Search for High-Ionization Nebular Emission (SHINE). I. A Systematically Selected [Ne V] Sample at z > 3 with JWST/NIRSpec PRISM

Shobita Satyapal, Sara Doan, Daniel Schaerer, Rui Marques-Chaves, Anurag Sawarkar, William Matzko, Camilo Vazquez, C. Daoutis, D. Korber, I. Morel, Mengyuan Xia

First listed 2026-09-24 | Last updated 2026-09-22

Abstract

We conduct the first systematic Search for High-Ionization Nebular Emission (SHINE) using [Ne V]$\lambda3426$ in JWST/NIRSpec PRISM spectroscopy at $z>3$. From more than 9,000 galaxies, we identify 25 [Ne V] emitters spanning $z=3.059$-$9.444$ and $\log_{10}(L_{\rm [Ne\,V]}/erg\,s^{-1})=40.70$-$42.86$. Their [Ne V] luminosities overlap with those of local [Ne V]-selected active galactic nuclei (AGNs) and exceed those of local metal-poor [Ne V] emitters, although such low-luminosity systems would fall below our sensitivity. The population is diverse, spanning compact and extended morphologies and a broad range of continuum properties and stellar masses, including very low-mass hosts. It includes sources with broad Balmer emission as well as others whose higher-resolution spectra do not require a broad component. The [Ne V] emitters overlap only partly with conventional AGN diagnostics: most do not satisfy conservative high-redshift AGN criteria based on strong rest-optical narrow-line ratios, and only a minority have adopted broad-line classifications. [Ne V] upper limits for independently selected broad-line AGNs and little red dots are too shallow to establish a population-wide [Ne V] deficit. Six sources have secure Chandra counterparts, showing that X-ray weakness is not universal among luminous [Ne V] emitters. Strong [Ne V]/[Ne III] emission is associated with redder ultraviolet slopes and stronger Balmer breaks, but not with UV luminosity, possibly linking strong high-ionization emission to recent changes in star formation. The observed incidence of luminous [Ne V] emission shows no significant evolution over $z>3$, despite an increasing robust [Ne III] detection fraction. The high [Ne V] luminosities and hard line ratios favor black-hole accretion as the dominant power source, establishing [Ne V] emission as a complementary probe of early black-hole growth.

Short digest

SHINE presents the first systematic [Ne V]λ3426 search in JWST/NIRSpec PRISM spectra at z>3, identifying 25 luminous emitters from a parent sample of more than 9,000 galaxies across z=3.059–9.444. Their [Ne V] luminosities and hard line ratios favor black-hole accretion, yet the sample is heterogeneous in morphology, stellar mass, continuum shape, and broad-Balmer properties, and overlaps only partly with standard broad-line and rest-optical narrow-line AGN selections. Six secure Chandra counterparts demonstrate that X-ray weakness is not universal, while the lack of significant evolution in the observed luminous [Ne V] incidence supports [Ne V] as a complementary route to finding early accretion beyond conventional JWST AGN selections.

Key figures to inspect

• Figure 1. This establishes the selection’s luminosity regime: SHINE [Ne V] emitters overlap local [Ne V]-selected AGNs and sit above the known local metal-poor comparison population, framing why accretion is favored for the detected sources.
• Figure 4. The high-ionization diagnostic planes show where the 25 [Ne V] sources fall relative to JWST broad-line AGNs, LRDs, narrow-line selections, local AGNs, and metal-poor galaxies, directly illustrating the sample’s incomplete overlap with conventional AGN criteria.
• Figure 6. This comparison tests whether independently selected broad-line AGNs, LRDs, and S2–VO87 AGNs lack [Ne V] relative to [Ne III]; the shallow upper limits make clear why the paper cannot claim a population-wide [Ne V] deficit in those samples.
• Figure 10. This is the population-level conclusion figure: after imposing a common apparent [Ne V] luminosity threshold and sensitivity requirement, the observed luminous-emitter fraction shows no significant evolution across z>3, with the stated caveat that it is not completeness-corrected.