On the lack of thermal emission from the quiescent black hole XTE J1118+480: Evidence for the event horizon

A soft component of thermal emission is very commonly observed from the surfaces of quiescent, accreting neutron stars. We searched with Chandra for such a surface component of emission from the dynamical black hole candidate XTE J1118+480 (=J1118), which has a primary mass M(1)approximate to8 M-circle dot. None was found, as one would expect if the compact X-ray source is a bona fide black hole that possesses an event horizon. The spectrum of J1118 is well fitted by a simple power-law model that implies an unabsorbed luminosity of L(X)approximate to3.5x10(30) ergs s(-1) (0.3-7 keV). In our search for a thermal component, we fitted our Chandra data to a power-law model (with slope and N-H fixed) plus a series of nine hydrogen atmosphere models with radii ranging from 9/8 to 2.8 Schwarzschild radii. For the more compact models, we included the important effect of self-irradiation of the atmosphere. Because of the remarkably low column density to J1118, N(H)approximate to1.2x10(20) cm(-2), we obtained very strong limits on a hypothetical thermal source: kT(infinity)<0.011 keV and L-infinity,L-th<9.4x10(30) ergs s(-1) (99% confidence level). In analogy with neutron stars, there are two possible sources of thermal radiation from a hypothetical surface of J1118: deep crustal heating and accretion. The former mechanism predicts a thermal luminosity that exceeds the above luminosity limit by a factor of greater than or similar to 25, which implies that either one must resort to contrived models or, as we favor, J1118 is a true black hole with an event horizon. In addition to neutron stars, we also consider emission from several exotic models of compact stars that have been proposed as alternatives to black holes. As we have shown previously, accreting black holes in quiescent X-ray binaries are very much fainter than neutron stars. One potential explanation for this difference is the larger and hence cooler surface of an 8 M-circle dot compact object that might be masked by the interstellar medium. However, our upper limit on the total luminosity of J1118 of 1.3x10(31) ergs s(-1) is far below the luminosities observed for neutron stars. This result strengthens our long-held position that black holes are faint relative to neutron stars because they possess an event horizon.