Astrobites RSS

Artist's impression of a hot Jupiter exoplanet near its host star

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: Constraining the Planet Occurrence Rate around Halo Stars of Potentially Extragalactic Origin
Authors: Stephanie Yoshida, Samuel Grunblatt, and Adrian Price-Whelan
First Author’s Institution: Harvard University
Status: Published in ApJ

More than 5,000 planets have been discovered orbiting stars in the Milky Way, but astronomers have not yet confirmed detections of any planets in other galaxies. Most known exoplanets reside within a few thousand light-years of the solar system (less than the distance from us to the center of the Milky Way), so how can we find exoplanets at extragalactic distances? Today’s article describes a search for planets that formed in a dwarf galaxy that has since merged with the Milky Way.

(Not) Finding Extroplanets

A few extragalactic exoplanet, or “extroplanet,” candidates have been identified in the past, though none have been confirmed yet due to the observational difficulties of following up detections that far away. For example, in 2010, researchers announced the discovery of a planet orbiting HIP 134044, a star left over from a small galaxy that the Milky Way absorbed. Further study has since refuted this claim, noting errors in the analysis, and there is no longer evidence that such a planet exists. Today’s article continues the streak of not discovering extroplanets, but the authors invoke a statistical analysis to calculate how common planets may be around halo stars of extragalactic origin.

In the Milky Way’s outer halo, there is a unique population of stars with motions and elemental abundances that are different from stars that formed in the Milky Way. These stars are believed to have formed in a dwarf galaxy called Gaia–Enceladus that merged with the Milky Way 8–11 billion years ago. The authors of today’s article used measurements of stellar motions from the Gaia satellite’s second data release to identify stars moving in ways inconsistent with stars that formed in the Milky Way. These Gaia–Enceladus stars tend to have low or negative rotational velocity in the frame of the Milky Way, unlike typical Milky Way stars that rotate in the disk. The authors also set limits on stellar magnitude, color, and radius, combining Gaia and Transiting Exoplanet Survey Satellite (TESS) data to select low-luminosity red giant branch stars for this study. This allowed them to make direct comparisons to a previous study of similar stars with Kepler data.

TESS is searching for exoplanets that pass in front of their stars, periodically causing the stars to appear dimmer. The authors produced light curves for their sample of 1,080 stars from TESS images and searched them visually for any of these transit dips. No planet candidates were identified, so today’s article focused on using this non-detection to put a limit on how common planets could be around the stars in their sample.

Upper Limits

While a non-detection may sound disappointing at first, it can still teach us something. By calculating a study’s “completeness,” or what fraction of such objects it could detect, researchers can use a non-detection to place a constraint, or upper limit, on how common the objects may be.

The authors used an injection-recovery method to calculate the completeness of their search. They inserted simulated transit signals into their light curves (example in Figure 1), then used a Box Least Squares search to try to identify the signals within some precision in orbital period and transit depth. They found that roughly 30% of their injected signals were recovered, that the recovery rate was highest for planets with short periods and large radii, and that planets smaller than half the size of Jupiter were essentially undetectable.

comparison of a real light curve with no transit signal with a light curve with a synthetic transit signal inserted

Figure 1: Left: TIC455692967’s observed light curve. Right: TIC455692967’s light curve with a simulated transit signal (marked with the red arrow). [Yoshida et al. 2022]

Takeaways

The final upper-limit calculation found that less than 0.52% of low-luminosity red giant halo stars should host hot Jupiters (planets similar in size to Jupiter with orbital periods of 10 days or less). This finding agrees with a previous estimate that put the occurrence rate at roughly 0.5%. The occurrence rate of hot Jupiters correlates with stellar metallicity, generally measured as the ratio between iron and hydrogen abundances in the star. Halo stars typically have very low metallicities, suggesting that hot Jupiters should be about ten times rarer around halo stars than around other stars in the galaxy.

These upper limits are only the maximum possible occurrence rate, so these planets may be even less common than the percentages given in Figure 2. How will we actually find these rare planets of extragalactic origin? By studying more stars! The recent third data release from Gaia and ongoing TESS observations are moving astronomers in the right direction.

a table listing upper limits on planet occurence rates for low-luminosity red giant stars in the Milky Way halo

Figure 2: Upper limits on planet occurrence for low-luminosity halo red giants for ranges of planet radius and orbital period. Undefined limits are the result of zero injected signals in the given radius and period range being retrieved. [Yoshida et al. 2022]

Original astrobite edited by Sahil Hedge.

About the author, Macy Huston:

I am a fourth-year graduate student at Penn State University studying Astronomy & Astrophysics. My current work focuses on technosignatures, also referred to as the Search for Extraterrestrial Intelligence (SETI). I am generally interested in exoplanet and exoplanet-adjacent research. In the past, I have performed research on planetary microlensing and low-mass star and brown dwarf formation.

hubble image of SN1987a

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: 3D Hydrodynamics of Pre-supernova Outbursts in Convective Red Supergiant Envelopes
Authors: Benny T.-H. Tsang, Daniel Kasen, and Lars Bildsten
First Author’s Institution: University of California, Berkeley
Status: Published in ApJ

All right, I know what you’re thinking: “What do my digestive problems (for the hopefully very few of you) have to do with a star’s last, quite spectacular, goodbye?” Unlike most humans, stars don’t possess working intestines, and their outbursts — supernovae — are far more impressive than anything we can manage. There are many different kinds of supernovae, but they are broadly split into two categories: Type I and Type II. We typically see no emission lines of hydrogen in the spectra of Type I supernovae (here’s an example of why), while the spectra of Type II supernovae do contain hydrogen lines. We will talk about this latter type in today’s bite.

Red Supergiant Burps and Interacting Supernovae

Before red supergiants go supernova, they are prone to breathtaking belches called pre-supernova outbursts. These outbursts push huge amounts of gas from the star out into the so-called circumstellar medium — the material in the star’s direct neighborhood. If this neighborhood is filled with enough gas when a star dies, the expanding supernova will push against and interact with this material. This interplay between the material around the star and the supernova can actually be observed from Earth, giving rise to what is known as a Type IIn or interacting supernova.

How visible this interaction is depends mostly on how much material is in the stellar neighborhood. This depends again on how much gas the star decides to throw out, and also on how these pre-supernova outbursts (or burps) are actually formed. The authors of today’s article show that this has a lot to do with convection in the red supergiant.

Red Supergiant Boiling Pot

Simulating these red supergiant outbursts shortly before they go supernova is not new, so we already know the causes of these pre-supernova outbursts:

  • Increasingly unstable nuclear fusion in the core of the star causes powerful gravity waves (not to be confused with gravitational waves)
  • Large-scale convection in the red supergiant carries around material in the star, which can destabilize the nuclear fusion in the core, giving a very variable energy output
  • Pair instability can cause the core’s energy output to go through cycles of drops and spikes
  • A binary companion star can disturb the red supergiant enough to cause the star to temporarily become unstable

The bottom line is that some process releases a large amount of extra energy inside the star, which, depending on how the star reacts to this energy release, can lead to different outbursts of gas. Until now, the simulations of these outbursts have usually been spherically symmetric, meaning that the simulation of the outburst looks exactly the same from any direction. You can also see this as a simulation along a single line of sight from the outside of the star inwards (i.e., one-dimensional).

The problem with this approach is that you cannot simulate convection this way. To deal with convection, the authors of today’s article took the brute-force approach and did a fully 3D simulation. They simulated the region of the star outside the nuclear core (called the envelope) and started with a large energy release at the innermost part of their simulation. The authors considered different styles of energy release in the envelope. These included:

  • A large, sudden energy release, comparable to the energy needed to keep the star together by gravity. This can cause a mass ejection, quite like the Sun but on much larger scales.
  • A slow release of energy, which causes a much steadier stream of mass flowing away from the star instead of an explosive loss of mass.
  • Varying direction of energy release, which influences how (and where to) the pre-supernova outburst will occur.

A snapshot of the authors’ simulation is shown in Figure 1. Here, we see both the envelope density on the left and the velocity of the envelope gas in the radial direction on the right. In the velocity graph, we can see zones both moving away from the star and falling back towards the core. These are the same as convection cells we can find in daily life — like in a pot of boiling water.

two plots of model ouput, showing the density and radial velocity

Figure 1: Left: Density slice of the star’s outer layers, with radius (R) vs. the distance from the core to the pole (z). Right: Velocity in the radial direction (away from the core) slice with the same axes as on the left. [Tsang et al. 2022]

The convection cells leave “holes” or channels of lower density in the envelope from the outside to inner parts of the star. Through these channels, much more gas can escape than would be possible without convection.

We can also see this in Figure 2: the simulation in the left panel, which included the convection, resulted in much more mass loss than the simulation in the right panel, which did not. These channels of low density appear where most of the mass escapes in the convection simulation.

projected images of the simulated star's surface under different conditions

Figure 2: Two images of the star’s surface in Mollweide projection, showing how much mass has escaped. On the left is a model with convection, where the colors indicate the amount of mass lost per direction (or, specifically, solid angle). On the right is a simulation without convection. [Tsang et al. 2022]

This article shows the necessity of taking convection in 3D into account, where the loss of mass from the pre-supernova outbursts has mostly been underestimated. This increases the amount of gas in the neighborhood of the red supergiant, ultimately affecting how the interacting supernova will look to us on Earth.

Original astrobite edited by Sasha Warren.

About the author, Roel Lefever:

Roel is a first-year PhD student at Heidelberg University, studying astrophysics. He works on massive stars and simulates their atmospheres/outflows. In his spare time, he likes to hike/bike in nature, play (a whole lot of) video games, play/listen to music (movie soundtracks!), and to read (currently The Wheel of Time, but any fantasy really).

an image of the Sun's disk with a large group of sunspots

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: Modeling Stellar Surface Features on a Subgiant Star with an M-dwarf Companion
Authors: Maria C. Schutte et al.
First Author’s Institution: University of Oklahoma
Status: Published in AJ

The authors of today’s article search for polka dots of concentrated magnetic flux on the surface of a subgiant star.

A Menagerie of Astrophysical Phenomena

The surface of a star hosts a menagerie of astrophysical phenomena: flares that dramatically increase the brightness of the star, coronal mass ejections that spew plasma into space, and cold, dark regions that grow and decay called starspots (check out this astrobite to learn more about stellar activity). Understanding this stellar activity is not only imperative for understanding stellar evolution, but also for studying the planets that orbit these stars.

The Many Uses of a Light Curve

One of the main ways that exoplanet scientists search for planets is with a technique called the transit method. Scientists monitor the brightness of a star, looking for a dip in that brightness. This dip can be caused by a planet passing, or “transiting,” in front of the star as seen from Earth. However, this graph of the brightness of the star — called a light curve — can also be used to detect other celestial happenings. Astronomers have used light curves to study binary star systems, supernovae, and — in the case of today’s article — starspots!

A Light Curve of Interest

The authors of today’s article studied the light curve of a star called KOI-340, where KOI stands for Kepler Object of Interest. KOIs are stars that were observed by the Kepler space telescope and are suspected to host exoplanets. This particular KOI is a subgiant star, which means that it is brighter and larger than a normal main-sequence star, but it is not as bright or as large as a giant star. KOI-340 also hosts a smaller, colder M-dwarf companion, making it a binary star system.

The light curve of KOI-340 was of interest to the authors because the depth of the planet’s transit is shallower than it should be if the star did not have any starspots. They therefore used a modeling code called STarSPot (STSP), which is publicly available on github, to model 36 transits of the planet orbiting KOI-340 to show evidence of starspots on the surface of the star. They found that the average radius of the starspots is roughly 10% the radius of the star, which would make these starspots some of the largest ever recorded (Figure 1).

histogram showing starspot sizes relative to the size of the sun and KOI-340

Figure 1: This plot shows the radius distribution for the starspots on KOI-340 (grey), compared with the distributions for the sunpots when the Sun was at solar maximum (red) and at solar minimum (blue) over the same duration as the Kepler mission (four years). The black dashed lines show the size of the smallest starspot, found in Transit 19, and the main starspot, found in Transit 21. The dashed red line shows the size of the largest sunspot ever detected. [Schutte et al. 2022]

The authors even found a starspot as large as 16% of the radius of the star (Figure 2)!

modeling of a planet transiting in front of a large starspot

Figure 2: In today’s article, the authors modeled 36 transits of the planet orbiting KOI-340 to show evidence of starspots on the surface of the star. The top of this plot shows the location of the starspot found in Transit 21. The bottom of this plot shows the light curve with the fit from STSP (red line), compared to the fit if KOI-340 did not have any starspots (cyan line). The residuals are shown as blue points below the light curve. [Schutte et al. 2022]

In a different transit, the authors found an additional dip in their fit, indicating the presence of a bright spot on KOI-340 followed by a dark spot (Figure 3).

modeling of a planet transiting in front of a large starspot and a bright spot

Figure 3: Similar to Figure 2, the top of this plot shows the locations of the starspot and the bright spot found in Transit 11 (the black and red circles, respectively). The bottom of this plot shows the light curve with the fit from STSP, compared to the fit if KOI-340 did not have any starspots. The residuals are shown as blue points below the light curve. [Schutte et al. 2022]

This work shows that starspots are not simply a nuisance to exoplanet scientists; they are also a window into the life and evolution of the star itself. For example, the authors of today’s article were able to conclude that the increased activity on KOI-340 as compared to the Sun is likely due to KOI-340’s faster rotation and/or the increasing size of its convection zone as it evolves from a main-sequence star into a red giant star.

In general, today’s authors show us how multifaceted starspots can be, as well as the splendor of polka-dotted stars!

Original astrobite edited by Maryum Sayeed.

About the author, Catherine Clark:

Catherine Clark is a PhD candidate at Northern Arizona University and Lowell Observatory. Her research focuses on the smallest, coldest, faintest stars, and she uses high-resolution imaging techniques to look for them in multi-star systems. She is also working on a Graduate Certificate in Science Communication. Previously she attended the University of Michigan, where she studied Astronomy & Astrophysics, as well as Spanish. Outside of research, she enjoys spending time outdoors hiking and photographing, and spending time indoors playing games and playing with her cats.

Hubble Space Telescope image of the dwarf spiral galaxy NGC 5949

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: Wandering Black Hole Candidates in Dwarf Galaxies at VLBI Resolution
Authors: Andrew J. Sargent et al.
First Author’s Institution: United States Naval Observatory and The George Washington University
Status: Published in ApJ

How do you make a black hole billions of times the mass of the Sun? Even for the planet-building Magratheans, this seems like a tall order. Plenty of mechanisms have been proposed to explain the formation of these supermassive black holes found at the centers of most galaxies. Some involve the mergers of “seeds” — massive black holes weighing in at merely hundreds to hundreds of thousands of solar masses. A simple way to test these theories is to search for relic massive black holes, and low-mass dwarf galaxies are excellent targets. Since dwarf galaxies haven’t undergone many mergers, any massive black holes they harbor should have avoided being gobbled up by growing supermassive black holes.

Today’s article studies 13 possible massive black hole candidates in dwarf galaxies, some of which may have wandered to the edges of their hosts. What’s up with that — and are they really massive black holes? Let’s dive in!

Here’s a question: since supermassive black holes are usually found near the center of their galaxies, why might we expect some massive black holes in dwarf galaxies to lie further out? The answer has to do with gravity: since dwarf galaxies are much less massive than the galaxies that host supermassive black holes, their gravitational potential is lower, making it easier for massive black holes to “wander” away from their centers. This means that if you see a radio source that appears far to the side of a dwarf galaxy’s center, it could be an massive black hole — or it could be an accreting supermassive black hole (an active galactic nucleus) in a galaxy far, far away that by chance simply happens to lie behind the dwarf galaxy. These unwanted interlopers can pose a challenge for identifying massive black holes.

Another issue with finding massive black holes is that they’re faint. While massive black holes go through periods of accretion like supermassive black holes, their low masses mean that they don’t accrete as quickly, reducing their luminosities. By the early 2000s, only two accreting black holes had been found in dwarf galaxies. Fortunately, this changed with the advent of sky surveys like the now famous Sloan Digital Sky Survey (SDSS), which has been running since 2000 and has amassed detections of close to a billion unique sources.

The 13 massive black hole candidates, shown in Figure 1, were assembled in an article from 2020 by some of the same astronomers who authored today’s article. In the 2020 article, the team sifted through 43,707 low-mass dwarf galaxies from SDSS, looking for sources that had been detected at radio frequencies by the Very Large Array. After keeping the matches and eliminating the radio sources that were background active galactic nuclei or could be explained by processes related to star formation, the team ended up with 13 massive black hole candidates, many of which aren’t aligned with the centers of their host galaxies.

optical images of the 13 dwarf galaxies in the sample

Figure 1: The 13 dwarf galaxies hosting possible massive black hole candidates, as seen by the Dark Energy Camera Legacy Survey at optical wavelengths. The red crosses show the location of the compact radio sources that may be massive black holes. While some appear close to their host’s center, others are significantly farther away. [Reines et al. 2020]

In this more recent article, the authors performed follow-up observations using the Very Long Baseline Array (VLBA). The VLBA uses radio telescopes thousands of kilometers apart to reach high angular resolution and allow astronomers to see fine details. Unfortunately, the VLBA was only able to detect four of the 13 candidates — and those four, because of their luminosity and position, seemed most likely to be active galactic nuclei in galaxies far beyond the dwarfs the team was targeting. The detected candidates are shown in Figure 2.

radio emission detected from four sources in the sample

Figure 2: The four sources the team was able to detect with the VLBA. Here, S is flux density, a quantity that describes the intensity of radio emission. As these sources are actually background active galactic nuclei rather than massive black holes in the targeted dwarf galaxies, the physical scales in the lower right are inaccurate. [Sargent et al. 2022]

This seems like an enormous problem! Only four detections, all of which appear to be imposters? Fortunately, the situation isn’t as dire as it might seem. While the VLBA is good at resolving sources on small scales in the configuration the team used, it may not resolve large-scale sources — and the radio emission from accreting massive black holes might be in the form of larger structures like radio lobes, rather than central point sources.

Multiwavelength observations confirmed that two of the remaining nine candidates are likely accreting supermassive black holes near the center of their host galaxies, but the other seven remain unknown. Five of those seven candidates are too bright to be from star formation and, based on their positions, could be either more background active galactic nucleus interlopers or, tantalizingly, wandering massive black holes.

Where do we go next? Follow-up observations at other wavelengths could be useful. The group suggests the Hubble Space Telescope in particular as a means of figuring out what those seven sources truly are. Given the difficulties involved in detecting massive black holes, even one more could prove valuable as astronomers try to understand the formation of the largest black holes in the universe.

Original astrobite edited by Suchitra Narayanan.

About the author, Graham Doskoch:

I’m a graduate student at West Virginia University, pursuing a PhD in radio astronomy. My research focuses on pulsars and efforts to use them to detect gravitational waves as part of pulsar timing arrays like NANOGrav and the IPTA. I love running, hiking, reading, and just enjoying nature.

composite ultraviolet and infrared image of the triangulum galaxy

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 Panchromatic Hubble Andromeda Treasury: Triangulum Extended Region (PHATTER) II. The Spatially Resolved Recent Star Formation History of M33
Authors: Margaret Lazzarini et al.
First Author’s Institution: California Institute of Technology
Status: Published in ApJ

The Panchromatic Hubble Andromeda Treasury (PHAT) team has already done the impossible. Led by Professor Julianne Dalcanton (read our interview with her from #AAS233 here!), PHAT completely revolutionized observational astronomy by imaging over 117 million stars in the disk of the Andromeda Galaxy, otherwise known as Messier 31. Imaging Messier 31 took two weeks of Hubble Space Telescope time, which is a remarkable achievement considering many observational astronomers are lucky to get even a few precious hours on Hubble!

Now, the PHAT team is ready for round two. They have moved on to Messier 31’s neighbor and the third most massive galaxy in our Local Group: the Triangulum Galaxy, or Messier 33. And of course, this observing program wouldn’t be complete without a new, catchy acronym: the Panchromatic Hubble Andromeda Treasury: Triangulum Extended Region, or “PHATTER.” Studying Messier 33 in addition to Messier 31 is beneficial because Messier 33 has had more star formation overall and can therefore provide more insight into a new parameter space previously unexplored in Messier 31. Messier 33 also has a lower stellar surface density (i.e., lower star-to-area ratio), so resolving individual stars is much easier in Messier 33 than in Messier 31. The PHATTER team has generously made their data publicly available, providing photometry (i.e., the measured flux from astronomical objects) for over 22 million stars covering 38 square kiloparsecs (about 400 million square light-years) of Messier 33.

This article, the second in the PHATTER series (where the first described the observations and photometry), measured the star formation history of Messier 33. Measuring the star formation history of a galaxy can provide crucial information about the astrophysical phenomena that shape galaxy formation, such as how the structure of a galaxy changes over time.

To measure star formation rates of galaxies, astronomers have historically used two different methods. The first method involves studying ultraviolet emission from massive young stars. Because young stars primarily emit at ultraviolet wavelengths, ultraviolet flux is often used as a tracer for star formation within the last 200 million years. The second method involves studying H-alpha emission, which occurs when the electron in a hydrogen atom falls from the third energy level to the second. H-alpha emission often indicates that hydrogen is being ionized, usually by young O stars, and this emission traces star formation within the last 5 million years. However, both of these techniques are limited by dust extinction, which can be difficult to correct for.

The authors of today’s article use a novel method referred to as “CMD-based modeling” to measure the star formation history of Messier 33. The basic premise of this technique is that if you have high-accuracy photometry, you can use color–magnitude diagrams (CMDs, the observer’s version of the H-R diagram, where instead of plotting luminosity vs. temperature, you plot magnitude vs. color) to infer the star formation rates throughout history that would have produced a given observed population of stars. For example, younger stars spend less time in a given color–magnitude diagram zone than older red giant branch stars, and this information can be used to interpret the observed color–magnitude distribution of stars in a galaxy. Another useful benefit of the CMD-based modeling technique is that it simultaneously fits for the dust extinction, unlike the ultraviolet or H-alpha methods.

To measure the star formation history in bins across the face of Messier 33, the authors split their photometry into ~2,000 regions, each of which contained 4,000 stars on average. To measure the star formation history, the team fit color–magnitude diagrams in each region using the MATCH software, which finds the combination of stellar populations that best produces the observed color–magnitude diagram. Using this software, the authors were able to reconstruct Messier 33’s star formation history by measuring the star formation rate in ~50-million-year bins, up until 630 million years ago. While the CMD-based method requires high-resolution photometry, you can study the star formation rate throughout history, whereas the ultraviolet and H-alpha techniques only measure recent star formation.

The Structure of Messier 33

Detailed star formation histories can be used to measure how a galaxy’s stellar structure has changed over time. Messier 33 has typically been characterized as a flocculent spiral galaxy, meaning its spiral arms are less defined than those of a grand design spiral galaxy like Messier 101 (see Figure 1 for a comparison of the two). However, by studying the star formation rate throughout Messier 33’s history (as opposed to just the recent star formation), the authors were able to reconstruct the evolution of Messier 33’s spiral structure using the measured star formation rate in ~50-million-year time bins.

The authors found that while Messier 33 does indeed have flocculent spiral structure that formed about 79 million years ago, it previously had two distinct spiral arms. In short, the younger stellar populations (younger than 80 million years old) present as a flocculent spiral structure and the older stellar populations are primarily present in two distinct spiral arms.

In Figure 2, you can clearly see the split between these two stellar populations. The authors also clearly detect a bar in Messier 33 that is older than about 79 million years, which is significant because there has been a lot of recent debate in the literature about whether Messier 33 has a bar. The detection of bars in galaxies has strong implications for the galaxy formation history; bars force a lot of gas towards the galaxy’s center, fueling new star formation, building central bulges of stars, and feeding massive black holes. In particular for Messier 33, a small bar could explain discrepancies between models and observed gas velocities in the inner disk. The authors suggest that more modeling should be done to explain why the younger stellar populations did not form in a bar, whereas the older stellar populations did.

plots of Messier 33's star-formation rate during two time periods

Figure 2: The spiral structure clearly evolves from 79–631 million years ago to 0–79 million years ago, indicating a transition in the spiral structure of Messier 33 around 79 million years ago from a two-armed barred spiral structure (right) to the more flocculent spiral structure we observe today. [Lazzarini et al. 2022]

Finally, the authors compared their global star formation rate (which has units of solar masses per year and measures the total mass of stars being added to the galaxy each year) to that measured by the conventional methods using ultraviolet and H-alpha emission. The author found their measured value was about 1.6 times larger than the ultraviolet/H-alpha measurement, indicating that ultraviolet/H-alpha measurements may not capture the full star formation rate of a galaxy. In the future, the authors plan to extend this analysis by focusing on measuring the age gradient of Messier 33’s spiral arms and bar.

Original astrobite edited by Isabella Trierweiler.

About the author, Abby Lee:

I am a graduate student at UChicago, where I study cosmic distance scales and the Hubble tension. Outside of astronomy, I like to play soccer, run, and learn about fashion design!

photograph of the Keck telescopes

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: Long Dark Gaps in the Lyβ Forest at z<6: Evidence of Ultra Late Reionization from XQR-30 Spectra
Authors: Yongda Zhu et al.
First Author’s Institution: University of California, Riverside
Status: Published in ApJ

The who, what, when, and where of reionization are unresolved questions that have important implications for our understanding of the cosmos. This final phase transition of the universe from neutral to ionized encompasses a variety of dramatic changes, as large-scale structures formed and evolved and the first stars and galaxies began to light up the universe. While shining a light on the history of reionization would also help uncover the history of how objects in the universe emerged and grew, this epoch is fundamentally dark.

However, we do have the basics of the why down: before the epoch of reionization, the universe was mostly filled with neutral hydrogen gas. Then, as the first stars, galaxies, and quasars began to form and started shining, these objects emitted high-energy photons that ionized the neutral gas around them. The ionizing radiation kicked out electrons from neutral hydrogen atoms until eventually most of the gas in the universe became ionized.

One key part of gaining a complete understanding of reionization is the when — precisely when did it begin and end, and how rapidly? There are a few main methods for probing these transition points, all of which suggest the process occurred early on, with a midpoint at roughly redshift z ~ 8 (600 million years after the Big Bang) and an endpoint somewhere around z ~ 5.5–6 (1 billion years after the Big Bang).

Absorbing It All

Many of the techniques used to trace reionization, including those used in today’s article, involve the Lyman series transitions of hydrogen, especially the Lyman-α (n=2 to n=1) transition. One such technique relies on observations of quasars (active supermassive black holes in the centers of galaxies and among the brightest objects in the universe) during the epoch of reionization. By observing distant quasars, we can understand the gas content in the universe along that line of sight using the presence and absence of Lyman-α emission and absorption compared to typical quasar spectra, which are fairly well understood and have strong signals.

As emission from a quasar travels through material between the quasar and the observer, some emission gets intercepted by gas clouds along the way, which can absorb the emission and produce a Lyman-α absorption line at a wavelength determined by the redshift of the gas cloud. However, at redshifts before the end of the epoch of reionization, the primarily neutral gas in the way will be very opaque at this wavelength, as photons with wavelengths (energies) near Lyman-α struggle to pass through the neutral hydrogen, which is optically thick enough to suppress observed emission nearly completely. This optical thickness to high-energy photons results in a contiguous region of strong absorption in the spectrum that is known as a dark gap (see Figure 1).

diagram showing how neutral gas creates a gap in quasar spectra

Figure 1: Example spectrum of a distant quasar with ionizing radiation emitted from the quasar intercepted by neutral gas along the line of sight. Emission lines directly from the quasar are marked by dashed colored lines. An over-dense patch of neutral gas causes a gap in the Lyman-α emission as nearly all of the flux from the quasar is absorbed. [Adapted from Figure 7 in An Introductory Review on Cosmic Reionization by John H. Wise]

Filling In the Gaps

These dark gaps could be caused by several different processes within reionization. For one, the ionizing background radiation itself could have some fluctuations, though the authors explain that their results and other recent results disfavor this scenario as it doesn’t imply sufficient neutral gas at later times. Alternatively, the prevalence of so-called “islands” of neutral gas, like pockets of Lyman photon absorption, could be the cause. Lastly, reionization could simply end later in the history of the universe, meaning more neutral gas is available to absorb high-energy photons at lower redshifts.

These scenarios are difficult to disentangle with Lyman-α gaps alone. One solution, as presented in today’s article, is to use a Lyman transition with a slightly shorter wavelength that passes through neutral gas slightly more easily (i.e., the neutral gas has a lower optical depth at that wavelength). The authors use this technique of tracing long dark gaps in quasar spectra but apply it to another Lyman transition, Lyman-β (n=3 to n=1). In order to study the dark Lyman-β gaps, the authors analyze spectra of a sample of epoch of reionization quasars at z > 5.5. Within each spectrum, they map out the dark gaps of Lyman-β absorption, the length of the gaps, and the redshift evolution of the gaps. As shown in Figure 2, one quasar spectrum in particular had a uniquely long dark gap down to z ~ 5.5. Within that line of sight, the authors found a low-density region of galaxies, which supports the idea that highly opaque sight lines are associated with galaxy underdensities. This makes sense: in areas with fewer galaxies to ionize their surroundings, there is more remaining neutral gas.

spectrum of the quasar PSOJ025-11

Figure 2: The light blue line in the top panel shows the spectrum for the quasar, and the dashed curve shows the predicted emission from the quasar in the absence of absorption from gas in front of the quasar. The region over which Lyman-β gaps were searched for is labeled, as is the corresponding Lyman-α forest for the same redshift range. The bottom panel shows a zoom-in of the Lyman-β gap region, with gaps (flux lower than the dashed threshold) shaded in gray. The the dark gap between z = 5.53 and 5.61 is the longest gap detected in the survey. [Adapted from Zhu et al. 2022]

The authors also emphasize the uniqueness of this study: the reionization scenarios are difficult to disentangle with Lyman-α gaps, as their signatures look similar. However, given the lower optical depth of neutral gas at the wavelength of Lyman-β, Lyman-β is a more sensitive probe of neutral gas in the late intergalactic medium, and it’s a useful tool to better understand the end of reionization. By further comparing their observations of dark gaps to expectations from cosmological simulations of these scenarios, the authors determine which reionization scenarios remain possible given the evidence.

Given the distinction enabled by Lyman-β data, the authors propose the best-fit scenario for their dark gap sample is late reionization, with the epoch of reionization ending at z ~ 5.3. They demonstrate that rapid late-reionization models, specifically those with a fraction of neutral gas > 5% at z = 5.6, are consistent with the observations. Looking ahead, these dark Lyman-β gaps and future large samples of quasar spectra with gaps can help fill in the gaps in our knowledge of the timing of reionization.

Original astrobite edited by Jana Steuer.

About the author, Olivia Cooper:

I’m a second-year grad student at UT Austin studying the obscured early universe, specifically the formation and evolution of dusty star-forming galaxies. In undergrad at Smith College, I studied astrophysics and climate change communication. Besides doing science with pretty pictures of distant galaxies, I also like driving to the middle of nowhere to take pretty pictures of our own galaxy!

collapsar

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: Radio Constraints on r-process Nucleosynthesis by Collapsars
Authors: K.H. Lee et al.
First Author’s Institution: University of Florida
Status: Published in ApJL

Most elements in the periodic table originate in stars. Elements lighter than iron are stable and can be formed by nuclear fusion in stellar cores. Heavier elements have unstable nuclei and require additional sources of energy to form. These elements are believed to originate in supernovae: the explosive deaths of stars with masses more than ten times that of the Sun. However, even supernovae cannot produce the heaviest elements in the periodic table — the lanthanides and the actinides. The origin of these “heavy” elements (which include other metals like gold and platinum) is a long-standing mystery. The authors of today’s article use radio observations of collapsars — special types of supernovae — to constrain whether these elusive heavy elements can be formed in them.

r-process Nucleosynthesis

The reason why lanthanides and actinides are so difficult to produce is because they are formed by a rare process involving the rapid capture of neutrons that is known as the r-process. In this process, a free neutron is captured onto the nucleus of an atom, producing a nucleus with a higher atomic mass number. The problem is that free neutrons are inherently unstable particles that undergo the beta-decay process to form a proton, an electron, and an antineutrino on a timescale of a few minutes. To efficiently activate the r-process, neutrons need to be captured faster than they decay. Therefore, a neutron-rich environment is needed. One such environment is formed during the explosive mergers of two neutron stars. So far, only one of these events has ever been observed: GW170817, which was detected in both gravitational and electromagnetic waves. However, other astrophysical explosions have also been proposed to produce neutron-rich ejecta, creating the conditions for r-process nucleosynthesis. One such explosion is known as a collapsar.

CollapsaRs

When massive stars (heavier than about 10 solar masses) die, they can form neutron stars or black holes. If a neutron star is formed, most of the star’s outer layers are ejected at large velocities (~10,000 km/s!), producing an energetic explosion that we all know as a supernova. However, if a black hole is formed instead of a neutron star, most of the star can be gobbled up by the black hole and very little material will be ejected. This changes if the star that formed the black hole had a mass of 20–30 solar masses and was spinning really fast. Owing to the large angular momentum of this star, a larger fraction of its total mass can now be ejected. In fact, some of the ejected material can be accelerated to relativistic speeds (comparable to the speed of light), producing a beam of high-energy photons called a gamma-ray burst. In addition to the gamma-ray burst, the remaining ejected material — which is not relativistic but is still moving really fast (~20,000 km/s) — can produce a regular supernova-like explosion. This explosion of a rapidly spinning massive star is known as a collapsar.

The ejected material from collapsars can power gamma-ray bursts and supernovae. The material that falls into the black hole can also do interesting stuff. Because this material has large angular momentum, it can form a disk around the newly formed black hole. Most of this disk will be accreted by the black hole within a few seconds, destroying the disk material, but the accretion process itself can eject a fraction of the disk in the form of winds. It turns out that these disk winds are extremely rich in neutrons and are thus viable sites for r-process nucleosynthesis (i.e., formation of lanthanides and actinides). However, the signatures of this r-process are extremely difficult to observe as they can be masked by other features arising from the supernova explosion. To date, there has been no direct evidence of r-process nucleosynthesis in collapsars.

Radio

Radio observations provide one way to observe the signatures of this elusive r-process in collapsars. Several months after the supernova explosion, disk-wind ejecta that is rich in lanthanides and actinides should interact with the interstellar medium surrounding the black hole. This will produce a radio flare powered by a mechanism known as synchrotron emission, in which electrons spiral around magnetic field lines. This flare should peak several months after the explosion, evolve slowly, and last for several years (Figure 1). Observations of such a radio flare in collapsars several years post-explosion can provide constraints on the amount of r-process elements synthesized in them.

plot of theoretical radio emission from a collapsar

Figure 1: Expected radio emission from the collapsar GRB060505 that occurred in 2006. The theoretical models assume an r-process ejecta mass of 0.1 solar mass (solid lines) and 0.01 solar mass (dotted lines). Different colors represent different model parameters, specifically interstellar medium density profiles and electron distributions. The black arrow marks actual upper limits from Very Large Array observations that can be used to constrain the r-process ejecta mass. [Lee et al. 2022]

Putting It All Together!

The authors of today’s article searched for the late-time radio emission in collapsars. First, they selected 11 collapsars that exploded in the last decade based on the gamma-ray bursts detected by the Swift space satellite. They then looked for possible radio emission at the locations of these supernovae using data from the Karl G. Jansky Very Large Array radio telescopes. Unfortunately, they did not detect any late-time radio flares from these collapsars. However, based on the sensitivity of the data, the team was able to place upper limits on the observed late-time radio emission. They then used theoretical models to derive upper limits on the total r-process material ejected by the collapsar. They found that the collapsars could not have ejected more than 0.2 solar mass of r-process material. For reference, the neutron star merger GW170817 produced about 0.05 solar mass of r-process material. This means that collapsars do not cause significantly more r-process nucleosynthesis than the one neutron star merger we know of.

The authors note that the derived upper limits depend on assumptions of their models about the interstellar medium density profiles, the energy distribution of electrons in the interstellar medium, and the velocity of the disk-wind ejecta. Despite these caveats, their observations place meaningful constraints on the amount of r-process material synthesized in collapsars. Future, more sensitive radio observations can help confirm whether collapsars can synthesize the heaviest lanthanides or actinides or not. This will be important to identify whether collapsars and neutron star mergers can account for the observed r-process elements in the universe, or whether we need to look for more r-process factories.

Original astrobite edited by Sasha Warren.

About the author, Viraj Karambelkar:

I am a second-year graduate student at Caltech. My research focuses on infrared time-domain astronomy. I study dusty explosions and dust enshrouded variable stars using optical and infrared telescopes. I mainly work with data from the Zwicky Transient Facility and the Palomar Gattini-IR telescopes. I love watching movies and plays, playing badminton and am trying hard to improve my chess and crossword skills.

Illustration of a brown dwarf

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: Tracing the Top-of-the-Atmosphere and Vertical Cloud Structure of a Fast-Rotating Late T-dwarf
Authors: Elena Manjavacas et al.
First Author’s Institution: Space Telescope Science Institute and Johns Hopkins University
Status: Published in AJ

While they have a reputation as “failed stars,” brown dwarfs might have more in common with their gas giant planet cousins than with stars. Brown dwarfs have swirling, patchy clouds in their atmospheres, and the light curves of brown dwarfs have been seen to vary in amplitude as they rotate and bring different faces with variable cloud coverage into view. By observing brown dwarfs over their rotation periods with spectrophotometry, astronomers can simultaneously measure how much the atmosphere is changing in multiple wavelength bands. This technique means a 3D map of a brown dwarf’s atmosphere can be built, since different wavelengths probe different levels of pressure within an atmosphere. Although most spectrophotometric observations of brown dwarfs have used the Hubble Space Telescope, the authors of today’s article employed the ground-based Keck I telescope to study 2M0050–3322, a rapidly rotating T-type brown dwarf.

Seeing One Atmosphere Through Another

Since the levels of variability in brown dwarf atmospheres can be small, it is important to characterise any other non-brown dwarf sources of noise in the data. For these ground-based observations, particular care has to be taken to account for changes in Earth’s atmosphere over the course of the observations. Using the Multi-Object Spectrometer For Infra-Red Exploration (MOSFIRE), the authors observed 2M0050–3322 for two of its rotation periods (around 2.5 hours in total). They also observed several other nearby stars to help calibrate the impacts of things like the local humidity and temperature on the measurements. The light curves of all the objects were obtained at multiple different infrared wavelengths: J band, H band, and in a wavelength range slightly redder than the H band, which the authors call the CH–H2O band. By dividing each of 2M0050–3322’s light curves by the median light curve of the calibration stars, 2M0050–3322’s light curves could be corrected for any effects of Earth’s changing atmosphere to find the true variability of the brown dwarf, as shown in Figure 1.

Light curves of the brown dwarf in several wavelength bands

Figure 1: Light curves of 2M0050–3322 in J, H and CH4–H2O bands. The CH4–H2O band shows the biggest fluctuations, but all bands are best fit by a flat line. [Manjavacas et al. 2022]

Over the course of their observations, the authors found that 2M0050–3322 had a minimum to maximum fluctuation of ~1% in the J and H band and a higher 5% amplitude in the redder CH4–H2O band. This seemingly low level of variation was also confirmed by fitting flat and sinusoidal models to the light curves, with a flat line proving to best the preferred fit for all the observations.

Models to the Rescue?

With observations in hand, the authors then sought to compare their results to models of 2M0050–3322 to see if a similar lack of variation was present. General circulation models of the thermal flux of the atmosphere predict a slightly sinusoidal light curve with almost a 1% variation, which matches the amplitude seen in the J and H band observations! Meanwhile, models of the structure of clouds in the atmosphere show that 2M0050–3322 has various types of clouds at different pressures, meaning that each of the observation bands could be probing different clouds.

Figure 2 shows that the CH4–H2O band traces similar pressure levels as those where sodium sulphide (Na2S) and potassium chloride (KCl) clouds condense. This could explain why the CH4–H2O light curves show more variability than the other bands, which reach deeper into the atmosphere and therefore do not probe these clouds. While these modelling efforts begin to explain the observations of the brown dwarf, the authors caution that longer-term monitoring is likely needed to fully explain the mysteries of 2M0050–3322.

Visual representation of the cloud structure of 2M0050–3322

Figure 2: Visual representation of the cloud structure of 2M0050–3322. Three types of clouds are seen to form in the atmosphere, each at different pressure levels. KCl and Na2S clouds are seen to form at similar pressure levels as those probed by the CH4–H2O band, possibly explaining why this band shows more variation than the J and H bands. [Manjavacas et al. 2022]

Original astrobite edited by William Balmer.

About the author, Lili Alderson:

Lili Alderson is a second-year PhD student at the University of Bristol studying exoplanet atmospheres with space-based telescopes. She spent her undergrad at the University of Southampton with a year in research at the Center for Astrophysics | Harvard-Smithsonian. When not thinking about exoplanets, Lili enjoys ballet, film, and baking.

illustration of the hubble and gaia spacecraft working together

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: GaiaHub: A Method for Combining Data from the Gaia and Hubble Space Telescopes to Derive Improved Proper Motions for Faint Stars
Authors: Andrés del Pino et al.
First Author’s Institution: Center for Studies of Astrophysics and Cosmology of Aragón (CEFCA) and Space Telescope Science Institute
Status: Published in ApJ

Intro: What Is Proper Motion and How to Find It

Stars in the night sky seem fixed, but they are all traveling through the Milky Way just like the Sun. Since objects in the universe travel in 3D space, we can separate their velocities into three components, as shown in Figure 1: one component is the radial velocity, which points towards or away from Earth, and the other two components come from the proper motion, which refers to motion in the plane of the sky. Radial velocity is usually measured by finding the redshift of the object’s spectral lines, and it can reach down to several kilometers per second in accuracy. Proper motion is much harder to measure.

diagram illustrating radial velocity and proper motion

Figure 1: An illustration of radial velocity and proper motion. [ESA/ATG Medialab]

The measurement of accurate sky positions is called astrometry. Proper motion measurement relies on astrometry, since we are comparing observations from two epochs and calculating how much the position of the star has changed. This is the strong suit of the Gaia mission, which probes stars out to the halo of the Milky Way galaxy (see this astrobite). Gaia has led to many discoveries: new globular clusters in the Milky Way, stars moving so fast enough to escape the Milky Way, groups of stars that move together, and plenty more to come.

However, Gaia data have two important shortcomings. Firstly, it is a small telescope and works better for bright stars. For faint stars, the astrometric errors rise rapidly. But if we are interested in a faraway system (e.g., a satellite dwarf galaxy of the Milky Way), all the stars will be faint. Using Gaia data alone, the velocity errors far exceed the true variation in the galaxy. The second issue is the time baseline. Given a constant velocity, stars will shift more if you wait longer between two observations. That is why the time baseline has a huge impact on proper motion accuracy. Gaia has only been operating and recording positions for three years. If there is a way to increase the time baseline, that can also improve the proper motion measurements.

How to Measure Proper Motions Better?

Prior to the launch of the Gaia space telescope, the workhorse in astrometry studies was the Hubble Space Telescope. Hubble data can solve both of the issues mentioned above; it can observe much fainter stars, and it’s been taking data since 10–15 years before Gaia was even launched. If there is a way to combine these datasets, the time baseline for the proper motion measurements could be extended by a factor of 4–6. As shown in Figure 2, even adding one Hubble image can push down the errors by a lot for faint stars (G magnitude > 17). That is precisely what the authors of today’s article did.

plot of proper motion uncertainties for gaia data alone versus gaia and hubble data combined

Figure 2: The expected proper motion uncertainties as a function of the magnitude of stars. In both panels, nominal errors of Gaia Early Data Release 3 are shown by a black dashed curve. The top panel shows the impact of using one or more Hubble images, taken at the same epoch, June 2011. The bottom panel shows the impact of using just one Hubble image taken on different years (the typical baseline found in the data is ~11 years). [del Pino et al. 2022]

Combining Hubble and Gaia

The authors of today’s article developed a software called GaiaHub, which compares the positions of the stars measured with Gaia with those measured with Hubble. The first step is to measure positions of stars in Hubble data. This is a well-established process that takes into account the instrument distortions and time variations, and it achieves a typical accuracy of 0.25–0.5 milliarcseconds.

Then comes the hard part: the star positions need to be matched to Gaia measurements. Since the two datasets are more than 10 years apart, establishing a common reference frame between the two is the key challenge. The software offers three different algorithms: when there is a large number of randomly moving stars, it matches the average positions of all stars; when the stars have some coherent motion, the proper motion can be modeled iteratively so that the coherent motion is removed; or, finally, if there are many contaminant stars, the code can also set up the reference frame from co-moving stars. The improved accuracy with Gaia and Hubble data can be seen in Figure 2 as a function of the magnitude of the stars.

Results

So how does this software perform on real data? Figure 3 shows the drastic improvement you get from combining Gaia and Hubble data. In this example, proper motions are used to identify member stars of a globular cluster Palomar 4. The stars in a cluster should move together, which means they should all have similar proper motions. The left column in Figure 3 shows the proper motion in as a function of on-the-sky coordinates, right ascension (RA) and declination (Dec), measured by Gaia alone (top panel) and GaiaHub (bottom panel). The proper motion measurements from GaiaHub clearly have much smaller scatter and allow for a cleaner selection of member stars. This is confirmed by the right column, which shows the sky positions of the selected stars and their proper motion vectors. In the Gaia selection, the lines indicating the direction of motion point all over the place, while GaiaHub results show very coherent motion. Given that member stars should move together, GaiaHub successfully picks out the likely members of Palomar 4.

Comparison between the results obtained using Gaia and GaiaHub for the Palomar 4 globular cluster.

Figure 3: Comparison between the results obtained using Gaia and GaiaHub for the Palomar 4 globular cluster. Left column: proper motion in RA vs. Dec, measured by Gaia (top panel) and GaiaHub (bottom panel). Right column: sky positions of the stars with the projected proper motion vectors. [del Pino et al. 2022]

This is a huge improvement for proper motion measurements! Large uncertainties in proper motion mean that we get more scatter in the velocities, and that leads to artificially large velocity dispersion measurements for globular clusters. With GaiaHub’s new capabilities, the artificial scatter is reduced and we can recover the real internal velocity dispersions. The authors did this exercise for 40 globular clusters and published their results in this article. Along with radial velocities, we now have the full 3D velocity information. GaiaHub opens exciting new science involving analyzing velocity dispersions along each direction.

As with all research techniques, GaiaHub has its limitations. Due to the cross matching, GaiaHub relies on stars that overlap in both datasets. That means the field of view is limited by the smaller of the two, which is Hubble. The magnitude of the stars that can be detected by both telescopes is also limited, since bright stars are often saturated in Hubble images. Both of these factors mean that GaiaHub works best at an intermediate distance, where the Hubble field of view is large enough to cover the globular cluster and the brightest stars are faint enough to not be saturated.

To summarize, GaiaHub improves the proper motion measurements by a factor of ten. More precise proper motions at fainter magnitudes allow us to study the kinematics of many stellar systems around the Milky Way. This public software will be a great resource for the astronomy community!

Original astrobite edited by Katya Gozman.

About the author, Zili Shen:

Hi! I am a PhD student in Astronomy at Yale University. My research focuses on ultra-diffuse galaxies and their globular cluster populations. Since I came to Yale, I have worked on two “dark-matter-free” galaxies NGC1052-DF2 and DF4. I have been coping with the pandemic and working from home by making sourdough bread and baking various cookies and cakes, reading books ranging from philosophy to virology, going on daily hikes or runs, and watching too many TV shows.

photograph of the green bank telescope in front of rolling mountains

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: Searching for Broadband Pulsed Beacons from 1883 Stars Using Neural Networks
Authors: Vishal Gajjar et al.
First Author’s Institution: Breakthrough Listen, University of California Berkeley
Status: Published in ApJ

The search for extraterrestrial intelligence (SETI) is perhaps humankind’s most ambitious and forward-thinking endeavor. We’ve been asking ourselves the fundamental question of “Are we alone?” since the dawn of written history, but technological advancements in the last 100 years have allowed us to take our first steps toward finding an answer. Today’s article describes a reimagination of one of the most common search techniques to look for signatures of extraterrestrials (ETs), and while we haven’t found any alien signals just yet, our search capabilities only continue to get better!

The easiest way to find ETs would be to look for their technosignatures — the light waves emitted by the technology they use (check out this Astrobite for more on technosignatures). In particular, if an alien civilization wanted to be found by other intelligent life, they would want to send out a signal that wouldn’t be deflected or absorbed by the space between us, would travel as fast as possible, and would require the least amount of energy to produce. For these and other reasons, most SETI searches have involved searching for artificial radio signals coming from the vicinity of nearby stars.

But what specific kinds of signals should we search for? Will they be transmitted over a narrow frequency range, or will they be “broadband” signals covering a large range of frequencies? Will the signal be continuously transmitted, or will it pulse on and off at specific intervals to clearly demonstrate it’s made by intelligent life? There is no one satisfactory answer to these questions, and most previous searches have looked for narrowband signals that are always being emitted, since we would need less time to detect a signal of that type than other types of signals.

However, the authors of today’s article were able to prove that for a civilization generating these signals, it would cost less energy to produce broadband pulsed signals, as long as those signals were being sent out for longer than a few hundred seconds. They made the reasonable assumption that ETs will try for longer than a few minutes to get our attention and went about searching for broadband pulsed signals in radio data from the Green Bank Telescope.

A Very Small Needle in a Turbulent Haystack

The Breakthrough Listen collaboration, of which many authors of this article are a part, chose 1,883 stars (explained in this article) as targets for their observations. They chose every star within 5 parsecs (a little more than 16 light-years) of Earth — so that the distance between us would not attenuate the signals too much — as well as all stars within 5–50 parsecs (163 light-years) that fall on the main sequence or the early part of the giant branch. Stars on these earlier segments of the stellar evolutionary track are less volatile and, if they have planets orbiting them, create environments that are the most likely to aid life to grow. The authors took 233 total hours of observations, broken up into 5-minute segments, since that is approximately the observational length for which a 0.3-millisecond long broadband pulse would take less power to send than a continuous narrowband signal.

Luckily, we have lots of experience searching for repeating broadband radio pulses in the form of radio pulsars and fast radio bursts! Pulsars are useful physical tools for a wide range of astronomical applications (for more, see the astrobites here, here, here, here, here, here, here, and here), but today, we can use our experience in analyzing transient radio signals to predict how a broadband signal sent by ETs would be affected by the interstellar medium between us. Radio waves are scattered and dispersed by the interstellar medium, and broadband radio signals undergo a dispersion delay, where the lower-frequency part of a pulse will be delayed relative to the higher-frequency part due to the ionized medium it travels through. The authors of today’s article focus on this dispersion delay.

plot of a dispersed broadband pulse signal

Figure 1: The received signal from a dispersed broadband pulse, as a function of frequency and time. Note that the higher-frequency parts of the signal arrive before the lower-frequency parts. [Gajjar et al. 2022]

The “waterfall” plot in Figure 1 shows the intensity as a function of frequency and time for a single broadband pulse that has undergone dispersion. The dispersion measure of a signal, which is related to the time delay between two reference frequencies, can help us measure the amount of ionized material a signal has traveled through. Combined with detailed maps of the Milky Way, we can use the dispersion measure to estimate the distance between us and the origin of the signal!

Most importantly, the dispersion delay time between two frequencies always scales as the inverse of the frequency difference, squared. The authors of today’s article suggest that if an alien civilization were to send us a signal, the best strategy would be to artificially arrange it in some way so that we would not see a normal dispersion trend; rather, we would see some other pattern that does not occur in nature, proving that it comes from other sentient life.

These other types of artificial dispersion are shown in Figure 2. The authors searched for dispersed pulses from their original dataset, and they also created artificial datasets by flipping the frequency and time axes, both independently and simultaneously. By doing this, each type of artificially dispersed pulse shown in Figure 2 would look to the single-pulse-search software as a normally dispersed pulse, allowing the team to run the same search code on all four datasets. Searching all of these datasets resulted in a staggering 133,393 candidates!

plots of artificially dispersed signals

Figure 2: The three types of artificially dispersed signals that the authors searched for. From left to right, they are made by flipping the time axis, the frequency axis, and both simultaneously, to make artificially dispersed broadband pulses that are not seen in nature. [Gajjar et al. 2022]

How to Analyze 133,000 Candidates This Century

Of course, having a human sit down and examine that many candidates would be beyond unreasonable — thankfully, machine learning and graphics processing units allow us to quickly filter out many bad options. The authors filtered out candidates that looked too much like human-made radio frequency interference or didn’t show enough of a difference between their on-pulse and off-pulse energy distributions. Many other filters were used to weed out unpromising candidates, leading to a shortlist of only 2,948 candidates.

The best candidates in each class of artificial dispersion were examined more closely, but similar-looking signals were found in other 5-minute-long pointings in completely unrelated areas of the sky. It’s not easy for us to prove definitively that these signals actually come from the region of the stars we’re pointing at, rather than human-made radio frequency interference; it’s much more reasonable to conclude that these “signals” are bright human-made signals that have made their way into the telescope, rather than two extremely distant alien civilizations sending us the exact same signal.

The authors used these non-detections to place limits on the maximum signal strength any civilization in those areas could be sending, finding some signals as weak as a few hundred times stronger than our strongest airplane radar. That doesn’t sound like much of a limit, but that’s a signal we’d be detecting from a whole other star system — and each new search is another step towards better technology and better search methods to make a possible discovery!

Original astrobite edited by Lili Alderson.

About the author, Evan Lewis:

Evan is a third-year graduate student in astronomy at West Virginia University. His research focuses on transient radio sources, including pulsars, magnetars, and fast radio bursts. Outside of research, he enjoys playing percussion, hugging dogs, baking, and playing video games!

1 … 15 16 17 18 19 … 48