🔭 astrophysics

Something Bright at the Edge of Everything: A Uniquely JWST-Dark Radio Source in COSMOS

This paper reports the discovery of a uniquely "JWST-dark" radio source in the COSMOS field that is detected across a wide radio frequency range but remains undetected in all optical, infrared, X-ray, and sub-millimeter observations, suggesting it is either an extremely dust-obscured radio-loud galaxy from the epoch of reionization or a detached radio lobe.

Original authors: Mingyu Li

Published 2026-04-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mingyu Li

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, dark ocean. For decades, astronomers have been trying to find the brightest lighthouses (radio sources) hidden deep within the fog of the universe's "childhood," a time called the Epoch of Reionization.

Usually, when we look for these ancient lighthouses, we use powerful telescopes to see the light they emit. But there's a problem: some of these ancient objects might be so covered in thick cosmic dust that they are invisible to our most powerful eyes, like the JWST (James Webb Space Telescope), which sees in infrared light. They are like ghosts that can't be seen, but they can still be "heard" because they shout loudly in radio waves.

Here is the story of a unique discovery made by Mingyu Li, a researcher from Tsinghua University, who found exactly this kind of "ghost."

The Great Cosmic Cross-Check

Think of the COSMOS field as a massive, high-definition map of a specific patch of the sky. Astronomers have two main lists for this map:

The Radio List: A catalog of everything that "shouts" in radio waves (detected by telescopes like VLA and LOFAR).
The Optical/Infrared List: A catalog of everything that "glows" in visible or infrared light (detected by JWST and Hubble).

Mingyu Li took these two lists and compared them, looking for a specific type of mismatch: A radio source that has no glowing partner.

Out of nearly 3,000 radio sources, most had a matching galaxy in the JWST images. But one stood out. It was a radio source that was completely invisible to JWST, Hubble, and even other powerful telescopes like Chandra (X-rays) and ALMA (sub-millimeter waves).

The Analogy: Imagine walking into a crowded room where everyone is wearing a bright, glowing costume. You can see everyone clearly. But then, you hear a loud voice coming from a corner, and when you look there, you see absolutely nothing. No person, no costume, just empty space. That is what this object is: a loud voice in the dark.

The "Silent" Giant

This object, named COSMOSVLA J100019.21+021337.2, is a mystery for two main reasons:

It's a Radio Roar: It is very bright in radio waves, with a spectrum (the "color" of its radio sound) that is steep and unbroken. This suggests it is powered by a supermassive black hole eating material, creating powerful jets.
It's a Light Vacuum: It is so dark in every other wavelength that it doesn't show up in any of the deep images taken by the world's best telescopes.

What Could It Be?

The paper proposes two exciting possibilities, like two different endings to a mystery novel:

Theory A: The "Orphaned" Radio Lobe
Imagine a lighthouse that was built on a cliff, but the cliff collapsed, and the lighthouse floated away on a piece of debris. The light (radio waves) is still shining, but the building (the host galaxy) is far away or hidden behind a mountain. In this case, the radio source we see might be a detached piece of a jet from a galaxy that is actually located somewhere else in the sky, just out of our current view.
Theory B: The "Cosmic Dawn" Monster (The More Likely Option)
This is the more thrilling idea. The object might be a supermassive black hole from the very early universe (around 7 billion years after the Big Bang, or z ~ 7).
- The Dust Blanket: It is so heavily wrapped in thick, cosmic dust that the dust acts like a heavy, black blanket. It blocks all the light trying to escape, making the galaxy invisible to JWST.
- The Radio Escape: However, radio waves are like sound; they can pass through the dust blanket easily. So, the black hole is screaming in radio waves, but whispering in light.
- The Scale: If this is true, it is one of the most distant and powerful radio sources ever found. It would be a "compact jet" only a few thousand light-years across, hidden in the deepest darkness of the early universe.

Why Does This Matter?

Finding this object is like finding a needle in a haystack, but the needle is made of pure energy and the haystack is made of dust.

It proves that there are still hidden worlds out there. For years, we thought JWST would find the counterparts to almost every radio source. This discovery shows that JWST has a blind spot: the most extreme, dust-obscured objects of the early universe.

By combining the "ears" of radio telescopes with the "eyes" of JWST, we are opening a new door to the edge of the observable universe. This object might be the first of many, helping us understand how the first supermassive black holes formed and how they shaped the early cosmos.

In short: We found a cosmic ghost that is loud in radio but invisible in light. It might be a monster from the dawn of time, hidden behind a wall of dust, waiting for us to listen closely enough to hear its story.

1. Problem Statement

The search for bright radio sources within the Epoch of Reionization (EoR, $z \gtrsim 6$ ) has been a long-standing challenge in astrophysics. High-redshift radio galaxies (HzRGs) are critical for two main reasons:

They host the most massive black holes and reside in overdense environments, serving as tracers of early galaxy evolution.
They act as potential background sources for detecting the "21 cm forest" (H I absorption by the neutral intergalactic medium), which would revolutionize our understanding of reionization.

While ultra-steep-spectrum (USS) radio sources have been identified up to $z \sim 5$ , and some candidates pushed to $z \sim 7$ , the population at $z > 6$ remains poorly characterized. A major obstacle is that the most distant, powerful radio sources are expected to reside in heavily dust-obscured galaxies, rendering them extremely faint or invisible at optical and infrared wavelengths. With the advent of deep JWST imaging, which can detect counterparts for nearly all radio sources, the most promising candidates for the highest-redshift radio galaxies are now those that remain undetected in all JWST bands ("JWST-dark").

2. Methodology

The author conducted a systematic cross-match between radio and infrared source catalogs in the COSMOS field to identify objects that are bright in radio but invisible to JWST.

Radio Sample: The study started with the complete sample of sources detected at $>5\sigma$ significance in the 3 GHz VLA-COSMOS catalog (V. Smolčič et al. 2017) within the JWST COSMOS-Web footprint. This yielded 2,744 sources.
Cross-Matching: These sources were cross-matched against the COSMOS2025 JWST source catalog (M. Shuntov et al. 2025) using a 1-arcsecond radius.
- This process identified 117 unmatched objects.
Filtering & Visual Inspection:
- 43 sources were discarded as transients or spurious detections (detected only at 3 GHz with faint fluxes and no counterparts at other radio frequencies like VLA 1.4 GHz, superMIGHTEE 650 MHz, or LOFAR 144 MHz).
- 73 sources were re-associated with galaxies via extended radio emission or found to have JWST counterparts that were missed by the automated catalog due to adjacent bright sources.
Target Selection: Only one source remained: COSMOSVLA J100019.21+021337.2, which had no clear counterpart in any JWST band.
Multi-wavelength Verification: The candidate was analyzed across the electromagnetic spectrum using:
- Radio: LOFAR (144 MHz), superMIGHTEE (650 MHz), VLA (1.4 GHz and 3 GHz).
- Optical/IR: HST/ACS, JWST/NIRCam, and JWST/MIRI (from COSMOS-Web, PRIMER, COSMOS-3D, and MINERVA programs).
- X-ray: Chandra.
- Far-IR/Sub-mm: Herschel and ALMA (including CHAMPS, Ex-MORA, and A3COSMOS surveys).

3. Key Results

The identified source, COSMOSVLA J100019.21+021337.2, exhibits unique properties:

Radio Properties:
- Flux Densities: Detected with integrated fluxes of $1950 \pm 210$ $\mu$ Jy (144 MHz), $504 \pm 19$ $\mu$ Jy (650 MHz), $216 \pm 11$ $\mu$ Jy (1.4 GHz), and $106 \pm 6$ $\mu$ Jy (3 GHz).
- Spectral Index: A power-law fit yields a steep, unbroken spectrum of $\alpha = -0.99 \pm 0.03$ ( $S_\nu \propto \nu^\alpha$ ), with no evidence of spectral curvature or low-frequency turnover.
- Morphology: At 3 GHz, the source is marginally resolved (FWHM $\approx 1.1'' \times 0.8''$ ) but lacks a clear double-lobed or core-jet structure at current resolution.
Non-Detections:
- The source is undetected in all available HST, JWST (NIRCam and MIRI), Chandra, Herschel, and ALMA imaging.
- Upper limits are extremely deep, covering UV to mm wavelengths.
Physical Interpretation:
- Hypothesis 1 (Detached Lobe): The radio emission could be a detached lobe from a host galaxy located elsewhere. However, this is unlikely given the compact morphology and the lack of any extended low-surface-brightness emission or a counter-lobe in the radio data.
- Hypothesis 2 (High- $z$ Obscured AGN): The source is likely a highly dust-obscured, radio-loud Active Galactic Nucleus (AGN) at high redshift.
  - Compared to the known obscured AGN COS-87259 ( $z=6.853$ ), this source is $\gtrsim 5$ times brighter in radio yet completely invisible from UV to mm.
  - Assuming $z \approx 7$ , the rest-frame 1.4 GHz luminosity is $\log(L_{1.4 \text{ GHz}}/\text{W Hz}^{-1}) = 26.1$ , placing it among the most luminous radio quasars at $z \gtrsim 6$ .
  - The non-detection implies extreme dust attenuation ( $A_V \gtrsim 5-10$ ) and/or a redshift $z \gtrsim 7$ .
  - The X-ray non-detection provides a mild constraint of $\log L_{2-10} \lesssim 43.9$ (assuming $\Gamma=1.8$ ).

4. Significance and Contributions

Discovery of a Unique Class: This paper identifies the only known object in the COSMOS field that is clearly detected from LOFAR (144 MHz) to VLA (3 GHz) but remains completely undetected in every JWST band.
Probing the Cosmic Dawn: If confirmed at $z \gtrsim 7$ , this object would rank among the most distant radio sources known. It represents a critical test case for the population of radio-loud AGNs at the edge of the observable Universe.
Methodological Validation: The study validates the strategy of cross-matching deep radio surveys with deep JWST imaging to find "dark" sources. It demonstrates that the intersection of these two datasets opens a new discovery space for the most heavily obscured, high-redshift objects that traditional optical/IR surveys miss.
Implications for 21 cm Forest: The existence of such bright, high-redshift radio sources supports the feasibility of using them as background beacons for 21 cm forest experiments, which require bright, compact radio sources at $z > 6$ to map the neutral intergalactic medium.

Conclusion

Mingyu Li presents a compelling candidate for a high-redshift, dust-obscured radio-loud AGN. While alternative explanations (such as a detached radio lobe) are considered, the extreme radio brightness combined with total infrared/optical invisibility strongly suggests a source at the cosmic dawn ( $z \gtrsim 7$ ) with extreme dust obscuration. This discovery highlights the power of combining deep radio and JWST surveys to uncover the most extreme environments in the early Universe.