[mesa-users] Unrealistic metallicity ([Fe/H] < -13) in the main-sequence evolution of a low-mass star

[Aaron Dotter]

Hi Earl,

What you observe in your calculation is normal and "correct" for the
physics you have assumed.  It's entirely possible for "diffusion" to reduce
the surface layer of a low-mass star to nearly pure H if the lifetime is
long enough and the surface convection zone is small enough.  That is the
case for your model.

A couple of things are missing:

1. Radiative levitation, which MESA will calculate but you need to use the
OP monochromatic opacities.  It is time consuming but may be worth it,
depending on your needs.

2. Some form of inhibited diffusion near the surface.  At least two options
for this in MESA.  I think they both have turbulent or turbulence in their
names.  Alternatively you could (I think) set the minimum temperature for
diffusion well inside the star.

If you want a better appreciation of these matters, I would
highly recommend papers by Georges Michaud and collaborators.

Aaron


On Sunday, January 24, 2016, Earl Bellinger <bellinger at mps.mpg.de> wrote:

> Hello everyone,
>
> I ran a track using MESA r7623 with
>     M = 0.921686
>     Y = 0.31379
>     Z = 0.0009
>     αMLT = 1.93045
> with an Eddington atmosphere, atomic diffusion enabled, a tiny mesh
> coef (0.2) and a tiny timestep (1d5 Gyr). Inlist attached.
>
> I calculate [Fe/H] as
>     log10(10**log_surf_z / surface_h1 / Z_div_X_solar)
> where
>     Z_div_X_solar = 0.02293.
> In this track, [Fe/H] is -1.24 at ZAMS and reaches a value of less
> than -13 at its main-sequence terminal age (Xc=0.001, age=2.45 Gyr).
> This is many orders of magnitude more metal-poor than any star ever
> observed. Is something going wrong in this evolution?
>
> This is not the only track where this happens. This behavior is common
> even in tracks with more metal-abundant initial conditions. For
> example, a [Fe/H] of -6.4 is achieved on the main sequence with
>     M = 1.273248
>     Y = 0.314728
>     Z = 0.014244
>     αMLT = 1.828888.
>
> The diffusion controls I use were taken from the solar-calibration
> inlist and are as follows:
>
>       do_element_diffusion = .true.
>       diffusion_use_isolve = .true.
>
>       diffusion_class_representative(1) = 'h1'
>       diffusion_class_representative(2) = 'he3'
>       diffusion_class_representative(3) = 'he4'
>       diffusion_class_representative(4) = 'c12'
>       diffusion_class_representative(5) = 'n14'
>       diffusion_class_representative(6) = 'o16'
>       diffusion_class_representative(7) = 'ne20'
>       diffusion_class_representative(8) = 'mg24'
>
>       diffusion_class_A_max(1) = 2
>       diffusion_class_A_max(2) = 3
>       diffusion_class_A_max(3) = 4
>       diffusion_class_A_max(4) = 12
>       diffusion_class_A_max(5) = 14
>       diffusion_class_A_max(6) = 16
>       diffusion_class_A_max(7) = 20
>       diffusion_class_A_max(8) = 10000
>
> I tried with and without diffusion_use_isolve, but it didn't have a
> major impact.
>
> Should I be using different diffusion settings to get more realistic
> evolutionary tracks?
>
> Thanks for any assistance,
> Earl Bellinger
> Stellar Ages & Galactic Evolution Group
> Max-Planck-Institut für Sonnensystemforschung
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20160124/cdd2aca1/attachment.html>