Follow-Up Chandra Observations of Three Candidate Tidal Disruption Events

Abstract

Large-amplitude, high-luminosity soft X-ray flares were detected by the ROSAT All-Sky Survey in several galaxies with no evidence of Seyfert activity in their ground-based optical spectra. These flares had the properties predicted for a tidal disruption of a star by a central supermassive black hole. We report Chandra observations of three of these galaxies taken a decade after their flares that reveal weak nuclear X-ray sources that are from 240 to 6000 times fainter than their luminosities at peak, supporting the theory that these were special events and not ongoing active galactic nucleus (AGN) variability. The decline of RX J1624.9+7554 by a factor of 6000 is consistent with the (t-t_D)^(-5/3) decay predicted for the fall-back phase of a tidal disruption event, but only if ROSAT was lucky enough to catch the event exactly at its peak in 1990 October. RX J1242.6-1119A has declined by a factor of 240, also consistent with (t-t_D)^(-5/3). In the H II galaxy NGC 5905 we find only resolved, soft X-ray emission that is undoubtedly associated with starburst activity. When accounting for the starburst component, the ROSAT observations of NGC 5905, as well as the Chandra upper limit on its nuclear flux, are consistent with a (t-t_D)^(-5/3) decay by at least a factor of 1000. Although we found weak Seyfert~2 emission lines in Hubble Space Telescope spectra of NGC 5905, indicating that a low-luminosity AGN was present prior to the X-ray flare, we favor a tidal disruption explanation for the flare itself.