Changing-Look Active Galactic Nuclei

• Changing-Look AGNs are defined by dramatic optical and X-ray spectral changes, notably the emergence or disappearance of broad Balmer lines on timescales of months to years.
• Observational surveys report CLAGN event rates from ~1.7% to 25%, linking these transitions to accretion state changes, inner disk instabilities, and varying obscuration.
• Theoretical models attribute CLAGN behavior to accretion rate fluctuations, disk instability, and disk-wind dynamics, offering new insights into SMBH-galaxy coevolution and AGN unification.

Changing-Look Active Galactic Nuclei (CLAGNs) are a rare and phenomenologically distinct subset of AGN that display dramatic, order-of-magnitude changes in their optical and/or X-ray spectral appearance on timescales of months to decades. These transitions typically involve the sudden emergence or disappearance of broad Balmer emission lines and a factor of several change in blue/UV continuum flux, reflecting changes in the accretion flow structure or line-of-sight obscuration. CLAGNs challenge the traditional, purely orientation-based AGN unification paradigm and are now a focal point for studies of accretion-disk physics, SMBH-galaxy coevolution, and AGN variability.

1. Operational Definition and Observational Diagnostics

A CLAGN is empirically identified by a change of spectroscopic type between classical Type 1 (broad Hα and Hβ, FWHM $\sim$ 10³–10⁴ km s⁻¹) and Type 1.9/2 (narrow lines only), or vice versa, over humanly accessible timescales (months to $\sim$decade). The canonical diagnostic is the full appearance or disappearance of broad emission lines, quantified by the ratio

$$R = \frac{[\text{flux(total}\ H\beta\ \text{broad+narrow)}]}{\text{flux}([O\ III]\ \lambda5007)}$$

with specific thresholds delineating Seyfert 1‒2 subclasses (Hon et al., 2022). Additional criteria include continuum variation of $\gtrsim$0.5 mag in the blue optical/UV and, in some cases, correlated changes in the X-ray continuum. The transition timescales, ranging from $\sim$1 month to $\sim$10 years (Zeltyn et al., 2022, Amrutha et al., 30 Oct 2024), far exceed typical BLR light-crossing times (tens of days) yet are orders of magnitude shorter than viscous timescales for a standard thin disk ($t_{\rm visc} \sim 10^4$ yr at $R \sim 10^3 R_g$). These rapid changes cannot be explained by stochastic variability or major mergers alone, and often cannot be reproduced by simple obscuration scenarios (Dodd et al., 2020, Zeltyn et al., 2022, Jin et al., 2021, Titarchuk et al., 14 Nov 2024).

2. Demographics, Selection, and Event Rates

CLAGNs are rare in statistical samples. Spectroscopic monitoring of large, unbiased parent samples finds:

• Incidence $f_{\rm CLAGN}\sim$ 1.7% (turn-on) and 9.6% (turn-off) over 15 years in a volume-limited $z<0.1$ sample (Amrutha et al., 30 Oct 2024); completeness-adjusted rates approach $\sim$10% per 15 yr for turn-off events, after accounting for orientation bias.
• In deeper samples, a total CLAGN rate of $\sim$25% over 15 yr is reported for low-luminosity AGN, substantially higher than the $\lesssim$0.1% found among luminous quasars due to lower host dilution and preferentially larger fractional accretion-rate swings at lower black-hole mass (Hon et al., 2022, Guo et al., 2023).
• Large spectroscopic and photometric surveys (SDSS, LAMOST, DESI, VLASS, ASKAP, ZTF, SkyMapper) employing tailored variability metrics (MIR amplitude, color change, optical light-curve structure function, etc.) have expanded the CLAGN catalog to several hundred confirmed cases, enabling robust measurement of event rates as a function of accretion rate, black hole mass, and host galaxy properties (Wang et al., 13 Nov 2025, Dong et al., 14 Aug 2024, Birmingham et al., 2 Jul 2025).

CLAGN transition types cluster around two paradigms: "turn-on" events (Type 2/1.9 $\rightarrow$ Type 1.x) and "turn-off" events (Type 1.x $\rightarrow$ Type 1.9/2). Most CLAGNs are found in lower-Eddington-ratio systems (median $\lambda_{\rm Edd} \sim 0.01$–0.05) compared to the parent AGN population (Jin et al., 2021, Dong et al., 14 Aug 2024, Lusso et al., 24 Feb 2025). Empirical rates are summarized below:

Survey/Method	Turn-On Rate	Turn-Off Rate	Timescale
6dFGS+ATLAS (Amrutha et al., 30 Oct 2024)	1.7% (15 yr)	9.6% (15 yr)	2–15 yr
SkyMapper+6dFGS (Hon et al., 2022)	12% (15 yr)	$\sim$12% est.	$<$3 mo–15 yr
SDSS+LAMOST (Dong et al., 14 Aug 2024)	---	---	$\sim$1–10 yr

3. Physical Drivers: Accretion-State Transitions, Disk Instability, and Obscuration

3.1. Accretion-State Transition Paradigm

The dominant physical interpretation is that CLAGNs represent transitions between two accretion modes—radiatively efficient, thin-disk (Shakura-Sunyaev) and radiatively inefficient, hot flow (RIAF/ADAF)—driven by moderate changes in $\dot{M}$ (Liu et al., 2022, Titarchuk et al., 14 Nov 2024, Lusso et al., 24 Feb 2025).

• At $\lambda_{\rm Edd} \gtrsim 0.01$–$0.02$, the disk remains optically thick with strong UV output, sustaining a BLR via photoionization: broad Balmer lines appear ("on" state).
• Below $\lambda_{\rm Edd} \lesssim 0.01$, the inner disk transitions to an ADAF/RIAF, suppressing UV/soft X-ray emission, quenching the BLR: broad lines vanish ("off" state) (Titarchuk et al., 14 Nov 2024, Dong et al., 14 Aug 2024).
• The threshold $\lambda_{\rm MIR}\approx 0.004$–$0.005$ (as measured in WISE bands) is consistently observed in both optical and X-ray CLAGN samples where broad lines appear/disappear (Bing et al., 2022).
• The X-ray photon index ($\Gamma$) bifurcates with Eddington ratio: positive $\Gamma$–$L_{2-10\,\mathrm{keV}}/L_{\rm Edd}$ for $\lambda_{\rm Edd} \gtrsim 10^{-3}$; negative for $\lambda_{\rm Edd} \lesssim 10^{-3}$, matching the "soft" and "hard" accretion states (Liu et al., 2022).

3.2. Disk Instability and Disk Tearing

Rapid (months-to-years) CL transitions cannot be accounted for by standard viscous timescales, but may be triggered by:

• Thermal or Magnetorotational Instabilities: Turbulent fluctuations and ionization-front propagation can move at the sound speed or thermal speed, producing order-of-magnitude luminosity changes on $\mathcal{O}$(yr) scales (Guo et al., 2023).
• Disk Tearing: GRMHD simulations of tilted thin disks demonstrate that strong Lense-Thirring torque can break the inner disk into precessing segments, launching shocks and causing quasi-periodic, large-amplitude continuum and broad-line variations. Simulated CLAGN events show $>10\times$ changes in both continuum and broad-line luminosity over months–years, corresponding to observed CLAGN phenomenology (Kaaz et al., 12 Nov 2025).

3.3. Obscuration vs Accretion Change

While in rare cases extremely rapid (weeks–months) state changes with matched continuum and line suppression can be reproduced by variable extinction (single-law dust screens with $A_V\sim1$–2), physical models favor accretion-driven scenarios given the physical implausibility of such rapid, coordinated obscuring clouds at parsec scales (Zeltyn et al., 2022). Typical CLAGN transitions show weak evidence for strong $N_H$ variation, and disappearance of the BLR is not matched by corresponding changes in narrow-line flux as would be expected for pure obscuration (Lusso et al., 24 Feb 2025, Titarchuk et al., 14 Nov 2024).

Exception: Some CLAGN display transitions best explained by temporary, large-scale, high-velocity (possibly nuclear) dust obscuration, particularly for events with $<2$ month timescales and continuum/line dimming in lock-step, but these are a minority (Zeltyn et al., 2022).

3.4. Disk-Wind Scenario and X-ray Obscuration

A significant subset of CLAGNs also show large, correlated changes in X-ray column density ($N_{H,\rm los}$) in both Compton-thin and –thick regimes. The strong anti-correlation between $N_{H,\rm los}$ and $L_X/L_{\rm Edd}$ in multiple objects supports a disk-wind scenario in which rising $\dot{M}$ both increases ionizing flux (turning on BLR) and physically drives wind material out of the line of sight, lowering X-ray obscuration (Lyu et al., 16 Jan 2025). This unifies optical "changing-state" and X-ray "changing-obscuration" phenomena as different facets of the same disk instabilities and wind physics.

4. Host Galaxy, Black Hole, and Population Properties

4.1. Host Morphology and Stellar Populations

• CLAGN hosts preferentially occupy "green valley" galaxies, lying $1$–$3\sigma$ below the star formation main sequence, indicative of moderate ongoing/quenching star formation (Dodd et al., 2020, Jin et al., 2021). SFR–$M_*$, and H$\alpha$ EW–H$\delta_A$ analyses show the majority are not in full post-starburst phase, in contrast to TDE hosts.
• Morphologically, CLAGN hosts have elevated Sérsic indices ($n\gtrsim4$) and high bulge-to-total ratios ($B/T\gtrsim0.5$), among the highest $\sim$8% of AGN hosts (Dodd et al., 2020). Both pseudo-bulge and classical bulge morphologies are represented, with MaNGA showing $80\%\pm16\%$ pseudo-bulges.
• Host asymmetry $A\lesssim0.05$ is low, which argues against recent major mergers. Minor mergers, secular bar-driven inflows, or gas dynamical instabilities are favored as fueling channels.

4.2. Stellar Populations and Star Formation Histories

Spectral synthesis on CLAGN "turn-off" states reveals: stellar light fractions with median $x_Y\sim8\%$ (young, $<55$ Myr), $x_I\sim62\%$ (intermediate, $100$ Myr–$1.3$ Gyr), and $x_O\sim30\%$ (old, $>1.5$ Gyr). Intermediate ages are enhanced relative to normal AGN. A correlation is observed: recent (last $<0.5$ Gyr) starburst activity correlates with higher current $\lambda_{\rm Edd}$ (Jin et al., 2021).

4.3. Black Hole Masses and Scaling Relations

• CLAGN SMBH masses cover $M_{\rm BH} \sim 10^{6.5}$–$10^{8.7}\,M_\odot$ (Dong et al., 14 Aug 2024, Dodd et al., 2020, Jin et al., 2021).
• All lie on the standard $M_{\rm BH}$–$\sigma_*$ relation, $$\log(M_{\rm BH}/M_\odot) = 8.49 + 4.38 \log(\sigma_e/200\,{\rm km\,s}^{-1}),$$ arguing for normal coevolution with the host bulge (Dodd et al., 2020, Yu et al., 2020).
• Eddington ratios are systematically lower than Type 1 AGN; typical CLAGN operate at $\lambda_{\rm Edd}\sim0.001$–$0.13$, with "turn-on" events at the higher end and "off" states approaching $\sim 10^{-3}$ (Dong et al., 14 Aug 2024, Jin et al., 2021).

4.4. Kinematics and Orientation

CLAGNs are predominantly found in face-on host disks ($b/a>0.7$), which can explain the visibility of BLR and may play a role in modulating the observed incidence of state changes (Yu et al., 2020). Up to 20% show stellar–gas counter-rotation, a marginal excess over control AGN ($\sim$2%), possibly reflecting minor merger or external gas accretion triggers.

5. Multiwavelength Variability, Timescales, and Line Variations

5.1. Optical, MIR, and X-ray Light Curves

• The observed structure function (SF) amplitude for CLAGNs, $A\sim0.2$ mag, and the shallow slope $\gamma\sim0.2$, are distinct from both regular Type 1 AGN ($\gamma\sim0.3$–$0.4$) and Type 2 AGN ($A\sim0.06$ mag), placing CLAGNs in an intermediate-variability mode (Wang et al., 13 Nov 2025).
• The optical color–magnitude relation, $\Delta(g - r) = k\,\Delta g$, for CLAGNs is $k\sim0.61\pm0.21$, "bluer when brighter", but with both "on" and "off" states clustering together, not splitting into Type 1/2 loci (Wang et al., 13 Nov 2025, Guo et al., 2023).
• Mid-IR (WISE) and optical continuum light curves are tightly cross-correlated, with dust-reverberation lags

$$\tau = (48 \pm 5\,\text{days}) \left(\frac{L_{\rm bol}}{10^{44}\,\text{erg\,s}^{-1}}\right)^{0.52\pm0.04}$$

measured in CLAGNs occupying $L_{\rm bol}\sim10^{44}$ erg s$^{-1}$, $\tau\sim30$–70 days (Bing et al., 2022).

5.2. Broad-Line Variability Sequence

In multi-epoch spectra, the emergence/disappearance of BELs proceeds in a reproducible pattern: H$\beta$ appears or vanishes before H$\alpha$, matching expectations from BLR radius stratification (H$\beta$-emitting gas at smaller radii than H$\alpha$) (Dong et al., 14 Aug 2024). The transition timescale, as constrained by two-epoch spectra, ranges from $\sim$244 to 5762 days rest-frame, with the bulk of events occurring in $<5$ yr (Guo et al., 2023).

5.3. Radio Properties and Jet Activity

Population radio monitoring (ASKAP VAST, VLASS, VLA) reveals that:

• CLAGNs have higher radio-detection rates (VAST $\sim$16%, VLASS $\sim$13%) and higher radio variability, but lower fraction of radio-loudness ($f_{\rm RL}\sim0.5$) compared to control AGN ($f_{\rm RL}\sim0.9$) (Birmingham et al., 2 Jul 2025).
• Only a minority of CLAGNs develop young radio jets, and major compactsymmetric-jet birth is not the rule. Individual events such as 1ES 1927+654 show nascent jet formation following a CL event, but most population-level radio variability can be attributed to scintillation and short-lived, low-power jets.

6. Theoretical and Unified Picture

The accumulated population and multiwavelength variability data robustly support the view that CLAGNs represent bona fide, accretion-rate and disk-state transitions near the critical $\lambda_{\rm Edd}\sim0.01$ boundary, where the BLR disappears/appears due to changes in the inner disk structure and/or the disk wind. This transition is analogous to state switching observed in X-ray binaries (hard/soft states, disk truncation) (Liu et al., 2022, Titarchuk et al., 14 Nov 2024, Bing et al., 2022).

Theoretical mechanisms invoked include:

• Accretion-rate fluctuations (thermal/ionization/MRI-driven), propagating at thermal timescales.
• Disk tearing and warping from Lense-Thirring precession around rapidly rotating SMBHs, generating precession, shocks, and rapid inflow/outflow events (Kaaz et al., 12 Nov 2025).
• Radiation or magnetically-driven disk winds, which clear the line of sight and modulate both obscuration and BLR survival (Lyu et al., 16 Jan 2025).
• Occasional dust obscuration or blowout, particularly for very rapid transitions with quasi-grey dimming (Zeltyn et al., 2022).

7. Implications and Future Directions

CLAGNs provide a unique probe of time-dependent accretion physics, the disk–corona–BLR interplay, and the coupling between SMBH growth and host galaxy evolution. Key implications and future avenues include:

• AGN unification models must incorporate both orientation and accretion-state axes; the changing-look phenomenon is the most direct indicator of state transitions in AGN (Bing et al., 2022, Lusso et al., 24 Feb 2025).
• Demographic and population studies will benefit from systematic, cadence-optimized spectroscopic and photometric monitoring, as being deployed in SDSS-V, LSST, and SKA (Kaaz et al., 12 Nov 2025, Amrutha et al., 30 Oct 2024).
• Empirical selection criteria (e.g., host $n\gtrsim4$, $B/T\gtrsim0.5$, green-valley colors) can isolate high-yield, pre-targeted samples for spectroscopic follow-up in current/future synoptic surveys (Dodd et al., 2020).
• Theoretical frameworks must integrate multi-scale physics: thermal/magnetic disk instabilities, Lense-Thirring effects, BLR wind structure, and host inflow modulations.
• Connections to jet launching: While some CLAGNs trigger compact, transient jets during accretion state transitions, this is not universal; disentangling the jet-disk connection requires contemporaneous optical and radio monitoring (Birmingham et al., 2 Jul 2025).

Open questions remain regarding the prevalence of disk tearing physics, the physical nature and duty cycles of accretion transitions, the exact structural changes in the inner disk/BLR/torus, and the stochastic versus deterministic processes driving CLAGN events. Ongoing high-cadence, multiwavelength monitoring and spatially resolved host studies are essential to settling these issues and leveraging CLAGNs as benchmarks for SMBH accretion theory.