Irregular Repeating Tidal Disruption Events due to Diffusive Tides

Abstract

A repeating partial tidal disruption event (rpTDE) is typically modelled as a star in a bounded orbit getting disrupted by a massive black hole at each pericenter passage. For the disruption to occur, the pericenter distance should be close to or within the characteristic tidal radius, such that the tidal field can overcome the star's binding force to trigger mass loss. However, a binary with a pericenter distance several times the tidal radius can build up its tidal perturbation over multiple orbits via a diffusive process, eventually triggering a nonlinear instability that may also eject mass and power an eruption. This leads to repeated disruptions that recur stochastically. In this letter, we propose that such a mechanism can produce a subclass of rpTDEs with large variations in the recurrence time (e.g., J0456-20), which we dub ``diffusive-tide rpTDEs''. We show that diffusive tidal growth can occur for a white dwarf or main-sequence star orbiting a massive black hole when the pericenter distance is a few times the tidal radius, provided the orbital period is shorter than the tidal energy dissipation timescale. These diffusive-tide rpTDEs may account for a significant fraction of all rpTDEs.

Figures (4)

• Figure 1: Left: The $f$-mode energy evolution of a WD, and two different MSs with $\alpha = 0.3\text{ and } 0.8$, orbiting a MBH. The initial ($D_p/R_t$, $P_0$) are (2.9, 10 hours) for the WD system and (2.0, 45 days) for the MS system. The red horizontal line indicates the wave-breaking threshold, and the red dots show where wave breaking occurs, triggering mass ejection and powering the rpTDEs. The range of the time between mass ejections (within the plotting range) is labeled in each case. The mass loss parameters ($\sigma_1$, $\lambda$) are taken to be ($10^{-3}$, $0$) and ($2\times10^{-2}$, $0$) for the WD and MS respectively. Right: The evolution of $D_p/R_t$ after each mass ejection.
• Figure 2: Left: The time between successive mass ejection episode against episode number occurred for the WD system and MS system with $\alpha = 0.8$. The non-monotonic pattern observed in J0456-20 is found in each case, as indicated by the points enclosed by the boxes. Right: The probability density of the recurrence time.
• Figure 3: (Left): The orbital parameter space of WD-MBH system where diffusive-tide rpTDEs are expected is colored in yellow. The black cross indicates the fiducial orbital parameters (see Table \ref{['tab:fiducial']}). The black solid line sets the boundary of diffusive growth assuming $Q = 10^6 = \omega_a P_0$, and the dotted line (with $Q = 2\times 10^3$) gives the marginal boundary for the fiducial model. The magenta line represents the threshold for diffusive growth driven by GW backreaction Lau_2025. The red solid (dotted) line is the threshold for diffusive growth due to nonlinear anharmonicity (linear tidal backreaction). The blue line corresponds to where prompt disruption is expected. (Right): Similar plot for an MS-MBH system with $\alpha=0.8$. A potential evolution pathway is shown by the arrowed lines: the MS might be captured on a wide orbit via the Hills mechanism (black star) and circularize, with the diffusive tide suppressed by large dissipation. When the fiducial point (cross) is reached, diffusive tides start to power rpTDEs. The MS becomes denser as it loses mass, eventually quenching the eruptions (diamond).
• Figure 4: (Left): Parameter space plot showing the evolution of the MS-MBH binary ($\alpha = 0.8$) for different initial $D_p/R_t$ and $\lambda$. The brown contours indicate the regions where the final mass of the star ($M_{*,f}$) reaches 50 % and 20 % of the initial mass ($M_{*, 0}$). The pink and cyan regions correspond to orbital circularization and unbinding, respectively. The red dotted line is the approximate boundary of whether the star or the orbit is disrupted first (Eq. \ref{['eq:Dp_crit']}). The prompt disruption region is colored gray. (Right): Parameter space plot of the evolution for different initial $D_p/R_t$ and $\alpha$, with $\lambda = 0$. The pink region now corresponds to $e_f < 0.95$, rather than $0.90$. There is also no region corresponding to $e_f > 0.99$.