Continuum-Fitting the X-ray Spectra of Tidal Disruption Events

We develop a new model for X-ray emission from tidal disruption events (TDEs), applying stationary general relativistic ``slim disk'' accretion solutions to supermassive black holes (SMBHs) and then ray-tracing the photon trajectories from the image plane to the disk surface, including gravitational redshift, Doppler, and lensing effects self-consistently. We simultaneously and successfully fit the multi-epoch XMM-Newton X-ray spectra for two TDEs: ASASSN-14li and ASASSN-15oi. We test explanations for the observed, unexpectedly slow X-ray brightening of ASASSN-15oi, including delayed disk formation and variable obscuration by a reprocessing layer. We propose a new mechanism that better fits the data: a ``Slimming Disk'' scenario in which accretion onto an edge-on disk slows, reducing the disk height and exposing more X-rays from the inner disk to the sightline over time.For ASASSN-15oi, we constrain the SMBH mass to $4.0^{+2.5}_{-3.1} \times 10^6M_\odot$. For ASASSN-14li, the SMBH mass is $10^{+1}_{-7}\times 10^6M_\odot$ and the spin is $0.998^{+0}_{-0.7}$. For both TDEs, our fitted masses are consistent with independent estimates; for ASASSN-14li, application of the external mass constraint narrows our spin constraint to $0.998^{+0}_{-0.15}$. The mass accretion rate of ASASSN-14li decays slowly, as $\propto t^{-1.1}$, perhaps due to inefficient debris circularization. Over $\approx$1100 days, its SMBH has accreted $\Delta M \approx 0.17 M_\odot$, implying a progenitor star mass of $> 0.34 M_\odot$, i.e., no ``missing energy problem.'' For both TDEs, the hydrogen column density declines to the host galaxy plus Milky Way value after a few hundred days, suggesting a characteristic timescale for the depletion or removal of obscuring gas.