Such IMBHs have long been the topic of
speculation and searches. Two classes of
black holes are observationally well-established: the stellar-mass black holes discovered in Galactic X-ray binaries, and the supermassive ones in the centers of galaxies,
with a large mass gap separating the two
classes. The best IMBH candidate so far is in
ESO 243-49 HLX-1, for which the recently-inferred black hole mass does enter the upper
end of the IMBH range (Webb et al., 2012),
depending on the definition adopted.
The remaining wide gap between stellarmass and supermassive black holes is frustrating for those hunting them, since many
theorists assume that today’s supermassive
black holes formed via “seed” IMBHs. If no
IMBHs exist in the present-day universe, it
would throw doubt on that scenario. Whilst
not detecting IMBHs is not the same as proving they are not present — black holes are,
after all, not intrinsically bright objects — a
direct detection would be very welcome.
There have been indirect inferences of the
presence of IMBHs in globular clusters, but
the arguments are not universally accepted.
Circumventing the Eddington Limit
One way around the apparent Eddington
limit would be if the emitted radiation was
non-spherical, i.e., if the luminosity of ULXs
was preferentially directed towards us. This
option cannot be excluded in all cases; how
would Galactic microquasars such as SS 433
or GRS 1915+105 appear if we were looking
directly down their jets? However, measurements of the energy which is deposited into
nebulae around ULXs suggest that the power output of typical ULXs is unlikely to be
significantly smaller than the value which is
derived using the assumption that the emission is spherically-symmetric (see, e.g., Pakull
and Mirioni, 2003).
4
Another easy-looking option would be to
discard the Eddington limit (which assumes
spherical symmetry) on the plausible-seeming grounds that accretion through a disk is
not spherically-symmetric. The simple version of this argument fails, however, because,
at luminosities approaching the Eddington
luminosity, the inner parts of the accretion
disk are expected to become radiation-pressure dominated. Without some additional
unknown mechanism, the inner disk would
consequently thicken and the accretion geometry would become quasi-spherical.
More complicated ways of circumventing
the Eddington limit have been proposed.
These have tended to invoke a mechanism
for transferring energy from the inner accretion disk to a corona surrounding the black
hole. The concept connects naturally with
the fact that the spectra of many ULXs are
dominated by a power-law component (see,
e.g., Gladstone, Roberts, and Done, 2009).
That power law is normally identified with
a “Comptonisin