X-ray Spectral-Timing Properties of Tidal Disruption Events

TL;DR

This study presents the first systematic X-ray timing analysis of tidal disruption events (TDEs) on minute-to-hour scales using XMM-Newton data, comparing thermal disk-dominated states with corona-dominated states across 18 TDEs. By applying Fourier timing techniques, the authors quantify variability via power spectral densities, fractional variability, and energy-dependent Fourier-resolved spectra, revealing that corona-bearing TDEs are more variable and exhibit steeper PSDs than thermal TDEs, with hard X-ray variability distinct from AGN. The work shows that thermal TDEs display Wien-tail–driven, energy-dependent variability consistent with local temperature fluctuations, while newborn coronae imply higher optical depths and altered disk-corona coupling, yielding intermediate PSD slopes between thermal TDEs and AGN. These results illuminate the microphysics of nascent accretion flows around supermassive black holes and demonstrate that TDE coronae differ fundamentally from AGN, offering a new avenue to probe turbulence and disk-jet–corona interactions in rapidly evolving systems.

Abstract

We perform the first systematic study of the minute-to-hours-timescale stochastic variability observed in the X-ray luminosity of tidal disruption events (TDEs) using XMM-Newton data and Fourier analysis methods. We measure the spectral properties, power spectral densities (PSDs), fractional variability amplitudes, and energy dependence of the variability for 18 TDEs spanning 54 observations, of which 27 occur in thermal disk-dominated states and 27 show a nonthermal hard X-ray corona. Compared to pure thermal sources, we find TDEs with coronae are more X-ray variable and show steeper PSDs indicating longer correlation timescales. This state-transition behavior is qualitatively similar to X-ray binaries, which show higher fractional variability in the hard state than in the soft state. However, newborn TDE coronae show systematically flatter PSDs and softer energy spectra than their long-lived AGN counterparts. We also show that the variability amplitude of thermal TDEs increases with photon energy, consistent with variations sourced by local temperature fluctuations and exponentially enhanced in the Wien tail. Our work demonstrates that combining spectral and timing properties of X-ray TDEs can probe the microphysics of newly formed accretion flows around supermassive black holes, and that the coronae formed in TDEs fundamentally differ from those in AGN.

Figures (15)

• Figure 1: (a) time-series light curve with 100 second bins, (b) power spectral density, (c) Fourier-resolved spectrum, and (d) energy spectrum of the TDE AT2022lri.
• Figure 1: All XMM-Newton 0.3-2 keV light curves for the 18 sources/54 observations in our sample, binned to 200 seconds for visual clarity. We retained only the longest continuous segments in each observation after filtering for background flare contamination. Observations with coronae are shown with square markers, while thermal observations are shown with vertical line markers.
• Figure 2: Normalized light curves (with offsets of 0 or 1), energy spectra, spectral residuals, and Fourier-resolved spectra of some representative TDEs. ASASSN-14li depicts the rising FRS typical of thermal TDEs. AT2020ocn, which showed a state transition from a pure thermal spectrum to one with a corona, showed a corresponding change in the slope of its FRS. For AT2022lri only, we have removed the gaussian emission component near 1 keV for visualization purposes only, such that the line is visible in the spectral residuals panel. We observe a deficit in the FRS coinciding with the spectral emission component, indicating a physically distinct emission and variability process than the disk or corona continuum.
• Figure 2: All XMM-Newton spectra fit with tbabs$\times$zashift$\times$simpl$\times$tdediscspec. Dashed lines show the thermal component only, and solid lines show the combined thermal + corona continua.
• Figure 3: Probability densities of 0.3-2 keV fractional variability RMS amplitude ($F_{\rm var}$), power-law index of the Fourier-resolved spectrum ($\Gamma_{\rm FRS}$), power-law index of the PSD ($\alpha_{\rm PSD}$), power-law index of the X-ray spectrum in sources with a corona ($\Gamma_{\rm spec}$), and $\log(M_{\rm BH}/M_\odot)$, for samples of thermal TDEs (red), TDEs with coronae (purple), and low-mass AGN (blue). We use Gaussian kernel density estimation with a bandwidth $\sigma N^{-1/4}$, where $\sigma$ is the standard deviation of a parameter within a given sample and $N$ is the sample size. AGN $M_{\rm BH}$ uncertainties uniformly were assumed to be 0.3 dex. AGN PSDs were drawn from Gonzales12 and photon indices from Liu16; for both samples, we retained only sources with $\log(M_{\rm BH}/M_\odot)<7.5$ to construct a comparable-mass sample to the TDEs.