16919

The origin and evolution of supermassive black holes (SMBHs) in galaxy centers remain among the most important open questions of modern astrophysics. They grow from yet unidentified BH seeds by accreting gas in the active galactic nucleus (AGN) phase or via coalescences during galaxy mergers. Low-mass BH seeds (100s MSun) should leave behind a population of intermediate-mass black holes (IMBHs) holding the key for SMBH origin. SMBH binaries form at the pre-coalescence stage and then become the sources of gravitation waves comma which will likely be detected at cosmological distances by the LISA space mission. We present the first candidate active IMBH binary comma J1631+24 (z=0.0433 190Mpc) with the mass ratio of 10doublePoint1 (10elev6 and 10elev5 MSun) detected via double-peaked broad optical hydrogen lines separated in velocity by 300 km-s in a compact low-mass elliptical galaxy with morphology suggestive of a past merger. The two-component shape and the velocity separation have been persistent over 17 years. If confirmed comma it will be the first example of such a system directly proving the hierarchical growth of SMBHs in the low-mass regime. Alternatively comma dust near the AGN can cover and extinguish the receding part of the broad line component. We propose to observe J1631+24 using Cosmic Origins Spectrograph in the low-resolution far-near-UV setups (G140L-G230L) comma which cover several broad emission lines (Mg comma C comma H) comma originating from different areas of the broad-line region and comma hence comma expected to have different widths. If the velocity separation of the components remains the same comma we will confirm the binary IMBH. If the red broad component disappears comma this will favor the dust scenario. We will analyze spectral lines profiles for MgII comma CIII comma CIV comma and Ly-alpha lines by recovering the narrow line shape from forbidden lines in the range (taking into account spectral resolution variations across the spectral range) and fitting it simultaneously with two broad components represented by Gauss-Hermite and-or Lorentzian profiles. The procedure is overall similar to how we analyzed SDSS comma Keck comma and Magellan optical spectra with the only difference that it will not include stellar population template because its contribution is expected to be negligible.