Context. The quest to detect dormant stellar-mass black holes (BHs) in massive binaries (i.e. OB+BH systems) is challenging; only a few candidates have been claimed to date, all of which must still be confirmed. Aims. To search for these rare objects, we study 32 Galactic O-type stars that were reported as single-lined spectroscopic binaries (SB1s) in the literature. In our sample we include Cyg X-1, which is known to host an accreting stellar-mass BH, and HD 74194, a supergiant fast X-ray transient, in order to validate our methodology. The final goal is to characterise the nature of the unseen companions to determine if they are main-sequence (MS) stars, stripped helium stars, triples, or compact objects such as neutron stars (NSs) or stellar-mass BHs. Methods. After measuring radial velocities and deriving orbital solutions for all the systems in our sample, we performed spectral disentangling to extract putative signatures of faint secondary companions from the composite spectra. We derived stellar parameters for the visible stars and estimated the mass ranges of the secondary stars using the binary mass function. Variability observed in the photometric TESS light curves was also searched for indications of the presence of putative companions, degenerate or not. Results. In 17 of the 32 systems reported as SB1s, we extract secondary signatures, down to mass ratios of ∼ 0.15. For the 17 newly detected double-lined spectroscopic binaries (SB2s), we derive physical properties of the individual components and discuss why they have not been detected as such before. Among the remaining systems, we identify nine systems with possible NS or low-mass MS companions. For Cyg X-1 and HD 130298, we are not able to extract any signatures for the companions, and the minimum masses of their companions are estimated to be about 7 M . Our simulations show that secondaries with such a mass should be detectable from our dataset, no matter their nature: MS stars, stripped helium stars or even triples. While this is expected for Cyg X-1, confirming our methodology, our simulations also strongly suggest that HD 130298 could be another candidate to host a stellar-mass BH. Conclusions. The quest to detect dormant stellar-mass BHs in massive binaries is far from over, and many more systems need to be scrutinised. Our analysis allows us to detect good candidates, but confirming the BH nature of their companions will require further dedicated monitorings, sophisticated analysis techniques, and multi-wavelength observations.
