[mesa-users] Odd problem with semiconvective regions

Bill Paxton paxton at kitp.ucsb.edu
Thu Jan 22 17:01:57 EST 2015


On Jan 21, 2015, at 11:51 AM, Pablo Marchant wrote:

> Hi Bill, could you help me a bit with this odd issue I have?
> 
> I'm dealing with some massive binaries that accrete plenty of mass, which causes the superadiabatic core to increase in size. However, as there is a mu gradient due to the receding convective core before accretion, semiconvection is needed for rejuvenation to actually proceed. I'm running some systems to model this and I find that when the semiconvective region is just reduced to a few zones, MESA stops considering those as semiconvective mixing regions.
> 
> I attach a simple saved model + inlist to show this (MESA version 7385). The saved profile for model number 3002 (the saved model starts at 3001) shows that between zones ~920-940 there is a region that is ledoux stable and swarzschild unstable, so it should be marked as semiconvective. However most of it is marked as a zone with no mixing, and the mixing coefficient is zero there. Is there any odd control that should produce this result? I couldnt find any... Anyhow, I'm not that familiar on debugging the mlt module to check these things.
> 



This is a nice one, so I"m Cc'ing mesa-users.

The question is about why we don't get semiconvection in a region where we expect to see it.

Here's the mu gradient:
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mu.pdf
Type: application/pdf
Size: 25735 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150122/79299798/attachment.pdf>
-------------- next part --------------

Here's the plot of mixing coefficients showing a no-mixing gap where we'd expect to see semiconvection.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mixing_Ds.pdf
Type: application/pdf
Size: 27004 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150122/79299798/attachment-0001.pdf>
-------------- next part --------------

Note the abrupt cutoff in semiconvection -- turns out there is yet another knob in mesa for setting a min for D.


         !### remove_small_D_limit

         ! If MLT diffusion coeff D (cm^2/sec) is less than this limit,
         ! then set D to zero and change the point to `mixing_type == no_mixing`.

      remove_small_D_limit = 5
      

When I change that and rerun, we get something more like what we expected:

    remove_small_D_limit = 1d-6

-------------- next part --------------
A non-text attachment was scrubbed...
Name: mixing_Ds.pdf
Type: application/pdf
Size: 28877 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150122/79299798/attachment-0002.pdf>
-------------- next part --------------


However note that we've also widened the gap --- or at least it is wider in terms of cells.

It seems it might be a good idea for me to drop the default value for remove_small_D_limit -- wonder what that will cause to break?

Cheers,
Bill





More information about the Mesa-users mailing list