[Mesa-users] Evolution of pop III stars up to core-collapse with rotation

Fri Apr 9 02:49:21 UTC 2021

hi falk,

if ondrea and you decide to share, mesa on zenodo is the place:

https://zenodo.org/communities/mesa/

fxt


> On Apr 8, 2021, at 5:08 PM, Ramandeep Gill via Mesa-users <mesa-users at lists.mesastar.org> wrote:
> 
> Dear Falk,
> 
> Thanks for your very detailed email. I'm still going over your recent
> work with Clarkson and I'll not claim that I fully understand all
> that's presented there at this point. However, I did notice that you
> were able to take some models to core-collapse. I can greatly benefit
> from using your setup for those models. Would you be willing to share
> your inlists?
> 
> Thanks,
> Ramandeep
> 
> On Wed, Apr 7, 2021 at 3:16 PM Falk Herwig <fherwig at uvic.ca> wrote:
>> 
>> Hi Ramandeep,
>> It seems you are running into a range of different types of problems. But Pop III stars pose some real physical challenges that higher-Z stars don’t. As stellar modellers we need to distinguish between numerical convergence problems (often because of user error setting input parameters incorrectly) when asking a simulation code to do an entirely reasonable thing, and complaints we get from a simulations code when we ask it to do the impossible. As you are surely aware there are many ways in which you can ask a 1D stellar evolution code to do the impossible, or to calculate something that is just a very bad approximation of what the star would actually do.
>> 
>> As discussed in great detail in https://ui.adsabs.harvard.edu/abs/2021MNRAS.500.2685C/abstract and also discovered and documented by many previous works, Pop III stars have due to their compactness very flat entropy profiles which make it almost impossible to avoid convective shell interactions at some point. Especially the H-He shell interactions have the potential to churn out a substantial fraction of the binding energy of the H-He shell layer which renders the 1D spherically symmetric stellar evolution approximation “indicative” in the best scenario. We encountered such conditions in the majority of our models, and in such case continuing the simulation (which we did not do) would probably best done by some severe stellar engineering intervention that tries to implement your best imagination as to what would happen in the 3D non-spherical, violent convective shell interaction and how your star looks like afterwards (or what is left of it). Without having access to your actual model output it is near-impossible to diagnose the exact mode of failure, but from our paper (link above) I would expect many (most) configurations to encounter such a near-death experience, either during He-core burning or after the end of He-core burning involving the He shell.
>> 
>> With regard to starting with previous inlists, that is always a good strategy. As you have encountered one problem is always that these inlists work best with the MESA version they originate from. We have developed specifically for this purpose (reproducibility of science) the utility NuDocker (https://github.com/NuGrid/NuDocker) which provides a docker environment (for Mac and Linux) to run previous MESA versions with ease. Docker images are already available and NuDocker has been tested to go back to version 4942. It is well documented and if you can run MESA you should be able to get NuDocker going in under and hour to run older MESA versions inside a Docker environment.
>> 
>> Another option to do a quick test with a previous MESA version is the “Public & Outreach” hub on https://astrohub.uvic.ca  which is tested to run several versions between 5329 and 12115 (see Install_Run_Analyze_MESA.ipynb notebook in home dir).
>> 
>> As for Kippenhahn diagrams etc as has been said before there is a multitude of tools around. What we use is NuGridPy. It can be pip installed, but I recommend the tagged version 0.8-beta1 https://github.com/NuGrid/NuGridPy/tags
>> You can try it out also on”Public & Outreach” hub  Astrohub, check Star_explore notebook.
>> 
>> 
>> All the best,
>> Falk
>> 
>> —
>> Dr Falk Herwig, Professor
>> Dept of Physics & Astronomy
>> University of Victoria
>> 
>> https://www.ppmstar.org
>> Email: fherwig at uvic.ca
>> Tel: +1 (250) 721-7743
>> Twitter: @fherwig
>> 
>> 
>> 
>> On Apr 5, 2021, at 4:14 PM, Ramandeep Gill via Mesa-users <mesa-users at lists.mesastar.org> wrote:
>> 
>> Hi Rob,
>> 
>> Thanks for your input on how to make the Kippenhahn diagrams in MESA.
>> It was very helpful.
>> 
>> I'm attaching below the results for the 60 M_sun model that was run
>> using v15140. All files are labeled with the MESA version used. To
>> remind you, this model crashes during core Carbon burn due to the
>> timestep becoming too small. The attached figures may provide more
>> insight into what might be going on.
>> 
>> I also took the time over the last few days to experiment with other
>> published inlists to see if they can be applied to my problem. I found
>> one prepared by Matteo Cantiello for the paper Perna, Lazzati,
>> Cantiello, 2018, ApJ, 859, 48. In that work, they evolve massive [40
>> M_sun] stars with Z = 1e-5 from zams to iron core formation using
>> MESA_v9793. First I tried the same inlist with the current MESA
>> version, it didn't work due to several changes to the code. I was able
>> to install v9793 successfully which I then used to reproduce Matteo's
>> results from that paper. In the next step, I tried to lower the
>> metallicity. I was able to get a solution for Z = 5e-7 until the
>> formation of the Iron core, which I also attach [notice the different
>> version v9793 in the attached figs].
>> 
>> The problem here was that with the adopted mass-loss prescription due
>> to radiative winds [Dutch with scaling factor 0.8], the star loses too
>> much mass [from 60 M to 21 M] even with such low Z. Other works that
>> evolve Z=0 stars find only minimal mass loss due to rotation. Lowering
>> the Dutch scaling factor or turning it off completely crashes in MESA.
>> 
>> If we can find a way to avoid very small time steps then some progress
>> can be made. Let me know your thoughts on this. Another point that's
>> worth considering is the mass loss prescription. Many Z=0 works use
>> the Kudritzki model which is not one of the options in MESA and so I
>> couldn't use it. Perhaps there's another way to implement it.
>> 
>> Best,
>> Ramandeep
>> 
>> On Thu, Apr 1, 2021 at 12:53 AM Rob Farmer <robert.j.farmer37 at gmail.com> wrote:
>> 
>> 
>> Hi,
>> There are python codes available for making kippenhans, but for now the simplest is to let mesa make them for you.
>> 
>> Add:
>> mixing_regions 40
>> burning_regions 40
>> to your history_columns file
>> 
>> and add:
>> kipp_win_flag=.true.
>> in your pgstar inlist
>> 
>> Rob
>> 
>> 
>> 
>> On Thu, 1 Apr 2021 at 01:13, Ramandeep Gill <rsgill.rg at gmail.com> wrote:
>> 
>> 
>> Hi Rob,
>> 
>> Thanks for writing back. Let me first answer your questions and then I
>> can provide some more details.
>> 
>> The model fails during Carbon ignition in the core. For your reference
>> I've attached two screenshots of the output just before the code
>> crashes. Fig.1 shows the Carbon fraction with C_cntr = 0.000131 and
>> this the lowest I've obtained with the current setup. The figure also
>> shows the current time step which is very small. The code crashes very
>> soon after this and that's shown in Fig. 2, just to give you some idea
>> of the error messages before it crashes. The time step is being
>> determined by max_increase, and I don't really know what that means.
>> 
>> I'll have to look into how to produce a Kippenhahn diagram. A quick
>> search on google revealed that many have written python scripts to do
>> just that. So, I might adopt one of them.
>> 
>> Finally, I looked at the MESA instrument paper II, which presents the
>> evolution of a 40 M_sun star with rotation for a low Z = 1e-5. I was
>> able to get the inlist for this paper from MESA marketplace. This
>> seemed like a perfect example for my case. I tried it with the most
>> recent version of MESA, and after fixing a few flags that have been
>> changed in the recent version, I was able to run it. However, it
>> crashed very quickly while core H ignition. Perhaps there's a way to
>> run this example with the latest version. Or I can try to install the
>> same version, if it's still available, that was used for that paper
>> and then run it.
>> 
>> Thanks,
>> Ramandeep
>> 
>> NB: In my original message I attached inlist_CC but it's missing some
>> flags that initialize the rotation. It's not the inlist that was used
>> to produce the attached HR diagram. So, I'm attaching the proper
>> inlist here. Sorry for the confusion.
>> 
>> 
>> On Wed, Mar 31, 2021 at 1:20 AM Rob Farmer <robert.j.farmer37 at gmail.com> wrote:
>> 
>> 
>> Hi,
>> 
>> The envelopes of massive stars are difficult things to get right. I would not worry (for now) so much about the envelope, instead worry about what the star is doing internally. How close to core collapse did your model get? Has it ignited carbon or later fuels? Try looking at a kippenhan diagram. If you look over the recent timesteps before the model crashes, what is limiting the timestep?
>> 
>> Rob
>> 
>> 
>> 
>> On Tue, 30 Mar 2021 at 22:13, Ramandeep Gill via Mesa-users <mesa-users at lists.mesastar.org> wrote:
>> 
>> 
>> Hello MESA experts,
>> 
>> I'm trying to obtain the pre-core-collapse structure of massive pop
>> III stars to study the launching and propagation of GRB jets in these
>> stars. In the end, I'm looking to obtain the density and angular
>> momentum radial profile of the star.
>> 
>> To this end, I've tried a few inlists, starting from the bare minimum
>> one used for introductory MESA tutorials as well as more advanced ones
>> from the MESA test suite and any that I could find online. The end
>> result in all is that after core He depletion, the code crashes most
>> of the time or, if I manage to somehow stabilize it, settles onto a
>> spurious solution with a lot of wild oscillations.
>> 
>> I attach an HR diagram from one of the runs for which I used the
>> inlist from the 'make_pre_ccsn_IIp' test suite example. The evolution
>> up to H core depletion, He core depletion, and that beyond He core
>> depletion is color coded. You can see that the last part of the
>> evolution results in spurious oscillations, after which the code many
>> times crashes or keeps chugging along at smaller and smaller time
>> steps.
>> 
>> To reproduce these results, I'm attaching two inlists. One produces
>> the evolution up to ZAMS and the second is designed to take it from
>> there up to the core collapse (CC) stage. I'm using version r15140 of
>> MESA.
>> 
>> At this point, I can use some help from MESA experts. What I'm looking
>> for is help in creating a basic inlist without any advanced effects
>> that can evolve the star to the core collapse stage. Advanced effects
>> can come later once I have a working solution for massive pop III
>> stars.
>> 
>> I was hoping to reproduce the results of Yoon, Dierks, & Langer (2012)
>> shown in their Fig. 1 & 2. They show the evolution only up to He core
>> depletion. In my case, I need to go beyond this to the core collapse
>> stage with and without rotation.
>> 
>> I'll really appreciate any help that you can offer. Please let me know
>> if you need any further info from me and forgive me if I haven't been
>> so clear in describing the problem.
>> 
>> Thanks,
>> Ramandeep
>> --
>> Ramandeep Gill
>> Postdoctoral Fellow
>> George Washington University
>> The Open University of Israel
>> http://www.ramandeepgill.com
>> _______________________________________________
>> mesa-users at lists.mesastar.org
>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>> 
>> 
>> 
>> --
>> Ramandeep Gill
>> Postdoctoral Fellow
>> George Washington University
>> The Open University of Israel
>> http://www.ramandeepgill.com
>> 
>> 
>> 
>> 
>> --
>> Ramandeep Gill
>> Postdoctoral Fellow
>> George Washington University
>> The Open University of Israel
>> http://www.ramandeepgill.com
>> <hr_002800_60Msun_v15140.png><conv_002800_60Msun_v15140.png><trho_profile_002800_60Msun_v15140.png><abund_002800_60Msun_v15140.png><conv_001450_60M_v9793.png><hr_001450_60M_v9793.png><trho_profile_001450_60M_v9793.png><abund_001450_60M_v9793.png>_______________________________________________
>> mesa-users at lists.mesastar.org
>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>> 
>> 
> 
> 
> -- 
> Ramandeep Gill
> Postdoctoral Fellow
> George Washington University
> The Open University of Israel
> http://www.ramandeepgill.com
> _______________________________________________
> mesa-users at lists.mesastar.org
> https://lists.mesastar.org/mailman/listinfo/mesa-users