Cosmic-dawn 21-cm signal extraction: reconciling EDGES, SARAS 3 and HERA with foreground modelling
The EDGES signal is unconfirmed and SARAS 3 contradicts it, but the path (independent instruments, better calibration, deeper HERA data) is clear, so mid-40s. Leverage 44: a detection measures the first stars and early X-ray sources and tests EDGES-type new physics. Application value 8.
The problem
Detect the faint radio fingerprint of hydrogen from when the first stars lit up, and test a surprising result.
- 1Hydrogen gas emits a 21 cm radio line
- 2First stars change the line's strength
- 3Signal is 100,000 times weaker than the Galaxy
- 4EDGES saw a dip; SARAS 3 did not
- 5Better antennas and models needed
The signal. Neutral hydrogen emits and absorbs at 21 cm (1.42 GHz) through its hyperfine transition. The brightness temperature of the 21-cm line against the CMB depends on the "spin temperature" of the gas, which is coupled to the gas temperature by Lyman-α photons from the first stars, and heated by X-rays from early black holes and stellar remnants. Redshifted from z of about 13 to 27, the line falls at 50 to 100 MHz. Cosmologists try to measure the sky-averaged (global) signal, and the power spectrum of its fluctuations at z of 8 to 10.
Built from
earlier problems → this oneThe problems this one is built from: progress or tools it needs (builds on), special cases that come first (generalizes), and weaker results on the way (implies). On the right, what builds on it.
- Start with
- Then
- Cosmic-dawn 21-cm signal extraction: reconciling EDGES, SARAS 3 and HERA with foreground modellingThis problem
Path to understanding
foundations → problemThe research entries this problem rests on, in the order to read them. Each links to its explainer.
- Start with
- Step 2
- Step 3
- Radio interferometry and aperture synthesisAlgorithm · host
- Then
- Cosmic-dawn 21-cm signal extraction: reconciling EDGES, SARAS 3 and HERA with foreground modellingThe problem
State of the art
EDGES reported an absorption profile centred at 78 MHz (arXiv:1810.05912, Nature, March 2018). SARAS 3, an independent instrument, did not detect the EDGES profile and rejected its best-fit shape at 95.3% confidence (arXiv:2112.06778, December 2021). The SARAS 3 non-detection alone disfavours radio-luminous galaxies in 4×10⁵–10⁷ M☉ halos at z ≈ 20 and a synchrotron radio background more than about 6% above the CMB at 1.42 GHz (Bevins et al., arXiv:2212.00464, Nature Astronomy), and a joint SARAS 3 + HERA analysis tightens these constraints (Bevins et al., arXiv:2301.03298). HERA Phase I gives upper limits on the reionisation-era power spectrum (arXiv:2108.02263) and improved limits on X-ray heating (arXiv:2210.04912, October 2022). The first HERA Phase II results (arXiv:2511.21289, November 2025), from a two-week subset, give 2σ limits in eight bands over 5.6 ≤ z ≤ 24.4, the best being 1.78×10³ mK² at k = 0.70 h/Mpc and z = 7.05, with mutual coupling between antennas now the dominant systematic. EDGES-3 has been commissioned (arXiv:2508.02577) and the EDGES pipeline and earlier raw data were released in 2026 (arXiv:2605.16643); EDGES-3 sky results so far appear only in non-refereed project memos and were not used here. REACH and MIST have published instrument papers but no sky-averaged constraints. The EDGES signal is neither confirmed nor excluded at high significance.
Realistic applications
A detection of cosmic-dawn 21-cm signals measures the first stars, early X-ray sources and the thermal history of the IGM, and any confirmation of the EDGES amplitude would indicate new physics (dark-matter–baryon scattering or an excess radio background). The calibration methods (very precise beam modelling) transfer to other radio cosmology. No practical application.
What counts as solved
Resolved when two independent instruments with different systematics detect a global 21-cm absorption signal at ≥5σ with matching shape (or exclude the EDGES profile at ≥5σ), and HERA or a comparable array detects the power spectrum at z ≈ 8–10.
How it gets verified
ObservationSettled by new measurements or observations.
Needs two independent instruments with different systematics detecting a global 21-cm absorption signal at 5 sigma or more with matching shape, or excluding the EDGES profile at 5 sigma, and HERA or a comparable array detecting the power spectrum at z of about 8 to 10.
Claims under verification
+ add a claimWhere the work is
Instrument calibration and long observing campaigns dominate, with foreground modelling.
Bottlenecks
ranked by votes- #1EngineeringBeam and bandpass calibration at 10⁻⁵
Chromatic antenna response and ground effects must be known to better than 1 part in 10⁵ over 50–100 MHz.
Community grades
Sign in to add your own grade.
Connected problems
Progress
- Nov 26, 2025HERA Phase II first limitssource
Limits over 5.6 ≤ z ≤ 24.4 from two weeks of Phase II data; mutual coupling is the dominant systematic.
- Dec 13, 2021SARAS 3 non-detectionsource
SARAS 3 disfavours the EDGES profile at 95.3% confidence.
- Mar 1, 2018EDGES absorption profilesource
Absorption trough at 78 MHz with amplitude about 0.5 K reported in Nature on 1 March 2018 (arXiv posting October 2018).