TBD LBD: The nature of `little blue dots'

Abstract: Previous Sirocco radiative-transfer models of gas-cocooned AGN predicted lower-column counterparts to little red dots (LRDs): compact, X-ray-weak sources with bluer continuum slopes and Balmer jumps rather than Balmer breaks. The recently identified population of little blue dots (LBDs) closely resembles this predicted phase. Here we explore these lower-column-density cocoons in which nebular recombination emission remains visible while strong Balmer-continuum absorption is avoided. We find that a sequence of increasing column density connects more classical AGN spectra, Balmer-jump LBD-like spectra at $$N_{\rm H}!\sim!{\rm few}\times10^{24} \mathrm{cm<sup>{-2}}$$, and Balmer-break LRD-like spectra at higher columns. In this sequence, electron scattering produces exponential line wings and suppresses X-ray emission before strong Balmer absorption features, characteristic of higher column densities, appear. We therefore propose that LBDs are lower-column analogues of LRDs within a common gas-cocooned AGN sequence. This interpretation predicts that Balmer-jump emission, X-ray weakness, permitted lines with exponential wings, He II $λ$4686 emission, smaller H$α$ FWHM values and equivalent widths than in LRDs, and weak or absent absorption features are characteristic of LBDs. We compare to three example LBD spectra and identify Balmer-jump signatures in them.

• The paper shows that lower-column gas cocoons with hydrogen columns of a few ×10²⁴ cm⁻² can reproduce the blue continua and Balmer-jump emission observed in little blue dots, while higher columns produce little red dot spectra.
• The paper uses Sirocco Monte Carlo radiative-transfer models to connect electron-scattering line wings, reduced Hα widths, weak absorption, He II emission, and Compton-thick X-ray suppression within one AGN sequence.
• The paper finds Balmer-jump signatures at highly significant levels in Nexus 5819, Rubies 50052, and GS 3073, supporting column density as a possible physical or evolutionary link between little blue dots and little red dots.

Overview and motivation

This paper by Sneppen, Watson, Matthews, and Nikopoulos (2606.12509) addresses the physical relationship between two JWST-discovered populations of compact AGN: the well-studied "little red dots" (LRDs) and the more recently identified "little blue dots" (LBDs). Building on prior Sirocco Monte Carlo radiative-transfer modelling of LRDs as supermassive black holes embedded in dense, translucent ionised cocoons, the authors investigate whether LBDs correspond to the lower-column-density regime of the same cocooned-AGN sequence. The central claim is that a single parameter — hydrogen column density $N_{\rm H}$ through the cocoon — organises both classes: at $$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$, nebular recombination produces visible Balmer-jump emission without strong neutral absorption (LBD-like spectra), whereas at higher columns ($\sim10^{25}\,\mathrm{cm^{-2}}$), absorption in a partially neutral outer layer converts the jump into an LRD-like Balmer break.

The paper is positioned against an alternative orientation-based interpretation in which LBDs are pole-on views of intrinsically blue, super-Eddington accretion flows whose equatorial sightlines are self-shadowed. Both frameworks unify LBDs and LRDs within one rapidly accreting population; the distinction lies in whether colour is set primarily by inclination or by gas column.

Methods

The models use Sirocco, a Sobolev Monte Carlo ionisation and radiative-transfer code that self-consistently computes the radiation field, ionisation state, continuum shapes, line profiles, and line ratios across the X-ray to near-infrared SED. The fiducial configurations are cocoons with $$N_{\rm H}\sim(1{-}15)\times10^{24}\,\mathrm{cm^{-2}}$$, sub-solar metallicity ($Z=0.1\,Z_\odot$), spatial scales of $R\sim10^{16}$–$10^{17}$ cm, characteristic velocities of a few hundred $\mathrm{km\,s^{-1}}$, and a central blackbody illuminating source with $T_{\rm BB}=30{,}000$ K. For the panchromatic comparison, a narrow-line Seyfert 1 X-ray template with $k_{\rm bol,X}=50$ is adopted to assess Compton-thick suppression. Model spectra are convolved to observed resolution and reddened modestly with $$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$0 under a Calzetti attenuation law — somewhat larger than the $$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$1 inferred for GS 3073, but consistent with the weak mid-infrared dust emission of the class. The authors note that these spectra are typical of this part of the simulation grid rather than extreme cases, and that the density sequence varies local density rather than geometric scale.

Balmer-jump emission in LBD spectra

A key observational test is the presence of a continuum offset across the Balmer limit at 3645 Å, diagnostic of nebular recombination emission. Comparing three example LBD spectra — GS 3073, Nexus 5819, and Rubies 50052 — against a low-column Sirocco model, the paper measures continuum-normalised flux ratios redward of the edge after fitting the continuum blueward of H$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$2. The results are statistically significant for all three objects:

Object	Combined windows	$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$3	Significance
Nexus 5819	combined	$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$4	19.5$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$5
Rubies 50052	combined	$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$6	6.0$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$7
GS 3073	combined	$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$8	10.7$$N_{\rm H}\sim{\rm few}\times10^{24}\,\mathrm{cm^{-2}}$$9

Nexus 5819 shows the most prominent offset, detected independently in both wavelength windows at up to 17.5$\sim10^{25}\,\mathrm{cm^{-2}}$0. The feature is expected to be smoothed rather than abrupt because high-order hydrogen lines fill in the continuum, analogous to Paschen jumps in LRDs. The authors also consider and reject \ion{Fe}{ii} pseudo-continuum contamination as an alternative explanation, noting the absence of other \ion{Fe}{ii} lines and the smooth continuation of the blue continuum down to 2000 Å. This constitutes direct evidence that LBDs exhibit the Balmer-jump signature predicted for the low-column cocoon regime.

A shared electron-scattering regime

The paper argues that the same ionised column responsible for bound-free emission also produces two further observables: exponential broad-line wings from electron scattering and Compton-thick X-ray suppression. All fitted broad H$\sim10^{25}\,\mathrm{cm^{-2}}$1 profiles in LBDs are better described by exponential than multi-Gaussian forms — in contrast to the general JWST/X-ray AGN population — which the authors treat as strong evidence for the electron-scattering interpretation. In the panchromatic model comparison with GS 3073, the column that broadens H$\sim10^{25}\,\mathrm{cm^{-2}}$2 simultaneously suppresses the emergent X-ray luminosity below the 7 Ms Chandra Deep Field limit at $\sim10^{25}\,\mathrm{cm^{-2}}$3. The implication is that X-ray weakness in LBDs carries the same physical information as in LRDs: it reflects the scattering column rather than an absence of accretion power.

Predicted properties along the column-density sequence

Across the model grid, several robust predictions emerge for Balmer-jump objects relative to LRDs:

• Exponential wings on permitted hydrogen lines, present in both regimes but narrower at lower columns.
• Larger core-to-broad flux ratios and smaller broad-component FWHMs, since electron-scattering optical depth grows with column.
• Weak or absent Balmer and He absorption, because the partially neutral absorbing layer has not yet become optically significant.
• Intrinsically blue UV and optical slopes, matching the LBD colour space defined by rest-frame slope measurements at 2160–3100 Å and 5360–6270 Å.
• Reduced H$\sim10^{25}\,\mathrm{cm^{-2}}$4 equivalent widths, since more of the intrinsic continuum escapes transmission through a translucent cocoon.
• Visible high-ionisation lines such as \ion{He}{ii}\,$\sim10^{25}\,\mathrm{cm^{-2}}$54686, broader than the hydrogen lines because they arise preferentially in the innermost cocoon regions; this matches observations of GS 3073.

Notably, the authors report that no model spectrum in their explored space exhibits simultaneous Balmer jumps and Balmer-line absorption, implying such configurations should be observationally rare — a falsifiable prediction. They also find that the jump-to-break transition remains primarily organised by $\sim10^{25}\,\mathrm{cm^{-2}}$6 even when density normalisation and geometric scale are varied independently, strengthening the case that column density, not geometry, is the controlling variable.

Relation to orientation-based unification models

The discussion contrasts this column-driven picture with the proposal that LBDs are face-on counterparts of LRDs viewed through a geometrically thick, self-shadowed accretion flow. Both interpretations connect the populations within one rapidly accreting AGN family, but they differ in the primary ordering parameter. Under the cocoon interpretation, LRDs and LBDs trace different columns of dense circumnuclear gas — possibly different evolutionary stages — rather than viewing angles. GS 3073 serves as a useful test object because it combines canonical LBD phenomenology (compactness, X-ray weakness, non-Gaussian broad lines) with the specific Balmer-jump behaviour predicted here, including prominent \ion{He}{ii} emission broader than H$\sim10^{25}\,\mathrm{cm^{-2}}$7.

Limitations and open questions

Several caveats bear directly on the strength of the conclusions. First, the comparison to observations is qualitative and restricted to three illustrative spectra; full spectral fitting and population-level inference against larger LBD samples remain outstanding, partly because each Sirocco spectrum costs of order $\sim10^{25}\,\mathrm{cm^{-2}}$8 CPU hours. Second, the precise onset column for absorption signatures depends on the assumed ionisation structure of the cocoon, so the boundary between LBD-like and LRD-like behaviour is not sharply determined. Third, the models predict broad Ly$\sim10^{25}\,\mathrm{cm^{-2}}$9 emission from the cocoon, but the Sobolev approximation precludes quantitative prediction of its emergent profile without dedicated Ly$$N_{\rm H}\sim(1{-}15)\times10^{24}\,\mathrm{cm^{-2}}$$0 radiative transfer. Fourth, the adopted input SED (a 30,000 K blackbody), metallicity, and modest Calzetti reddening are assumptions whose influence on detailed spectral morphology has not been fully marginalised. Finally, the competition between the column-based and orientation-based unification pictures is not settled by the evidence presented here; discriminating tests using larger samples are required.

Conclusion

This paper extends gas-cocooned AGN radiative-transfer modelling to lower column densities and identifies the resulting spectral phase with the emerging LBD population. The unified sequence predicts blue continua, compactness, X-ray weakness, exponential broad lines, Balmer-jump emission, weaker absorption, smaller H$$N_{\rm H}\sim(1{-}15)\times10^{24}\,\mathrm{cm^{-2}}$$1 FWHMs and equivalent widths, and detectable \ion{He}{ii}\,$$N_{\rm H}\sim(1{-}15)\times10^{24}\,\mathrm{cm^{-2}}$$24686 — properties already partially confirmed in GS 3073, Nexus 5819, and Rubies 50052, with Balmer-jump offsets measured at up to 19.5$$N_{\rm H}\sim(1{-}15)\times10^{24}\,\mathrm{cm^{-2}}$$3 significance. If borne out by systematic fitting of larger samples, LBDs and LRDs would constitute a single physical class ordered by circumnuclear gas column rather than distinct phenomena, with column density serving as a possible tracer of evolutionary stage in early SMBH growth.