A correlation between H α trough depth and inclination in quiescent X-ray transients: evidence for a low-mass black hole in GRO J0422+32

Casares, J.; Muñoz-Darias, T.; Torres, M. A. P.; Mata Sánchez, D.; Britt, C. T.; Armas Padilla, M.; Álvarez-Hernández, A.; Cúneo, V. A.; González Hernández, J. I.; Jiménez-Ibarra, F.; Jonker, P. G.; Panizo-Espinar, G.; Sánchez-Sierras, J.; Yanes-Rizo, I. V.
Bibliographical reference

Monthly Notices of the Royal Astronomical Society

Advertised on:
We present a new method to derive binary inclinations in quiescent black hole (BH) X-ray transients (XRTs), based on the depth of the trough (T) from double-peaked H α emission profiles arising in accretion discs. We find that the inclination angle (i) is linearly correlated with T in phase-averaged spectra with sufficient orbital coverage (≳50 per cent) and spectral resolution, following i(deg) = 93.5 × T + 23.7. The correlation is caused by a combination of line opacity and local broadening, where a leading (excess broadening) component scales with the deprojected velocity of the outer disc. Interestingly, such scaling allows to estimate the fundamental ratio M1/Porb by simply resolving the intrinsic width of the double-peak profile. We apply the T-i correlation to derive binary inclinations for GRO J0422+32 and Swift J1357-0933, two BH XRTs where strong flickering activity has hindered determining their values through ellipsoidal fits to photometric light curves. Remarkably, the inclination derived for GRO J0422+32 (i = 55.6 ± 4.1○) implies a BH mass of $2.7^{+0.7}_{-0.5}$ M⊙ thus placing it within the gap that separates BHs from neutron stars. This result proves that low-mass BHs exist in nature and strongly suggests that the so-called 'mass gap' is mainly produced by low number statistics and possibly observational biases. On the other hand, we find that Swift J1357-0933 contains a $10.9^{+1.7}_{-1.6}$ M⊙ BH, seen nearly edge on ($i=87.4^{+2.6}_{-5.6}$ deg). Such extreme inclination, however, should be treated with caution since it relies on extrapolating the T-i correlation beyond i ≳ 75○, where it has not yet been tested.