JWST preprint finds shallower faint‑end at z≈7.4 and a power‑law UV luminosity function at z≈10

A new James Webb Space Telescope preprint reports the largest samples yet of candidate galaxies from roughly 700 million and 500 million years after the Big Bang, and argues that the early universe’s galaxy census may look different from many previous estimates. In the manuscript, posted Oct. 23 on arXiv, researchers Bangzheng Sun and Haojing Yan say the ultraviolet luminosity function — a count of how many galaxies exist at different brightness levels — appears shallower than expected at about redshift 7.4 and is not well described by the standard Schechter form at about redshift 10.

The paper, “Galaxy Luminosity Functions at z~7-10 from Four Years of JWST Observations,” is a preprint, not yet peer-reviewed, and is marked as submitted to Monthly Notices of the Royal Astronomical Society. Using four years of public JWST NIRCam imaging across nine fields totaling about 1,045 square arcminutes, the authors searched for very distant galaxies with the dropout technique, which identifies objects that disappear in one filter because their light has been redshifted. They analyzed seven blank fields and two galaxy-cluster fields, selecting 3,733 F090W dropouts for the z≈7.4 sample and 2,396 F115W dropouts for the z≈10 sample. In the abstract, the authors say those “constitute the largest samples of candidate galaxies at z≈7.4 and z≈10 to date.”

To help estimate false positives, the researchers used subsamples that already have secure spectroscopic redshifts from JWST’s NIRSpec instrument, which splits light to measure distance more directly. They say 338 of the F090W dropouts and 99 of the F115W dropouts have such confirmations, and that those objects were used to derive magnitude-dependent contamination rates. The reported luminosity functions were then statistically corrected for both incompleteness and contamination.

Their two headline results point in different ways to tension with earlier expectations. At z≈7.4, the authors say the luminosity function can still be fit with a Schechter function, the standard formula often used to describe galaxy populations, but the faint-end slope is α = -1.70 ± 0.16. In the abstract, they write: “We find that the luminosity function at z≈7.4 can be well described by a conventional Schechter function but with a faint-end slope of α = −1.70 ± 0.16...” At z≈10, though, they say the standard Schechter form fits poorly and that the data are better described by a power law, with a best-fit slope of β = -2.36 ± 0.16. The abstract says: “The luminosity function at z≈10 ... is more consistent with a power law ... with the best-fit slope of β = −2.36 ± 0.16.”

That matters because the ultraviolet luminosity function is a basic input for models of early galaxy formation and cosmic reionization, the era when radiation from the first stars and galaxies ionized hydrogen in intergalactic space. If the z≈7.4 faint end is truly shallower than many earlier estimates, it would imply fewer very faint galaxies than researchers often assumed. Those faint galaxies have frequently been invoked as major contributors to reionization.

Before JWST, many Hubble Space Telescope studies at about z≈7 to z≈8 often found steeper faint-end slopes, around α≈-2. JWST, launched in late 2021 and operating scientifically since 2022, has expanded both the depth and area available for this kind of work through NIRCam imaging and NIRSpec confirmations. But JWST-era measurements have not all lined up neatly, and the new preprint adds another data-rich result to an active debate. For z≈10, Sun and Yan say the departure from the Schechter form is largely driven by high number densities of bright galaxies at ultraviolet magnitudes brighter than about -21.8, a trend they say is supported by existing spectroscopic confirmations.

The paper stops well short of presenting these results as final. The authors call for deeper imaging and spectroscopy around 30th magnitude to test the z≈7.4 faint-end measurement, and more spectroscopy of bright sources at about 26th magnitude or brighter to strengthen the z≈10 result. For now, the work stands as a notable new preprint based on a large public JWST data set — one that could sharpen, but not settle, the picture of how galaxies populated the universe’s first few hundred million years.