Repeating Changing-Look AGNs

• Repeating Changing-Look AGNs are a unique subset of active galactic nuclei characterized by cyclic transitions in broad emission lines that reveal rapid accretion state changes.
• They are identified through multi-epoch spectroscopy and robust flux calibration from surveys like SDSS, LAMOST, and DESI, with transition timescales of 2–5.3 years.
• Their study provides critical insights into disk instability, accretion physics, and host galaxy evolution, challenging traditional AGN unification models.

Repeating Changing-Look Active Galactic Nuclei (RCL AGNs) are a distinctive subset of AGNs in which the broad emission lines—definitive signatures of a broad-line region (BLR)—repeatedly appear and disappear over different observational epochs. This cyclical transformation of spectral type between broad-line (type 1) and narrow-line/absent broad-line (type 2 or type 1.9/1.8) states occurs on timescales of years or less, offering a dynamic laboratory for accretion physics and AGN evolution. The underlying physical drivers, transition mechanisms, and host environment properties for RCL AGNs are central to constraining models of disk instability, AGN duty cycles, and the connection between accretion and circumnuclear gas dynamics.

1. Phenomenology and Sample Properties

RCL AGNs are observationally defined by the repeated appearance and disappearance of broad Balmer and/or UV emission lines in multi-epoch spectra. Spectroscopic confirmation requires at least three distinct spectra per object to document more than one state change. In the latest comprehensive paper, a sample of 25 RCL AGNs (19 newly identified) was extracted out of 331 known changing-look AGNs (CL AGNs), yielding an occurrence rate of ~8% for repeating transitions within CL AGNs (Dong et al., 21 Oct 2025).

Detection relies primarily on quantifying the relative BEL flux variation using a criterion such as

$R_s = \frac{S_b - S_d}{S_b} > 0.3,$

where $S_b$ and $S_d$ are the integrated broad emission line fluxes in the bright and dim states, respectively. The selection is validated by visual inspection and is supported by robust flux calibration using constant narrow lines ([O III] λ5007), and by comparing spectral states across surveys (SDSS, LAMOST, DESI) complemented by densely sampled mid-infrared (MIR) light curves.

The rapid expansion of the RCL AGN sample (from fewer than ten to now at least 25 well-documented cases (Dong et al., 21 Oct 2025), with additional candidates in other works (Wang et al., 21 Oct 2024, Dong et al., 14 Aug 2024, Hon et al., 2022)) provides statistically meaningful constraints on phenomenology and transition timescales.

2. Transition Timescales and Physical Interpretation

The measured "turn-on" and "turn-off" transition timescales (as determined by MIR light curve extrema bracketing spectral changes and dense sampling from WISE/NEOWISE) are tightly constrained in most RCL AGNs to the range 2–5.3 years (rest frame) (Dong et al., 21 Oct 2025). Typical dim or "plateau" states last 4–7 years, with transitions often occurring smoothly and symmetrically in rising and falling phases (Wang et al., 21 Oct 2024). Some extreme cases—such as those identified in (Hon et al., 2022)—exhibit full type changes on timescales as brief as a few months; these are exceptionally valuable for probing accretion physics.

Significantly, studies consistently find no correlation between the observed transition timescale and black hole mass (when measuring $M_{\mathrm{BH}}$ virially via BEL widths and $R-L$ relations):

$$M_{\mathrm{BH}} = f\,\frac{R_{\mathrm{BLR}}\Delta V^2}{G}$$

where $R_{\mathrm{BLR}}$ is from the continuum luminosity, $\Delta V$ is the line FWHM, and $f$ is a geometric factor (Dong et al., 21 Oct 2025). Statistical analyses yield negligible Pearson and Spearman coefficients (p ≳ 0.5), challenging purely viscous/accretion timescale-based models.

This indicates that the physical mechanisms enabling repeated CL transitions act on timescales unrelated to the standard global disk viscous timescale (which is typically ≳ 10⁴–10⁶ years for optical-emitting disks). Instead, more rapid modes—e.g., thermal or magnetic instabilities, front propagation, and local state changes in the disk—are implicated.

3. Physical Mechanisms and Disk Instability Models

Current theoretical models for RCL AGNs converge on intrinsic accretion-rate variability as the primary driver, largely discounting the dominance of line-of-sight obscuration or transient external events (such as tidal disruption events) except in isolated cases (Dong et al., 21 Oct 2025, Jana et al., 13 Nov 2024, Wang et al., 21 Oct 2024, Ricci et al., 2022). The abrupt (yet repeated) accretion changes result in major shifts in the ionizing continuum, modulating the ionization of the BLR and therefore the visibility of BELs.

Radiation pressure instability in a narrow annulus at the interface between the outer thin disk and inner optically thin ADAF is a leading mechanism for the cyclicity. The instability triggers limit-cycle oscillations in the mass inflow, yielding periods of strong irradiation and BLR formation, followed by disk evaporation and BLR disappearance (Śniegowska et al., 2019, Sniegowska et al., 2020). Large-scale magnetic fields, when present, can further decrease the limit-cycle period and increase outburst intensity by driving disk winds and enhancing angular momentum loss (Pan et al., 2021). In the radio-quiet regime, the rapid collapse of the inner ADAF into a thin disk via radiative cooling is proposed as an efficient driver of turn-on transitions on timescales that can be orders of magnitude shorter than the thin-disk viscous time (Li et al., 4 Jul 2025).

The relevant timescales are more consistent with heating/cooling front propagation within the disk (Wang et al., 21 Oct 2024, Dong et al., 21 Oct 2025), given by

$$t_{\mathrm{front}} \approx 11\,\text{yr} \times (h/R/0.05)^{-1} (\alpha/0.03)^{-1} (M_{\mathrm{BH}}/10^8 M_\odot)(R/100r_g)^{3/2}$$

with $h/R$ as the scale-height ratio and $\alpha$ the Shakura-Sunyaev viscosity parameter.

4. Host Galaxy Environments and Duty Cycle

Host galaxies of RCL AGNs are structurally and evolutionarily distinct from typical AGN hosts. They are preferentially found in the "green valley"—the transitional zone between star-forming and quiescent galaxies—with high Sérsic indices (≈4–5), high bulge-to-total light ratios, and low asymmetry (Dodd et al., 2020). This suggests that secular processes (bars, spirals, minor mergers) dominate the inward transport of gas fueling episodic accretion, as opposed to major mergers.

High central stellar concentrations facilitate conditions for both episodic and repeating gas inflows. Accordingly, the recurrent dormancy and reawakening of AGNs on timescales of a few years—a hallmark of RCL AGNs—imply a much more dynamic duty cycle for SMBH accretion, with the possibility of short-lived, repeated accretion episodes embedded within longer cycles of host galaxy evolution (Wang et al., 21 Oct 2024).

5. X-ray, Mid-Infrared, and Multiwavelength Signatures

RCL AGNs display multiwavelength signatures consistent with rapid intrinsic accretion-state changes. X-ray and MIR light curves closely track each other, and the emergence or disappearance of BELs is tightly synchronized with optical and MIR continuum excursions (Jana et al., 13 Nov 2024, Wang et al., 21 Oct 2024, Dong et al., 21 Oct 2025). These changes are not accompanied by significant new obscuration features or absorption column density changes, further supporting an intrinsic (rather than obscuration-driven) mechanism.

Eddington ratio ($$\lambda_{\mathrm{Edd}} = L_{\mathrm{bol}} / L_{\mathrm{Edd}}$$) consistently shows transitions around a critical threshold ($\lambda_{\mathrm{Edd}} \approx 0.01$), echoing the state-change physics of black hole X-ray binaries (Jana et al., 13 Nov 2024). For example, both optical and X-ray luminosity drop significantly in the dim states, and the recurrence of BELs is observed as the luminosity and Eddington ratio increase, often exceeding a threshold consistent with disk-wind BLR formation models (Wang et al., 2020, Wang et al., 2022). There is also a bluer-when-brighter trend in the continuum during transition (Yang et al., 2017, Guo et al., 2023).

6. Methodological Advances

The identification and characterization of RCL AGNs rely on:

• Repeated, well-calibrated multi-epoch spectroscopy across large surveys (SDSS, LAMOST, DESI, SkyMapper, and others) with flux scaling using constant narrow lines ([O III] λ5007).
• Systematic photometric monitoring with WISE/NEOWISE in the MIR and various optical transient surveys (CRTS, ASASSN, ATLAS, ZTF).
• Automated, yet rigidly checked, spectral decomposition (via packages such as PyQSOFit, QSfit, QSOFITMORE), supported by principal component analysis for host–nucleus separation.
• Quantitative selection criteria for BEL flux changes and flux ratios combined with extensive visual validation (Dong et al., 21 Oct 2025, Dong et al., 14 Aug 2024).

Dense MIR sampling is vital for constraining transition epochs to within a fraction of a year, enabling robust measurement of transition timescales and their possible dependencies.

7. Theoretical and Observational Implications

The robust occurrence rate (~8% of all CL AGNs) and the non-negligible fraction of AGNs with demonstrably repeating state changes establish RCL AGNs as a significant and non-pathological population. The lack of a dependence of transition timescale on black hole mass, the critical role of changes in Eddington ratio, and the inability of standard viscous timescales or pure orientation scenarios to explain the data collectively overturn simplified AGN unification models and require the inclusion of (magneto-)thermal disk instabilities, rapid front propagation, and dynamic disk–BLR coupling.

These findings argue for a fully time-dependent, multiwavelength, and multi-component approach to AGN accretion, BLR physics, and AGN/host coevolution in both observation and modeling. The accelerating growth of RCL AGN samples due to synoptic sky surveys and multi-epoch spectroscopy (SDSS-V and beyond) will further resolve the fundamental duty cycle and triggering mechanisms for episodic supermassive black hole accretion in the low-redshift universe.