whee, we make stars go splat again!
what exactly does happen when stars collide,
just ordinary low mass main sequence stars
Glebbeek and Pols A&A 2008 - v 488 p 1007 and p 1017
BSE: binary stellar evolution code - Hurley et al 2002 from Tout et al 1997
- mass is wrong,
- lifetime is wrong,
- luminosity is wrong
If you are using the BSE prescription for merged stars.
Other than that it is pretty good...
We are adding two stars, M1 and M2 with M1+M2
assume near parabolic collisions with velocities at infinity small compared to stellar surface escape velocities;
get few percent mass loss during near parabolic collisions
you get some core mass and some envelope mass
core is ideally H free
in detail depending on state of the core of the stars hitting, the core may have
a compositon gradient rather than a composition discontinuity
the very low mass stars are convective, so cores don't have sharp composition discontinuities
famously, we expect collision product to layer itself in accordance with initial specific entropy, as a function of stellar radius - ie shocks are weak and have a low volume fraction so there is little entropy generation during the collision, relatively speaking
if there is H in the core, or the H free core is small enough, then there is core hydrogen burning over some remaining main sequence lifetime - at a rate appropriate to an evolved main sequence star of this mass
ie there is little or no mixing of the envelope and original partially burned cores
assuming the star becomes completely mixed on collision is generally a poor assumption.
So... is the opposite true? Or do we have to allow for some mixing of envelope into the core some of the time?
Can you for example get rotational mixing during oblique collisions and H enrich the core to any substantial extent.
Yes, that is a question.
Also, how evolved are the original cores - do we generally spend any significant time on the main sequence or jump directly to burnt out cores and shell burning tracks?
So we want an algebraic prescription for where we end up when we take two low mass main sequence stars with partially evolved cores, and collide them, building a new star layered in isentropic shells - so where on the main sequence does the resulting star comes out, and what is its subsequent evolution.
Voila.
For non-rotating head-on collision products, with no magnetic fields natch.
don't forget - luminosity depends on mean molecular weight also, like μ4 or so
ok, it makes a difference
(from Glebbeek and Pols '08)
PS: damn, that's annoying - I do a search for nice images of blue stragglers, and the highest ranked results are my own darned blog...
- Log in to post comments
Scientific American had an interesting and somewhat scary article on what would happen if a white dwarf collided with the sun. Dwarf essentially unscathed; sun ruined. This article later formed the basis of part of a Jack MacDivett science fiction novel.
Stars may just happen to be by-products of highly pinched plasmas. Given the average distance between stars in a galaxy, and the small relative motions driven more by electromagnetic fields than gravity, chances of a collision are remote at best. I cannot recall any papers regarding the actual collision of two stars, but I have just started reading such papers and don't know a lot yet. Are there any actual observations of stellar collisions? Direct, glancing, all sizes...any at all? Would be fascinating to read what happens.
A number of papers have been written on collisions of stars, some cited right here, some written by me.
There are places in the universe considerably more crowded than the galactic neighbourhood, and stellar collisions do occur interestingly often, we think.
The reason most of us estimate that gravity dominated dynamics on large scales is because electric charge comes in with both positive and negative signs, and net charge tends to be small on large objects - in fact capacitance is surprisingly small for large objects, and neutralization is rapid and complete for most plausible ways for building up any net charge.
Magnetic fields to play a role, both in formation and evolution of stars, and are intensively studied and observed. It is even possible for magnetic pressures to be comparable to or larger than equipartition in some circumstances, and so for electromagnetic forces to dominate.
You might want to read up on Frank Shu's classic discussion of the role of magnetic fields in star formation.
Stars themselves are mostly plasma, and they give off stellar winds which are also mostly plasma. I assume also that the majority of stars have internal magnetic fields. So inside astrospheres, electromagnetic forces may well dominate. But most of the galaxy is outside of astrospheres, and there you will find large fractions of neutral gas, which will ignore any magnetic field (but it tends to get photoionized if it gets too close to a star). So I, too, am confused by JAJ's comment that electromagnetic fields are the primary driver of relative stellar velocities.
In the case of the Sun (the one astrosphere for which we have in situ measurements), the interplanetary magnetic field strength is typically a few nanotesla at 1 AU, falling to ~100 pT near the termination shock. The latter value is several times stronger than the estimated strength of the galactic magnetic field in the local interstellar medium (~30 pT), so it is not clear to me how the galactic magnetic field would significantly affect the motion of the Sun. And as Steinn points out, large objects are quasineutral, so there would not be a whole lot of charge to work with even if the galactic magnetic field could have an effect.