Astrophysics papers — 2026-09-15

We are starting the day with a major look at why the universe's star formation has slowed down over the last several billion years. By stacking the signals from over eight thousand star-forming galaxies in the COSMOS field, researchers have finally measured the average atomic gas content at a redshift of about one.

The data shows these galaxies hold an average hydrogen mass of about 15.5 billion solar masses. This suggests that the decline in star formation seen in the later universe is likely driven by a lack of fresh gas being pulled in from the cosmic web to replenish what stars are consuming.

This connection between gas supply and cosmic evolution is mirrored in our understanding of dark matter. New observations from the James Webb Space Telescope are providing a clearer picture of how matter is distributed.

By analyzing hundreds of galaxies from the cosmic noon epoch, researchers have reconstructed the first direct dark matter density profiles using galaxy rotation curves. The results reveal a surprising universal feature where the central density of dark matter cores remains remarkably constant even as the surrounding rotation velocities change.

This suggests that the density of these dark matter cores is somewhat decoupled from the evolution of visible stars and gas. While we are learning more about the structure of dark matter, we are also refining our search for the force driving the universe's expansion.

New constraints on dark energy have emerged from studying the gamma-ray spectra of a specific supernova. By comparing the energy of iron nuclei in that supernova to laboratory values, scientists found no significant deviation, which limits how much a hypothetical dark energy field could be changing over time.

This helps narrow down the possibilities for what dark energy actually is, keeping the cosmological constant as a leading candidate. We also have direct evidence that the early universe was much more chemically messy than previously thought.

By looking at how carbon absorption lines cluster around star-forming galaxies during the Epoch of Reionization, researchers found that metals like CIV and CII are not just trapped inside massive galaxies. Instead, these elements are spread out far beyond those galaxies, suggesting they were blown into the diffuse intergalactic medium or seeded by a vast population of tiny, faint dwarf galaxies.

The spatial reach of this metal pollution is telling. While the warm-ionized CIV gas extends out to about one megaparsec, the cooler CII gas drops off at around half a megaparsec, showing that the ionized phase of the cosmic web is more extended than its cooler counterpart.

This widespread distribution of matter also helps us understand how individual objects behave, such as the potentially hazardous asteroid 2024 YR4. While it poses no threat to Earth or the Moon for at least a century, new thermal modeling from JWST data has pinned its size at about 61 meters and revealed it is a fast-rotating object with high thermal inertia.

The complexity of measuring physical properties continues even when we look at the extreme environments around black holes. In studies of stellar-mass black holes like GRS 1716-249, researchers found that determining spin is difficult because different physical configurations can produce nearly identical X-ray spectra.

This means we still need much broader, coordinated observations to be certain. We might eventually be able to use dwarf galaxies as giant thermometers to catch dark matter in the act.

By simulating how dark matter candidates like axions or sterile neutrinos deposit heat into gas, researchers found that this energy injection alters the thermal structure of a galaxy's circumgalactic medium without changing its star formation rate. For certain heating rates, the neutral hydrogen and metal column densities drop drastically, creating a signature that could be detected using quasar absorption-line spectroscopy.

This ability to use cosmic structures as detectors is mirrored in how we refine our distance ladder. New multi-chromatic observations from the CHARA Array have allowed us to measure the limb-darkening of Cepheid variables across the R, H, and K bands for the first time.

By using these precise diameters to calibrate surface brightness-color relations, we can reduce scatter in these distance indicators and improve the accuracy of the Baade-Wesselink method. While we refine our cosmic yardsticks, we are also looking at how individual stars evolve after dramatic events.

One candidate post-merger giant star, TYC 4144-329-2, is showing signs of weak coronal activity in X-rays alongside variable accretion seen in its hydrogen profiles. This suggests the merger was recent enough that the star has not yet developed a deep convection zone, offering a glimpse into the immediate aftermath of stellar collisions.

If we want to solve the Hubble tension, we might need to look at both dark energy and neutrinos simultaneously. Using Bayesian physics-informed neural networks to analyze cosmological data suggests that a model with both a dynamical dark energy component and massive neutrinos can ease the discrepancy between Planck and SH0ES measurements.

Specifically, this approach finds neutrino masses between 0.16 and 0.28 eV, which brings the tension down to about 0.83 sigma for SH0ES, though it does not fix the Planck side of the problem entirely. This need for more complex models is mirrored in how we understand the lifecycle of stars and their contributions to cosmic dust.

New theoretical limits on dust masses show that oxygen-rich silicates dominate the budget, with masses reaching up to 1.43 solar masses for high-mass progenitors. However, these yields are surprisingly unpredictable because the amount of silicate dust produced can vary by a factor of two to five for the exact same type of star due to random stellar evolution events.

The chaos inherent in stellar deaths also affects the survival of heavy elements. When protomagnetars explode, they create intense outflows that should theoretically synthesize ultraheavy nuclei, but these nuclei must survive a gauntlet of high-energy photons.

Depending on whether the outflow is a spherical wind or a directed jet, these particles might be destroyed by photodisintegration before they can escape the stellar envelope and enrich the galaxy. We are also seeing how these energetic outflows evolve into much quieter, older structures over time.

A newly discovered radio source called J1248+4826 appears to be a re-energized remnant lobe from an inactive galaxy. It appears that moderate shocks within its host galaxy group are breathing new life into old plasma, creating complex, diffuse shapes.

We are finally seeing the power of neural networks to clean up our maps of the early universe. A new method called NERV uses a convolutional neural network to reconstruct the baryon acoustic oscillation signal in the BOSS DR12 galaxy sample, effectively undoing the blurring caused by nonlinear structure growth.

By treating the survey as a collection of local patches, it significantly improves the precision of our cosmic distance measurements. This push for better precision is mirrored in the study of how galaxies interact with their surroundings.

Using DESI Year 1 data, researchers tracked the cool gas surrounding 800,000 luminous red galaxies and found that the amount of this gas increases at higher redshifts. They also noticed that in the inner regions of these galaxies, higher stellar mass seems to suppress the cool gas content, while a more energetic gas component becomes dominant in the outer reaches.

While we map large-scale structure, we are also getting better at understanding the physics within individual nebulae. A new way to recover turbulent velocity statistics from noisy spectroscopy allows us to extract reliable data from intermediate-resolution observations like those from VLT MUSE.

By fitting a parametric model to the velocity structure function, researchers successfully recovered turbulent parameters in the Orion Nebula that match high-resolution echelle data. This proves we can still get the physics right even when the spectral resolution is relatively poor.

We finally have a way to bridge the gap between speed and accuracy when modeling the stellar streams that trace dark matter. A new basis-expansion code called KRIOS reproduces complex N-body cluster models much more accurately than standard methods while using a fraction of the time.

The mismatch between these models is most obvious when the progenitor cluster is tightly bound to its host galaxy, where tidal forces are strongest. This ability to model complex structures is echoed in new work attempting to solve the circularity problem in using gamma-ray bursts to map the expansion of the universe.

By using artificial neural networks to calibrate luminosity relations, researchers have bypassed the need to assume a specific cosmological model upfront. This approach confirms that the Amati relation remains consistent with previous low-redshift calibrations.

The search for high-energy signals is also seeing a shift toward quantum-assisted processing. A new pipeline uses a hybrid Quantum Vision Transformer to identify fast radio bursts from raw telescope data, achieving a ninety-four percent accuracy rate.

This approach shows that we can achieve performance similar to classical models while testing quantum processors in large-scale surveys. We also need a better way to measure the expansion of the universe, and a new Bayesian framework for the tip of the red giant branch is a major step toward that goal.

By modeling stellar catalogs as inhomogeneous Poisson point processes, researchers can account for contamination from asymptotic giant branch stars and photometric noise. When applied to Hubble Space Telescope data of the galaxy NGC 4258, this method yielded an absolute magnitude for the tip of-4.073, which is slightly brighter than previous studies.

This precision in distance scales is mirrored by efforts to pin down the beginning of the universe through inflation. Using a combination of Planck, BICEP/Keck, and DESI data, new constraints on single-field slow-roll inflation have ruled out simple monomial-potential models.

Instead, the data points toward concave potentials, like the Starobinsky model, though the predicted tensor spectral index is likely too tiny to be measured with CMB data alone anytime soon. While we look at the largest scales, we are also refining our understanding of the dark matter that fills the gaps.

The HAWC observatory has used improved event reconstruction to study dwarf spheroidal galaxies. Even with more data and better sensitivity, they found no signal, setting new upper limits on the dark matter annihilation cross-section.

The search for unseen components continues in the gas between the stars as well. By comparing 21-cm hydrogen observations with numerical simulations, researchers found that the neutral interstellar medium is more complex than a simple two-phase model.

Their data suggests a significant amount of gas resides in a thermally unstable intermediate phase, which aligns more closely with TIGRESS-NCR simulations than with older models. We finally have a way to account for the fact that we observe the universe on a light cone, where the geometry of our view is fundamentally two-dimensional.

By treating Fourier vectors as derivatives projected onto a sphere, a new analytical approach to the angular bispectrum avoids the mathematical cancellations that usually make Limber's approximation fail. This method provides a more accurate way to calculate the covariance between different types of galaxy clustering statistics.

The precision of our cosmic models is also being pushed by a new way to track how elements like europium and barium are produced. By looking directly at the ages of stars at solar metallicity, researchers found that about 60 percent of the europium and barium seen today comes from strongly delayed processes.

This finding is significant because the production rates for these elements rise with delay time, suggesting neutron-capture production might come from something other than neutron star mergers. In the realm of high-energy astrophysics, we are learning that uncertainties in cosmic background light are no longer the main limitation.

New modeling shows that for the local universe, we can now make robust inferences about the spectra of gamma rays and the composition of cosmic rays without being limited by how well we know the background photon fields. Moving from the deep cosmos to our own neighborhood, we are seeing new patterns in how galaxies grow.

Observations of Seyfert galaxies show that central fast shocks are a common feature, often appearing perpendicular to the light from the central black hole. These shocks, which likely come from jets or winds hitting the surrounding gas, help clarify how much an active nucleus influences star formation in its host galaxy.

Closer to home, we are seeing unexpected behavior in the most reliable cosmic clocks. The millisecond pulsar PSR J0437-4715 has shown two consecutive events where its pulse profile changed in a specific, localized way.

Because these changes are tied to the pulsar's own magnetic field rather than the interstellar medium, they provide a new way to test models of how pulsars emit radiation. On a much smaller scale, we are getting better at reading the solar atmosphere.

A new web application called the Riemann Map Operator helps observers identify different types of magnetic shocks by testing whether observed brightness changes obey the laws of physics. It can even distinguish between different types of shock waves in extreme-ultraviolet data to help us understand solar eruptions.

Finally, we are seeing the first results from new spectroscopic surveys of young star clusters. Using the WEAVE instrument, researchers are beginning to map out the populations of massive stars in crowded regions by using new tools to strip away the signatures of interstellar gas.

Today's papers

The papers

Important terms

Redshift
A measurement used to determine how far away objects are and how fast they are moving by looking at shifts in light. It helps scientists study the evolution of galaxies and the expansion of the universe over time.
Dark Matter Density Profiles
A map showing how dark matter is distributed within a galaxy. Researchers use these to understand if dark matter stays constant in its central core or changes as stars and gas evolve around it.
Epoch of Reionization
A period in the early universe when the first stars and galaxies formed, stripping electrons from hydrogen atoms. Studying this helps scientists track how metals like carbon were spread throughout space by early galaxies.
Baryon Acoustic Oscillations
Regular fluctuations in the density of visible matter in the universe. These act as a cosmic yardstick, helping astronomers measure distances and understand how large-scale structures like galaxy clusters have grown over time.