AT2019qiz: TDE–QPE Bridge

• AT2019qiz is a nearby nuclear transient identified as an optical tidal disruption event (TDE) that later exhibited repeating soft X-ray quasi-periodic eruptions (QPEs).
• Its early evolution featured a t² luminosity rise, rapid photospheric expansion, and strong signatures of a dense reprocessing outflow observed in optical, UV, and infrared wavelengths.
• Subsequent observations revealed delayed coronal lines, IR dust echoes, and accretion disk evolution, linking the event to remnants of a recently faded active galactic nucleus.

AT2019qiz is a nearby nuclear transient at $z \approx 0.015$ that was identified as a spectroscopically confirmed optical tidal disruption event (TDE) and later became the first standard optical TDE with unambiguous repeating X-ray quasi-periodic eruptions (QPEs). Across the literature it is treated as a bridge object between optical TDEs and QPEs, and as an unusually complete laboratory for studying outflow-dominated optical emission, UV broad-absorption-line phenomenology, spectropolarimetric geometry, late-time accretion-disk evolution, infrared dust echoes, delayed coronal-line reverberation, and the circumnuclear environment of a possibly recently faded active galactic nucleus (AGN) (Nicholl et al., 2020, Hung et al., 2020, Nicholl et al., 2024, Wu et al., 17 Jul 2025).

1. Discovery, host galaxy, and basic system parameters

AT2019qiz was discovered in September 2019 by ATLAS and ZTF, is coincident with the nucleus of 2MASX J04463790-1013349, and peaked in early October 2019. It was classified as a TDE from early optical spectroscopy while still rising. The source offset from the nucleus was measured as $15\pm46$ pc from imaging, with Gaia astrometry giving $12\pm32$ pc, consistent with a nuclear origin. The host is described as a face-on barred spiral or barred galaxy with a centrally concentrated stellar distribution and weak signs of AGN-like activity in late-time nebular diagnostics (Nicholl et al., 2020, Short et al., 2023, Xiong et al., 25 Mar 2025).

Published black-hole mass estimates cluster around $10^6\,M_\odot$, but the exact value depends on method. Stellar-velocity-dispersion measurements gave $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$ in one study and $72\pm1~{\rm km\,s^{-1}}$ in another, while light-curve and late-time disk modeling yielded $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$, $6.14^{+0.09}_{-0.10}$, and $6.3^{+0.3}_{-0.2}$ in different analyses. Early TDE light-curve modeling favored disruption of a star of approximately $1\,M_\odot$, with $15\pm46$0, placing the event in the regime of a shallow encounter and possibly a severe partial disruption (Nicholl et al., 2020, Kovács-Stermeczky et al., 2023, Nicholl et al., 2024).

Within the optical TDE population, AT2019qiz has generally been described as nearby, fast-evolving, and of intermediate luminosity between the bulk of optical TDEs and the faint-fast event iPTF16fnl. That combination of proximity and early discovery made it one of the best-observed optical TDEs from pre-peak rise through late-time disk and QPE phases (Nicholl et al., 2020, Nicholl et al., 2024).

2. Early optical, UV, and spectroscopic evolution

The early optical/UV evolution is one of the defining empirical results. The rise began $15\pm46$1 days before maximum light, and the bolometric luminosity evolved as $15\pm46$2. At the same time, the blackbody radius increased approximately linearly and implied a photospheric expansion speed of $15\pm46$3, more specifically $15\pm46$4. Near maximum, the temperature was initially close to $15\pm46$5 K, the peak bolometric luminosity was reported as $15\pm46$6, and the blackbody radius reached $15\pm46$7 cm. After peak, the photosphere first cooled at roughly constant radius and then contracted at roughly constant temperature, while the luminosity declined steeply (Nicholl et al., 2020).

Optical spectroscopy showed broad H and He features with early blueshifts and line-forming velocities of $15\pm46$8–$15\pm46$9. The fastest optical ejecta approached the velocity inferred from radio detections, $12\pm32$0, supporting a unified outflow picture in which the optical rise and radio emission are associated with the same mass outflow. In this interpretation, the early optical continuum was dominated by a quasi-spherical, expanding reprocessing layer rather than by direct compact-disk emission (Nicholl et al., 2020).

The ultraviolet spectroscopy added a second, more unusual layer of information. The first HST spectrum at $12\pm32$1 d showed a fast FeLoBAL system, the first reported FeLoBAL in a TDE. Broad absorption at high ionization and low ionization evolved over roughly 50 days into a more standard HiBAL-like TDE UV spectrum. The broad UV troughs implied outflow speeds of $12\pm32$2 at early times, shifting to $12\pm32$3 by $12\pm32$4 and $12\pm32$5 d. Over the same interval, broad H$12\pm32$6 narrowed from $12\pm32$7 to $12\pm32$8, reinforcing the interpretation that broad Balmer emission formed in an evolving TDE outflow (Hung et al., 2020).

Taken together, the $12\pm32$9 rise, the expanding blackbody radius, the blueshifted line profiles, the UV broad absorption lines, and the weak early X-rays support the view that AT2019qiz spent its early observable phase inside a dense reprocessing outflow. A plausible implication is that direct EUV/X-ray power from early accretion was present but strongly reprocessed into the optical/UV band (Nicholl et al., 2020, Hung et al., 2020).

3. Reprocessing geometry and spectropolarimetric constraints

Optical spectropolarimetry provided a direct constraint on geometry. On day $10^6\,M_\odot$0 relative to peak brightness, the continuum polarization was

$10^6\,M_\odot$1

consistent with zero within systematics. By day $10^6\,M_\odot$2, it had increased to

$10^6\,M_\odot$3

Because interstellar polarization should not vary in time, this change demonstrates intrinsic TDE polarization evolution (Patra et al., 2022).

The physical interpretation advanced for these measurements is that the last-scattering surface was nearly circularly symmetric at peak and became moderately aspherical one month later. The authors argued that these data are incompatible with a naked eccentric disk lacking substantial mass outflow. Instead, they favored a nearly spherical, optically thick electron-scattering photosphere at peak, which then receded and exposed a more aspherical interior. The characteristic size scale was large: the electron-scattering photosphere was estimated to be $10^6\,M_\odot$4 au at maximum brightness, compared with a thermalization radius of order $10^6\,M_\odot$5–40 au and a tidal radius of order $10^6\,M_\odot$6 au (Patra et al., 2022).

The same study found that the H$10^6\,M_\odot$7 peak was depolarized to

$10^6\,M_\odot$8

on day $10^6\,M_\odot$9, compared with a continuum polarization near $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$0. That was interpreted as evidence that the H$\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$1-forming region lay near the electron-scattering photosphere rather than deep inside the continuum-forming region. The $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$2-$\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$3 plane also showed greater scatter at late time, which was taken as evidence for clumpiness or inhomogeneity in the receding scattering layer (Patra et al., 2022).

These spectropolarimetric results sharpen the outflow picture suggested by the photometry and spectroscopy. They imply that low early polarization did not mean the TDE was globally spherical; rather, it meant that the visible last-scattering surface was nearly spherical while more aspherical inner structures remained hidden (Patra et al., 2022).

4. Late-time accretion disk and the emergence of QPEs

The most consequential later development was the appearance of bona fide QPEs. Roughly 1500 days after discovery, Chandra, NICER, Swift, and AstroSat revealed nine X-ray eruptions from AT2019qiz. The recurrence intervals ranged from 39 to 54 hours, with a mean recurrence time of $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$4 hours and a standard deviation of 7.2 hours. Individual eruptions typically lasted 8–10 hours and showed asymmetric fast-rise, slow-decay profiles. A null test with random peak times gave such regularity in $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$5 of trials, so the peaks were interpreted as true QPEs rather than stochastic flares (Nicholl et al., 2024).

The QPEs were spectrally soft. During eruption, the peak bolometric luminosity was

$\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$6

the peak blackbody temperature was

$\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$7

and the characteristic emitting radius was of order $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$8. The late-time quiescent disk, modeled separately, had $\sigma = 69.7 \pm 2.3~{\rm km\,s^{-1}}$9 and luminosity around $72\pm1~{\rm km\,s^{-1}}$0, corresponding to $72\pm1~{\rm km\,s^{-1}}$1 of Eddington. The same multiwavelength disk fit gave $72\pm1~{\rm km\,s^{-1}}$2 and an initial disk mass of $72\pm1~{\rm km\,s^{-1}}$3. No optical QPE-like flares were detected, despite extensive monitoring (Nicholl et al., 2024).

The preferred interpretation in the discovery paper was a collision between an orbiting body and the late-time TDE disk. For the measured black-hole mass, a 48.4 hour Keplerian period corresponds to an orbital radius of about $72\pm1~{\rm km\,s^{-1}}$4, and the modeled late-time disk extended to radii large enough to intersect such an orbit. On that basis, an orbiting body colliding with the disk was argued to explain the recurrence timescale, asymmetric flares, and compact expanding X-ray photosphere more naturally than radiation-pressure or magnetic-pressure disk-instability models, which were found to predict incompatible timescales (Nicholl et al., 2024).

Subsequent theory papers complicated that picture. One argued that magnetized TDE disks should spread rapidly enough that QPEs ought to have started earlier than first detected in AT2019qiz, suggesting either missed earlier eruptions or additional triggering conditions (Alush et al., 5 Mar 2025). Another argued that once the finite mass and angular momentum budget of a TDE disk is enforced, simple steady-state AGN-disk collision models fail badly for AT2019qiz: collisions with an orbiting black hole were ruled out, collisions with the surface of a stellar EMRI were also ruled out, and only a model involving a star puffed up to its Hills sphere with a trailing debris stream remained viable (Mummery, 30 Apr 2025). AT2019qiz is therefore not only the clearest TDE–QPE bridge object, but also the most constraining test case for TDE-disk collision models (Nicholl et al., 2024, Mummery, 30 Apr 2025).

5. Infrared dust echoes, torus geometry, and the hidden-energy problem

AT2019qiz is one of the rare optical TDEs with a strong, long-lived mid-infrared echo. Archival WISE and NEOWISE monitoring provided a long baseline from 2010 to 2024. After quality cuts and half-year binning, the source was quiescent before the TDE and then brightened suddenly in the IR on 2020 February 7, about 120 days after the optical peak. Nine IR echo epochs were identified. The W1 and W2 light curves showed an early bump, then a steady continued rise through the penultimate epoch, with only the final epoch suggesting a plateau. The dust temperature declined until the fourth IR epoch and then remained approximately constant over the last five epochs (Wu et al., 17 Jul 2025).

A late-time reverberation fit using a thin inclined torus or “convex dust ring” model yielded a robust lower limit of about $72\pm1~{\rm km\,s^{-1}}$5 pc for the inner dust radius, with $72\pm1~{\rm km\,s^{-1}}$6–$72\pm1~{\rm km\,s^{-1}}$7 pc nearly independent of grain type and size. For $72\pm1~{\rm km\,s^{-1}}$8, the mean temperatures over the final five epochs were about 679 K for silicate, 837 K for SiC, and 557 K for graphite. The inferred covering factor,

$72\pm1~{\rm km\,s^{-1}}$9

is torus-like and much larger than the $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$0 covering factors often inferred for ordinary optical TDE hosts. Because a normal AGN sublimation radius for a $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$1 black hole at around Eddington luminosity is only $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$2 pc, this large radius was interpreted as a torus remnant left behind after the inner torus disappeared in a recently faded AGN (Wu et al., 17 Jul 2025).

The same IR echo was used as a bolometer for the hidden TDE luminosity. Assuming optically thin IR emission and rapid thermal equilibrium, the minimum peak bolometric luminosity required to heat the dust at $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$3 pc was estimated as $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$4, $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$5, and $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$6 for silicate, silicon carbide, and graphite grains, respectively. All exceed the observed optical/UV blackbody peak luminosity, quoted there as $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$7, by factors of roughly $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$8 to $\log_{10}(M_\bullet/M_\odot)=5.89^{+0.05}_{-0.06}$9. This was presented as further evidence that much of the missing TDE energy was emitted in the unobservable EUV (Wu et al., 17 Jul 2025).

A separate 2025 study proposed a different interpretation. It emphasized that the IR echo rose continuously and approximately linearly for roughly 1000 days, reached dust temperatures of $6.14^{+0.09}_{-0.10}$0–$6.14^{+0.09}_{-0.10}$1 K, and then began to plateau. In that model, the original $6.14^{+0.09}_{-0.10}$2-day TDE flare and a $6.14^{+0.09}_{-0.10}$3 remnant disk could not explain both the long rise and the high dust temperature at $6.14^{+0.09}_{-0.10}$4 pc. Instead, a dust shell reprocessing “bright” QPE emission with peak bolometric luminosities $6.14^{+0.09}_{-0.10}$5 and recurrence-time-averaged luminosity $6.14^{+0.09}_{-0.10}$6 was argued to reproduce the observations, provided the dust covering fraction is high, $6.14^{+0.09}_{-0.10}$7. That model predicts IR QPE variability at the $6.14^{+0.09}_{-0.10}$8 level and a roughly linear IR decline over $6.14^{+0.09}_{-0.10}$9 days if the QPEs shut off (Pasham et al., 17 Feb 2025).

These two IR interpretations are physically different: one treats the late-time IR echo as a TDE-powered reverberation signal from a torus remnant at $6.3^{+0.3}_{-0.2}$0 pc, while the other treats it as a QPE-powered shell at $6.3^{+0.3}_{-0.2}$1 pc. The present observational limitation is that the IR decline phase has not yet been well sampled, so the dust geometry remains underconstrained (Wu et al., 17 Jul 2025, Pasham et al., 17 Feb 2025).

6. Delayed coronal lines, extended emission-line gas, and the recently faded AGN hypothesis

AT2019qiz later developed a delayed coronal-line spectrum that directly linked an optical TDE to the extreme coronal line emitter (ECLE) phenomenon. Optical spectra at 428, 481, and 828 rest-frame days after optical peak showed strong [Fe VII], [Fe X], [Fe XI], and [Fe XIV], together with increasing narrow H$6.3^{+0.3}_{-0.2}$2 and H$6.3^{+0.3}_{-0.2}$3, a decline in broad H$6.3^{+0.3}_{-0.2}$4, and a luminous IR echo. The coronal lines had $6.3^{+0.3}_{-0.2}$5–300 km s$6.3^{+0.3}_{-0.2}$6, intermediate between broad and narrow line widths, and were therefore interpreted as arising from a region between the broad-line and narrow-line emitting gas (Short et al., 2023).

Between the early optical flare and the late-time Swift observation, the X-ray spectrum softened dramatically. The 0.3–1 keV X-ray flux increased by a factor of $6.3^{+0.3}_{-0.2}$7, while the hard X-ray flux decreased by a factor of $6.3^{+0.3}_{-0.2}$8. At the same time, late UVOT UVW2 photometry was consistent with the host only, whereas the IR fluxes continued to rise. The delayed coronal lines, the softened X-ray spectrum, and the IR echo were therefore interpreted as reverberation of pre-existing circumnuclear gas and dust by a hard EUV/soft-X-ray continuum associated with the TDE and its later accretion state (Short et al., 2023).

Integral-field spectroscopy of the host added a larger-scale environmental context. VLT/MUSE data revealed a bi-conical extended emission-line region (EELR) approximately $6.3^{+0.3}_{-0.2}$9 in scale, with a Seyfert-like central region of $1\,M_\odot$0 and LINER-like diffuse outer gas. The ionizing luminosity required for all AGN-classified spaxels was estimated as $1\,M_\odot$1, while the central Seyfert zone alone required $1\,M_\odot$2. The current nucleus, with $1\,M_\odot$3, is close to the latter but insufficient for the full EELR, implying either a recently faded AGN or a delayed response to historical nuclear activity. The host also showed a post-starburst stellar population, with an old component around 10 Gyr and a prominent star-formation episode around 1 Gyr ago, plus disturbed gas kinematics relative to the stars (Xiong et al., 25 Mar 2025).

In combination with the parsec-scale IR echo and the late-time coronal lines, the kpc-scale EELR has made AT2019qiz a central example in the “recently faded AGN” or “embers of AGN” picture. In that framework, TDEs and later QPEs preferentially occur in nuclei that retain fossil circumnuclear structure from a previous AGN phase, including dust, ionized gas, and possibly orbital architectures favorable to EMRI-like bodies (Wu et al., 17 Jul 2025, Xiong et al., 25 Mar 2025).

7. Model dependence, open questions, and broader significance

AT2019qiz has also become a standard cautionary example for parameter inference. A comparison between the semi-analytic TiDE code and MOSFiT showed that similar optical light curves can be reproduced with very different physical parameters. For AT2019qiz, the preferred visual TiDE solution used $1\,M_\odot$4, $1\,M_\odot$5, $1\,M_\odot$6, and $1\,M_\odot$7, whereas a TiDE solution forced toward the MOSFiT black-hole mass used $1\,M_\odot$8, $1\,M_\odot$9, $15\pm46$00, and $15\pm46$01. The MOSFiT values quoted there were $15\pm46$02, $15\pm46$03, $15\pm46$04, and $15\pm46$05 d. The broader conclusion was that light-curve fits alone do not robustly determine TDE parameters unless the fallback and accretion prescriptions are themselves well constrained (Kovács-Stermeczky et al., 2023).

The object’s broader legacy lies in how often it now serves as a reference system. Later work on AT2019vcb cited AT2019qiz as the benchmark case making the TDE–QPE connection “almost direct” (Bykov et al., 2024). The discovery of QPEs in AT2022upj made AT2019qiz one of three optically selected TDEs with QPEs and one of two such systems also showing coronal-line emission, which led to a Bayesian estimate that the fraction of optical TDEs producing QPEs within five years may be $15\pm46$06 (Chakraborty et al., 24 Mar 2025).

Several major questions remain open. The true origin of the late IR luminosity is unresolved because the observed light curve has not yet cleanly entered decline; the physical engine of the QPEs remains contested between orbit-crossing and disk-based scenarios; the role of a pre-existing or recently faded AGN in setting up the TDE and QPE conditions is suggestive but not fully established; and the hidden EUV/bolometric output of the original TDE is still limited by uncertainties in grain composition, geometry, and radiative transfer. The literature accordingly emphasizes continued high-cadence X-ray monitoring, JWST mid-IR spectroscopy, and future IR monitoring by facilities such as NEO Surveyor and Roman as the most direct routes to resolving the remaining degeneracies (Wu et al., 17 Jul 2025, Pasham et al., 17 Feb 2025, Nicholl et al., 2024).

AT2019qiz therefore occupies a distinctive place in TDE studies. It began as a nearby, fast optical TDE with early evidence for an outflow-powered reprocessing layer; evolved into a source with delayed coronal lines, a strong and long-lived IR echo, and a host environment suggestive of faded AGN activity; and then entered a late phase of repeated soft X-ray eruptions that anchored the empirical TDE–QPE connection. Few SMBH transients have constrained so many separate components of the tidal-disruption problem within a single system (Nicholl et al., 2020, Nicholl et al., 2024, Wu et al., 17 Jul 2025).