When mass transfer is taking place from one star to the other in a binary system, there is a considerable amount of gas flowing in the Roche lobe of the star that is receiving the mass. The splash as the gas stream strikes the star is probably sufficient to produce a flow of gas around the whole system as well. The gas streams can have unfortunate effects when we attempt to derive the orbital elements of the system. The gas stream can, for example, produce confusing spectral lines, whose Doppler shifts are unrelated to the orbital motion. Worse still, these lines may just blend in with the stellar spectral lines and shift them systematically. The main effect of this is to make the orbit look eccentric, whereas it may in fact be perfectly circular! In addition stellar light reflected from the gas or light generated by friction in the gas flow itself can interfere with the eclipse light curve of the binary star and induce spurious results. Even greater confusion will arise if the gas streams also give rise to additional eclipses.

We now have a confession to make: Algol is a semi-detached binary system. We should perhaps further confess that it is in fact a triple system with the two eclipsing stars, of orbital period 2.8 days, orbiting around a third body in a period of 1.8 years. The more massive star in the system, which is also the brighter star and the one eclipsed at primary minimum, lies well inside its Roche lobe, and is, therefore, almost spherical. Since this star contributes most of the light in the system, the light curve outside eclipse is almost flat even though the fainter companion is highly distorted. Therefore, strictly, to produce an Algol-type light curve it is only necessary for the brighter star to be undistorted. The fainter companion is less massive and fills its Roche lobe. The effect of the resulting gas flow is not very marked.

On the other hand, in the binary system (3 Lyrae, the brighter star is the less massive of the two and also overfills its Roche lobe. This star contributes most of the light in the system, and because it is so distorted gives rise to the characteristic ? Lyrae-type light curve. The effects of the gas streams in ? lyrae are quite considerable and caused much confusion for many years. It now seems clear that there is a disc of gas flowing around the fainter star, which is inside its Roche lobo, as well as streams from the brighter star to the disc through the inner Lagrangian point and from the disc back to the brighter star.

Perhaps the most spectacular example of gas flow in a binary system is seen in the dwarf novae. These are very close binary systems with orbital periods of a few hours; the prototype of them is U Geminorurn. This star system consists of a main-sequence star filling its Roche lobe and transferring mass to a heavier, but more compact, white dwarf companion. The gas flow does not strike the white dwarf directly but, rather, forms a disc around it. The inner parts of the disc orbit the white dwarf more rapidly than the outer parts. (This is similar to orbits in the Solar System, where just the inner planets have shorter years than the outer ones). The disc in a dwarf nova is therefore rotating differentially. Friction in the disc between the layers rotating at different speeds causes the whole disc to light up. Often the disc itself is brighter than the central white dwarf. In addition a lot of energy is released where the mass stream crashes into the outer part of the disc, producing a highly luminous hot spot. To an optical astronomer, this hot spot can be the brightest object in the system. That is, in some dwarf novae, more light comes from the hot spot than from either of the two stars! As you may imagine, an attempt to analyse the eclipse light curve of such a system by the classical methods described above can be somewhat frustrating.

Comments are closed.