The Photons That Reionized the Universe Are Missing. Hubble May Be the Only Way to Find Them.

During the epoch of reionization, the universe was a fog. Hydrogen atoms between galaxies were neutral, and their electrons had not yet been stripped away. Then the fog lifted. The photons that did the lifting, energetic ultraviolet light from the first galaxies, are almost impossible to see directly at the time they mattered, because the same neutral hydrogen absorbed them.

A white paper posted to arXiv on June 2, 2026, titled HST's Deep Blue, by Ilias Goovaerts of the Space Telescope Science Institute and colleagues argues that the missing piece is a single number: the escape fraction of ionizing photons. That number, often written f_esc, is the share of hydrogen-stripping photons that leave a galaxy instead of being swallowed by its gas and dust.

The Number That Decides Who Reionized the Universe

Cosmic reionization transformed the intergalactic medium, the diffuse gas between galaxies, from neutral to ionized. JWST can now see galaxies that lived during that era, but it cannot directly measure how many ionizing photons they let escape. During the epoch of reionization, intervening neutral hydrogen absorbs Lyman continuum (LyC) photons, the ultraviolet light energetic enough to ionize hydrogen. The signal is erased before it reaches us.

The workaround is to study lower redshifts, where the universe is more transparent. The paper focuses on galaxies at redshifts 2 to 4. Redshift measures how cosmic expansion has stretched light. The redder filters in the proposed survey reach within about a billion years of the end of reionization, while bluer filters catch galaxies farther from the epoch, where intergalactic absorption is milder. By measuring LyC escape at these lower redshifts, astronomers can calibrate indirect tracers that JWST can apply inside the epoch itself.

The stakes are high. Recent years have produced a photon budget problem: models sometimes find too few ionizing photons from early galaxies, and sometimes too many, to match the timeline of reionization. Those debates have unfolded without a solid measurement of f_esc for ordinary high-redshift galaxies. Existing detections at z > 2 mostly come from extreme, bright, high-escape systems, valuable but not representative.

Why Hubble Must Look Closer Than the Epoch Itself

The paper argues that Hubble is the only observatory that can do the required job. HST's Wide Field Camera 3 has an ultraviolet-visible channel, WFC3/UVIS, sensitive to the wavelengths where LyC from 2 < z < 4 galaxies appears. Being above the atmosphere gives it access to ultraviolet light that ground-based telescopes cannot see, and its spatial resolution helps reject low-redshift interlopers that can mimic LyC signals.

Four filters, F225W, F275W, F336W, and F390W, target LyC at progressively higher redshifts. Bluer filters catch galaxies farther from the epoch, where intergalactic absorption is milder. Redder filters reach closer to reionization but suffer more absorption. Each choice trades sample size against cosmic relevance.

The paper estimates that individual fields can require tens to 1,000 orbits in a given filter. A 20-field program would start at about 1,000 orbits total with F275W and rise for redder filters. Such an effort would need a large multi-cycle treasury or director's discretionary allocation, not a normal PI-led proposal. The authors also argue for continued support of WFC3/UVIS, with Hubble's ACS useful for cheaper rest-frame ultraviolet imaging.

The central observational demand is breadth. At z > 2, patches of neutral hydrogen in the intergalactic medium absorb LyC along each line of sight. The exact transmission is unknown. The paper cites work showing that 20 independent fields are needed to average over enough sightlines and reduce that uncertainty.

Those fields should already have deep spectroscopy. The Multi-Unit Spectroscopic Explorer (MUSE) on the Very Large Telescope in Chile can confirm redshifts and measure Ly-alpha emission for galaxies above z about 2.9. JWST's NIRCam and NIRISS wide-field slitless spectroscopy can supply redshifts at lower redshifts by spreading every object's light into a spectrum. Redshift confirmation is essential: without it, a faint signal in a Hubble filter might come from a nearby galaxy, not a distant LyC emitter.

The goal is a sample comparable to the Low-redshift Lyman Continuum Survey, which has about 50 Lyman continuum emitters. At high redshift, the authors want to reach escape fractions as low as 5 percent for faint galaxies, the rough population average often invoked for reionization. Individual detections may reach tens of percent, while stacking many galaxies, combining their faint light, can push toward the 5 percent regime.

What a Deep Blue Survey Could Reveal—and What It Is Not

With hundreds of galaxies, the survey could build an ionizing luminosity function, a census of how much LyC galaxies of different brightness produce. Comparing it with the familiar ultraviolet luminosity function would show how the ionizing and non-ionizing light are related, and would help constrain the ultraviolet background, which is presently unconstrained.

It could also test tracers of escape. The paper highlights Ly-alpha line properties, such as double-peaked line separation and the fraction of light in a diffuse Ly-alpha halo. Other promising tracers include star formation rate surface density, the ultraviolet slope, and the ratio of [OIII] to [OII] emission lines. These relationships are debated and have not been verified at high redshift. If they hold, they become tools for JWST inside the epoch of reionization.

Strong gravitational lensing offers another route. Massive galaxy clusters can magnify faint background galaxies, allowing the same analysis to reach fainter galaxies that would otherwise be out of reach. The paper notes that well-studied clusters and moderate magnifications can keep lens-model uncertainties manageable.

This is a white paper, not a new set of observations. Its numbers are forecasts based on models, existing surveys, and simulated intergalactic transmission. The relationships between Ly-alpha, the ultraviolet slope, and escape fraction remain uncertain; some studies have already challenged them. The paper's recommended program would take years and a large allocation of Hubble time.

Still, the scientific gap is concrete. JWST is central to studying early galaxies, but it cannot directly answer how many ionizing photons escaped them during reionization. That calibration has to come from a telescope that can see ultraviolet light from the right distance and with enough sensitivity.

If the Deep Blue program is carried out, it would not simply add more galaxies to a catalog. It would supply the missing conversion factor between what JWST sees inside the reionization epoch and what those galaxies were doing to the gas around them. The cosmic fog lifted once. Measuring how it happened means catching the faint blue light that slipped through.