Revisiting the unification of tidal disruption events with polarimetry

TL;DR

Tidal disruption events exhibit optically bright emission whose origin remains debated. By expanding optical polarimetry to 19 TDEs (including 9 newly analysed), the study tests a viewing-angle–dependent unification model: most non-relativistic TDEs show continuum polarisation $P_V\sim1$–$2\%$ early on, with line depolarisation indicating electron-scattering reprocessing in aspherical envelopes. Multi-epoch polarimetry reveals rapid evolution of geometry, with polarisation often declining toward zero after $\sim70$ days, implying early disk formation and rapid circularisation in many cases, though a subset deviates from the unification predictions. The work also links polarimetric behavior to multi-wavelength data (X-ray, infrared, radio), finding broad agreement with unification for many events while highlighting diversity that invites time-dependent modeling and alternative emission scenarios for a complete picture.

Abstract

Tidal disruptions of stars by supermassive black holes produce multi-wavelength emission, of which the optical emission is of ambiguous origin. A unification scenario of tidal disruption events (TDEs) has been proposed to explain the different classes of X-ray and optically selected events by introducing a dependence on the viewing angle and geometry. This work aims to test the unification scenario among optically bright TDEs using polarimetry. By studying the optical linear polarisation of 19 TDEs (of which 9 newly analysed in this work), we place constraints on their photosphere geometry, inclination, and the emission process responsible for the optical radiation. We study how these properties correlate with the relative X-ray brightness. We find that 14/16 non-relativistic events can be accommodated by the unification model. Continuum polarisation levels of optical TDEs lie most often in the range P ~ 1-2% (13 events), and for all except one event, remain below 6%. For those optical TDEs that have multi-epoch polarimetry, the continuum polarisation decreases after peak light for 5/10 events, increases for 3/10 events, and stays nearly constant for 2/10 events. When observed after +70 days (7/16 events), they become consistent with P = 0% within uncertainties (5/7 events). This implies the photosphere geometries of TDEs are at least initially asymmetric and evolve rapidly which, if tracing the formation of the accretion disk, suggests efficient circularisation. The polarisation signatures of emission lines of 7 TDEs directly support a scenario in which optical light is reprocessed in an electron-scattering photosphere. [...] However, a subset of events deviates from the unification model to some extent, suggesting this model may not fully capture the diverse behaviour of TDEs. Multi-epoch polarimetry plays a key role in understanding the evolution and emission mechanisms of TDEs.

Figures (16)

• Figure 1: Polarisation spectra of two epochs of AT 2021blz after correcting for the ISP and the host galaxy dilution. In each panel, left-hand axes refer to the polarisation angle $\theta$, the Stokes parameters $q$ and $u$, and the polarisation degree $P$ (corrected for the polarisation bias), all shown in red. The right-hand axes refer to the flux spectra at the time of polarimetry (in units of $10^{-16}$ erg cm$^{-2}$ s$^{-1}$ Å$^{-1}$), which are shown in black across emission lines and in grey across the assumed continuum. Regions of strong telluric absorption are marked as grey shaded regions. The identified line species are shown as vertical coloured lines, e.g., the Balmer lines (dark blue), N III emission lines (olive), and the He II emission line (green).
• Figure 2: Same as Figure \ref{['fig:specpol_AT2021blz']} but for AT 2022dsb.
• Figure 3: Dominant axis fits to the data of AT 2022dsb at +2 days and AT 2023mhs at +9 days in the Stokes $q,u$ plane. For each fit, we include the reduced $\chi^2$ statistic followed by the number of degrees of freedom in parentheses, the coefficient of determination $R^2$, and the ellipse axial ratio $b/a$.
• Figure 4: Same as Figure \ref{['fig:specpol_AT2021blz']} but for AT 2022bdw.
• Figure 5: Same as Figure \ref{['fig:specpol_AT2021blz']} but for AT 2023mhs.