Spectropolarimetry of the changing-look active galactic nucleus NGC 1566 and its potential link to supermassive black hole binaries

Context. The active galactic nucleus (AGN) NGC 1566 is known to present dramatic and regular spectral shape changes, associated with the appearance and disappearance of broad emission lines. Aims. The underlying mechanism responsible for such changes is yet to be identified, but occultation, eccentric accretion disks, turbulent disk-dominated broad line regions (BLRs) or binary supermassive black holes have been hypothesized. Methods. Because the scenarios used to explain the variable spectral shapes of NGC 1566 each have a specific geometric configuration, we used the Very Large Telescope (VLT) FOcal Reducer/low dispersion Spectrograph 2 (FORS2) instrument to obtain nine 3500─10 000 Å polarized spectra of the source between August 2 and September 21, 2025. Results. We caught the AGN in a type-2 state, i.e., without any broad component in total nor polarized fluxes. Its low (0.22─0.24%) and wavelength-independent polarization degree (and angle) above 4000 Å argues against occultation of the BLR and is consistent with a significant weakening or disappearance of the BLR. The polarized spectrum reveals a strong rise of polarization in the blue band, likely echoing the 2018 outburst of the AGN. The temporal variability of the total flux continuum but the steadiness of the line profiles demonstrate that the object is viewed close to pole-on, irrespective of its spectral type at the time of observation. Relative to archival data, NGC 1566 shows significant variability in polarization degree, angle, and wavelength dependence, transitioning from gray to chromatic polarization. Even more surprisingly, NGC 1566 behaves opposite to the basic predictions of the unified model: its polarization angle is perpendicular to the AGN polar axis and its polarization degree is higher when in a brighter, type-1 phase. Conclusions. The results reported above contradict occultation and binary supermassive black hole hypotheses. They support, rather, accretion-driven photoionization and/or structural changes in the internal accretion flow and the BLR, but their geometry can only be probed by new spectropolarimetric campaigns during a type-1 phase of NGC 1566.