Close Binaries?

[Evil Dr Ganymede]

I know that in most cases, a white dwarf in a close enough orbit around a Giant is how you get a nova system. The dwarf pulls matter off the giant, which ignites when it reaches a critical density as it spirals in.

I think (if I'm correctly remembering what happens at Algol) that when you get two giants that are both bigger than their roche lobes, you still get mass transfer occuring that can actually significantly change the spectral type and size of the stars.

But what happens when you get a main sequence star orbiting a giant that's bigger than its roche lobe? Do you just get mass transfer again, or can you get nova systems?

[maaf]

I'm not an expert about close binary systems. But I belive that you would get something between a nova and a normal mass transfer between giants stars.

Supposing that the main sequence star is a star like our Sun, which has a critical scape velocity of nearly 600 km/s. The infalling gas from the giant star would crash into the G2V main sequence star surface at about 600 km/s, heating to millions Kelvins (20,000,000K if my computations aren't wrong), perhaps hot enough to cause some nuclear fusion, but since the density is very low this nuclear fusion would probably ocurr to a very little extent, not enough to cause a violent explosion (at least for the Sun). The problem is that we have to figure out precisely how much would the infalling gas be heated, since the speed of nuclear reactions is very sensitive to temperature changes, and 20,000,000K is precisely where the speed of hydrogen nuclear reactions increases fast with temperature. (At the sun's core, hydrogen is slowly converted to helium at a density of 148 g/cm^3 and a temperature around 16,000,000K. The mean time required for the hydrogen to react in these conditions is around 50% to 70% of the Sun's predicted lifetime (that is around 10 Gyr) ).
But if nuclear reactions goes at a first moment (in the very thin photosphere (density around 10^-7 g/cm^3), it wouldn't last long, since the outer layers of the main sequence star would soon expand (due to the heat released by the infalling gas from the giant star). So probably we would end up with a very luminous double (emmiting strongly in the UV and soft X-rays, during the mass transfers), but this probably wouldn't be strong enough to cause it to become a nova.
By the way, the critical escape speed of a giant star is much lower than a main sequence star, being around 20 - 100 km/s. Thus when the infalling gas from a giant star crashes in (the very undefined limits) of another giant star, it heat's up only to some tens of thousand Kelvins.

This is surely a very interesting subject that deserves more attention than given in this simple reply.

Astrophysical greetings,
M?rcio.

[maaf]

But what happens when you get a main sequence star orbiting a giant that's bigger than its roche lobe?

To be precise, the main sequence star doesn't orbit aroung a giant star, both orbit around their center of mass, since their masses are similar, unless the main sequence star is "orbiting" a supergiant much more massive. What makes a star giant or supergiant is not it's mass, but is it's radius and luminosity.

Masses of stars vary in a very narrow range, from 0,08 MSol to about 120 MSol. (Msol = 1,99*10^30 kg = mass of the Sun). But stellar radius can be as big as 50 RSol for a giant star and (guessing) 1,000 RSol for a supergiant.

Since volume is proportional to r^3 (radius power three), and a mass of a giant star is not too larger than a mass of a main sequence star (being even smaller in some cases), it happens that giant and supergiant stars are very fluffly, with a mean density usually much smaller than the density of the air we breathe.

Giant and supergiant stars are stars that evolved off the main sequence, being in their death throes.

[granthutchison]

But what happens when you get a main sequence star orbiting a giant that's bigger than its roche lobe? Do you just get mass transfer again, or can you get nova systems?

You get mass transfer, which significantly affects both stars. The main sequence star has its mass increased and therefore moves off the main sequence a little sooner than it would otherwise have done; it also becomes enriched with anomalous elements from the red giant. "Barium stars" like Alphard form in this way - barium formed in the helium-burning outer shell of the giant is dumped on to the surface of its less-evolved companion star, which otherwise wouldn't show a barium spectrum at that stage in its evolution.
The giant star of course loses mass in this process, and can partially abort the usual progression to a white dwarf - instead of burning almost all its He to C and O, the depleted giant can collapse with some He unchanged, forming a "helium dwarf".

Grant

[Evil Dr Ganymede]

Presumably the ultimate end of the mass transfer is if/when the masses are equalised? If so, what happens when you end up with two stars of equal mass that are both overflowing the roche lobes? And what happens if they want to get even bigger due to their evolution?

Is there also some kind of tidal interaction going on that pushes the stars apart as they exchange mass?

[granthutchison]

Presumably the ultimate end of the mass transfer is if/when the masses are equalised?

Feels right, but I don't know.

If so, what happens when you end up with two stars of equal mass that are both overflowing the roche lobes? And what happens if they want to get even bigger due to their evolution?

They merge with an increasingly thick "neck" at the inner Lagrange point - there's a contact binary in UMa that's the type specimen for this, but I can't recall its variable letters. Eventually if they get big enough they start to "leak" from an outer Langrange, like Beta Lyr.

Is there also some kind of tidal interaction going on that pushes the stars apart as they exchange mass?

Umm ... shouldn't be, should there? If they're that close they'll be synchronous already, and if the total mass of the system remains the same, the orbital period will remain the same, and so synchrony will be preserved, despite a shift in CofM.
But if they leak mass from an outer Lagrange then there will be orbital shifts.

Grant

[Evil Dr Ganymede]

Well, I did some hunting on the net and came up with some stuff about binary star evolution - apparently "contact binaries".

There's a potentially nifty program called Starlight Pro that shows how the stars change shape and merge when they grow bigger than their roche lobes, but it's somewhat flawed - it can get really slow, and once the stars become contact binaries the globes that are representing the stars can kinda 'implode'. And if you set any number to zero (when you're editing them) it crashes. But it gives you an idea anyway.

But anyway - I get the impression that when both stars are bigger than their roche lobes, you get a full-on contact binary, which is a single peanut-shaped envelope containing two cores. Then the stars actually start to spiral in because of friction between their envelopes, and may even merge if they live long enough.

The only other case of merging stars I've heard of is ER Vulpeculae binaries - at least, that's what an article in Astronomy magazine that I read sometime in the 1990s said. Though apparently that was due to the stars' magnetic field braking their orbits.

Thanks for reminding me about "leaking" from the other lagrange point... I found up some stuff on Beta Lyrae - this artist's impression shows it quite nicely - that's how you get the streamer coming out from the "back" of the pair, right? Apparently Beta Lyrae's orbital period is also slowing down too, presumably from the mass loss.

[granthutchison]

They merge with an increasingly thick "neck" at the inner Lagrange point - there's a contact binary in UMa that's the type specimen for this, but I can't recall its variable letters.

Ah, found it - they're W Ursae Majoris binaries.

Grant

[Evil Dr Ganymede]

Yeah, I forgot to say that I'd also found that out while I was looking for contact binary info. Rather fascinating stuff though.

[brunetto_64]

Hi at all!!!
the site where there is the software "starlight" is very interesting , there are many xls files and more....
http://www.cox-internet.com/ast305/binstar.html

it's good!!!

Brunetto