We have studied the candidate optical counterparts and the stellar population in the star-forming complex around the bright ULX in the western part of the spiral galaxy NGC 4559 , using HST /WFPC2 , XMM-Newton /Optical Monitor , and ground-based data .
We find that the ULX is located near a small group of OB stars , but is not associated with any massive young clusters nor with any extraordinary massive stars .
The brightest point source in the Chandra error circle is consistent with a single blue supergiant of mass \approx 20 M _ { \odot } and age \approx 10 Myr .
A few other stars are resolved inside the error circle : mostly blue and red supergiants with inferred masses \approx 10 – 15 M _ { \odot } and ages \approx 20 Myr .
This is consistent with the interpretation of this ULX as a black hole ( BH ) accreting from a high-mass donor star in its supergiant phase , with mass transfer occurring via Roche-lobe overflow .
The observed optical colors and the blue-to-red supergiant ratio suggest a low metal abundance for the stellar population : 0.2 \la Z / Z _ { \odot } \la 0.4 ( using the Padua tracks ) , or 0.05 \la Z / Z _ { \odot } \la 0.2 ( using the Geneva tracks ) .
The age of the star-forming complex is \la 30 Myr .
H \alpha images show that this star-forming region has a ring-like appearance .
We propose that it is an expanding wave of star formation , triggered by an initial density perturbation , in a region where the gas was only marginally stable to gravitational collapse .
We also suggest that the most likely trigger was a collision with a satellite dwarf galaxy going through the gas-rich outer disk of NGC 4559 less than 30 Myr ago .
The culprit could be the dwarf galaxy visible a few arcsec north-west of the complex .
If this is the case , this system is a scaled-down version of the Cartwheel galaxy .
The X-ray data favour a BH more massive ( M > 50 M _ { \odot } ) than typical Milky Way BH candidates .
The optical data favour a “ young ” BH originating in the recent episode of massive star formation ; however , they also rule out an association with young massive star clusters ( none are present in the X7 field ) .
We speculate that other mechanisms may lead to the formation of relatively massive BHs ( perhaps M \sim 50 – 100 M _ { \odot } ) from stellar evolution processes in low-metallicity environments , or when star formation is triggered by galactic collisions .