In the present paper we show that within all the uncertainties that govern the process of Roche lobe overflow in Case Br type massive binaries, it can not be excluded that a significant fraction of them merge and become single stars. We demonstrate that at least some of them will spend most of their core helium burning phase as hydrogen rich blue stars, populating the massive blue supergiant region and/or the massive Be type star population. The evolutionary simulations let us suspect that these mergers will explode as luminous hydrogen rich stars and it is tempting to link them to at least some super luminous supernovae.

Liberal evolution of interacting binaries has been proposed previously by several authors in order to meet various observed binary characteristics better than conservative evolution does. Since Algols are eclipsing binaries the distribution of their orbital periods is precisely known. The distribution of their mass ratios contains however more uncertainties. We try to reproduce these two distributions theoretically using a liberal scenario in which the gainer star can lose mass into interstellar space as a consequence of its rapid rotation and the energy of a hot spot. In a recent paper (Van Rensbergen et al. 2010, A&A) we calculated the liberal evolution of binaries with a B-type primary at birth where mass transfer starts during core hydrogen burning of the donor. In this paper we include the cases where mass transfer starts during hydrogen shell burning and it is our aim to reproduce the observed distributions of the system parameters of Algol-type semi-detached systems. Our calculations reveal the amount of time that an Algol binary lives with a well defined value of mass ratio and orbital period. We use these data to simulate the distribution of mass ratios and orbital periods of Algols. Binaries with a late B-type initial primary hardly lose any mass whereas those with an early B primary evolve in a non-conservative way. Conservative binary evolution predicts only ~ 12 % of Algols with a mass ratio q above 0.4. This value is raised up to ~ 17 % using our scenario of liberal evolution, which is still far below the ~ 45 % that is observed. Observed orbital periods of Algol binaries larger than one day are faithfully reproduced by our liberal scenario. Mass ratios are reproduced better than with conservative evolution, but the resemblance is still poor.

In a liberal evolutionary scenario, mass can escape from a binary during eras of fast mass transfer. We calculate the mass lost by binaries with a B-type primary at birth where mass transfer starts during hydrogen core burning of the donor. We simulate the distribution of mass-ratios and orbital periods for those interacting binaries. The amount of time the binary shows Algol characteristics within different values of mass-ratio and orbital period has been fixed from conservative and liberal evolutionary calculations. We use these data to simulate the distribution of mass-ratios and orbital periods of Algols with the conservative as well as the liberal model. We compare mass-ratios and orbital periods of Algols obtained by conservative evolution with those obtained by our liberal model. Since binaries with a late B-type primary evolve almost conservatively, the overall distribution of mass-ratios will only yield a few Algols more with high mass-ratios than conservative calculations do. Whereas the simulated distribution of orbital periods of Algols fits the observations well, the simulated distribution of mass-ratios produces always too few systems with large values.

The observed distribution of periods and mass ratios of Algols with a B type primary at birth was updated. Conservative evolution fails to produce the large fraction with a high mass ratio: i.e. q in [0.4-0.6]. Interacting binaries thus have to lose mass before or during Algolism. During RLOF mass is transferred continuously from donor to gainer. The gainer spins up; sometimes up to critical velocity. Equatorial material on the gainer is therefore less bound to the system. The material coming from the donor through the first Langrangian point impinges violently on the surface of the gainer or the edge of the accretion disc, creating a hot spot in the area of impact. The sum of rotational energy (fast rotation) and radiative energy (hot spot) depends on the mass-loss rate. The sum of both energies on a test mass located in the impact area equals exactly its binding energy at some critical value. As long as the mass transfer rate is smaller than this value the gainer accepts all the mass coming from the donor: RLOF happens conservatively. But as soon as the critical rate is exceeded the gainer will acquire no more than the critical value and RLOF runs into its liberal era. Low mass binaries never achieve mass-loss rates larger than the critical value. Intermediate mass binaries evolve mainly conservatively but mass will be blown away from the system during the short era of rapid mass transfer soon after RLOF-ignition. Binaries with 9+5.4 solar masses (P in [2-4] d) evolve almost always conservatively. Only during some 20,000 years the gainer is not capable of grasping all the material that comes from the donor. During this short lapse of time a significant fraction of the mass of the system is blown into interstellar space. The mass ratio bin [0.4-0.6] is now much better represented.

We present a set of evolutionary computations for binaries with a B-type primary at birth. Some liberal computations including loss of mass and angular momentum during binary evolution are added to an extensive grid of conservative calculations. Our computations are compared statistically to the observed distributions of orbital periods and mass ratios of Algols. Conservative Roche Lobe Over Flow (RLOF) reproduces the observed distribution of orbital periods decently but fails to explain the observed mass ratios in the range 0.4-1. In order to obtain a better fit the binaries have to lose a significant amount of matter, without transferring too much angular momentum.