[Mesa-users] problems with trace elements diffusion

Tue Mar 22 21:22:17 UTC 2022

It might also help if you could describe more precisely what your goal is for including u238 diffusion in a MESA model. It’s not clear to me that MESA is the right tool for simulating the type of diffusion described in the Horowitz & Caplan paper. My understanding is that they discuss crystals growing until they reach a point of criticality while they are still very small (a few mg), so MESA won’t resolve that sort of crystal nucleation and diffusive growth. It’s essentially sub-grid physics from the point of view of a 1D stellar evolution model.

Cheers,
Evan

> On Mar 22, 2022, at 5:05 PM, Evan Bauer <evan.bauer.astro at gmail.com> wrote:
> 
> Hi Ana,
> 
> I’d be happy to take a closer look if you can send the exact set of inlists that produced the flat abundance profile. From the diffusion velocities reported in the profile, it looks like the diffusion timescale in the interior is on the order of 10^10 years, while your model is less than 10^8 years old. So it isn’t necessarily surprising that the abundance profile would stay mostly flat over the timescale that your model has run, but it shouldn’t stay precisely constant if diffusion is happening. I think I’ll have to run the model to say anything more specific.
> 
> Cheers,
> Evan
> 
>> On Mar 21, 2022, at 2:40 PM, Ana Antonini <anacarolina.srantonini at gmail.com <mailto:anacarolina.srantonini at gmail.com>> wrote:
>> 
>> Hi Evan, Hi Frank, 
>> sorry it took me so long to get back to you!
>> 
>> Regarding Evan's questions:
>>  
>>  We were looking at the abundance profile to monitor diffusion, so by "diffusion isn't happening" I meant  "the abundance profile is a flat horizontal line". By adding edv and diffusion_D to the profile_columns list, MESA reports that for both the uranium isotopes, the diffusion velocity is not zero. I'm attaching one of the final profile.data files for further detail. 
>> 
>> On Franks suggestions:
>> Increasing the mass fraction for both the isotopes does not cause diffusion to happen. 
>> 
>> Now, for the "adding Ne22" suggestions things get a little weird.   
>> When adding Ne22 to the models, diffusion happens not only for the Ne22 but also for U238 (U235 doesn't diffuse though). The range of diffusion velocities for U238 and U235 remains pretty similar to the case with no Ne22. I'm also attaching one of the final profile.data for the case with Ne22  
>> 
>> Also, we accidentally discovered that if Ne22 is present in the add_isos list U238 will diffuse (even if de initial abundance of Ne22 is zero and it isn't produced in any of our reactions).
>> 
>> Once again, any insights on what might be happening and why adding Ne22 helps with U238 diffusion (but not U235) would be deeply appreciated. 
>>  
>> Kind Regards, 
>> Ana Antonini
>> 
>> Evan Bauer <evan.bauer.astro at gmail.com <mailto:evan.bauer.astro at gmail.com>> escreveu no dia segunda, 21/02/2022 à(s) 18:31:
>> Hi Ana,
>> 
>> Just to follow up on Frank’s question, what exactly does it mean that diffusion is not happening? Is the diffusion velocity reported by MESA exactly 0? I.e. what gets reported in your profile output if you add the following to your profile_columns.list file?
>> 
>> 	edv u238
>> 	diffusion_D u238
>> 
>> I think MESA should be able to do diffusion in this regime, but the diffusion velocity for an element with such a high charge might be very slow, so it will be helpful to know the exact numbers that MESA is reporting.
>> 
>> Cheers,
>> Evan
>> 
>>> On Feb 19, 2022, at 8:29 AM, Ana Antonini via Mesa-users <mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>> wrote:
>>> 
>>> Hi Frank, 
>>> 
>>> That's not something we've thought of trying before, so I'll have to test it out and get back to you. 
>>> But regarding the last question, yeah, models are running long enough. 
>>> 
>>>  Thanks for the suggestions!
>>> 
>>> A. Antonini
>>> 
>>> Francis Timmes <fxt44 at mac.com <mailto:fxt44 at mac.com>> escreveu no dia quinta, 17/02/2022 à(s) 22:25:
>>> hi ana,
>>> 
>>> some suggestions sparked by this one:
>>> 
>>> > Diffusion is definitely happening, just not for the uranium.
>>> 
>>> if ne22 is added to the composition in the file "u238",
>>> then neon diffuses as expected but uranium does not?
>>> 
>>> does increasing the uranium mass fractions in the file "u238"
>>> by a factor of, say, 1e6 cause uranium to diffuse as expected?
>>> 
>>> are the models being run long enough relative to uranium's diffusion timescale?
>>> 
>>> fxt
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> > On Feb 17, 2022, at 11:58 AM, Ana Antonini via Mesa-users <mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>> wrote:
>>> > 
>>> > Hi Morgan, 
>>> > 
>>> > we set "do_element_diffusion" to true on our run_star_extras file so that it only starts once the star is in the wd cooling track. 
>>> > Diffusion is definitely happening, just not for the uranium.
>>> >  
>>> > thanks for taking the time to try to help though! 
>>> > 
>>> > Morgan Taylor <taylormorgan32 at gmail.com <mailto:taylormorgan32 at gmail.com>> escreveu no dia quinta, 17/02/2022 à(s) 15:23:
>>> > Hi Ana,
>>> > 
>>> > I noticed that !do_element_diffusion =.true. is commented out in your inlist.
>>> > In r-15140, the default is set to false.  Does your inlist_extra file set it to true?
>>> > If not, diffusion won't be turned on, and your other diffusion controls will be ignored.
>>> > 
>>> > Hope this helps,
>>> > 
>>> > Morgan
>>> > 
>>> > On Thu, Feb 17, 2022 at 10:53 AM Ana Antonini via Mesa-users <mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>> wrote:
>>> > 
>>> > Dear Mesa users, 
>>> > 
>>> > We've been working on a project using MESA version r-15140 and have run into a problem we can't seem to solve.
>>> > 
>>> > We've been trying to investigate if the results from Horowitz & Caplan (2021) can be reproduced with stellar evolution models, but despite trying a broad combination of controls and parameters, we can't get our uranium isotopes to actually diffuse. 
>>> > Some of the things we tried: 
>>> >       •  setting a more rigorous tolerance for "diffusion_X_tot_atol" and "diffusion_X_tot_rtol"
>>> >       • forcing diffusion by setting "diffusion_class_factor" higher for the uranium species
>>> >       • limiting the timestep by using "diffusion_dt_div_timescale"
>>> >       • adding the timestep controls used in the wd_diffusion test_suite to our work 
>>> >       • both using and not using the full net 
>>> >       • setting the class_factor to zero for all classes except uranium  
>>> > I'm attaching our inlist as well as other pertinent files and would greatly appreciate any ideas you might offer or even just a confirmation that  MESA is indeed capable of doing diffusion for species with such a low abundance. 
>>> > 
>>> > Kind  Regards, 
>>> > A. Antonini 
>>> > _______________________________________________
>>> > mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>
>>> > https://lists.mesastar.org/mailman/listinfo/mesa-users <https://lists.mesastar.org/mailman/listinfo/mesa-users>
>>> > 
>>> > _______________________________________________
>>> > mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>
>>> > https://lists.mesastar.org/mailman/listinfo/mesa-users <https://lists.mesastar.org/mailman/listinfo/mesa-users>
>>> > 
>>> 
>>> _______________________________________________
>>> mesa-users at lists.mesastar.org <mailto:mesa-users at lists.mesastar.org>
>>> https://lists.mesastar.org/mailman/listinfo/mesa-users <https://lists.mesastar.org/mailman/listinfo/mesa-users>
>>> 
>> 
>> <profiles.zip>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20220322/3abdb5d8/attachment.htm>