We present optical spectroscopic measurements of the eclipsing High Mass X-ray Binary XMMU J013236.7+303228 in M 33 .
Based on spectra taken at multiple epochs of the 1.73 d binary orbital period we determine physical as well as orbital parameters for the donor star .
We find the donor to be a B1.5IV sub-giant with effective temperature is T = 22 , 000 - 23 , 000 K. From the luminosity , temperature and known distance to M33 we derive a radius of R = 8.9 \pm 0.5 R _ { \odot } .
From the radial–velocity measurements , we determine a velocity semi-amplitude of K _ { opt } = 63 \pm 12 km s ^ { -1 } .
Using the physical properties of the B-star determined from the optical spectrum , we estimate the star ’ s mass to be M _ { opt } = 11 \pm 1 M _ { \odot } .
Based on the X-ray spectrum , the compact companion is likely a neutron star , although no pulsations have yet been detected .
Using the spectroscopically derived B-star mass we find the neutron star companion mass to be M _ { X } = 2.0 \pm 0.4 M _ { \odot } , consistent with the neutron star mass in the HMXB Vela X-1 , but heavier than the canonical value of 1.4 M _ { \odot } found for many millisecond pulsars .
We attempt to use as an additional constraint that the B star radius inferred from temperature , flux , and distance , should equate the Roche radius , since the system accretes by Roche lobe overflow .
This leads to substantially larger masses , but from trying to apply the technique to known systems , we find that the masses are consistently overestimated .
Attempting to account for that in our uncertainties , we derive M _ { X } = 2.2 ^ { +0.8 } _ { -0.6 } M _ { \odot } and M _ { opt } = 13 \pm 4 M _ { \odot } .
We conclude that precise constraints require detailed modeling of the shape of the Roche surface .