COMMON ENVELOPE SIMULATIONS

Putting in the MESA equation of state

Work done

  1. Modified my own version of the code on bluehive to incorporate changes made by Yisheng and Jonathan
    1. Substituted Yisheng's /scratch/afrank_lab/EOS/astrobear/src/objects/tables.f90, src/physics/EOS.f90, src/physics/abundances.f90, src/hyperbolic/riemann_solvers.f90 and src/module_control.f90 (the only change in the latter is to comment out the Jean's length refinement criterion), src/Makefile (only differences is that it changes the order of compilation of EOS.o and tables.o to avoid a compilation error).
    2. Other files kept my own: important differences include src/particle/particle_control.f90, where I comment out creation of new particles. Certain other files are different between our versions but I decided to keep my own: src/objects/ambients.f90 (not used by me anyway), src/particle/particle_info_ops.f90 (apparently recent changes were made to the Bondi-Hoyle-Lyttleton accretion rate algorithm, but this does not affect simulations where subgrid accretion is turned off), src/data/data_info_ops.f90 (new lines of code, not sure of significance), src/distribution/distribution_control.f90 (Intent OUT had been changed to INOUT for pointer newgrids—-not sure of significance), src/hyperbolic/sweep/sweep_declarations.f90 (small differences, not sure whether significant).
  2. Compiled and ran the version of the code which was basically the same as Run 143 (fiducial RGB run) but with iEOS=6 in physics.data (tabular EoS) on BlueHive (110 cores, one frame of 2e4 seconds—same as for fiducial Run 143 which lasted 173 frames—takes about 24 hours). This run is called Run 207.
  3. Analyzed the AstroBear initial condition from Run 207 to see if it agrees with expectations.

Results

  1. Initial conditions
    1. The ​density profile from frame 0 of Run 207 is the same as what was inputted, as expected. Red=Run 143 (fiducial ideal gas EoS), Purple=Run 207 (MESA EoS).
    2. The ​internal energy density profile from frame 0 of Run 207 is different from that of Run 143, as expected.
    3. The internal energy density profile from frame 0 of Run 207 matches very well the internal energy profile directly obtained from MESA. See ​figure. Thus, the code passes this initial test!
    4. However, we see that there is an inversion of the internal energy density at large radius. This happens in the region where the pressure is almost constant (~1e5 dyne/cm2) and the density had decreased to quite low values (~7e-9 g/cc). Yisheng plotted the profile of specific internal energy (i.e. E_int per unit mass) in the rho-P plane, and found that the specific internal energy turns up at small density, which is consistent with what we are getting. See ​figure1 and ​figure2. We are not sure what causes this in the EoS, but it is not necessarily a problem. One could in principle get rid of it by using a lower ambient pressure (but this would require increasing the resolution near the stellar surface).
    5. ​mass fractions. The ratios of number densities of He and H and of metals and H are specified in physics.data. These are used to calculate the mass fractions X and Z and then the MESA EoS table with this X and Z are selected. I went over this calculation to make sure that everything is consistent. But what about ionization??
    6. ​free electron number and mean molecular weight per gas particle (ions + free electrons).
  2. Simulation results
    1. Density profile for new run at frame 1 (0.23 d): ​whole star with mesh and ​zoom-in on secondary and ​zoom-in on primary core.
    2. Same thing but for fiducial RGB run 143: ​whole star with mesh and ​zoom-in on secondary and ​zoom-in on primary core.

Next steps

  • Modify mu for ideal EoS since current choice of 2.21 seems too high. But we are mostly not in the ideal gas regime, so should not matter much if at all. Need to redo tables to reflect this change. Keep gamma=5/3 for ideal gas.
  • Plot radiation pressure and gas pressure to get sense of contribution of radiation pressure (it will be small, but worth plotting).
  • Better understand what is causing the difference between initial E_int profiles in MESA EoS and ideal gas EoS (latent recombination energy? radiation?)
  • Get thermodynamic variables like pressure and temperature to be outputted in chombos and plot them; also compare with ideal gas case.
  • Compute and output gamma
  • Re-read Nandez+Ivanova, Ohlmann+2017, Reichardt+2020
  • Putting recombination energy into our simulations

Plan

  • What is the plan for this paper?
    • Convection?
    • Compare fiducial RGB run with MESA EoS RGB and recombination run?
    • Improve fiducial RGB?
    • Compare fiducial AGB run with MESA EoS AGB and recombination run?
    • Which analysis to do?
      • Unbinding
      • Orbital evolution
      • Energy transfer
    • Do we try to compute opacities, optical depths, and diffusion times to assess whether recombination energy can be thermalized locally, or just assume that it can be?

Comments

No comments.