Motivation
Introduction
BG Geminorum (hereafter BG Gem) was discovered by Hoffmeister (1933) and was initially thought to be RV Tauri star. Benson et al. (2000) discovered an eclipsing and ellipsoidal variation in BG Gem, with an optical amplitude of 0.5 mag and a period of 91.645 days. Combined with a sinusoidal radial velocity curve for the secondary this suggested a binary system, with a K0 giant/supergiant secondary feeding material into an accretion disc around the unseen primary. Using low resolution spectra, Kenyon et al. (2002) calculated the mass function to be (3.5 ± 0.5) M_sun; this is greater than the theoretical minimum mass for a black hole (3 M_sun, Rhoades and Ruffini, 1974). However, because the primary cannot be observed directly (the inclination of the system is close to 90 degrees), it could also be a B-star (Kenyon et al., 2002). If confirmed to be a black hole, BG Gem would be the black hole binary system with the largest orbital period, by an factor of 3, as well as being the only known eclipsing black hole binary system in the Galaxy.
UV & optical spectroscopic evidence for a black hole primary
One method to investigate the nature of the primary is to acquire higher quality optical spectra, to determine the maximum rotational velocity of the accretion disc surrounding the primary. Large velocities would favour a black hole primary as the maximum velocity for a disc surrounding a B star is ∼ 500 -- 700 km/s (Kenyon et al., 2002). Based on low quality spectra, Kenyon et al. (2002) claimed evidence for the presence of rotational velocities approaching 2000 km/s. We acquired higher resolution optical spectra with the ALFOSC spectrograph at the Nordic Optical Telescope (NOT) in La Palma. The resolution was ∼ 1.5 Å in the wavelength range 3300 Å ≤ λ ≤ 10500 Å; however, these spectra revealed velocities for the Hα and Hβ lines of approximately 500 km/s, low enough that the primary could still be a B-star (see Fig. 1). Unforunately, due to the poorly constrained ephemeris, the orbital phase at which these observation were taken could not be determined, as such the primary could have been eclipsed, shielding any possible high velocity material from our view.
BG Gem was also observed by HST in the UV in an effort to further constrain the nature of the primary star. Grism spectra were obtained with STIS in 2 wavelength ranges, FUV: 1200 - 1700 Å; NUV: - - 1700 - 3000 Å using the G140L & G230L grisms respectively, the resulting summed spectrum is displayed in Fig.2. We note that if the primary was in fact a B star we would expect to observe absorption lines of Si II & Si III in addition to a strong continuum, neither of which is evident in the spectrum. Instead, the observed spectrum exhibits a number of strong emission lines from ions of Si, Al and Mg in addition to a weak continuum. The velocities of these emission lines are displayed in Table 1, it can be seen that the lines are of a high velocity. In particular we note; for a B8V star an emission line at a velocity of 700 km/s (1000 km/s) is emitted from a radius of 2.3 R_sun (1.13 R_sun), assuming Keplerian rotation. This is less than the radius of a B8V star (3.0 R_sun, Cox 2000) and strongly suggestive of a black hole primary.
Left : Hα and Hβ emission lines normalised to the local continuum, and offset for added clarity. Right :Combined spectrum for the FUV and the NUV data acquired on 2001 December 30. The insets show the Si IV (Å) and S I (λ1914.70 Å) lines plotted against velocity of emission.
