---
title: 'NGC 3822: Changing-Look AGN'
url: https://www.emergentmind.com/topics/ngc-3822
type: topic
---

# NGC 3822: Changing-Look AGN

NGC 3822 is a nearby active galactic nucleus at redshift $z = 0.019$ whose nucleus exhibits changing-look (CL) behavior on months-to-years timescales. A 17-year multiwavelength study spanning 2008 to 2025 combines X-ray, ultraviolet, and optical observations and shows that the source alternates between narrow-line Seyfert states and a Type 1 state in which broad Balmer emission becomes prominent. In the adopted framework, the black hole mass is $M_{\rm BH} = 2.70\times10^{7}\ M_\odot$, and the bolometric luminosity varies between $(1.32-17)\times10^{43}\ {\rm erg\ s^{-1}}$ [2510.05599].

## 1. Source identity and historical classification

The nucleus of NGC 3822 is classified as Seyfert, and its optical type is not stationary. The historical record summarized for the 1993–2025 interval shows repeated transitions in the visibility of broad Balmer emission, which is the defining phenomenology of its changing-look designation [2510.05599].

| Epoch | Optical type | Spectroscopic state |
|---|---|---|
| 1993 | Seyfert 2 | no broad lines |
| 1994–1995 | Seyfert 1.9 | broad H$\alpha$ weakly present |
| 2018 | Seyfert 1.9 | narrow Balmer only |
| 2022 | Seyfert 1 | broad H$\alpha$ and H$\beta$ present |
| 2024–2025 | Seyfert 1.9/2 | broad lines absent; only narrow components |

This sequence establishes that NGC 3822 is not merely variable in continuum flux but undergoes spectroscopic state changes in the broad-line region (BLR) tracer lines themselves. A plausible implication is that the object belongs to the low-redshift CL AGN population in which BLR visibility is intermittent rather than permanently obscured.

## 2. Observational campaign and reduction strategy

The monitoring campaign uses Swift/XRT and UVOT from 2008 to 2024, with XRT observations grouped into XRT1–XRT7 and the dense 2022 interval subdivided into XRT5a–XRT5d. Typical XRT exposures are $\lesssim 5$–$12\ {\rm ks}$, and the low signal-to-noise ratio motivates Cash statistics with $\geq 5$ counts per bin. XMM-Newton provides a single EPIC-pn and Optical Monitor epoch on 2010-06-01 (ObsID 0655380101) with an exposure of approximately $14\ {\rm ks}$; the OM observation uses UVM2 at $2310\ \AA$. NuSTAR contributes two broadband epochs, 2016-01-12 (ObsID 60061332002), simultaneous with Swift, and 2022-06-04 (ObsID 90801611002), quasi-simultaneous with Swift [2510.05599].

Optical spectroscopy is drawn from VLT/X-Shooter and the Himalayan Chandra Telescope (HCT). X-Shooter UVB/VIS observations were obtained on 2018-06-11, 2022-06-17, and 2022-07-11, with total wavelength coverage of $3000$–$25000\ \AA$, UVB resolving power $R\approx 3200$, and VIS resolving power $R\approx 5000$. HCT/HFOSC observations were obtained on 2024-03-06, 2024-04-01, 2024-05-02, 2024-06-02, 2025-01-01, and 2025-03-03 using Gr7 ($3800$–$6840\ \AA$, $R\approx 1330$) and Gr8 ($5800$–$8350\ \AA$, $R\approx 2190$).

The UVOT photometric bands are $V$ ($5468\ \AA$), $B$ ($4392\ \AA$), $U$ ($3465\ \AA$), UVW1 ($2600\ \AA$), UVM2 ($2246\ \AA$), and UVW2 ($1928\ \AA$). Galactic extinction corrections were applied with $E(B-V)=0.0479$, $R_V=3.1$, and $A_\lambda(V,B,U,UVW1,UVM2,UVW2)=0.15, 0.19, 0.24, 0.34, 0.45, 0.39$ mag. XMM/EPIC reduction used SAS v18 with pile-up mitigated through annular extraction and spectra binned to $\geq 25$ counts per bin. NuSTAR reduction used NuSTARDAS v2.1.2 with spectra grouped to $\geq 10$ counts per bin. For optical spectroscopy, the host stellar population was modeled and subtracted using pPXF with the MILES library, emission-line profiles were fit using pyQSOFIT, and instrumental resolution corrections were applied.

## 3. Multiwavelength variability

NGC 3822 shows strong long-term variability from X-rays through the optical/UV, with the variability amplitude increasing toward shorter wavelengths. In the $0.3$–$10\ {\rm keV}$ band, the source entered a high state in 2013 with $F_{0.3-10} = (23.71\pm3.03)\times10^{-12}$ in the table’s cgs units. A second high state occurred in 2022, with an average flux of approximately $12\times10^{-12}$, a minimum of $5.70\times10^{-12}$ on 2022-06-26, and a maximum of $16.37\times10^{-12}$ on 2022-11-13. Lower states were recorded in 2010 at $(6.75\pm0.18)\times10^{-12}$, in 2015 at $(8.44\pm3.13)\times10^{-12}$, and during the decline through 2023–2024, when the X-ray flux dropped to approximately $1\times10^{-12}$ and below [2510.05599].

The UV and optical bands track the same long-term behavior but with reduced amplitude. UVW1 peaked in 2013 at approximately $4.86\times10^{-15}$, corresponding to an approximately threefold increase over 2016. In 2022 the source underwent a nuclear outburst across the X-ray to UV/optical bands, followed by a monotonic UV decay through 2023–2024. The decay law is consistent with $F_{\rm UV} \propto t^{-0.41\pm0.03} \approx t^{-5/12}$.

The fractional variability amplitudes quantify the wavelength dependence. Over 2008–2024, $F_{\rm var}$ is approximately $66.8\pm8.7\%$ for XRT, $40.1\pm5.5\%$ for UVW2, $39.8\pm5.1\%$ for UVM2, $33.2\pm4.3\%$ for UVW1, $28.9\pm4.4\%$ for $U$, $5.75\pm1.18\%$ for $B$, and $3.33\pm1.20\%$ for $V$. During 2022 alone, the corresponding values are approximately $20.2\pm5.2\%$, $22.4\pm4.5\%$, $19.4\pm3.7\%$, $15.0\pm2.3\%$, $8.67\pm1.76\%$, $4.89\pm1.11\%$, and $3.33\pm1.20\%$. The wavelength dependence is modeled as $F_{\rm var} \propto \lambda^{-c}$ with $c \approx 2.51$ for 2008–2024 and $c \approx 1.91$ for 2022, indicating stronger variability at shorter wavelengths.

Explicit hardness-ratio trends were not reported. Instead, spectral hardness is inferred from the photon index $\Gamma$, which is harder at low states with $\Gamma\sim1.4$–$1.5$ and softer during the outburst with $\Gamma\sim1.8$–$2.1$. This suggests a continuum-state evolution coupled to luminosity changes.

## 4. X-ray continuum, absorption, and coronal properties

The baseline $0.3$–$10\ {\rm keV}$ spectral model is $\text{Constant} \times \text{TBabs} \times \text{Pcfabs} \times \text{Powerlaw}$. For broadband $0.3$–$60\ {\rm keV}$ epochs, the analysis adds Cutoffpl or CompTT. No significant Fe K line between $6$ and $7\ {\rm keV}$ is detected, no high-energy reflection hump above $10\ {\rm keV}$ is detected, and the reflection component is judged absent or insignificant during the broadband epochs. The high-energy cutoff $E_{\rm cut}$ is not constrained and is fixed at $300\ {\rm keV}$; CompTT fits assume spherical geometry and a seed photon temperature $kT_s = 100\ {\rm eV}$ [2510.05599].

Epoch-by-epoch fits show that intrinsic absorption is absent in 2008, 2010, 2013, 2015, late 2022, 2023, and 2024, but present in 2016 and several 2022 epochs. In 2016 (XRT4+NU1), the absorber has $N_{\rm H}=(1.20^{+1.01}_{-0.71})\times10^{22}\ {\rm cm^{-2}}$ and covering fraction $C_f=0.62^{+0.09}_{-0.12}$ with $\Gamma=1.80\pm0.11$. In 2022, absorption is found in XRT5a with $N_{\rm H}=(1.92^{+1.03}_{-0.68})\times10^{22}$ and $C_f=0.67^{+0.13}_{-0.23}$, in XRT5b+NU2 with $N_{\rm H}=(1.44^{+0.66}_{-0.53})\times10^{22}$ and $C_f=0.73^{+0.09}_{-0.15}$, and in XRT5c with $N_{\rm H}=(1.75^{+0.93}_{-0.70})\times10^{22}$ and $C_f=0.77^{+0.11}_{-0.25}$. By XRT5d in late 2022 the absorption is absent again. Across the full campaign, the inferred intrinsic absorber is therefore a partially covering component with $N_{\rm H}\approx(1.2$–$2.0)\times10^{22}\ {\rm cm^{-2}}$ and $C_f\approx0.5$–$0.8$.

The $2$–$10\ {\rm keV}$ luminosity spans $\log L_X \approx 42.0$–$43.1$. Representative values are $\log L_X=43.10\pm0.09$ in 2013, $\log L_X=42.73\pm0.03$ in XRT5b+NU2 during 2022, and $\log L_X=42.02\pm0.11$ in 2023. Unabsorbed fluxes illustrate the same range: on 2013-07-26, $F_{2-10} = 15.56\pm3.00$, $F_{0.3-2} = 8.16\pm0.99$, and $F_{0.3-10} = 23.71\pm3.03$; on 2022-11-13, $F_{2-10} = 10.81\pm2.21$ and $F_{0.3-10} = 16.37\pm2.24$; on 2023-06-12, $F_{2-10} = 0.81\pm0.39$ and $F_{0.3-10} = 1.11\pm0.40$, in units of $10^{-12}$ cgs.

The two NuSTAR-supported epochs also constrain coronal parameters. In 2016, CompTT gives $kT_e=50^{+21}_{-19}\ {\rm keV}$ and $\tau=0.97^{+0.57}_{-0.34}$, while in 2022 it gives $kT_e=88^{+51}_{-32}\ {\rm keV}$ and $\tau=0.77^{+0.46}_{-0.31}$. These values correspond to an optically thin-to-marginal corona in 2016 and a hotter, optically thinner corona in 2022. Because the line-of-sight column remains at the $\sim10^{22}\ {\rm cm^{-2}}$ level and reflection dominance is absent, Compton-thick obscuration is ruled out at these epochs. The favored physical picture is that clumpy clouds move into and out of the line of sight on BLR or torus scales.

## 5. Optical spectroscopy and broad-line state changes

The optical CL diagnosis depends on careful removal of host-galaxy starlight. The stellar continuum was modeled with pPXF using the MILES library, and the AGN spectrum was obtained after subtraction of stellar absorption features. Emission-line profiles were then measured with pyQSOFIT, a step that is especially important for robust H$\beta$ characterization [2510.05599].

The 2018 X-Shooter spectrum shows only narrow components, with H$\beta$ narrow $\mathrm{FWHM}=444\pm107\ {\rm km\ s^{-1}}$, [O III] $\lambda5007$ $\mathrm{FWHM}=444\pm107\ {\rm km\ s^{-1}}$, and H$\alpha$ narrow $\mathrm{FWHM}=367\pm84\ {\rm km\ s^{-1}}$; no broad Balmer component is detected, yielding a Type 1.9 classification. In contrast, the 2022-06-17 X-Shooter spectrum shows broad H$\beta$ with $\mathrm{FWHM}=5311\pm1222\ {\rm km\ s^{-1}}$ and broad H$\alpha$ with $\mathrm{FWHM}=3454\pm795\ {\rm km\ s^{-1}}$, while the narrow lines remain nearly unchanged, with [O III] narrow at approximately $435\pm102\ {\rm km\ s^{-1}}$. The 2022-07-11 spectrum continues to show broad Balmer emission, with broad H$\beta$ at $5110\pm1176\ {\rm km\ s^{-1}}$ and broad H$\alpha$ at $4578\pm1054\ {\rm km\ s^{-1}}$, so the source is Type 1 during the outburst.

The HCT spectra from 2024 to 2025 show that the broad components disappear again. Over these epochs, narrow H$\beta$ has $\mathrm{FWHM}=543$–$893\ {\rm km\ s^{-1}}$ and narrow H$\alpha$ has $\mathrm{FWHM}=543$–$684\ {\rm km\ s^{-1}}$, with no broad-line detection; the source is therefore back in a Type 1.9/2 state. A weak blue-wing hump is seen in [O III] $\lambda5007$ in the X-Shooter data, with a fitted Gaussian $\mathrm{FWHM}\sim391\pm90$ in 2018, $480\pm110$ in 2022-06, and $461\pm106$ in 2022-07, which is suggestive of mild outflow. The feature is not detected in the HCT spectra because of resolution and signal-to-noise limitations.

Optical continuum luminosities, line fluxes, and equivalent widths are not tabulated. The optical analysis is therefore centered on line-profile widths and the presence or absence of broad components rather than on full reverberation-style or single-epoch virial diagnostics.

## 6. Accretion-state interpretation and SED diagnostics

The bolometric correction is computed from the Eddington-ratio-dependent relation $\log k_{\rm bol} = C(\log \lambda_{\rm Edd})^2 + B(\log \lambda_{\rm Edd}) + A$, with $C = 0.054\pm0.034$, $B = 0.309\pm0.095$, and $A = 1.538\pm0.063$, and with $L_{\rm bol} = k_{\rm bol}L_{2-10\ {\rm keV}}$. The Eddington ratio is $\lambda_{\rm Edd} = L_{\rm bol}/L_{\rm Edd}$. For internal consistency the study adopts $L_{\rm Edd} = 1.95\times10^{46}\ {\rm erg\ s^{-1}}$, although it also notes the canonical form $L_{\rm Edd}=1.26\times10^{38}(M_{\rm BH}/M_\odot)\ {\rm erg\ s^{-1}}$, which would give approximately $3.4\times10^{45}\ {\rm erg\ s^{-1}}$ for $M_{\rm BH}=2.7\times10^{7}\ M_\odot$ [2510.05599].

Within this framework, the broad-line states track accretion power. Broad emission lines are present in 2022 at $\lambda_{\rm Edd}\approx 3.8\times10^{-3}$, corresponding to $\log \lambda_{\rm Edd}\approx -2.42$ to $-2.52$ across the 2022 epochs. Broad lines are absent in 2018 and 2024–2025 at $\lambda_{\rm Edd}\approx (0.6$–$0.9)\times10^{-3}$, exemplified by XRT6 with $\log \lambda_{\rm Edd}=-3.17\pm0.11$ and XRT7 with $\log \lambda_{\rm Edd}=-3.06\pm0.07$. The resulting interpretation is that the optical CL transitions are governed primarily by changes in accretion rate rather than by persistent obscuration. The X-ray absorbers are real but transient, whereas the broad-line visibility follows the higher- versus lower-$\lambda_{\rm Edd}$ state.

The UV decline after the 2022 flare is consistent with a tidal disruption event-like disk-cooling fallback law, $F_{\rm UV}\propto t^{-5/12}$, following the observed slope of $-0.41\pm0.03$. Broad Balmer features during the outburst resemble TDE-H class behavior noted in AGN hosts, and a plausible implication is that a TDE temporarily boosted accretion and re-illuminated the BLR. At the same time, NGC 3822 is already an active AGN with recurrent CL transitions, so the flare is interpreted within an accreting-Seyfert context rather than as a standalone dormant-galaxy disruption event.

The study also places the source in a disk-corona coupling context. A weak positive $\Gamma$–$\lambda_{\rm Edd}$ correlation is reported in the low-luminosity regime, with $\Gamma\approx1.4$–$1.9$ while $\lambda_{\rm Edd}\sim10^{-3}$. The optical-to-X-ray slope is defined as $\alpha_{\rm ox} = -0.384\log(F_{2\,{\rm keV}}/F_{2500\ \AA})$, using UVW1 as the nearest band to $2500\ \AA$. There is a strong positive correlation between $F_{2\,{\rm keV}}$ and $F_{2500\ \AA}$ with $\mathrm{PCC}\approx0.98$, while $\alpha_{\rm ox}$ anti-correlates with UV luminosity with $\mathrm{PCC}\approx-0.73$ and $p\approx0.04$, and also strongly anti-correlates with $\lambda_{\rm Edd}$. This is consistent with reduced coronal efficiency at higher accretion rates.

Standard BLR radius-luminosity scaling, $R_{\rm BLR}\approx A(L_{5100}/10^{44}\ {\rm erg\ s^{-1}})^{\beta}$ light-days with commonly used $A\approx30$–$40$ and $\beta\approx0.5$, is noted but not directly applied because $\lambda L_\lambda(5100\ \AA)$ is not provided. Dust sublimation is summarized through the scaling $R_{\rm sub}\propto L^{1/2}$. The threshold-like BLR behavior at low $\lambda_{\rm Edd}$ is therefore discussed qualitatively as being consistent with BLR suppression in low-luminosity CL AGN samples.

## 7. Caveats and broader significance

Several technical limitations condition the interpretation. Many Swift/XRT spectra are low signal-to-noise and require Cash statistics with minimal binning, so weak spectral features, including Fe K structure, should be treated cautiously. The high-energy cutoff is not constrained and is fixed at $300\ {\rm keV}$ for pragmatic fitting. Optical continuum luminosities, line fluxes, and equivalent widths are not reported, the black hole mass is adopted from external work rather than derived in situ, and no single-epoch virial mass estimate is produced because the required continuum luminosity calibration is absent [2510.05599].

The reported absence of hard X-ray reflection may also reflect data quality and sensitivity rather than a strict physical null result, and deeper hard X-ray coverage would improve those constraints. Likewise, the adopted value $L_{\rm Edd} = 1.95\times10^{46}\ {\rm erg\ s^{-1}}$ differs from the canonical numerical expectation for the stated black hole mass, although the published $\lambda_{\rm Edd}$ values are internally consistent with the adopted normalization.

Within those limits, NGC 3822 is a well-defined CL AGN in which broad Balmer emission appears during an accretion-enhanced 2022 outburst and disappears again by 2024–2025. Its X-ray spectra reveal transient, partially covering intrinsic absorption episodes in 2016 and mid-2022, but these do not account for the optical state changes. The source therefore occupies the class of CL AGN where BLR visibility tracks accretion power, while cloud-driven X-ray obscuration acts as a secondary and transient phenomenon.

Source: https://www.emergentmind.com/topics/ngc-3822