Dynamical Evidence for a Black Hole in the Eclipsing X-ray Binary M33 X-7

ATel #977; Jerome A. Orosz (San Deigo State University), Jeffrey E. McClintock (CfA), Charles D. Bailyn (Yale), Ronald A. Remillard (MIT), Wolfgang Pietsch (MPI Garching), & Ramesh Narayan (CfA)
on 17 Jan 2007; 20:02 UT
Credential Certification: Jerome A. Orosz (orosz@sciences.sdsu.edu)

Tweet

We have obtained optical spectroscopy of the counterpart to the eclipsing high-mass X-ray binary M33 X-7 (Pietsch et al 2004, 2006) using the GMOS instrument on the Gemini-North Telescope. We used the B1200 grating (lambda_c=4650 Angstroms) and a 0.5 arcsecond slit rotated to a position angle of 215.6 degrees (which is the angle defined by M33 X-7 and its neighbor star 0.9 arcseconds to the southwest). Twenty four 40-minute spectra were acquired in service mode between 2006 August 18 and November 16 in good seeing (always less than 0.8 arcseconds). The two observations obtained on 2006 September 17 are suspect and will not be considered here.

The data were reduced using the GMOS package in IRAF. The overlap of the spatial profiles of M33 X-7 and the neighbor star was modest. The spectral extraction aperture was adjusted so that light from the neighbor star was not included, which resulted in about a 20% loss of light from M33 X-7. Fortunately, M33 X-7 is bright (B = 18.9), and the final extracted spectra had signal-to-noise ratios of 20 or more per 0.47 Angstrom pixel near H beta. The wavelength range covered is 3910 to 5370 Angstroms, with a spectral resolution of about 80 km/sec.

Numerous nebular lines are seen in emission. These include the Balmer lines H beta through H epsilon, [O III] near 4363, 4969 and 5007 Angstroms, and weak He I lines near 4026, 4471, 4921 and 5015 Angstroms. The He II line near 4686 Angstroms and the N III lines near 4640 are also in emission. The average heliocentric velocity of the brightest [O III] line is -131.2 km/sec (the velocity of M33 itself given in the NED database is -179 +/- 3 km/sec). The standard deviation of the 22 measurements is 1.5 km/sec, which indicates good wavelength stability.

Two strong absorption lines from the secondary star, He II 4200 and He II 4541, are apparent in all of the spectra and are not contaminated by nebular emission or interstellar absorption. Radial velocities of the secondary were derived using the FXCOR package in IRAF. A synthetic spectrum derived from the OSTAR2002 grid of Lanz & Hubeny (2003) with T=37,500K and log(g)=3.75 was the template, and the cross-correlation region was 4140-4300, 4521-4578, and 4708-4835 Angstroms, which includes the two strong He II lines noted above plus a few weaker metallic lines. Good cross correlation peaks were found for each of the 22 spectra.

Using the orbital period of 3.453014 days determined from the observed X-ray eclipses (Pietsch et al. 2006), a sine fit to the 22 velocities yields K = 108.9, chi^2 = 10.92, and an rms of 21.44 km/sec. The uncertainties in the radial velocities were scaled to yield a reduced chi^2 of 1, and the following parameters were derived: K = 108.9 +/- 6.4 km/sec, gamma = -152 +/- 5 km/sec, and T_0 = HJD 2,453,967.157 +/- 0.048. Here, T_0 is the predicted time of mid-X-ray eclipse, which is in full agreement with that of Pietsch et al. (2006) -- they differ by 95.0005 +/- 0.0140 orbital cycles. The optical mass function is f(M) = 0.46 +/- 0.08 solar masses.

Given the above value of the mass function, the mass of the compact object can be less than 3 solar masses only if the mass of the secondary is less than 4.7 solar masses, which can be ruled out since the absolute V magnitude of -6.1 and the blue colors found by Pietsch et al. (2006) indicate that the secondary is an O-type giant or supergiant. For a plausible rock-bottom secondary mass of 20 solar masses, the mass of the compact object exceeds 6.9 solar masses. We therefore conclude that the compact object in M33 X-7 is a black hole.

Lanz & Hubeny 2003, ApJS, 146, 417

Pietsch et al. 2004, A&A, 413, 879

Pietsch et al. 2006, ApJ, 646, 420