Ultraluminous X-ray sources and star formation

Chandra observations of the Cartwheel galaxy reveal a population of ultraluminous X-ray sources (ULXs) with lifetimes less than or similar to10(7) yr associated with a spreading wave of star formation which began some 3x10(8) yr ago. A population of high-mass X-ray binaries provides a simple model: donor stars of initial masses M(2)greater than or similar to15 M-circle dot transfer mass on their thermal time-scales to black holes of masses M(1)greater than or similar to10 M-circle dot. For alternative explanations of the Cartwheel ULX population in terms of intermediate-mass black holes (IMBHs) accreting from massive stars, the inferred production rate greater than or similar to10(-) yr(-1) implies at least 300 IMBHs, and more probably 3x10(4), within the star-forming ring. These estimates are increased by factors of eta(-1) if the efficiency eta with which IMBHs find companions of greater than or similar to15 Mcircle dot within 10(7) yr is <1. Current models of IMBH production would require a very large mass (greater than or similar to 10(10) M-circle dot) of stars to have formed new clusters. Further, the accretion efficiency must be low (less than or similar to 6x10(-3)) for IMBH binaries, suggesting super-Eddington accretion, even though intermediate black hole masses are invoked with the purpose of avoiding it. These arguments suggest either that, to make a ULX, an IMBH must accrete from some as yet unknown non-stellar mass reservoir with very specific properties, or that most if not all ULXs in star-forming galaxies are high-mass X-ray binaries.