Prospects for Measuring Black Hole Masses using TDEs with the Vera C. Rubin Observatory

TL;DR

The paper demonstrates that black hole masses inferred from Rubin-observed TDEs can achieve typical uncertainties of about $0.26$ dex when using MOSFiT in conjunction with Rubin cadence simulations, with improved accuracy ($ oughly 0.25$ dex) if pre-detection data are sufficiently deep. It emphasizes that while the disrupted-star mass is highly degenerate with accretion efficiency and the ability to classify full versus partial disruptions is limited (about 57% accuracy), Rubin's large sample will constrain the black hole mass function and BH–galaxy co-evolution, including spins, provided that TDE-rate variations and observable degeneracies are thoroughly understood. The work also highlights the need for complementary observables (e.g., UV data) and modeling of TDE rates across BH masses to fully exploit Rubin’s forthcoming data set. Overall, Rubin TDEs offer a powerful, complementary channel to AGN-based methods for mapping SMBH demographics and their connection to galaxy evolution, especially for higher-mass systems and potential spin inferences via population statistics.

Abstract

Tidal Disruption Events (TDEs) provide an opportunity to study supermassive black holes that are otherwise quiescent. The Vera C. Rubin Legacy Survey of Space and Time will be capable of discovering thousands of TDEs each year, allowing for a dramatic increase in the number of discovered TDEs. The optical light curves from TDEs can be used to model the physical parameters of the black hole and disrupted star, but the sampling and photometric uncertainty of the real data will couple with model degeneracies to limit our ability to recover these parameters. In this work, we aim to model the impact of the Rubin survey strategy on simulated TDE light curves to quantify the typical errors in the recovered parameters. Black hole masses $5.5< \log M_{\rm BH}/M_\odot < 8.2$ can be recovered with typical errors of 0.26 dex, with early coverage removing large outliers. Recovery of the mass of the disrupted star is difficult, limited by the degeneracy with the accretion efficiency. Only 57\% of the cases have accurate recovery of whether the events are full or partial, so we caution the use this method to assess whether TDEs are partially or fully disrupted systems. Black hole mass measurements obtained from Rubin observations of TDEs will provide powerful constraints on the black hole mass function, black hole -- galaxy co-evolution, and the population of black hole spins, though continued work to understand the origin of TDE observables and how the TDE rate varies among galaxies will be necessarily to fully utilize the upcoming rich data set from Rubin.

Figures (10)

• Figure 1: Eddington fraction vs. black hole mass for the simulated TDEs we generate with MOSFiT. We generate a sample of 300 synthetic TDEs, as described in the text. The majority of these have peak Eddington fractions $L_{\rm bol}/L_{\rm Edd} > 30$, which is higher than typically considered in TDE simulations and beyond what can be accurately modeled with MOSFiT. We restrict our analysis going forward to only the 87 events with input $L_{\rm bol}/L_{\rm Edd} < 30$. The majority of simulated TDEs around black holes with $\log M_{\rm BH}/M_\odot<5.5$ are highly super-Eddington and excluded from our sample. Because of this effect, we stress that our resulting population of simulated TDEs is no longer a complete model population covering the full range of likely black hole masses.
• Figure 2: Black hole masses (left), star masses (center), and scaled impact parameters ($b$) used to generate synthetic TDEs. The unfilled (filled) histogram shows the sample before (after) our cut on Eddington fraction. The dotted line in each case shows the scaled distributions from which we select each parameter.
• Figure 3: Four example synthetic TDE light curves, constructed as described in §\ref{['sec:methods']}. The detections, limits, seasonal gaps, and cadence reflect realistic observing conditions that will limit our ability to recover TDE parameters. Best-fit MOSFiT results are shown as solid lines, obtained as discussed in §\ref{['sec:results']}.
• Figure 4: Recovery of black hole mass by MOSFiT fitting of synthetic TDE observations, colored by the number of detections (left) and maximum depth achieved (in any band) in the 30 days prior to the first detection (right). Grey circles are cases with no limits $>1$ mag below the first detection in the 30 days prior to that first detection. The black hole masses have a median difference of 0.26 dex between the input and recovered values. The total number of detections primarily scales with the black hole mass, as the higher black hole mass events are brighter and longer-lived. The synthetic TDEs with good early coverage are the most accurately recovered. If we consider only synthetic TDEs with limits of $>1$ mag below the first detection in the 30 days prior, the black hole mass recovery is accurate to 0.25 dex and we remove cases with extreme $>1$ dex differences from the input black hole mass.
• Figure 5: Left: recovery of disrupted star mass by MOSFiT fitting of synthetic TDE observations, colored by the best-fit accretion efficiency. Right: ratio of recovered to input star mass vs. accretion efficiency. The mass of the disrupted star is highly degenerate with the efficiency of luminosity produced by mass accretion $\epsilon$Mockler2021. Because we fix the efficiency of the synthetic TDEs to be $\epsilon = 0.1$, while the efficiency is a free parameter in our MOSFiT recovery fits, the resulting best-fit star masses are degenerate with the efficiency. Cases where the recovered star mass is higher than the input are typically those with lower recovered efficiencies $\epsilon < 0.1$, and vice versa.