Little Red and Blue Dots: AGN-excited narrow lines, Lyman-$α$ emission, and resemblance to standard quasars

Sophia Geris, Roberto Maiolino, Xihan Ji, Guido Risaliti, Giorgio Lanzuisi, Francesco D'Eugenio, Yuki Isobe, Gareth Jones, Anishya Harshan, Matilde Brazzini, Ignas Juodžbalis, Jan Scholtz, Pierluigi Rinaldi, Hannah Übler, William Baker, Andrew J. Bunker, Marcella Brusa, Stefano Carniani, Stephane Charlot, Mirko Curti, Andrea Comastri, Emma Curtis Lake, Roberto Gilli, Kevin Hainline, Piero Madau, Stefano Marchesi, Giovanni Mazzolari, Lorenzo Napolitano, Eleonora Parlanti, Laura Pentericci, Cristina Ramos Almeida, Brant Robertson, Maddie S. Silcock, Roberta Tripodi, Giacomo Venturi, Cristian Vignali, Fabio Vito, Yongda Zhu

First listed 2026-06-23 | Last updated 2026-06-19

Abstract

We present an analysis of a sample of 36 Little Red and Blue Dots (LRDs and LBDs) at $2.26<z<7.89$, identified by JWST in the GOODS fields. While both categories are selected to have broad Balmer lines, both of them are extremely X-ray weak. Both classes share the same location on various diagnostic diagrams, consistent with AGN excitation (with some deviations which can be ascribed to low metallicity), although their weak HeII emission suggests a generally softer ionizing spectrum than ordinary AGN. LRDs display Ly$α$ emission stronger than normal star-forming galaxies, and with a broad component consistent with the broad component of H$α$. Overall, these findings indicate that LRDs and LBDs are both powered by growing black holes and their ionizing radiation escapes to ionize the surrounding interstellar medium (ISM). The broad Balmer lines ($Hα_b$ and $Hβ_b$) have different apparent properties: LBDs have EW(H$α_b$) and $Hα_b/Hβ_b$ broadly consistent with normal AGN, while LRDs have higher values of both quantities, although still in the tail of the quasars distribution. LRD models in which a gas envelope completely encases the black hole, are inconsistent with these results -- these scenarios need modification to include clumpiness, or a (classical) equatorial geometry, letting ionizing photons reach the ISM. The different broad Balmer properties imply that LBDs cannot simply be LRDs with more galaxy contribution. Scenarios in which LRDs are simply dust-obscured LBDs seem broadly consistent with the observations. Finally, these results indicate that LRDs' bolometric luminosities estimated assuming isotropic emission and complete covering by the absorber are inadequate. The few X-ray-detected LRDs suggest no deviation from the standard AGN bolometric corrections, once absorption is accounted for.

Short digest

Using 36 JWST-selected Little Red and Blue Dots in GOODS at 2.26<z<7.89, this paper asks whether the two classes are fundamentally different or instead variants of the same accreting black-hole population. The main result is that both LRDs and LBDs occupy narrow-line diagnostic loci consistent with AGN excitation despite being extremely X-ray weak, while weak HeII points to a softer ionizing spectrum than in standard quasars. LRDs additionally show Lyα stronger than ordinary star-forming galaxies, with a broad Lyα component matching broad Hα, implying that ionizing radiation escapes to the surrounding ISM rather than being fully trapped in a closed cocoon. The broad-line comparison is the key discriminator: LBDs look broadly quasar-like, whereas LRDs shift to higher broad-Hα equivalent width and Balmer decrement, disfavoring fully enclosing gas-envelope models and favoring obscured, clumpy, or equatorial-geometry interpretations in which LRDs can be dustier analogs of LBDs and standard isotropic bolometric estimates break down.

Key figures to inspect

• Figure 1. Use this as the selection-definition figure. It shows exactly how the authors separate LRDs, LBDs, and more classical reddened AGN in rest-frame UV versus optical slope space, while also marking the stacked spectra and the two Rosetta Stone objects that recur throughout the paper.
• Figure 3. This is the cleanest overview of the paper's X-ray weakness result. It places both the individual sources and the stacks in bolometric-to-X-ray luminosity space relative to local quasars, NLS1s, SEAMBHs, and the few X-ray detections, making the mismatch with standard expectations and the bolometric-correction discussion immediately visible.
• Figure 7. This figure carries the core narrow-line evidence that both LRDs and LBDs are AGN excited. The multi-panel BPT and related diagnostic diagrams show that the stacks and many individual objects land in AGN-consistent regions once low-metallicity offsets are allowed for, while also emphasizing that some line ratios differ from the local AGN locus.
• Figure 13. This is the most important comparison figure for the paper's interpretation of the two classes. By plotting broad-Hα equivalent width against the broad Balmer decrement for the stacks, Rosetta Stones, and SDSS quasar distribution, it makes clear that LBDs are broadly quasar-like while LRDs occupy a more extreme but still quasar-overlapping tail, which is central to the obscured-LBD and non-fully-enclosed-geometry conclusions.
• Figure 18. Use this as the strongest Lyα-physics figure. The matched red-wing velocity profiles of Lyα and Hα in the LRD stack argue that the broad Lyα component is linked to the same broad-line region emission rather than being produced solely by resonant scattering, directly supporting the claim that AGN radiation escapes to ionize the surrounding gas.