The Extreme Rarity and Physical Properties of Low-redshift AGNs with Balmer Absorption

Jinyi Shangguan, Chang-Hao Chen, Luis C. Ho, Jiwei Liao, Yanqing Liu, Chengzhou Wu, Ruancun Li, Kohei Inayoshi, Linhua Jiang

First listed 2026-06-04 | Last updated 2026-06-03

Abstract

Balmer absorption lines are increasingly observed in the little red dots (LRDs) discovered by the James Webb Space Telescope, potentially tracing dense circumnuclear gas around rapidly accreting black holes. Motivated by this connection, we search for Balmer absorption using homogeneously analyzed spectra of a representative parent sample of 14,584 low-redshift ($z<0.35$) type 1 active galactic nuclei selected from the Sloan Digital Sky Survey. We identify seven sources with robust Balmer absorption (occurrence $\sim 0.05\%$) and model them with a partially covering absorber model, accounting for the spectral resolution. By fitting H$α$, H$β$, and H$γ$ simultaneously and tying their optical-depth ratios to theoretical values, we constrain optical depth at the line center ($τ_0$) and the covering factor ($C_f$). All sources with robust modeling require optically thick H$α$ absorption and typically moderate covering factors ($C_f\approx 0.2-0.6$), while the LRD analog J1025 shows $C_f \gtrsim 0.8$ consistent with recent measurements of high-redshift LRDs. The absorbers have modest velocity offsets ($\sim 150-850\,\mathrm{km\,s^{-1}}$) and narrow intrinsic widths ($\sim 20-200\,\mathrm{km\,s^{-1}}$). Multi-epoch spectroscopy of three sources reveals Balmer-absorption variability on both year and month timescales. Three objects exhibit exceptionally weak Fe II emission, high Eddington ratio, and low gas-phase metallicity, an atypically rare combination of properties that might elevate the incidence of Balmer-absorption to $\sim$10%. We argue that low-metallicity conditions may suppress disk winds and help retain dense neutral gas along the line-of-sight in systems of high accretion rate.

Short digest

Using homogeneously analyzed SDSS spectra of 14,584 z<0.35 type 1 AGNs, the authors find just seven robust Balmer-absorption systems, implying an extreme local occurrence of about 0.05%. Simultaneous partially covering fits to Hα, Hβ, and Hγ show optically thick Hα absorption in every well-modeled case, usually with covering factors of about 0.2-0.6, modest velocity offsets of roughly 150-850 km s^-1, and narrow intrinsic widths of about 20-200 km s^-1, while the local LRD analog J1025 reaches Cf ≳ 0.8, much closer to recent high-z LRD measurements. Multi-epoch spectroscopy of three sources shows Balmer-absorption variability on both month and year timescales, pointing to compact circumnuclear gas rather than a static large-scale screen. The most suggestive subset is three weak-Fe II, high-Eddington, low-metallicity AGNs where the Balmer-absorption incidence may rise to about 10%, supporting the idea that metal-poor conditions suppress disk winds and help preserve dense neutral gas along the line of sight to rapidly accreting black holes.

Key figures to inspect

• Figure 3. This is the key evidence figure for the absorber modeling, because it shows how changes in optical depth, velocity dispersion, and covering factor map onto the observed broad Hα plus [N II] profiles across the sample. It lets the reader see that the Balmer troughs are not just decomposition residuals and that the inferred absorber parameters are tied directly to the line-shape distortions, including the resolution-dependent comparison between the SDSS and DESI spectra of J1545.
• Figure 15. This figure places the seven Balmer-absorption AGNs inside the full 14,584-object parent sample using broad Hβ width, Fe II strength, and [O III] equivalent width. It is central for understanding the paper's rarity argument and for identifying the three unusual systems that combine weak Fe II and strong narrow-line signatures, the corner of parameter space that motivates the proposed enhanced Balmer-absorption incidence and low-metallicity connection.
• Figure 18. This is one of the strongest physics-diagnostic figures because it connects the measured Hα optical depth and Balmer-absorption equivalent widths to Eddington ratio, while explicitly removing J2220 from the formal correlation tests because its measurements are uncertain. It shows how absorption strength tracks accretion state and also highlights the differing behavior of Hα and Hβ, which matters for interpreting optical-depth effects rather than treating the troughs as a single phenomenological feature.
• Figure 19. This figure is especially important for the paper's bottom-line claim about absorber geometry, because it shows how covering factor varies with both Eddington ratio and Hα optical depth. It also isolates J1025 as the local LRD analog with an unusually high covering factor comparable to values reported for high-redshift LRDs, making this the clearest bridge between the low-z sample and the JWST-selected Balmer-absorption population.
• Figure 20. The SED comparison figure matters because it places these local Balmer-absorption AGNs in the broader LRD phenomenological context using far-UV to mid-IR data. By comparing each source to the average SDSS SED and to the LRD RUBIES-EGS-49140, it shows that the sample generally has red UV continua associated with obscured or embedded accretion, while also making clear that the local systems are usually less extreme than a canonical JWST little red dot.