halos.

Histograms of the distances between RR Lyrae stars and the centers of their host galaxies

Figure 2: Histograms of RR Lyrae stars and their distances from the center of their ultra-faint dwarf (UFD) host galaxy in units of Rh. Blue represents all RR Lyrae stars, and orange represents a cleaned sample that has removed stars from ultra-faint dwarfs for which the authors are concerned about contamination (e.g., stars in the Bootes dwarf galaxies due to contamination from the Sagittarius Stream). The right panel is a zoom-in of the left panel, focusing on distances of up to 15 Rh. [Tau et al. 2024]

The authors note that there are opportunities for further studies in this area. One thing that remains unclear is why only some ultra-faint dwarfs are observed to host RR Lyrae stars, and why some don’t contain RR Lyrae stars at distances that would suggest the existence of a stellar halo. The authors propose that gravitational interactions with the Milky Way may tidally disrupt the dwarfs, changing their shape and potentially leading to extended stellar populations. This means that dwarfs that have close encounters with the Milky Way may be more likely to host a stellar halo. In addition to further observational surveys, cosmological simulations of dwarf galaxy formation may help shed light on how and why extended stellar populations develop.

Original astrobite edited by Alexandra Masegian.

About the author, Brandon Pries:

I am a graduate student in physics at Georgia Institute of Technology (Georgia Tech). I do research in computational astrophysics with John Wise, using machine learning to study the formation and evolution of supermassive black holes in the early universe. I’ve also done extensive research with the IceCube Collaboration as an undergraduate at Michigan State University, studying applications of neural networks to event reconstructions and searching for signals of neutrinos from dark matter annihilation.

1 … 6 7 8 9 10 … 46