RUBIES: The Evolution of the Ionization Parameter from 0 < z < 9 Facilities: JWST (NIRSpec), SDSS, ESO/VLT (VIMOS), Keck (MOSFIRE) Astropy ( 3 ) , linmix ( 41 ) , Matplotlib ( 34 ) , NumPy ( 31 ) , pandas ( 81 ) , Prospector ( 37 ) , SciPy ( 126 )

Nikko J. Cleri, Zach J. Lewis, Joel Leja, Jakob M. Helton, Emilie Burnham, Olivia Curtis, Anna de Graaff, Michaela Hirschmann, Harley Katz, Michael V. Maseda, Ian McConachie, Adéle Plat, Lucie Scharré · 2026

Read RUBIES: The Evolution of the Ionization Parameter from 0 < z < 9 Facilities: JWST (NIRSpec), SDSS, ESO/VLT (VIMOS), Keck (MOSFIRE) Astropy ( 3 ) , linmix ( 41 ) , Matplotlib ( 34 ) , NumPy ( 31 ) , pandas ( 81 ) , Prospector ( 37 ) , SciPy ( 126 ) by Nikko J. Cleri, Zach J. Lewis, Joel Leja, Jakob M. Helton, Emilie Burnham, Olivia Curtis, Anna de Graaff, Michaela Hirschmann, Harley Katz, Michael V. Maseda, Ian McConachie, Adéle Plat, Lucie Scharré (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-30410.txt.

Introduction

Nikko J. Cleri and 12 co-authors track the ionization parameter U in galaxies across cosmic time. They compare JWST spectra of 434 RUBIES galaxies at high redshift with low-redshift sets, to infer U from oxygen lines. They find U rises with redshift and star formation rate, and falls with stellar mass. Model spread gives an error of about a third of a dex in logU. A formula to guess U from redshift, mass, and star formation rate where O32 is missing finds U grows about fourfold from redshift 2 to 6.

Catalog record

CREATORNikko J. Cleri, Zach J. Lewis, Joel Leja, Jakob M. Helton, Emilie Burnham, Olivia Curtis, Anna de Graaff, Michaela Hirschmann, Harley Katz, Michael V. Maseda, Ian McConachie, Adéle Plat, Lucie Scharré
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43088
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30410

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

RUBIES: The Evolution of the Ionization Parameter from 0 < z < 9 Facilities: JWST (NIRSpec), SDSS, ESO/VLT (VIMOS), Keck (MOSFIRE) Astropy ( 3 ) , linmix ( 41 ) , Matplotlib ( 34 ) , NumPy ( 31 ) , pandas ( 81 ) , Prospector ( 37 ) , SciPy ( 126 ) Nikko J. Cleri, Zach J. Lewis, Joel Leja, Jakob M. Helton, Emilie Burnham, Olivia Curtis, Anna de Graaff, Michaela Hirschmann, Harley Katz, Michael V. Maseda, Ian McConachie, Adéle Plat, Lucie Scharré Abstract High-redshift galaxies have smaller radii, harder ionizing continua, and higher ionizing photon production efficiencies than lower redshift systems, which implies a corresponding evolution in the nebular conditions. A key metric to quantify gas properties is the ionization parameter, q q , the ratio of the local ionizing photon flux to the local hydrogen density. The ionization parameter is often inferred from observed emission line ratios, namely O32 ≡ [O III] / [O II] \equiv\hbox{[{O~III}]}/\hbox{[{O~II}]} . Prior to JWST, statistical samples of ionization parameter-sensitive emission lines in the rest-frame optical remained inaccessible for high-redshift galaxies. We investigate the dimensionless ionization parameter, U ≡ q / c U\equiv q/c at 3 ≲ z ≲ 9 3\lesssim z\lesssim 9 , inferred using Cloudy photoionization models from the O32 ratios for 434 galaxies in the RUBIES survey with JWST/NIRSpec PRISM and G395M spectroscopy, constituting the largest high-redshift population study of the ionization parameter to date. We compare to lower-redshift samples from SDSS, LEGA-C, and KBSS to probe the evolution of the ionization parameter from 0 ≲ z ≲ 9 0\lesssim z\lesssim 9 . We find that the ionization parameter increases with redshift and specific star formation rate, and decreases with stellar mass. We combine the predictive power with multivariate relations to estimate log U from redshift, stellar mass, and specific star formation rate for use in cases where O32 is not available from spectroscopy and show that the ionization parameter increases with redshift even at fixed stellar mass and specific star formation rate by a factor of ∼ 4 \sim 4 from z = 2 z=2 to z = 6 z=6 . Crucially, and in contrast to previous linear best-fit calibrations, our inference results in a systematic uncertainty in log U of ∼ 0.3 \sim 0.3 dex at zero measurement uncertainty due to the wide range of photoionization models that predict the same O32 ratio without informative priors. Finally, we discuss future modeling frameworks to accept many observed emission lines, unlocking the potential to simultaneously constrain gas-phase abundances, densities, ionizing sources, and ionization parameters to high accuracies for individual galaxies. show][email protected] I Introduction Spectroscopy of strong rest-frame UV/optical emission lines can be used to determine a wide range of physical properties of the observed galaxy, from determining sources of ionizing photons ( 4 ; 125 ; 48 ; 50 ; 121 ; 122 , e.g.,) , to deriving the conditions of the interstellar medium (ISM) such as gas-phase metallicities, temperatures, and densities ( 77 ; 133 ; 130 ; 45 ; 46 ; 120 ; 24 ; 64 ; 63 ; 95 ; 96 ; 109 ; 110 ; 112 ; 113 ; 103 ; 104 ; 105 ; 119 ; 118 ; 18 ; 91 , e.g.,) , to estimating star formation rates ( 43 ; 44 ; 42 ; 47 , e.g.,) and dust attenuation ( 11 ; 10 ; 92 , e.g.,) . The ionization parameter is a key metric to quantify nebular conditions, though it is rarely the direct focus of high-redshift galaxy studies due to the intrinsic degeneracies of the conversion from observed emission line ratios (e.g., O32) to the ionization parameter (e.g., 98 , and see Section IV.2 for discussion) . The ionization parameter is defined as \displaystyle q \displaystyle\equiv\frac{\Phi_{H^{0}}}{n_{H}} (1) where \Phi_{H^{0}} is the local ionizing photon flux incident on the face of the cloud and n_{H} is the local hydrogen density. This ionization parameter has units of cm s -1 , and can be interpreted to first order as the speed at which the ionization front expands into the surrounding neutral medium ( 50 , e.g.,) . Throughout this work, we refer to the dimensionless ionization parameter, U\equiv q/c . To constrain the ionization parameter through observations, studies often use ratios of strong optical emission lines which trace different “ionization zones”: regions of a nebula where ions of similar ionization energies are present ( 112 ; 113 ; 45 ; 50 ; 5 ; 78 , e.g.,) . The four-zone model ( 5 ) defines the zones by the ionization energies needed to produce N + (“low”: 14.53-29.60 eV), S 2+ (“intermediate”: 23.33-34.79 eV), O 2+ (“high”: 35.12-54.94 eV), and He 2+ (‘‘very high’’: >54.42 eV). One such line ratio is O32 1 1 1 The O32 line ratio may also assume the definition \mathrm{O32}\equiv\frac{\hbox{[{O~III}]}~\lambda\lambda 4960,5008}{\hbox{[{O~II}]}~\lambda\lambda 3727,3730} , though each convention can be rescaled to match the other through the fixed \hbox{[{O~III}]}~\lambda 5008/\hbox{[{O~III}]}\lambda 4960=2.985 ( 111 , e.g.,) , where \displaystyle\mathrm{O32} \displaystyle\equiv\frac{\hbox{[{O~III}]}~\lambda 5008}{\hbox{[{O~II}]}~\lambda\lambda 3727,3730} (2) The [O III] emission lines arising from doubly-ionized oxygen ( O 2+ ) probe ionization energies 35.1-54.9 eV, and [O II] emission lines from singly-ionized oxygen ( O + ) require 13.6-35.1 eV. The O32 ratio represents an “intermediate” ionization parameter as a ratio between a “high” and “low” ionization species ( 45 ; 95 ; 5 , e.g.,) , which is the energy regime traced by the most common abundance, density, and ionizing source diagnostics. A primary advantage of O32 is the insensitivity to relative elemental abundances compared to other similar-ionization strong line ratios accessible in the rest-frame optical, such as \hbox{[{O~III}]}~\lambda 5008 / H \beta and \hbox{[{Ne~III}]}~\lambda 3870 / [O II] ( 5 , e.g.,) . O32 is observable in JWST/NIRSpec at 1\lesssim z\lesssim 9 , which makes it an accessible predictor of the ionization parameter across much of cosmic time. Previous works have suggested that ionization parameter-sensitive line ratios such as O32 evolve with redshift ( 98 , e.g.,) . An increasing ionization parameter in star-forming galaxies may cause them to be misidentified as AGN in commonly-used diagnostics ( 50 ; 16 , e.g., “BPT” diagrams) , and may cause additional systematics in derived quantities from spectral energy distribution (SED) fitting ( 9 ; 37 ; 57 , e.g.,) . Studies of the ionization parameter itself beyond cosmic noon ( z\sim 2 ) have been limited, primarily by sample size and selection function ( 98 , e.g.,) . This is in part due to the availability of rest-frame optical ionization parameter-sensitive line ratios at high- z being limited to JWST spectroscopy, and the model dependence of the conversion between observed line ratios and the ionization parameter. In this work, we use JWST spectroscopy and low-redshift comparison samples with large photoionization model suites to demonstrate the evolution of the ionization parameter and discuss the implications of such on other inferred properties. We leverage PRISM and G395M spectroscopy from the Red Unknown: Bright Infrared Extragalactic Survey ( 19 , RUBIES;) combined with spectroscopy of lower redshift galaxies ( z\lesssim 2 ) from ground-based surveys to trace the full evolution of the ionization parameter from 0\lesssim z\lesssim 9 . The remainder of this Letter

Read the complete text