Bowen Fluorescence Flares

Updated 6 July 2026

Bowen Fluorescence Flares (BFFs) are luminous nuclear flares from SMBHs characterized by prominent Bowen-excited emission lines (e.g., N III 4640 and He II 4686) and a blue continuum.
They are identified through time-resolved spectroscopy that reveals slow photometric evolution, narrow Balmer lines, and recurring features such as rebrightenings and infrared dust echoes.
The physical origin of BFFs remains debated, with interpretations ranging from enhanced accretion onto active SMBHs to TDE-like events in AGN, necessitating multiwavelength follow-up studies.

Searching arXiv for the cited BFF and related papers to ground the article. arxiv_search: query="Bowen Fluorescence Flares accreting supermassive black holes AT 2017bgt AT 2021loi AT2019aalc AT2021aeuk J012026" max_results=10 Bowen Fluorescence Flares (BFFs) are luminous nuclear flares from accreting supermassive black holes whose spectra show prominent Bowen-excited emission, especially the blend around $4660$ \AA\ associated with $\mathrm{N\,III}\,\lambda 4640$ and $\mathrm{He\,II}\,\lambda 4686$ , often together with a blue continuum and strong high-ionization coronal lines. In the current literature, BFFs are treated as a distinct observational class of SMBH-related flares that overlaps with, but is not reducible to, ordinary AGN variability, changing-look behavior, or standard optical tidal disruption events (TDEs). The class emerged from the AT 2017bgt, F01004-2237, and OGLE17aaj phenomenology, and subsequent work has increasingly framed BFFs as a key interface between flaring AGN and TDE-like accretion episodes in active nuclei (Trakhtenbrot et al., 2019).

1. Spectroscopic definition and Bowen-fluorescence basis

Bowen fluorescence, in the sense relevant to BFFs, is a line-coincidence pumping process driven by an intense EUV/soft-X-ray radiation field. In the AT 2017bgt framework, photons from the $\mathrm{He\,II}$ Lyman- $\alpha$ -like transition at $303.783$ \AA\ excite certain O III and N III states because of near-resonant wavelength coincidence; the resulting cascades produce broad optical and near-UV emission including $\mathrm{O\,III}\,\lambda 3133$ , $\mathrm{O\,III}\,\lambda 3341$ , $\mathrm{O\,III}\,\lambda 3429$ , $\mathrm{O\,III}\,\lambda 3444$ , and $\mathrm{N\,III}\,\lambda 4640$ 0, together with strong $\mathrm{N\,III}\,\lambda 4640$ 1 (Trakhtenbrot et al., 2019). In the later BFF literature, the key observational manifestation is the line complex around $\mathrm{N\,III}\,\lambda 4640$ 2 \AA, especially the blend of $\mathrm{N\,III}\,\lambda 4640$ 3 and $\mathrm{N\,III}\,\lambda 4640$ 4, often called the Bowen fluorescence complex (Veres, 11 Dec 2025).

The literature does not provide a single formal threshold for BFF identification. Instead, a practical spectroscopic pattern recurs: a flare-like optical/UV brightening from a galactic nucleus, a blue continuum, prominent Bowen-excited lines, and frequently strong, variable high-ionization coronal lines such as $\mathrm{N\,III}\,\lambda 4640$ 5. Many BFFs also show Balmer emission lines that are narrower than in classical optical TDEs. The papers repeatedly emphasize that BFF spectra are not explained by ordinary AGN broad-line phenomenology alone: broad $\mathrm{N\,III}\,\lambda 4640$ 6 and O III Bowen lines are unusual in standard AGN composites, and their presence implies a strong ionizing source plus dense reprocessing gas.

A central terminological point is that “BFF” is a spectroscopic and time-domain designation for SMBH-related nuclear flares, not a generic synonym for Bowen fluorescence in astrophysics. Bowen fluorescence also appears in X-ray binaries, compact binaries, and other transient environments, but the BFF literature uses the term for a specific nuclear-flare phenomenology associated with accreting SMBHs.

2. Emergence of the class and its observational boundaries

Trakhtenbrot et al. presented AT 2017bgt as a “new class of flares from accreting supermassive black holes,” arguing that the event was too persistent and too spectroscopically peculiar to be ordinary AGN variability, too unlike known TDEs, and distinct from changing-look AGN because its defining feature was not a type transition but the emergence of an unusually UV-bright state that powered strong broad Bowen-fluorescence lines (Trakhtenbrot et al., 2019). In that paper, AT 2017bgt, F01004-2237, and OGLE17aaj were grouped as the first members of the class. Later work adopted “Bowen Fluorescence Flare” as the class label and expanded the sample.

AT 2021loi was especially consequential because it occurred in a previously known unobscured broad-line AGN. Its spectra showed a double-peaked $\mathrm{N\,III}\,\lambda 4640$ 7 \AA\ feature interpreted as blended $\mathrm{N\,III}\,\lambda 4640$ 8 and $\mathrm{N\,III}\,\lambda 4640$ 9, the appearance of $\mathrm{He\,II}\,\lambda 4686$ 0, and a rebrightening about one year after the main peak. Because the pre-flare nucleus was already a Type 1 AGN, the event strengthened the interpretation that at least some BFFs are episodes of enhanced accretion onto already active SMBHs rather than flares in quiescent hosts (Makrygianni et al., 2023).

Operationally, survey work distinguishes BFFs from TDEs by a recurring pair of properties: their spectral lines are typically much narrower, and their light curves decline much more slowly. The ZTF nuclear-transient search paper treated BFFs as an explicit target class alongside TDEs and AGN flares, describing them as blue nuclear flares with H and He lines that were initially confused with TDEs but are observationally separated by narrower line widths and slower photometric evolution (Dgany et al., 2023). That same paper also showed that TDE-oriented filtering can be incomplete for BFFs, because cuts designed to suppress prior AGN activity may remove exactly the pre-existing active nuclei in which BFFs often occur.

3. Multiwavelength phenomenology

Across the current sample, BFFs are characterized less by a single luminosity or line-ratio threshold than by a recurrent bundle of photometric, spectroscopic, and host-dependent properties. The optical/UV flare is typically superposed on a nucleus that is either already active or at least plausibly weakly active. Many light curves are not smooth declines: rebrightenings, bumps, and in some cases multiple distinct major flares are common. Infrared dust echoes are frequent, consistent with dusty circumnuclear environments. Soft X-ray behavior is diverse: some BFFs show a strong soft X-ray excess or recurrent soft X-ray flares, while others have marginal or absent direct X-ray detections even though the Bowen and coronal lines imply a strong ionizing field (Veres, 11 Dec 2025).

The class also overlaps substantially with extreme coronal line emitters (ECLEs). Strong and variable coronal lines are common in BFF discussions, and several objects are explicitly classified as both BFFs and ECLEs. This overlap is important because it ties BFFs to gas-rich, dust-rich nuclear environments in which reprocessing, obscuration, and delayed infrared emission are expected.

Object	Host context	Distinctive reported properties
AT 2017bgt	already accreting SMBH / AGN	UV increase by $\mathrm{He\,II}\,\lambda 4686$ 1, broad He II, N III, O III Bowen lines, little evolution for at least 14 months
AT 2021loi	previously known unobscured AGN	$\mathrm{He\,II}\,\lambda 4686$ 2 \AA\ feature strengthens by factor of $\mathrm{He\,II}\,\lambda 4686$ 3, $\mathrm{He\,II}\,\lambda 4686$ 4, rebrightening $\mathrm{He\,II}\,\lambda 4686$ 5 days after peak
AT2019aalc	broad-line Seyfert 1 AGN	two optical flares separated by $\mathrm{He\,II}\,\lambda 4686$ 6 years, Bowen and coronal lines, bright UV, dust echoes, multiple X-ray flares
J012026	eROSITA-discovered nuclear transient	He II and Bowen lines, strong Fe II and coronal lines, triple-peaked H $\mathrm{He\,II}\,\lambda 4686$ 7, slow multi-year evolution, BFF and ECLE classification
AT 2021aeuk	radio-loud NLSy1	three optical flares within five years, pre-burst Bowen features, BFF-like but not securely confirmed without on-burst spectroscopy

These sources also illustrate the class diversity. AT 2017bgt was defined by persistence and a strong UV excess. AT 2021loi combined a known AGN host with clear Bowen strengthening and a delayed second optical/UV peak. AT2019aalc added repeated major flares, dust echoes, radio activity, and strong coronal-line behavior. J012026 linked the class to soft-X-ray selection, ECLE phenomenology, and disk-like Balmer structure. AT 2021aeuk showed that BFF-like behavior can occur in a nucleus with pre-burst Bowen features and triple flares, while still resisting secure classification because the decisive flare-state spectra are missing (Baldini et al., 7 Jul 2025).

4. Physical interpretations and major controversies

No single engine has been established for all BFFs. The literature instead supports several overlapping interpretations. One is enhanced accretion onto a pre-existing AGN. This was the preferred reading of AT 2017bgt and is strongly reinforced by AT 2021loi, where the flare occurred in a previously known unobscured AGN with broad Balmer lines already present in the archival spectrum. In this picture, the flare reflects a sudden change in the accretion flow and ionizing SED, which then amplifies line emission in dense BLR-scale gas (Trakhtenbrot et al., 2019).

A second, increasingly prominent interpretation is that at least some BFFs are TDEs in AGN, or more broadly TDE-like accretion episodes inside active nuclei. The synthesis paper centered on AT2019aalc argues that BFFs may occupy an observational and physical middle ground between ordinary AGN variability and classical TDEs in quiescent galaxies, making them strong candidates for a hidden TDE population in active nuclei (Veres, 11 Dec 2025). The evidence cited in favor includes blue continua, Bowen fluorescence, strong coronal lines, soft X-ray emission in some objects, delayed radio flares, IR echoes, repeated major flares suggestive of partial disruptions, QPO-like behavior, and disk-like or multi-peaked Balmer profiles.

The same literature also emphasizes why a blanket identification of all BFFs with TDEs is premature. Their optical declines often do not follow the canonical $\mathrm{He\,II}\,\lambda 4686$ 8 fallback expectation; Balmer lines can be much narrower than the $\mathrm{He\,II}\,\lambda 4686$ 9 broad components common in classical optical TDEs; some systems show cooling during decline; and pre-existing AGN emission can dominate or modify the continuum. Ordinary or triggered accretion-disk instabilities remain viable alternatives in at least some objects.

J012026 sharpened this debate by combining nearly all of the defining BFF traits with a simultaneous ECLE classification. Its soft X-ray flare, strong He II and Bowen lines, strong Fe II and coronal lines, slow evolution, and dusty nuclear environment all support a TDE-like origin in a gas-rich setting. At the same time, its hour- to year-scale X-ray flaring, triple-peaked H $\mathrm{He\,II}$ 0, and kinematic separation between Balmer and Bowen/coronal gas show that the class can be dynamically and geometrically complex rather than a simple narrow-line analog of a standard TDE (Baldini et al., 7 Jul 2025).

AT2019aalc has also motivated disk-instability models. A dedicated 2025 study discussed radiation-pressure instabilities in the pre-existing AGN accretion disk as a possible origin for its recurrent flares and line variability, while still noting that much of the observed behavior links the source to the broader BFF class and sets it apart from canonical TDEs (Śniegowska et al., 30 Apr 2025). This suggests that “BFF” is best treated as an observational class whose members may not all share the same trigger.

5. Time-domain diagnostics: line response, reverberation, and polarimetry

Time-resolved spectroscopy has become one of the most informative diagnostics of BFF behavior. In AT 2021loi, the emission feature around $\mathrm{He\,II}$ 1 \AA\ seen in the pre-flare spectrum strengthened by a factor of $\mathrm{He\,II}$ 2 around optical peak and became clearly double peaked, supporting identification as a blend of $\mathrm{He\,II}$ 3 and $\mathrm{He\,II}$ 4. The appearance of $\mathrm{He\,II}$ 5 during the flare further strengthened the Bowen classification, and the event later showed a rebrightening about one year after the main peak (Makrygianni et al., 2023).

AT2019aalc provides the clearest current example of short-timescale line variability inside a BFF. During the relatively slow dimming after its second optical flare, the UV/optical light curve exhibited minor rebrightening events, while the Bowen fluorescence and coronal lines varied roughly in tandem with these bumps. The authors reported that the broad $\mathrm{He\,II}$ 6 complex and $\mathrm{He\,II}$ 7 can strengthen on week-to-month timescales and may slightly precede the broadband maxima. This constrains the line-emitting gas to compact scales and shows that BFF classification cannot rely only on a single near-peak spectrum; multi-epoch spectroscopy captures the class-defining response pattern (Śniegowska et al., 30 Apr 2025).

Related time-domain work outside the narrow BFF sample provides a methodological framework for interpreting such behavior. Reverberation mapping of $\mathrm{He\,II}$ 8 in NGC 4593 detected day-scale variability with fractional rms amplitude $\mathrm{He\,II}$ 9 and a delay of $\alpha$ 0 days relative to Ly $\alpha$ 1, used as a proxy for Ly $\alpha$ 2, establishing the first Bowen-fluorescence line reverberation mapped. Although that study is not a BFF paper, it shows directly that a Bowen-powered line can respond on day timescales to variations in its pumping radiation field. This suggests that future BFF campaigns should track the pump-line field and the Bowen line simultaneously rather than treating Bowen features as generic recombination responses (Ochmann et al., 17 Dec 2025).

Polarimetry adds a geometrical dimension. A systematic optical polarimetry study of TDE-like transients included three spectroscopically identified BFFs—AT2019aalc, AT2020afhd, and AT2022fpx—and found that all three show polarisation-angle variability. The BFF-specific conclusion was that BFFs tend to exhibit sustained late-time evolution in the optical polarisation angle $\alpha$ 3, likely at least partly because they fade more slowly and therefore remain measurable to later times. The same study stressed that no universal trend appears when time is normalized by the model-dependent fallback time $\alpha$ 4, and that the observed phenomenology favors evolving, non-axisymmetric geometries and/or shocks rather than a static axisymmetric reprocessor (Floris et al., 2 Mar 2026).

BFF discovery is still strongly selection-limited. In a systematic real-time search of ZTF public alerts for SMBH-related nuclear transients, 345 rising nuclear events with no history of previous activity were manually vetted, 223 were spectroscopically classified, and only one BFF was found, corresponding to $\alpha$ 5 of the classified sample. The same study showed that a strict blue cut, $\alpha$ 6, which improved TDE purity, did not recover the one BFF in the sample. The authors therefore concluded that TDE-optimized search strategies are not automatically BFF-optimized, especially because cuts that exclude prior activity can remove AGN-associated BFFs while cuts that relax AGN filtering greatly increase contamination (Dgany et al., 2023).

X-ray selection is emerging as an important complement to optical discovery. J012026 was found in eROSITA as a soft X-ray transient more than an order of magnitude brighter than the previous survey limit, and its optical counterpart was weaker and slower than the X-ray trigger. This directly demonstrates that systematic soft-X-ray surveys can uncover BFFs that would be less conspicuous in optical alert streams and that future facilities such as eROSITA and Einstein Probe are likely to enlarge the class sample (Baldini et al., 7 Jul 2025).

The most important unresolved issues are classification, physical unity, and causality. It remains uncertain whether BFFs constitute a physically unified class or a phenomenological grouping that contains multiple engines. It is not yet known what fraction are true TDEs in AGN, what fraction are enhanced-accretion or disk-instability events, and how much of the observed diversity is driven by viewing angle, obscuration, pre-existing BLR structure, dust geometry, or accretion-disk precession. The literature repeatedly calls for dense photometric and spectroscopic follow-up, infrared monitoring for dust echoes, X-ray monitoring to distinguish absorbed from intrinsically faint ionizing continua, and polarimetric campaigns to constrain geometry (Veres, 11 Dec 2025).

A cautious synthesis is therefore warranted. BFFs are securely an observational class of SMBH-related nuclear flares defined by strong Bowen fluorescence, slow and structured time evolution, and frequent association with active nuclei. They are not well described as ordinary AGN flickering, yet they are also not straightforwardly equivalent to classical TDEs in quiescent galaxies. Current evidence most strongly supports a transitional regime in which intense EUV/soft-X-ray episodes illuminate dense circumnuclear gas in already active or weakly active nuclei, with different events likely weighting enhanced accretion, TDE-like disruption, reprocessing, shocks, and obscuration differently.