[mesa-users] MESA for giant planet initial thermal evolution

[Sergei Nayakshin]

Dear Bill,

Thank you for this, the EOS coverage is very broad indeed. Perhaps Ana-Marie's problem is not with the EOS but with the fact that a pre-collapse planet may be hydro-dynamically unstable, or at least MESA encounters an unstable region while iterating for a solution. After all this is why Core Accretion instability (in planet formation) is an instability. 

Your advice on inflating the planet makes physical sense, but I doubt one can get to the H2-dominated pre-collapse planets this way. Pre-collapse giant planets are basically Laron (1969) first cores that however do not accrete gas at huge rates like proto-stars do, and instead cool radiatively until they reach T_c ~ 2500 K when H2 dissociates and then hydrodynamical collapse leads to a proto-star like initial condition. My understanding is that MESA uses the hydrostatic balance assumption explicitly, while the transition from T_c ~ 2x10^4 K planets to T_c ~ 2500 K planets (or in the opposite direction) is dynamical. 

I am actually experimenting with the "hottest start possible" giant planets, and I was able to find those by inflating pre-exisiting MESA's planets, as you suggest. There is however a maximum planet radius, or alternatively entropy, after which MESA finds no stable configuration. For 1 Jupiter mass planet R_max is about 15 R_Jupiter. This maximum R is not that much off found by Spiegel & Burrows 2012, fig. 3.

Don't get me wrong: I would not expect a stellar evolution code to handle this transition unless specifically designed to capture such hydro-transitions. I'm truly amazed by the robustness of the code and how user-friendly the documentation is. I'm sure I'll be a grateful user and a fun for years to come. (I have my own code to deal with the H2-dominated stage of planet formation, after which I can turn to MESA, I think).

Cheers,
Sergei.

On 28 Apr 2014, at 17:14, Bill Paxton wrote:

> 
> On Apr 27, 2014, at 11:09 PM, Sergei Nayakshin wrote:
> 
>> Hi Ana-Maria,
>> 
>> I'm new to MESA myself, but I also come in from the planet formation side, and I imagine that the problems you have are physical rather than numerical. If I understand it correctly you're trying to simulate the pre-collapse stage of planet formation, i.e., a planet so fluffy and cool that it is dominated by molecular H. The gas density at the Bondi radius would be of the order of the accretion disc density, which is probably 10^-10 g/cm^3, and temperature T ~ 100 K, although these numbers could vary with the separation of the planet from the star.
>> 
>> Perhaps MESA creators could comment on this further as I am out of my depth on how MESA would deal with this parameter space.
> 
> Hi Sergei,
> 
> Attached are plots show ingthe current mesa coverage for opacities and eos.   as you can see we blend together pieces from various sources to cover the large parameter space.   The numbers you mention are definitely pushing the limits.  I'd again suggest testing how far you can get with the current mesa by doing stellar engineering to re-inflate the model -- that will progressively take you farther into the lowT-low Rho region so you can see how far you get.   
> 
> And of course keep in mind that given interest in planet formation, the mesa input physics can be extended as necessary.
> 
> to get the Freeman opacities set kappa_lowT_prefix = 'lowT_Freedman11' in the &star_job inlist.
> the eos automatically includes SCVH.
> 
> -Bill
> 
> 
> 
> opacities
> <logK.pdf>
> 
> eos
> <blends.pdf>