[Paper Review] A spectroscopic analysis of the ionizing photon production efficiency in JADES and CEERS: implications for the ionizing photon budget

Anthony J. Pahl, Michael W. Topping+7 more|arXiv (Cornell University)|Jul 3, 2024

Radiation Therapy and Dosimetry5 citations

TL;DR

The paper combines JWST/NIRSpec data from CEERS and JADES to measure the ionizing photon production efficiency (xi_ion) of 163 galaxies from z ~1 to 6, correcting for dust via Balmer decrements and examining its redshift, UV luminosity, and nebular-metallicity correlations to reassess the ionizing photon budget for reionization.

ABSTRACT

We have used a combined sample of JADES and CEERS objects in order to constrain ionizing photon production efficiency ($ξ_{ m ion}$) from JWST/NIRSpec and JWST/NIRCam data. We examine 163 objects at 1.06 < z < 6.71 with significant (3$σ$) spectroscopic detections of H$α$ and H$β$ in order to constrain intrinsic H$α$ luminosities corrected from nebular dust attenuation via Balmer decrements. We constrain dust-corrected UV luminosities from best-fit spectral-energy distribution modeling. We find a sample median log$_{10}$($ξ{ m ion,0}$/erg Hz$^{-1}$) = $25.29^{+0.29}_{-0.37}$, assuming f$_{ m esc}$=0 for the escape fraction of Lyman continuum emission. We find significant correlation between $ξ_{ m ion,0}$ and z, with 17 objects at z > 4.64 having median log$_{10}$($ξ_{ m ion,0}$/erg Hz$^{-1}$) = $25.38^{+0.38}){-0.38}$, with those below having log$_{10}$($ξ_{ m ion,0}$/erg Hz$^{-1}$) = $25.24^{+0.30}_{-0.33}$. We also find significant, positive correlations between $ξ_{ m ion,0}$ and LUV; Wλ([O iii]); [O iii]λ5007/[O ii]λλ3726, 3729; and inverse correlations with metallicity. In contrast with some previous results, we find no trends between $ξ_{ m ion,0}$ and stellar mass, stellar dust attenuation, or UV slope. Applying a multivariate fit to $ξ_{ m ion,0}$, z, and MUV to an empirically-motivated model of reionization, and folding in f$_{ m esc}$ estimates from direct observations of the Lyman continuum at z ~ 3 from the Keck Lyman Continuum Spectroscopic survey, we find that the number of ionizing photons entering the IGM causes reionization to end at z ~ 5 - 7.

Motivation & Objective

• Constrain intrinsic ionizing photon production efficiency (xi_ion) for a spectroscopic sample spanning z ~1–6.7 using Balmer-decrement corrected H-alpha and UV measurements.
• Investigate how xi_ion,0 evolves with redshift and UV luminosity, and how Nebular lines and metallicity correlate with xi_ion.
• Assess implications for the ionizing photon budget and reionization history by integrating xi_ion with empirical f_esc constraints.
• Provide a baseline multivariate relation for xi_ion,0 as a function of z and M_UV to inform reionization models.

Proposed method

• Use JWST/NIRSpec spectroscopy from CEERS and JADES to detect H-alpha and H-beta with >3 sigma significance for 163 galaxies (1.06 ≤ z ≤ 6.71).
• Correct intrinsic H-alpha luminosities for nebular dust using the Balmer decrement and Case B recombination assumptions; derive xi_ion from n_dot_ion and L_UV,int.
• Estimate L_UV,int from best-fit SEDs and apply dust corrections with Calzetti or SMC extinction curves depending on z and M_*, following FAST fits.
• Calculate uncertainties via 10,000 Monte Carlo realizations perturbing L_UV, A_V, H-alpha, and H-beta; report xi_ion,0 as the median.
• Derive metallicities from strong-line diagnostics (R23, O32, O3, O2, Ne3O2) using a chi-squared combination, with Monte Carlo error propagation.
• Explore correlations of xi_ion,0 with z, M_UV, W_lambda([OIII]), O32, and metallicity; fit a multivariate relation for xi_ion,0 (Equation 5) and compare with reionization models.

Experimental results

Research questions

• RQ1 What is the dust-corrected ionizing photon production efficiency (xi_ion,0) for a JWST-based galaxy sample spanning 1.06 < z < 6.71?
• RQ2 How does xi_ion,0 correlate with redshift, UV luminosity, and nebular line properties such as W_lambda([OIII]) and O32?
• RQ3 What is the role of metallicity in xi_ion,0, and do xi_ion,0 show trends with stellar mass or UV slope?
• RQ4 Can a multivariate relation combining xi_ion,0, z, and M_UV reproduce reionization-era photon budgets under empirical f_esc assumptions?
• RQ5 How does xi_ion,0 evolve with redshift and UV brightness when accounting for sample completeness?

Key findings

• Median log10(xi_ion,0 / erg Hz^{-1}) = 25.29^{+0.29}_{-0.37} for the full sample at f_esc = 0.
• Significant positive correlation between xi_ion,0 and redshift (Spearman rs = 0.19, p = 0.01) overall, with higher-z subsamples showing higher medians.
• Significant, positive correlations between xi_ion,0 and L_UV, and between xi_ion,0 and W_lambda([OIII]); a positive correlation with O32 is also found, while metallicity is anti-correlated.
• No observed trends of xi_ion,0 with stellar mass, stellar dust attenuation, or UV slope in this sample.
• A multivariate fit yields log10(xi_ion,0/Hz erg^{-1}) = (0.02±0.01) z + (-0.04±0.02) M_UV + 24.38±0.41, illustrating joint dependence on z and M_UV.
• Using xi_ion,0, z, and M_UV in a reionization framework with f_esc from KLCS suggests ionizing photons entering the IGM could end reionization around z ~ 5–7.