- Evidence for the Black Hole Event Horizon
Abstract: Astronomers have discovered many candidate black holes in the universe and have studied their properties in ever-increasing detail. Over the last decade, a few groups have developed observational tests for the presence of event horizons in candidate black holes. The talk will discuss one of these tests, which indicates that the supermassive black hole at the center of our Galaxy must have a horizon.
You can find the recording at PIRSA: 09020024.
Black hole formation is a prediction of General Relativity (GR). We know that stars that have masses more than a few times the solar mass can not, once their nuclear power is burned out, stabilize at a finite radius and the gravitational pressure of their own mass will cause them to completely collapse. In this process, the density of the object increases, and the gravitational force on the surface gets stronger. If the gravitational force on the surface gets so strong not even light can escape, we call this surface an event horizon. It is the characteristic feature of black holes. Classically, nothing can ever leave the region behind the event horizon.
Since the early 90s, evidence has mounted for astrophysical black holes. These come in two rough categories: solar size black holes, with masses of a few times the mass of our sun that form directly from collapse of stars, and the so-called supermassive black holes, with masses about a million to a billion times the solar mass, that form through accretion in densely populated areas, mostly in the center of galaxies.
The cheap way to label an object a black hole is to measure its mass (eg from the motions of nearby stars) and its radius (eg by determining the source area of its emission). For a black hole, we know the relation between both, R = 2 GM, where G is the gravitational constant, R is the radius, and M is the mass. The radius of a black hole of about solar mass would be roughly 3 km, and that of a supermassive black hole is then consequently some million to billion kms. If one has data that allows to estimate mass and radius, if there is too much mass in an observed region of spacetime, one can conclude it has to be a black hole. (Keep in mind this is astrophysics, so observables typically have large errorbars and it takes some effort to pin down conclusions.)
This is however somewhat unsatisfactory. What one would really like to know is whether the object does have an event horizon, which is the defining feature of a black hole. The question is then, what observables can help us to determine whether we are dealing with a compact object that has a surface, or with an object that has an event horizon?
First let me emphasize that compact objects of the masses we are concerned with here that have a radius close by but not quite the radius of a black hole are not possible in GR. These objects can't be stabilized. But if one modifies GR, one can get away with this. People have looked into such modifications but these are not very convincing options. The reason is simple: To avoid collapse, one needs a mechanism to stabilize matter at a density that allows the matter to just not form a black hole. That is, one needs a deviation from the standard theory at densities of about M/R3, and inserting the black hole radius this goes as ~ 1/M2. This means, the more massive the black hole is, the smaller is the density at which you need deviations from the standard theory.
And this density can be arbitrarily small. It can be as small as densities we deal with every day. Take a supermassive black hole with 109 times the mass of the sun, which has a radius of about 109km. This gives a density of about 1039kg per 1027km3, or 1 kg per dm3, which is about the density of water. Not exactly a very extreme condition, and one that we have quite some experience with. From Einstein's field equations we further know the density scales like the background curvature. This means if you want to generally avoid the formation of black holes, you need modifications of GR in the arbitrarily small curvature regime. In this regime, the theory is extremely well tested, and we have not seen any deviations whatsoever.
But still, one would like to have observational evidence for the presence of the horizon (after all, it could be a naked singularity, no?). The key to this is to compare the emissions of an object that does have a surface with that of an object that does not have a surface. Astrophysical black holes accrete matter, and that matter heats up, which leads to emissions. When the accreted matter hits the surface this also leads to emissions, that can - in the case of astrophysical black holes - be violent nuclear explosions. An object with an event horizon on the other hand will not have contributions to the emitted radiation from the surface. Both will thus differ in their luminosity, which is observable.
In his talk, Narayan summarized the observations of the luminosity of both solar mass black holes in our galaxy, and for Sgr A*, the supermassive black hole in the center of our galaxy. In both cases, the observed emission is much smaller than would be expected if the object had a surface, and thus clear evidence for the presence of an event horizon.
Related: Coincidentally, Moshe just today wrote a nice post on Frozen Stars.