🔭 astrophysics

Can tidal disruption event models reliably measure black hole masses?

This study validates that current tidal disruption event models can reliably estimate black hole masses within 0.3–0.5 dex using three repeating partial TDEs, while highlighting critical limitations in fallback model grids and the necessity of near-UV light curve coverage to prevent systematic underestimation in future large-scale surveys.

Original authors: C. R. Angus, A. J. Smith, D. Magill, P. Ramsden, N. Sarin, M. Nicholl, B. Mockler, E. Hammerstein, R. Stein, Y. Yao, T. de Boer, K. C. Chambers, M. E. Huber, C. -C. Lin, T. B. Lowe, E. A. Magnier, S.

Published 2026-08-05

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: C. R. Angus, A. J. Smith, D. Magill, P. Ramsden, N. Sarin, M. Nicholl, B. Mockler, E. Hammerstein, R. Stein, Y. Yao, T. de Boer, K. C. Chambers, M. E. Huber, C. -C. Lin, T. B. Lowe, E. A. Magnier, S. J. Smartt, R. J. Wainscoat

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the center of a galaxy as a cosmic playground where a giant, invisible monster—a supermassive black hole—lurks. This monster is so heavy that it bends space and time, but it's usually quiet, sleeping in the dark. To wake it up, you need a snack: a star that wanders too close. When a star gets too near, the black hole's gravity stretches it like a piece of taffy, tearing it apart in a spectacular event called a Tidal Disruption Event (TDE). This creates a brilliant flash of light, a cosmic beacon that tells us the monster is awake.

For a long time, scientists thought these flashes were like a "full course meal," where the black hole devoured the entire star in one go. By measuring how bright the flash was and how long it lasted, they tried to guess the monster's weight (its mass). But there's a catch: many of these stars might only get a "snack" taken from them, leaving a core that survives and circles back for another bite later. These are called "repeating partial TDEs." The big question is: if we use our old "full meal" recipes to measure the weight of a black hole that only took a bite, will our scales still work, or will we get the wrong answer? This is the puzzle a team of astronomers set out to solve.

The researchers decided to test their black hole scales using three specific cosmic events where a star was clearly only partially eaten and came back for more: TDE 2020vdq, TDE 2022dbl, and a new discovery they helped confirm, TDE 2023adr. Think of these three events as a perfect control group. Since the black hole's weight doesn't change between the first bite and the second, any good measuring tool should give the exact same weight for the monster every time it looks at the flash. If the tool gives different answers for the same monster, the tool is broken.

The team tried four different ways to measure the black holes, ranging from complex computer models that simulate how debris falls back into the black hole, to simpler rules of thumb based on how bright the flash is. They found that, generally, the tools work well enough. When they measured the black holes in these three events, the different methods agreed with each other and matched the weights estimated by looking at the host galaxy's properties. They could determine the black hole's mass to within about 0.3 to 0.5 "dex" (a logarithmic unit where 1 dex is a factor of 10), which is a decent estimate in the messy world of astronomy.

However, the study also found some cracks in the tools. When the models tried to figure out exactly how much of the star was eaten and how deep the black hole's "bite" went, they got confused. Often, the models insisted the star was completely destroyed, even though we know it survived to come back. It's like a scale that insists you ate the whole pizza when you only took a slice. This suggests that the current "recipes" for these models need a little more seasoning to handle partial bites correctly.

The paper also looked ahead to the future, specifically the massive Rubin Observatory, which will soon take pictures of the sky every few nights and find thousands of these events. The team ran simulations to see how well they could measure black holes with just the data from this new telescope. They found a worrying trend: without extra, high-quality follow-up observations, especially in ultraviolet light, the models will likely underestimate the black hole's mass by about 0.5 dex on average. It's like trying to guess the size of a shadow without seeing the object casting it; you might think the object is smaller than it really is.

In short, the paper concludes that while we can currently get a rough idea of a black hole's mass from these flashes, our tools aren't perfect yet. They work best when we have a lot of data covering the event from start to finish, especially in ultraviolet light. As we prepare to discover thousands of these events in the coming years, we'll need to be careful not to trust the numbers too blindly and will need to keep refining our cosmic scales to ensure we aren't underestimating the giants in the dark.

Technical Summary: Can tidal disruption event models reliably measure black hole masses?

Problem Statement
Tidal Disruption Events (TDEs) offer a unique, host-independent method for measuring the masses of quiescent supermassive black holes ( $M_{BH}$ ), particularly in the low-mass regime where scaling relations are poorly constrained. However, the reliability of current semi-analytic TDE models is uncertain. Most models assume "full" stellar disruptions, whereas theoretical expectations and recent observations suggest that "partial" disruptions (pTDEs), where a stellar core survives, may dominate the population. In partial disruptions, the degeneracy between black hole mass ( $M_{BH}$ ), stellar mass ( $M_*$ ), and the impact parameter ( $\beta$ ) increases, complicating parameter recovery. Furthermore, it is unclear whether models designed for full disruptions can accurately recover consistent $M_{BH}$ values when applied to the multiple flares produced by the same surviving star in repeating partial TDEs (rpTDEs).

Methodology
The authors test the robustness of current TDE models using a sample of three spectroscopically confirmed rpTDEs: 2020vdq, 2022dbl, and 2023adr. The study introduces AT 2023adr as the third spectroscopically confirmed rpTDE, establishing its nature through multi-epoch spectroscopy showing broad hydrogen, helium, and nitrogen emission features consistent with the "H+He" spectral class.

The authors independently model the light curves of each flare in these three events using four distinct approaches:

Fallback Accretion Models: Using the Modular Open Source Fitter for Transients (MOSFiT), which assumes luminosity tracks the fallback rate of debris, incorporating a reprocessing layer.
Stream-Stream Collision Scaling: Using TDEmass, which relies on peak luminosity and color-temperature relations derived from debris stream collisions.
Empirical Luminosity Scaling: Applying relations between peak $g$ -band luminosity and $M_{BH}$ , and between late-time plateau luminosities and $M_{BH}$ .
Cooling Envelope Models: Using the Redback package (Sarin & Metzger 2024), which models emission from a pressure-supported envelope formed by super-Eddington fallback.

The derived $M_{BH}$ values from each flare are compared for internal self-consistency. Additionally, these values are benchmarked against $M_{BH}$ estimates derived from host galaxy properties, including stellar velocity dispersion ( $\sigma_*$ ), total stellar mass ( $M_*$ ), and bulge mass ( $M_{bulge}$ ).

To assess the impact of observational limitations expected in future surveys, the authors simulate LSST (Legacy Survey of Space and Time) light curves for TDE 2022dbl. They test parameter recovery under Wide-Fast-Deep (WFD) and Deep Drilling Field (DDF) cadences, with and without rest-frame near-UV coverage, to quantify systematic biases.

Key Results

Model Consistency: Across the three rpTDEs, the different modeling frameworks generally return self-consistent $M_{BH}$ values between flares (within $\sim 1\sigma$ ) and are broadly consistent with host-galaxy proxies. The typical precision is estimated at 0.3–0.5 dex once model-specific systematics are accounted for.
Degeneracies and Systematics: While $M_{BH}$ is relatively robust, stellar mass ( $M_*$ ) and impact parameters remain highly degenerate. Notably, the MOSFiT fallback model frequently converges on solutions implying near-full or full disruptions ( $b \approx 1$ ) even for known rpTDEs, suggesting limitations in current stellar structure grids and fallback prescriptions for partial encounters.
Data Coverage Importance: The quality of constraints is heavily dependent on light curve coverage, particularly in the near-UV. Events lacking near-UV data (e.g., the first peak of 2020vdq) yield significantly weaker constraints on temperature and bolometric luminosity, leading to broader posteriors.
LSST Simulation Findings: Simulations of LSST-quality light curves reveal a systematic bias in $M_{BH}$ $M_{B H}$ recovery. When fitting simulated data with MOSFiT:
- DDF-like sampling (higher cadence) results in a mean underestimation of $\sim -0.12$ dex.
- WFD-like sampling (sparser cadence) results in a mean underestimation of $\sim -0.5$ dex.
- This bias is attributed to sparse sampling of the light curve rise and peak (compressing the inferred fallback timescale) and the lack of rest-frame near-UV coverage (leading to underestimated temperatures and bolometric luminosities).

Significance and Claims
The paper claims that while current TDE models can provide order-of-magnitude constraints on $M_{BH}$ that are consistent with host galaxy scaling relations, significant systematic uncertainties remain. The convergence of models towards unphysical stellar masses and impact parameters in rpTDEs highlights the need for improved stellar structure models and better handling of partial disruption physics.

Crucially, the authors argue that for the thousands of TDEs expected from LSST, relying solely on photometric light curve modeling without targeted follow-up will likely lead to a systematic underestimation of black hole masses. They conclude that robust $M_{BH}$ inference for large samples will require a complementary approach: using empirical plateau-luminosity scaling relations for low-redshift events where plateaus are accessible, and relying on improved light curve models that incorporate better stellar physics and multi-wavelength coverage (specifically near-UV) to break degeneracies in stellar properties and disruption depth.

More like this