Features RSS

image of the Sun with clouds crossing the Sun's disk

Even in the sunny spots where solar telescopes are stationed, clouds are a fact of life. Rejecting cloud-obscured solar images means losing a lot of data, but new machine-learning methods show that we don’t have to cast off cloudy images after all.

Contending with Clouds

Dunn Solar Telescope

The Richard B. Dunn Solar Telescope in Sunspot, New Mexico. Even this sunny site experiences clouds. [NSF/NSO/AURA; CC BY 4.0]

Space-based solar telescopes have the luxury of perfect “weather” every day, but ground-based solar observatories have to contend with clouds. A study undertaken at Udaipur Solar Observatory in India, for example, found that the Sun’s disk was unimpeded by clouds only 63% of the time. If clouds can’t be removed reliably from images taken the remaining 37% of the time, that means losing a ton of information about the Sun. Is there a way to salvage these images?

So far, existing methods of handling cloud-contaminated images fall short, either failing to remove the clouds fully or failing to reconstruct the obscured solar features underneath. Now, a study led by Zhenhong Shang (Kunming University of Science and Technology) has shown how a new machine learning method rescues clouded-out images of the Sun’s disk.

Machine Learning Method

The model developed by Shang’s team handles the two critical aspects of cloud removal: 1) identifying and characterizing cloudy images and 2) removing the clouds and recovering the features underneath.

examples of clear and cloudy solar images

Examples of (a) clear, (b) mildly cloudy, (c) moderately cloudy, and (d) severely cloudy images from various telescopes in the GONG network. [Shang et al. 2025]

First, the team’s method reports not only whether an image is cloudy but also how cloudy it is. This is an improvement over previous methods that returned only a yes or no to the question of cloud cover. Previous methods that searched only for variations in brightness across the solar disk could also be fooled by widespread cloud cover, whereas the new method also takes in information about the median brightness of the disk, identifying images that are nearly completely cloud covered.

For the second step of the procedure, Shang’s team used data from the Global Oscillation Network Group (GONG), six solar telescopes that maintain eyes on the Sun every hour of every day. The team selected mildly, moderately, and severely cloud-covered images from this dataset. To train the machine learning model to remove clouds, they used 760 pairs of clear and cloudy images from the GONG telescopes. When cloud cover is intermittent, these clear–cloudy pairs are constructed from images taken by the same telescope. When cloud cover is persistent, a cloudy image from one observatory is paired with a clear image from another observatory, taken just seconds later. This dataset was also augmented by resizing and rotating some of the images, making sure it can handle even more cloud-cover scenarios.

examples of cloud removal model performance

Examples of the model’s cloud-removal and feature-retrieval performance. This particular test used data from solar telescopes not in the GONG network. [Shang et al. 2025]

Clear Skies Ahead

After testing and validating their method against a separate set of GONG images, Shang’s team compared the model’s performance to that of existing methods. On nearly every quantitative and qualitative metric, the new machine learning method outperformed existing methods. The new method showed particular improvements in recovering details of the solar surface under dense cloud cover.

The team also demonstrated the method’s performance on data from other ground-based solar observatories, showing its generalizability and readiness to tackle the challenge of cloud removal.

Citation

“Cloud Removal in Full-Disk Solar Images Using Deep Learning,” Zhenhong Shang et al 2025 ApJS 276 56. doi:10.3847/1538-4365/ad93ca

Dimorphos

final full image from DART

The final full image from DART, taken 2 seconds before the spacecraft crashed into the asteroid moonlet Dimorphos. [NASA/Johns Hopkins APL]

In 2022, humanity performed its first test of its ability to defend itself against an oncoming asteroid impact. When NASA’s Dual Asteroid Redirection Test (DART) spacecraft slammed into Dimorphos, the 160-meter-wide moonlet of the asteroid Didymos, a spray of material escaped the moonlet’s surface, and Dimorphos’s path around its parent asteroid was altered — showing how a 610-kilogram spacecraft can change the course of a 4.8-billion-kilogram asteroid. Both the spray of ejecta and the change in orbit reflect the transfer of momentum from the spacecraft, which came in at a screaming 22,530 kilometers per hour.

Could an impact from a spacecraft like DART actually nudge a hazardous asteroid onto a safer course? The timeliness of this question recently ratcheted up a notch: the probability of the asteroid 2024 YR4 striking Earth on 22 December 2032 has risen above 2%.

Discovery images of asteroid 2024 YR4

Discovery images of 2024 YR4. [ATLAS]

Answering this question isn’t straightforward. Among the factors that influence the outcome of a mission like DART — or a future mission to reshape the trajectory of a hazardous asteroid — is the target asteroid’s structure and material strength. Today, we’re taking a look at three research articles that tackle simulations of an impactor striking an asteroid to understand the role of asteroid structure and strength, providing results relevant to both planetary defense and the study of asteroid properties.

Subsurface Structure and Impact Outcomes

First up, a team led by Kaiyi Dai (Macau University of Science and Technology; Imperial College London) and Xi-Zi Luo (Macau University of Science and Technology) simulated the outcome of an impactor striking an asteroid at an angle, just as DART did, rather than head on. Their goal was to understand how the presence of boulders — like those expected to make up “rubble-pile” asteroids like Didymos and Dimorphos — affects the transfer of momentum from the impactor to the target asteroid. Specifically, the team used a three-dimensional shock physics code to explore the effect of different subsurface arrangements of boulders.

Plots of simulated momentum transfer efficiency

Momentum transfer efficiency as a function of the impact angle for different boulder configurations. Click to enlarge. [Adapted from Dai et al. 2024]

In simulations that contained no boulders, Dai and Luo found that head-on collisions excavated the largest amount of ejecta and transferred momentum most efficiently, but oblique impacts launched the fastest-moving ejecta. The addition of boulders into the simulations altered the momentum of the ejected material: surface boulders provided an “armoring” effect, reducing the transfer of momentum and resulting in a smaller crater, whereas buried boulders had an “anti-armoring” effect, increasing the efficiency of momentum transfer and carving out a larger crater. The more deeply buried the boulders, the less they affected the outcome.

An impactor striking a rubble-pile asteroid doesn’t just dislodge material: it also melts some of it. The simulations showed that the amount of melted material, as well as whether the melted material remained on the surface or was ejected into space, varied substantially with the angle of the impact and the configuration of subsurface boulders. This suggests that the volume of melted material remaining on an asteroid’s surface could be used to discern the structure of the material beneath the surface.

Dai and Luo’s team noted that applying these simulation results to the outcome of the DART mission will require simulating a more realistic impactor, as the DART spacecraft’s shape and composition is more complicated than that of the spherical projectile used in these simulations.

Impact of Asteroid Strength and Porosity

Mallory DeCoster (Johns Hopkins University Applied Physics Laboratory) and collaborators also investigated the outcomes of high-velocity impacts on rubble-pile asteroids. The team focused on the effects of not only the boulders at the impact site, but also the “matrix material” of finer particles between the boulders. Because this matrix material effectively holds a rubble-pile asteroid together, the strength of the material plays a large role in determining the outcome of an impact.

Examples of matrix, large boulder, and small boulder configurations

Examples of matrix, large boulder, and small boulder configurations. Matrix material is shown as gray, the impactor is the small gray rectangle, and boulders are shown in different colors (the color of the boulder does not have any meaning). [Adapted from DeCoster et al. 2024]

The team carried out 52 two-dimensional shock physics simulations of impacts onto an asteroid that is 163 meters across — roughly the size of Dimorphos. The properties of the impactor were tuned to be similar to the DART spacecraft: a porous aluminum cylinder 2 meters long and 0.5 meter wide with a mass of 501 kilograms and an impact velocity of 6.6 kilometers per second. The porosity of the simulated cylinder reflects the fact that spacecraft aren’t solid, and this porosity changes the way the impactor interacts with the target.

DeCoster and collaborators placed boulders with diameters of 1–10 meters randomly within their simulated asteroids, filling 62% of each asteroid with boulders before topping off the rest of the volume with matrix material. The simulated asteroids were categorized as “matrix,” “large boulders,” and “small boulders” depending on the primary makeup of the impact site. The strength and porosity of the matrix material were varied between simulations.

The simulations showed that more porous matrix material resulted in larger impact craters, but this outcome was moderated by the distribution of boulders on and under the surface; for matrix-rich asteroids, the presence of weaker, more porous material boosted the crater width, while asteroids with many large or small boulders showed smaller differences in crater width with matrix strength and porosity. None of these factors particularly influenced the depth of the crater.

Matrix-rich asteroids tended to experience larger momentum transfer than boulder-rich asteroids, but this finding was complicated by certain boulder-rich simulations in which entire boulders were thrown into space, resulting in a large momentum transfer for those simulation runs. These results highlight the complex factors that determine how an asteroid responds to an impact.

Estimating Cratering on Dimorphos

The final of today’s three research articles focuses on simulations of the DART spacecraft impacting Dimorphos. After DART slammed into the asteroid, the Italian Space Agency’s Light Italian CubeSat for Imaging of Asteroids (LICIACube) surveyed the aftermath from a safe distance. LICIACube imaging showed fantastic sprays of ejecta and boulders tumbling through space, but the precise size of the crater excavated by DART’s impact is still unknown. That will change with the arrival of the European Space Agency’s Hera mission, which launched in October 2024. Upon reaching Dimorphos and Didymos in late 2026, Hera will assess the impact site and measure the size of the impact crater.

DART observations of Dimorphos and a simulated target constructed to match the observations

DART observations of Dimorphos before impact (panels a and b). Panels c and d show the rubble-pile target constructed to match the DART observations. Click to enlarge. [Stickle et al. 2025]

While they wait for Hera to finish its journey, scientists are fine-tuning their simulations. Angela Stickle (Johns Hopkins University Applied Physics Laboratory) and collaborators used information gleaned from the DART mission — namely, the change in Dimorphos’s orbit, the characteristics of the ejecta, and the appearance of the asteroid’s surface — to predict the size of the resulting crater.

Stickle’s team modeled Dimorphos as a collection of boulders held together by matrix material, with the sizes and positions of the surface boulders determined from pre-impact photographs of the asteroid. The sizes of the subsurface boulders were drawn from the same size–frequency distribution as observed for the surface boulders, and these boulders were given random non-spherical shapes. The impactor was a DART-like conglomeration of three spheres, with the large central sphere representing the body of the spacecraft and two smaller, lower-mass side spheres representing the spacecraft’s solar panels.

LICIACube observations of ejecta compared to simulated ejecta

Comparison of LICIACube observations of ejecta rays (panel a) to simulated ejecta (panels b through d). Click to enlarge. [Stickle et al. 2025]

Using three shock physics codes, Stickle’s team examined the transfer of momentum, the creation of a crater, and the production of filamentary ejecta rays like those seen by LICIACube. A rubble-rich structure seems to be necessary to create dramatic ejecta rays, supporting the classification of Dimorphos as a rubble-pile asteroid. The transfer of momentum — a marker of the ability of an impactor to deflect the asteroid from its initial trajectory — depended strongly on the material properties of the asteroid: less-cohesive asteroid material with a larger resistance to crushing correlated with larger momentum transfer.

Estimated crater diameters ranged from 15 to 60 meters, with the likeliest size for Dimorphos being in the range of 40–60 meters. The team notes that this calculation is complicated, as crater sizes continue to evolve beyond the end of the simulations in a way that depends on the strength of the material and the gravity of the asteroid.

The upcoming measurements from the Hera mission will provide an important test of these model results, providing further information on the outcome of the DART mission and helping to advance our understanding of humanity’s ability to deflect hazardous asteroids.

Citation

“Impact Momentum Transfer — Insights from Numerical Simulation of Impacts on Large Boulders of Asteroids,” Kaiyi Dai et al 2024 Planet. Sci. J. 5 214. doi:10.3847/PSJ/ad72eb

“Statistical Analysis of Near-Surface Structure and Material Properties on Momentum Transfer in Rubble Pile Targets Impacted by Kinetic Impactors,” Mallory E. DeCoster et al 2024 Planet. Sci. J. 5 244. doi:10.3847/PSJ/ad7cff

“Dimorphos’s Material Properties and Estimates of Crater Size from the DART Impact,” Angela M. Stickle et al 2025 Planet. Sci. J. 6 38. doi:10.3847/PSJ/ad944d

Trapezium Cluster

Silhouetted against the background glow of the Orion Nebula, the protoplanetary disk 114–426 provides an excellent opportunity to study a site of planet formation. The discovery of water ice in this disk suggests that ice can survive in disks in the dense star cluster environments where most stars form.

Clues to Our Planetary Past

artist's impression of a protoplanetary disk

Artist’s impression of a protoplanetary disk around a young star. [A. Angelich (NRAO/AUI/NSF)/ALMA (ESO/NAOJ/NRAO); CC BY 4.0]

To understand how our solar system came to be, researchers can search for clues close to home, piecing together the compositional and dynamical hints that persist today — or they can look farther afield, taking cues from distant young planetary systems in the throes of formation.

Protoplanetary disks — disks of gas and dust that encircle young stars and set the stage for planet formation — provide an opportunity to understand how our solar system may have formed. In today’s article, JWST provides a new perspective on an unusual protoplanetary disk.

A Disk’s Silhouette

The Orion Nebula is a prolific star-forming region that’s home to hundreds of young stars. At the nebula’s center is the Trapezium Cluster, a tight-knit grouping of massive stars that illuminates the surrounding nebula. Located near the Trapezium Cluster, the protoplanetary disk 114–426 (named according to its coordinates) is notable for both its size and its location. At more than 1,000 au across — about 25 times the distance from the Sun to Pluto — it’s one of the largest disks in the region. It’s also oriented edge-on from our perspective, showing off a narrow disk of gas and dust that is illuminated from behind by the bright glow of the Orion Nebula.

protoplanetary disk 114–426

JWST image of the protoplanetary disk 114–426 in Orion. In this image, 1.15- and 1.40-micron light is blue, 1.62-micron light is green, 1.82-micron light is yellow-green, and 2.77-micron light is red. [Adapted from Ballering et al. 2025]

Previous observations of 114–426 with the Hubble Space Telescope and the Atacama Large Millimeter/submillimeter Array (ALMA) allowed for detailed studies of the dust grains in the disk. Now, Nicholas Ballering (Space Science Institute and University of Virginia) and collaborators have used JWST’s Near Infrared Camera to scan the disk in 12 wavelength bands from 1 to 5 microns (1 micron = 10-6 meter).

The JWST observations showcase 114–426’s edge-on, dusty disk and two bright lobes of scattered light from the hidden star at the disk’s center. The two lobes are asymmetrical, which previous research suggests means that the inner part of the disk is tilted. A tilted disk could mean that the hidden central star is actually a binary system, or that the star is orbited by a massive planet.

Icy Interpretation

spectra of the disk 114–426

Spectra from several locations in the disk. Each spectrum shows a dip at 3 microns, indicative of water ice. Click to enlarge. [Ballering et al. 2025]

Looking at the portions of the disk that extend beyond the lobes of scattered light, Ballering’s team identified several wavelengths where the spectrum dips, indicating absorption of the background light. One such feature around 3 microns is due to water ice, which has been identified in other protoplanetary disks.

The team estimated that about half an Earth mass of ice and dust is present in the areas of the disk that are backlit by the Orion Nebula. Given that this disk is subjected to intense ultraviolet radiation from the nearby Trapezium Cluster, it may seem surprising that water ice can survive there. However, the team’s calculations showed that the disk is far too cold for ice to sublimate, and the ultraviolet radiation from nearby stars isn’t sufficient to remove the ice from the surface of dust grains.

This finding suggests that water ice is likely to survive in disks around stars in cluster environments like this one. The fact that ice persists is important for planet formation, as ice helps dust grains clump together into pebbles and eventually planets. Ice may also play a role in the transport of water from the cold outer regions of disks to the temperate inner regions where habitable planets may reside.

Citation

“Water Ice in the Edge-On Orion Silhouette Disk 114–426 from JWST NIRCam Images,” Nicholas P. Ballering et al 2025 ApJ 979 110. doi:10.3847/1538-4357/ad9b7a

galaxy merger

Over the course of its lifetime, the Milky Way has encountered many other galaxies — from quick flybys to full on collisions, each interaction is important to our galaxy’s evolution. A series of simulations has predicted how galaxy interactions have contributed to star formation across the lifetimes of Milky Way–mass galaxies. 

Merging Galaxies

As galaxies move through the universe, they interact and merge with other galaxies. Sometimes these interactions are quick flybys, leaving behind little evidence of the encounter. Other times, galaxies smash into each other, creating extended tails and messy morphologies. From observations and simulations, scientists have suggested that the merging of galaxies can also trigger increased star formation. However, it is unclear if this behavior is consistent for galaxies across cosmic time. 

In the local universe, observations have shown increased star formation rates for galaxies with nearby neighbors and lower star formation rates for galaxies that are more isolated, likely experiencing very few interactions and mergers. Earlier in the universe, mergers were more common as galaxies were closer together; however, directly observing evidence of galaxy mergers from the distant past is difficult, leading to an incomplete understanding of how ancient mergers impacted star formation rates. Fortunately, simulations can explore where our telescopes can’t quite reach, allowing scientists to track galaxy interactions over time. 

Galaxy Interaction Simulations

Simulation snapshot showing the gas (left) and stellar (right) surface density maps of a major merger event. Click to enlarge. [Li et al 2025]

In order to explore the impact galaxy mergers and interactions have on the star formation activity in Milky Way–mass galaxies, Fei Li (University of Toronto) and collaborators performed simulations using the Feedback in Realistic Environments (FIRE) cosmological simulation suite. In these simulations, the authors track major, minor, and mini interactions that a Milky Way–mass galaxy experiences over ~12 billion years. Over this time, the galaxy’s interaction history and star formation rate is monitored, allowing the authors to explore how interactions may play a role in the galaxy’s star formation. 

torque star formation rate comparison

Torque (blue) and star formation rate (orange) measurements during major merger interaction snapshots from the simulations ran in this study. Not all peaks in star formation rate correspond to peaks in torque, and not all peaks in torque cause a peak in star formation rate. Click to enlarge. [Li et al 2025]

The simulations show that in major interactions, where the interacting galaxies are closer in mass, there is a positive correlation between the torque exerted by the companion and the star formation rate — increased torque often causes a burst of star formation. However, the star formation history in the galaxy shows multiple starbursts, and most often, these starbursts are unrelated to any sort of merger or interaction with another galaxy. In the minor and mini interactions, there is no significant relationship between the torque and star formation rate for the galaxy. While major mergers can trigger some starburst activity, the overall pattern in star formation for Milky Way–mass galaxies is independent of interaction history — less stellar mass created early on can be directly attributed to mergers than previously assumed.

Observation and Simulation Comparison

How do the authors’ results compare to other simulations and observations that explore interaction impacts? Observations that compare the star formation activity of merging and non-merging galaxies find no significant difference between the star formation activity in these two groups of galaxies across the last ~12 billion years, which is consistent with the results of this simulation. Other simulations predict that massive galaxies like the Milky Way go from bursty to steady star formation across cosmic time, and that mergers and interactions do not leave a significant mark on this overall trend of star formation activity. 

This study finds that Milky Way-mass galaxies experience most major interactions during their early years, experience only a few major mergers, and their star formation activity across cosmic time is predominantly driven by internal dynamics within the galaxy itself. These findings provide predictions for what we may observe as advanced telescopes continue to peer into the distant past.

Citation

“The Effect of Galaxy Interactions on Starbursts in Milky Way-mass Galaxies in FIRE Simulations,” Fei Li et al 2025 ApJ 979 7. doi:10.3847/1538-4357/ad94ef

Hubble Space Telescope image of stars in the Milky Way's galactic bulge

Regulus, the brightest star in the constellation Leo, is known to float through space with three companion stars. New research shows that it may have yet another member in its stellar entourage.

The Growing Regulus Family

constellation Leo

The constellation Leo, showing the location of Regulus. Click to enlarge. [IAU/Sky & Telescope; CC BY 4.0]

Looking at a field of stars, it’s impossible to know at a glance which stars are associated with one another. Astronomers discern the connections between stars through careful measurements of stars’ distances and motions, as well as similarities in age or metal content.

Using these techniques, astronomers have discovered that Regulus, the nearest B-type star to Earth, has three companions. First to be identified was Regulus B, a 0.8-solar-mass K-type star. Regulus B itself was discovered to have a companion, a 0.32-solar-mass M-type star named Regulus C, in 1873.

In 2008, technological advances allowed for the discovery of an even smaller companion: a 0.31-solar-mass stellar core that sweeps around Regulus every 40 days. Although this companion has yet to be imaged directly, it was detected spectroscopically only a few years ago.

Now, researchers may have tracked down Regulus’s smallest companion yet.

A Potential Companion

SDSS J100711.74 +193056.2, or SDSS J1007+1930 for short, is a brown dwarf in Regulus’s vicinity. Brown dwarfs lie in between stars and planets, massive enough to temporarily experience fusion of deuterium and sometimes lithium, but not massive enough to initiate hydrogen fusion. SDSS J1007+1930’s mass is estimated to be just 0.06 solar mass. The brown dwarf sits 12.6 light-years from Regulus, and its proper motion across the sky is similar to that of Regulus, suggesting that the two objects might be associated.

near-infrared spectrum of SDSS J1007+1930

A near-infrared spectrum of SDSS J1007+1930 (gray is the high-resolution spectrum, black is the smoothed spectrum) along with the spectrum of an L9-type brown dwarf standard (purple). Click to enlarge. [Adapted from Mamajek & Burgasser 2025]

To explore this possibility, Eric Mamajek (NASA’s Jet Propulsion Laboratory) and Adam Burgasser (University of California San Diego) collected spectra of SDSS J1007+1930 using the 10-meter Keck II telescope. Their aim was to determine if SDSS J1007+1930’s properties were similar to those of Regulus or any of the stars in its retinue.

Assembling the Clues

Mamajek and Burgasser found that SDSS J1007+1930 bears a close resemblance to members of the Regulus system in several ways. The brown dwarf is less metal-rich than the Sun, with a metallicity similar to that of Regulus B. Its radial velocity is also similar to that of Regulus B and Regulus itself. Finally, spectral analysis suggests that SDSS J1007+1930 is unlikely to be young, and its age may instead be similar to the 1–2 billion year age of the system.

These clues leave open the possibility that SDSS J1007+1930 is associated with Regulus — but is it? So far, the available evidence isn’t conclusive, but it does suggest that SDSS J1007+1930 formed either within the Regulus system or in the same natal star cluster. Even if SDSS J1007+1930 is currently linked to the Regulus system, the association might be short lived. Given the brown dwarf’s distance from Regulus, it’s possible that the object is no longer gravitationally bound to the system, and a future interaction with a passing star or even a massive gas cloud could steal the brown dwarf away.

Citation

“SDSS J100711.74+193056.2: A Candidate Common Motion Substellar Companion to the Nearest B-Type Star Regulus,” Eric E. Mamajek and Adam J. Burgasser 2025 AJ 169 77. doi:10.3847/1538-3881/ad991b

Artist's impression of a pulsar

Neutron star crust is the strongest material in the universe, but it’s not infinitely strong. New research explores how the cracking of a neutron star’s crust might determine how fast these extreme stellar remnants can spin.

Stellar Speed Limit

Neutron stars are the remnant cores of massive stars that have expired as core-collapse supernovae. These stars are extraordinarily compact, typically packing more than the mass of the Sun into a sphere with a volume about a billion times smaller than Earth’s volume.

Neutron stars also rotate extremely quickly, with the fastest-spinning neutron stars zipping around hundreds of times each second. That’s fast — but it turns out to be only about half as fast as a neutron star could hypothetically spin before being ripped apart by its rotation, a speed known as the breakup rate. What’s stopping neutron stars from spinning even faster?

Crust-Covered Pasta

Jorge Morales and Charles Horowitz (Indiana University) have explored one possibility: that the cracking of a neutron star’s crust occurs at roughly half the star’s breakup rate, and that once the crust cracks, the star can spin no faster.

Neutron stars are composed of a strange substance called nuclear pasta, which is enclosed by a shell of atomic nuclei and electrons. Both of these layers are reportedly billions of times stronger than steel.

plot of strain versus rotation rate

The strain experienced by a neutron star’s crust as a function of the rotation rate divided by the breakup rate. [Morales & Horowitz 2025]

Morales and Horowitz modeled the strain that a neutron star’s crust encounters under different spin rates and neutron star masses. They showed that as a neutron star spins, the material around its equator bulges outward while the material near its poles draws inward. When the strain of this deformation surpasses the crust’s strength, the crust splits along the star’s equator, at the base of the crust where the crust meets the interior nuclear pasta. For reasonable estimates of the crust’s breaking strength, this happens at 58% of the star’s breakup rate — roughly the maximum spin speed drawn from observations.

Continuous Gravitational Waves

Why does the breaking of a neutron star’s crust inhibit faster rotation? This is especially relevant for neutron stars that are expected to “spin up” to faster speeds over time as they accrete matter from their companions; even these appear to obey the spin speed limit.

After the crust breaks, the pieces of the crust can shift around. This may transform the neutron star from one that is symmetric around its spin axis to one that is asymmetric. These objects might produce a continuous gravitational wave signal, and Morales and Horowitz proposed that any angular momentum added to the star — through accretion, for example — is carried away in the form of gravitational waves, preventing the star from spinning faster.

Morales and Horowitz noted that there are more aspects of this scenario to be explored, such as the impact of crustal magnetic fields, the importance of relativistic effects, and the prospects of detecting gravitational waves from these sources.

Citation

“Limiting Rotation Rate of Neutron Stars from Crust Breaking and Gravitational Waves,” J. A. Morales and C. J. Horowitz 2025 ApJL 978 L8. doi:10.3847/2041-8213/ad9ea7

cartoon of Thorne–Żytkow object formation

A Thorne–Żytkow object is a star within a star — a star with a neutron star at its core. These objects are theorized to form in close binary systems, but new research reveals complications in this proposed formation pathway.

A Star Within a Star

The term “star” encompasses a wide variety of objects, from our familiar Sun to roiling supergiants dozens of times as massive and hundreds of times as wide. Certain types of stars are only theorized, like those containing huge amounts of dark matter or with cores composed of strange quarks. One such theorized star — and the subject of today’s article — is a Thorne–Żytkow object, also known as a hybrid star.

illustration of an X-ray binary system

An artist’s impression of an X-ray binary, in which a compact object accretes material from a companion star and emits X-rays during intermittent outbursts. [ESO/L. Calçada; CC BY 4.0]

Thorne–Żytkow objects might form in close binary systems when a star engulfs its neutron-star companion. Because of their potential binary origin, it’s possible that Thorne–Żytkow objects arise in X-ray binary systems containing a neutron star that collects gas from its companion in a super-heated accretion disk.

After being engulfed by its companion, the neutron star is thought to sink to the star’s core. There, it is hypothesized to energize the surrounding star through accretion and nuclear fusion, creating a curious mix of elements that distinguishes a Thorne–Żytkow object from an ordinary star.

Diverging Paths

At least, that’s the theory. But as a recent research article led by Tenley Hutchinson-Smith (University of California, Santa Cruz; University of Copenhagen) shows, more work is needed to understand whether X-ray binary systems could truly evolve into Thorne–Żytkow objects. At the core of this question is how the inspiraling neutron star affects the companion that has engulfed it. Does the engulfing star hold on to its extended gaseous envelope, or does it lose its atmosphere and diverge from the Thorne–Żytkow object path? And how long would the Thorne–Żytkow phase last — could the neutron star remain at the center of its companion indefinitely, or does the neutron star eventually gain mass and collapse into a black hole?

The team used as the basis of their exploration the X-ray binary system LMC X-4, which contains a 1.57-solar-mass neutron star and an 18-solar-mass primary star. The stars are locked in a tight gravitational embrace, separated by only 14 solar radii and orbiting one another every 1.4 days.

simulation snapshots showing gas density

Simulation screenshots showing the density of gas as the neutron star (white circle) spirals in toward the core of its companion star. Click to enlarge. [Adapted from Hutchinson-Smith et al. 2024]

Total Collapse of the Heart

Using a three-dimensional fluid dynamics simulation, Hutchinson-Smith and collaborators followed the evolution of LMC X-4 as the primary star engulfed the neutron star. As the neutron star spiraled inward, the energy released ejected only a small amount of gas, and the neutron star accreted only a small amount of matter from the companion star. At that point, the formation of a Thorne–Żytkow object seemed inevitable, but the merger of the neutron star with the companion star’s core set the system on a different course.

plot of luminosity and duration of gamma-ray emission

Comparison of the luminosity and duration of the gamma-ray burst produced by the collapse of the neutron star in LMC X-4 with the properties of long gamma-ray bursts (LGRBs) and ultra-long gamma-ray bursts (ULGRBs). Click to enlarge. [Hutchinson-Smith et al. 2024]

As the neutron star melded with the companion’s core, it imparted angular momentum to the core. This created an accretion disk that fed the neutron star until it collapsed into a black hole. The collapse launched a relativistic jet and powered gamma-ray emission that was about as bright and as long-lasting as an ultra-long gamma-ray burst. Feedback from accretion onto the black hole ejected nearly all of the gaseous envelope, definitively halting the short-lived Thorne–Żytkow phase.

Thus, Hutchinson-Smith’s team has demonstrated that a Thorne–Żytkow object is unlikely to result from the evolution of an X-ray binary system like LMC X-4 — though this evolution may provide a path to powering ultra-long gamma-ray bursts. This suggests that accretion and feedback leading to the collapse of the neutron star and the ejection of the stellar envelope must be taken into consideration when exploring the formation of Thorne–Żytkow objects.

Citation

“Rethinking Thorne–Żytkow Object Formation: The Fate of X-Ray Binary LMC X-4 and Implications for Ultra-Long Gamma-Ray Bursts,” Tenley Hutchinson-Smith et al 2024 ApJ 977 196. doi:10.3847/1538-4357/ad88f3

Illustration of a Neptune-like exoplanet

With nearly 6,000 exoplanets discovered to date, it’s clear that not all of the planets in our galaxy resemble the planets in our solar system. Today’s Monthly Roundup explores three types of planets with no analogs in our solar system: a warm Neptune orbiting its tiny host star in less than 4 days; a polar Neptune orbiting perpendicular to its host star; and a famous sub-Neptune whose structure is a matter of debate.

Close-In Neptune, Breaker of Chains

LHS 3154b is a typical Neptune-mass exoplanet in an atypical place. This planet has a mass of at least 13.2 Earth masses, but it orbits an M-dwarf star just 11% the mass of the Sun, swinging around its tiny host star every 3.7 days. It’s not clear how such a setup came to be, as conventional theories of planet formation provide few avenues for such massive planets to form around low-mass stars.

Donald Liveoak and Sarah Millholland (Massachusetts Institute of Technology) have proposed an explanation for LHS 3154b: it’s several planets in a trench coat. In other words, rather than having formed as the single planet observed today, LHS 3154b is the product of a series of collisions between multiple planets that once existed in the system.

To test this theory, Liveoak and Millholland simulated the evolution of a system containing 11 small planets with an average mass of 2 Earth masses. The planets in this system are arrayed in what’s known as a resonant chain, in which the orbital periods of the planets are integer multiples of one another. This configuration is thought to arise naturally when young planets migrate within a protoplanetary disk. Liveoak and Millholland nudged the planets out of this stable configuration by having the planets begin to lose their atmospheres. With their orbital stability disrupted, the planets abandoned their carefully balanced orbits and collided.

masses and orbital periods of planets in systems with a close-in Neptune-mass planet

Masses and orbital periods of the planets remaining in the six simulated systems that generated a close-in Neptune-mass planet. The planetary system architectures are remarkably similar. [Liveoak & Millholland 2024]

Planets akin to LHS 3154b — defined here as planets with masses of 12–20 Earth masses and orbital periods less than seven days — arise in just 1.2% of simulations. This result shows that while it’s possible to create a planet like LHS 3154b this way, it’s not surprising that more planets like LHS 3154b have yet to turn up. The simulations did provide a clue to find evidence of this process in other systems, though: the systems yielding close-in Neptunes always had one or two companions. The Neptune-mass planet is always the closest planet to the star, and the remaining planets have orbital periods around 30 days. By searching for systems containing a short-period Neptune with an outer companion, researchers can study the chain-breaking process that may be responsible for LHS 3154b’s existence.

Stability of Polar Neptunes

The planets in our solar system orbit in nearly the same plane as the one defined by the Sun’s spin, but it’s clear that not all planetary systems in our galaxy are so orderly. A sizeable chunk of the exoplanet population orbits nearly perpendicular to their stars’ spins. Curiously, roughly half of this population are planets with masses similar to that of Neptune. These highly inclined exo-Neptunes share several characteristics, including slightly elongated orbits, puffy atmospheres with the possibility of past mass loss, and, in some cases, a massive planet in the same system, orbiting farther from the host star.

Emma Louden (Yale University) and Sarah Millholland (Massachusetts Institute of Technology) investigated whether disk-driven resonance could torque exo-Neptunes into their perpendicular orbits while also explaining the other properties of this exoplanet population. In this framework, a young exo-Neptune’s orbit would evolve under the gravitational influence of an outer giant planet and a dissipating planetary disk. Over time, the dissipation of the disk combines with a nodal resonance between the inner Neptune and the outer giant planet, eventually launching the inner Neptune into a highly inclined orbit. This mechanism clearly links polar-orbiting Neptunes with outer giant planets.

diagram showing the process of disk-driven resonance

A diagram that describes the disk-driven resonance for creating polar-orbiting planets. [Louden & Millholland 2024]

This process is thought to take place in the first 10 million years of a planetary system, and it’s not yet clear whether this setup is stable over billions of years. Louden and Millholland simulated the evolution of this setup under the influence of tides. Remarkably, the authors found that exo-Neptunes in polar orbits are stable over long stretches of time, though some planets with smaller orbital inclinations, in the 45–80-degree range, are not stable.

Louden and Millholland then considered two known polar Neptunes, WASP-107b and HAT-P-11b. Both of these planets have an outer giant planet in their system, and both planets show signs of mass loss due to tidal heating. Louden and Millholland showed that these planets’ orbital configurations are incredibly stable, on par with the orbital stability of Uranus and Neptune in our own solar system. While this study doesn’t provide proof of the disk-driven resonance hypothesis, it does demonstrate the feasibility of the concept, and further observational evidence can strengthen the hypothesis.

Is K2-18b Covered in a Supercritical Ocean?

Of the many sub-Neptune exoplanets — those with masses between the mass of Earth and the mass of Neptune — K2-18b is perhaps the most famous. Discovered in 2015 in data from the Kepler Space Telescope, K2-18b has a mass of 8.63 Earth masses and a radius of 2.6 Earth radii. The structure of planets of this size is a matter of debate: do sub-Neptunes have solid surfaces, like Earth, or are they primarily gaseous, like Neptune? Or are they somewhere in between, unlike any of the planets in our solar system?

Possible structures for K2-18b

Possible structures for K2-18b that have been explored in previous work and will be tested in this work (far right). Click to enlarge. [Luu et al. 2024]

K2-18b’s structure has been difficult to pin down, even with the help of JWST. Atmospheric spectroscopy enabled the detection of methane (CH4) and carbon dioxide (CO2) in K2-18b’s atmosphere. Water, carbon monoxide, and ammonia were not detected. Through modeling, some researchers have found these data to be consistent with the planet having a thin atmosphere with a habitable — even speculated to be inhabited — surface, while others have concluded that K2-18b is gas rich, with no solid surface.

Now, Cindy Luu (University of Texas at San Antonio) and collaborators have explored yet another possibility: that K2-18b is covered with an ocean of hot, supercritical water. A supercritical fluid is hot enough to become gaseous but is under enough pressure that it is left in a strange in-between state that shares properties with both its liquid and gaseous forms. Pure water goes supercritical at 647.15K (705℉/374℃).

Luu’s team used geochemical modeling to test the hypothesis that a global supercritical water ocean could explain the observed chemical abundances of K2-18b’s atmosphere. In this scenario, the boundary between the ocean and the atmosphere would be fuzzy, with a greater degree of mixing than would be found at a typical ocean–atmosphere boundary. The team found that an ocean with a temperature between 710K and 1050K could reproduce the observed chemical abundance ratios in K2-18b’s atmosphere. What’s more, this scenario naturally reproduces the non-detection of carbon monoxide.

While the current investigation shows that a supercritical water ocean is consistent with the existing atmospheric abundances measured with JWST, more work is needed to explore this exotic possibility further. More data on carbon monoxide and ammonia, in particular, will be critical to future research.

Citation

“Formation of Close-In Neptunes around Low-Mass Stars through Breaking Resonant Chains,” Donald Liveoak and Sarah C. Millholland 2024 ApJ 974 207. doi:10.3847/1538-4357/ad7383

“Polar Neptunes Are Stable to Tides,” Emma M. Louden and Sarah C. Millholland 2024 ApJ 974 304. doi:10.3847/1538-4357/ad74ff

“Volatile-Rich Sub-Neptunes as Hydrothermal Worlds: The Case of K2-18b,” Cindy N. Luu et al 2024 ApJL 977 L51. doi:10.3847/2041-8213/ad9eb1

pre-supernova star

As scientists excitedly await the first light of the Vera C. Rubin Observatory, a recent study has projected that this facility will aid in identifying hundreds of massive stars on the cusp of death.

Before Detonation

Throughout their lifetimes, stars burn through hydrogen in their cores — million-degree furnaces smashing atoms together to form new ones. Massive stars, many times the mass of our Sun, have very high temperatures and pressures in their cores, causing them to live fast and die young. When all the fuel is burned, the star no longer produces enough thermal pressure to balance gravity, and the star dies in a rapid and massive explosion known as a Type II (or core-collapse) supernova. 

But the final stages of a massive star’s life are not yet fully understood. Observations of Type II supernovae show narrow emission lines that indicate that their progenitor stars were surrounded by circumstellar material — material that was shed from the star as it evolved. However, the exact mechanism through which these stars lose mass is unclear, but if we can catch a star nearing its end but before its deadly detonation, we can better understand these massive stars’ elusive final days. 

Rubin Observatory

Artist’s illustration of the Rubin Observatory observing the sky searching for supernovae. The large field of view of the telescope captures large areas of sky in a single image. Click to enlarge. [NOIRLab/NSF/AURA/P. Marenfeld; CC BY 4.0]

Rubin’s Remedy

Looking to explore many facets of the universe, the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST), first light anticipated July 2025, will scan the Southern Hemisphere sky searching for transient events like supernovae. Recognizing the power of LSST, Alexander Gagliano (The NSF AI Institute for Artificial Intelligence and Fundamental Interactions) and collaborators run simulated observations to predict how many stars LSST will catch in their final days, before they explode as supernovae.

Previous observations of Type II supernovae reveal enhanced emission in the months to years prior to explosion, and as LSST monitors the sky, it will be able to capture this pre-explosion emission. The authors carefully model the expected light curves for various types of core-collapse supernova precursors based on the handful of pre-explosion emission events observed thus far. With these models, the authors simulate LSST observations, applying two methods that would allow for the detection of stars gearing up to explode. The first being single-visit observations in which the enhanced emission is detected using differential photometry prior to the star’s explosion, independent of detecting the subsequent supernova. The second method involves going back after a supernova has been detected. By performing binned photometry of the star with observations taken of it prior to its explosion, the preceding emission can be recovered. From here, the authors can predict how many events LSST will recover after it goes online. 

Power in Numbers

CCSN precursors from one year of synthetic LSST observations

All core-collapse supernova precursors for both detection methods from one year of synthetic LSST observations. Each color corresponds to each model light curve used. Click to enlarge. [Gagliano et al 2025]

Based on their analysis, the authors predict that LSST will detect ~150–240 Type II supernova progenitors per year with single-visit photometry. Over the course of the first three years of LSST, they anticipate 150–400 detections from the binned photometry. This projected frequency of detections will launch the study of late-stage stellar life to new levels, increasing the observed sample of Type II supernova progenitors astronomically. As these detections come in, the observations will reveal the properties and behaviors of these stars in their final days. This opens the door to understanding how end-stage massive stars lose mass and ignite into the most powerful events in the universe. 

Citation

“Finding the Fuse: Prospects for the Detection and Characterization of Hydrogen-rich Core-collapse Supernova Precursor Emission with the LSST,” A. Gagliano et al 2025 ApJ 978 110. doi:10.3847/1538-4357/ad9748

Ingenuity helicopter on Mars

Researchers have used an aircraft to measure the wind speed on Mars, marking the first time this method has been used on another planet. This groundbreaking measurement was made possible by the Ingenuity helicopter, which was active for nearly three years and spent, cumulatively, more than two hours in flight on the Red Planet.

A New Era of Aviation

photo taken by the Ingenuity helicopter

Ingenuity’s navigation camera snapped this photo of the Martian surface during the helicopter’s first flight. [NASA/JPL-Caltech]

In February 2021, Mars got two new inhabitants: the Perseverance rover and the Ingenuity helicopter, bundled together in NASA’s Mars 2020 mission. Ingenuity made history by being the first aircraft to carry out a powered and controlled flight on another planet.

The mission showed that flight is possible in the rarefied Martian air — more than 100 times thinner than Earth’s — and covered roughly 18 kilometers (11 miles) in total. Ingenuity, carrying no scientific instruments and weighing about as much as a Chihuahua, paved the way for future aircraft missions to other worlds, such as the highly scientifically capable, half-ton Dragonfly mission to Saturn’s moon Titan.

diagram of the Ingenuity helicopter with roll, pitch, and yaw labeled

Diagram showing the directions of roll, pitch, and yaw — the three dimensions used to describe the orientation of an aircraft. [Jackson et al. 2025]

Which Way the Martian Wind Blows

Now, almost exactly one year since its final flight, Ingenuity is still enabling more firsts. In a research article published today, a team led by Brian Jackson (Boise State University) described how they used information from Ingenuity to measure the speed and direction of Mars’s winds. Though Ingenuity did not carry any instruments capable of directly measuring the wind, the helicopter recorded its attitude, or orientation, as it flew.

Previously, Jackson had carried out field experiments on Earth with a small drone to show that wind parameters could be extracted from an aircraft’s attitude data. Building on that proof-of-concept study, Jackson’s team used models to understand how Ingenuity’s attitude would change in response to winds of varying speed and direction. From this modeling, the team reconstructed the winds that battered the tiny helicopter as it flew at altitudes spanning 3 to 24 meters (10 to 79 feet).

Comparing Conditions

The team calculated wind speeds ranging from 4.1 to 24.3 meters per second (9 to 54 miles per hour; that’s anywhere from a “gentle breeze” to a “strong gale,” to use the earthly terms). Compared to meteorological models, the measured speeds tended to be higher than expected and the wind directions did not always match. These differences might reflect the influence of localized geological features, like craters and scarps that whip the wind in highly variable directions, that the models do not fully capture.

plot of wind speed as a function of time

The wind speed calculated from Ingenuity’s motion (red circles) compared to the speed measured by the Perseverance rover’s instruments (blue line). Click to enlarge. [Jackson et al. 2025]

While the wind directions implied by the Ingenuity helicopter data generally agreed with measurements by the Perseverance rover, which measured the planet’s surface weather at an altitude of 1.5 meters (5 feet), Ingenuity measured higher wind speeds. Jackson’s team found it unlikely that the higher speeds measured at Ingenuity’s higher altitude were the result of random fluctuations; instead, they proposed a physical explanation rooted in the aerodynamic conditions upwind of the rover and helicopter.

This study highlights both the challenge and potential of measuring winds with an aircraft, and Jackson’s team plans for future work to refine the method. Accurate measurements of wind speeds on Mars can help scientists investigate our neighboring planet’s surface processes and dust transport, as well as help to plan safe entry, descent, and landing for future missions.

Citation

“Profiling Near-Surface Winds on Mars Using Attitude Data from Mars 2020 Ingenuity,” Brian Jackson et al 2025 Planet. Sci. J. 6 21. doi:10.3847/PSJ/ad8b41

1 … 3 4 5 6 7 … 109