A black hole by itself is invisible, detectable only through its mild effect on passing starlight, but place it where it can pull massive amounts of gas onto itself, and the black hole lights up as brilliantly as any object in the galaxy at all wavelength, extending from the radio into the gamma-ray. Nature provides a black hole with just this environment when the black hole is a member of a compact binary star system. When a black hole is in close orbit with a fusion-powered star, so that the center of gravity of the system lies within the atmosphere of the star, gas flows from the star onto the black hole, forming a hot disk orbiting and lighting up the black hole. We believe we are seeing this process in the x-ray binary star systems.
X-ray binary stars are among the most brilliant objects in the x-ray sky. They can release up to 1038 ergs s-1 of power, which is 5,000 times the power output of the Sun. They are visible across the galaxy, because x-rays can penetrate the clouds of gas and dust in the galactic plane, and some are visible in nearby galaxies such as the Magellanic clouds. They are highly variable, with low-luminosity states that generate a tenth of a percent of the power generated in the high-luminosity state. Many of these systems are composed of a neutron star in orbit with a fusion-powered star, but a handful of systems appear to be a black hole in orbit with a fusion-powered star. It has been estimated that the Galaxy contains about 108 black holes, but only several hundred to several thousand are currently in x-ray binary systems.
The best evidence for the presence of a black-hole candidate in some of these systems is the mass measure for the compact object in the system. This is not a simple undertaking, because the system cannot be resolved in a telescope. Under the best circumstances, the surface of a giant star can be resolved through interferometry; smaller stars remain points of light in the best optical telescopes. With a compact binary system, we have two objects separated by the diameter of a star, with a distance of at least a kiloparsec, and more often tens or hundreds of kiloparsecs. Such systems cannot be resolved, so our only information about the orbit of a compact binary system comes from measuring the Doppler shift of light from each object.
The problem is that the Doppler shift does not contain enough information to precisely derive a mass for each object in a close binary. We can derive a period for the system, and we can derive the velocity along the line of sight for each object in the system, but we cannot derive a velocity perpendicular to our line of sight. For instance, if we were looking along a system's rotation axis, we would see no Doppler shift, but if were were looking perpendicular to the rotation axis, we would see maximum and minimum Doppler shifts equal to the absolute velocity of each object in the system. At any other angle, the maximum velocity we measure would be less than the absolute velocity. The effect is that we cannot derive a mass, but we can derive a minimum value for the mass. Because we believe a compact object larger than about 3 solar masses must be a black hole, this spectroscopic measure of mass is sufficient for showing that an x-ray binary cannot contain a neutron star.
Some x-ray binary systems show eclipses of the accretion disk by the fusion-powered star. This effect enables us to refine estimates of a compact object's mass by giving us an estimate of the angle between our line of sight and the rotation axis of the system.
From spectroscopic measurements of mass, we know of several systems that contain black-hole candidates. Two sources in the Small Magellanic Cloud, SMC X-1 and SMC X-3, contain black-hole candidates with masses of about 6 solar masses for the former and greater than 7 solar masses for the latter. Within our own galaxy, the binary Cygnus X-1, which is about 2.5 kiloparsecs away, had a mass greater than 7 solar masses, with a most likely value of 16 solar masses. Two other Galactic binaries with the position-based names of 0620-003, and 2023+338 have masses greater than 7 and 8 solar masses. Under current theory, the compact object in each of these systems should be a black hole.
The interesting question, however, is can we prove that the large, compact objects in these systems are in fact black holes? Does the light emitted by these systems contain unambiguous signatures of general relativity?
The hurdle confronting astronomers who want to answer these questions is the similarity between neutron stars and black holes. A neutron star is not much larger than the last stable orbit of a black hole of the same mass, so the amount of energy released by an accretion disk around a neutron star is not dramatically different from that release by an accretion disk around a black hole of the same mass. The only difference in the accretion disk between the two types of compact object is that the inner edge of the disk around a neutron star interacts with the surface of the neutron star, while the inner edge of the disk around a black hole goes into free-fall to the event horizon.
Under this circumstance, the evidence is largely negative; one can observe the effects of a surface, but not of a last stable orbit. The strongest signature of a surface is a thermonuclear explosion of hydrogen and helium. Such events have been seen in many x-ray binaries containing neutron stars. Hydrogen and helium from the accretion disk flows onto the surface of the neutron star. If the gas flows at a high rate, the atmosphere of the neutron star will be hot enough to sustain continuous thermonuclear fusion of hydrogen and helium to heavier elements, but if the gas flow is at a low rate, hydrogen and helium build up in the atmosphere until a thermonuclear detonation occurs. This sudden outburst is easily identified as a thermonuclear event, and its observation is clear evidence that the compact object has a surface. If one of our large black-hole candidates would exhibit a thermonuclear flash, we would have evidence of a compact object larger than 3 solar masses that is not a black hole. So far, no such event has been seen.
Other signatures are more ambiguous. Both the energy liberated through steady thermonuclear fusion and the energy liberated as the gas from the accretion disk mixes with the atmosphere of a neutron star are converted to thermal radiation deep in the atmosphere. This radiation should escape the star as black-body radiation, which should be apparent in the spectrum of these systems. Many, although not all, x-ray binaries with neutron stars do show such emission, and none of the binaries with black-hole candidates show such thermal emission. Again, this provides an absence of evidence for a surface, which does not necessarily mean that no surface is present.
The x-ray binary provides us with the best opportunity to find black-hole candidates. Whether we can prove that these candidates are in fact black holes is yet to be seen. Certainly if one of our candidates suddenly produces thermonuclear bursts, we will have evidence against some piece of our physics that says black holes are larger than 3 solar masses. To this point, however, these systems tell us more about the physics of accretion disks and the evolution of stars in compact binaries than it does about black holes and general relativity.