We report deep Subaru { H } { \alpha } observations of the extended ultraviolet ( XUV ) disk of M83 Based in part on data collected at Subaru Telescope , which is operated by the National Astronomical Observatory of Japan . .
These new observations enable the first complete census of very young stellar clusters over the entire XUV disk .
Combining Subaru and GALEX data with a stellar population synthesis model , we find that ( 1 ) the standard , but stochastically-sampled , initial mass function ( IMF ) is preferred over the truncated IMF , because there are low mass stellar clusters ( 10 ^ { 2 - 3 } { M _ { \odot } } ) that host massive O-type stars ; that ( 2 ) the standard Salpeter IMF and a simple aging effect explain the counts of FUV-bright and { H } { \alpha } -bright clusters with masses > 10 ^ { 3 } { M _ { \odot } } ; and that ( 3 ) the { H } { \alpha } to FUV flux ratio over the XUV disk supports the standard IMF .
To reach conclusion ( 2 ) , we assumed instantaneous cluster formation and a constant cluster formation rate over the XUV disk .
The Subaru Prime Focus Camera ( Suprime-Cam ) covers a large area even outside the XUV disk – far beyond the detection limit of the HI gas .
This enables us to statistically separate the stellar clusters in the disk from background contamination .
The new data , model , and previous spectroscopic studies provide overall consistent results with respect to the internal dust extinction ( A _ { V } \sim 0.1 mag ) and low metallicity ( \sim 0.2 { Z _ { \odot } } ) using the dust extinction curve of the Small Magellanic Cloud .
The minimum cluster mass for avoiding the upper IMF incompleteness due to stochastic sampling and the spectral energy distributions of O , B , and A stars are discussed in the Appendix .