Finding “Scary Barbie”’s sisters: a new sample of ambiguous nuclear transients

Title: A systematically-selected sample of luminous, long-duration, ambiguous nuclear transients

Author(s): P. Wiseman, R. D. Williams, I. Arcavi, M. J. Graham, S. Hönig, M. Newsome, B. Subrayan, M. Sullivan, Y. Wang, D. Ilić, M. Nicholl, S. Oates, T. Petrushevska, and K. W. Smith

First Author’s Institution: School of Physics and Astronomy, University of Southampton, Southampton, SO17 1BJ, United Kingdom

Status: Submitted to MNRAS, available on arXiv

The first image of a black hole from a simulation (Jean-Pierre Luminet 1979).

Who is “Scary Barbie”?

Only the most luminous optical transient ever discovered! Scary Barbie was found last year, and its unique properties have kickstarted a search for other transients like her. With an energy release only explainable by accretion onto a supermassive black hole (SMBH), Scary Barbie was initially theorized to be a tidal disruption event (TDE) where a ~15 solar mass star wandered too close to a 108 solar mass SMBH and was torn apart by gravity.  The unusually large masses in the TDE model and the lack of a host galaxy detection challenging the AGN model suggested that a new class of transients may emerge, prompting a search for more ultra-luminous long-lived optical flares in galactic nuclei. Beyond TDEs, a new class of highly luminous flares coincident with the centers of galaxies has been identified, called ambiguous nuclear transients (ANTs). Almost all serendipitously discovered and with heterogeneous properties, ANTs are defined by their distance from the centers of the host galaxy, their light curves and spectra failing to match known sources, and long decays over several years. The properties that have been identified appear to fall in between those of discrete accretion events like “typical” TDEs and those of more long-term, continuous accretion, like in an active galactic nucleus (AGN). Variations in AGN optical luminosity occur on timescales ranging from seconds to years, typically described as a “damped random walk,” and ANTs show similar stochastic variability but superimposed onto a single, large flare. Some AGN show the disappearance or re-appearance of broad emission lines in their spectra, called “changing-look” AGN (CLAGN), which is poorly understood in the classic viewing-angle model for AGN. The authors of today’s paper aim to understand the relationship between AGNs, ANTs, and TDEs by searching for more examples of transients like Scary Barbie.

Figure 1: The ZTF and WISE (mid-infrared) filter light curves of the ten sources passing all cuts and manual inspection. The three unfilled curves with variable light curves are classified as ANTs spectroscopically, not via the light curve selection pipeline since it excludes sources that do not show a purely linear decline. In all these cases, optical flares with significant spread in rise, timesfall times, and variability precede long-lived, delayed MIR flares. Figure 3 from today’s paper.

Long Lost Sisters

To find more ANTs, the authors of today’s paper search the entire Zwicky Transient Facility archive. ZTF is a widefield survey that captures the entire night sky visible from California every ~two nights and has been operating for ~6 years. Its combination of sensitivity, cadence, and reference baseline is optimal for searching for ANTs, which have huge amplitude flares that can rise over two weeks to two years and fade away over 3+ years. To search for these transients, the authors identified all events less than 0.3 arcseconds from the center of a cataloged host galaxy and hostless “orphan” events like Scary Barbie. Among these, they performed several key data cuts to find sources similar to Scary Barbie. There must be at least fifty detections spanning at least a year beyond the initial detection, and the transient must brighten by at least one magnitude. Finally, the decay phase must be well-fitted by a power law or exponential decay (this looks like a straight line when plotted in magnitude over time). The resulting 59 candidates were manually inspected, and some were followed up spectroscopically for confirmation, resulting in a sample of nine new ANTs plus one unclear classification. The optical and mid-infrared light curves are shown in Figure 1.

Figure 2: A sub-set of the spectral range observed for each ANT, zoomed in to specific lines of interest. On the right, H-alpha (656.3nm) shows a variety of line profiles, from the narrow emission of Scary Barbie/AT2021lwx to the highly asymmetric profile of AT2019brs. This light curve most closely resembles that of Scary Barbie. On the left, H-beta shows a similar range in profiles. Also on the left, sources like AT2020adpi show some nitrogen (N III) and helium (He II) emission lines, indicators of Bowen fluorescence caused by a steep far-UV or X-ray source, like an accretion disk. Figure 5 in today’s paper.

The resulting sample of ANTs shows a wide range of properties. The rise times vary from as short as 17 days to 2 years and decay over many years. Although its peak luminosity is lower, the integrated energy release of ZTF19aamrjar exceeds that of Scary Barbie, attributed to its much longer rise time (see the first two upper-left panels of Figure 1). All show broad Balmer lines; see Figure 2. However, H-beta is comparably narrower than typically observed in TDEs. All ANTs in this sample except one have a large amplitude mid-infrared flare consistent with light “echoing” off a dusty torus a few tenths of a light-year around the central SMBH, a feature typically seen in AGN. 

The authors consider several explanations for these events. They argue that instability in a pre-existing accretion disk is unlikely to cause these flares since the dynamical timescales for a typical accretion disk surrounding an SMBH are three orders of magnitude too slow. There is no a priori reason a TDE could not occur in an existing AGN, a popular explanation for one of the first ANTs discovered, PS-16dtm. However, models for such a scenario are sparse, given the complexity. In the third scenario, the ubiquitous presence of MIR flares is invoked to argue that there is an overdensity of gas and dust near the SMBH. A three-body interaction could result in a shrinking orbit and disruption of a giant molecular cloud. Finally, “normal” TDEs of particularly massive stars may also contribute to a significant fraction of ANTs, with the rate of ANTs observed here consistent with predictions for mid-mass (1.5-10 solar mass) stars. However, to definitively determine the origin of these ANTs, a larger sample is necessary to disentangle the different formation pathways and test more complex models of the relationship between TDEs, ANTs, and AGNs. 

Featured Image Credit: NASA/JPL-Caltech

Edited by Nathalie Korhonen Cuestas

Author

  • Will Golay

    I am a graduate student in the Department of Astronomy at Harvard University and the Center for Astrophysics | Harvard & Smithsonian, advised by Edo Berger. I study radio emission from transient astrophysical objects like tidal disruption events.

    View all posts

Submit a Comment

Your email address will not be published. Required fields are marked *