Chandra spectroscopy of the TDE AT2021ehb: evidence of a relativistic outflow

ATel #15179; J. M. Miller, M. T. Reynolds, S.-B. Yun, A. Zoghbi (Univ. of Michigan), B. Cenko (NASA/GSFC), J. J. Drake (SAO), J. A. Irwin (Univ. Alabama), C. Miller (Univ. of Maryland), R. Mushotzky (Univ. of Maryland)
on 25 Jan 2022; 15:09 UT
Credential Certification: Jon Miller (jonmm@umich.edu)

Tweet

Following a rise in flux that met the trigger of our approved Cycle 22 TOO program, we observed AT 2021ehb (Munoz-Arancibia et al. 2021, Yao et al. 2021) with Chandra for 50 ks, divided between two exposures obtained on 2021 Nov 29. The dispersed spectra from the LETG were read using the HRC. The most interesting result of our analysis is that a power-law continuum shaped by a moderately ionized, relativistic wind (v = -0.15c) is preferred over alternatives, including a blackbody plus power-law continuum.

The Cash statistic was minimized in the fitting process, across the 5-25 Angstrom range, after binning to require a signal to noise of S/N = 3.5. A simple power-law continuum, modified by neutral absorption in the Milky Way and host galaxy, leaves emission-like residuals near to 16 A with flux decrements on either side (Gamma = 2.7). However, no single-temperature thermal plasma model significantly improves the fit to this region. The addition of a blackbody with a temperature of kT = 0.10 keV improves the fit close to 16 A, but does not fully describe the complex. Comparing this two-component model to the simple power-law using the Akaike Information Criterion gives delta(AIC) = 1.2, signaling the improvement is not significant.

Instead, allowing the power-law continuum to be modified by a partially covering wind component provides the best overall description of the spectrum. Simple fits with the "zxipcf" model give a Cash statistic of 24.4 for 37 degrees of freedom and 44 bins. The column density of the absorber is N_H = 8 (+7,-2) E+22 cm^-2, the ionization parameter is log xi = 2.0+/-0.1, the covering fraction is 0.77 (+0.06,-0.09), and the wind velocity is v = -0.15 +/- 0.02 (measured relative to the host galaxy at z = 0.017). Again utilizing the AIC, we find delta(AIC) = 5, signaling an improvement. The zxipcf model samples ionization very coarsely; replacing zxipcf with alternative absorption models confirms the results obtained with zxipcf.

Examining the 10 Swift/XRT monitoring spectra obtained closest to the Chandra observation, spanning October 30 - December 30 2021, we found that joint fits including a zxipcf component with a variable column density results in an improvement over a simple power-law, with delta(AIC) = 10. The jointly-determined outflow velocity is -0.24c, suggesting that the wind may be even faster on average. We caution that CCD data are ill-suited to this kind of spectral analysis.

The Chandra results imply a wind with a high covering factor. However, if the wind is active for a lengthy period - as may be implied in the Swift monitoring observations - a low volume filling factor is required if the total mass outflow is not to exceed the mass available from a disrupted solar-mass star. Alternatively, the TDE may have stimulated enhanced accretion from a pre-existing accretion disk or gas reservoir around the massive black hole. This may be supported by the fact that the X-ray rise phase occurred long after the discovery of AT2021ehb (March 3, 2021; Munoz-Arancibia et al. 2021).

We thank the Chandra Director, Pat Slane, the CDO, and the Mission Planning team for executing our observations, and the Neil Gehrels Swift Observatory team for monitoring AT2021ehb.

Munoz-Arancibia, A., et al., 2021, TNSTR, 651, 1
Yao, Y., 2021, TNSCR, 2291, 1