OB/WR mass loss on Post-AGB/WD cooling tracks

[jmerritt1]

For stars in the MZAMS range of ~6-9Msol, the Dominant_Mass_Loss_Type variable is recording OB(main sequence) and sometimes WR mass loss along white dwarf cooling tracks. The actual mass loss is often small but nonzero after envelope ejection. The specific abnormal mass loss types vary as you change --mass-change-fraction, and --radial-change-fraction.

^Color indicates Dominant Mass Loss Rate, not Stellar Type.

DMLR Key:
{ MASS_LOSS_TYPE::NONE, "NONE" },
{ MASS_LOSS_TYPE::GB, "GB" },
{ MASS_LOSS_TYPE::LBV, "LBV" },
{ MASS_LOSS_TYPE::OB, "OB" },
{ MASS_LOSS_TYPE::RSG, "RSG" },
{ MASS_LOSS_TYPE::VMS, "VMS" },
{ MASS_LOSS_TYPE::WR, "WR" }

You may not see this behavior with SSE, since evolution is halted at WD formation. With default change fractions the winds go from GB to WR during the post-AGB, and do not dip into MS winds.

Label the issue

urgency_low - This issue is not urgent

severity_minor - This is a minor bug with minimal impact

To Reproduce
Steps to reproduce the behavior:
./COMPAS --mode BSE -n 1 --initial-mass-1 6.5 --initial-mass-2 0.1 --semi-major-axis 1e20 --detailed-output True --mass-change-fraction 0.005 --radial-change-fraction 0.005

Expected behavior
Not sure what to expect, perhaps NONE or the new ZERO would be most appropriate for WD, and we could set this to be the case for all remnants. For our purposes, a physically realistic way to track mass loss here is probably not important, at least not after the removal of the envelope. It has been difficult to track down why these types (OB, WR) are applied, but the WDs don't seem to have the appropriate Teff, etc. for those winds at those times either.

Also note that we've changed the variable name to MASS_LOSS_TYPE, but in the output files it is still recorded as DOMINANT_MASS_LOSS_RATE.

Versioning (please complete the following information):

• OS: Ventura 13.6.9
• COMPAS 3.01.05

[ilyamandel]

It also looks like you sometimes have OB winds applied during the blue loop -- is that intentional?

In general, if winds that shouldn't be applied at all are sometimes labeled dominant, it makes me wonder whether they are being applied more often than intended.

According to BaseStar::CalculateMassLossRateFlexible2023(), OB winds are applied all the time, not just to MS stars, as long as the stars are not in the LBV, RSG, cool star, or VMS regimes. So yes, they will be applied to WDs, and to any other stars not in the regimes above (and in the HURLEY_ADD model, to LBV stars, too). Is that the intended behaviour?

[ilyamandel]

@jmerritt1 , @SimonStevenson :
Is it the intended behaviour to apply OB winds to all stars that are not in the LBV, RSG, cool stars, or VMS regimes (e.g., WDs, but also any other stars that don't satisfy the other criteria, whether or not they would be classified as OB stars)?

[jmerritt1]

Thanks, I've opened a PR to set mass loss to 0 for remnants.

-Regarding blue loops, for the brief time spent above 12.5kK, OB mass loss does seem like the most appropriate treatment we can give, despite their He burning?

-For post AGB mass loss: we don't have a specific prescription, and envelope ejection is forced as per Hurley. It does affect the lifetime and cause a convergence issue depending on timestepping and change fraction settings(mass loss on the asymptotic part is truncated earlier). For example: given 6.5msol at ZAMS, final WD mass is 1.38msol using default settings:

but with --mass-change-fraction 0.005 --radial-change-fraction 0.005, the final mass is 1.23msol due to the additional time spent losing mass after the initial envelope ejection:

We may not want OB mass loss stacking with the normal AGB envelope loss, even if the impact is minimal?

[ilyamandel]

@jmerritt1 :
I would agree on both points: we probably want OB mass loss during the blue loop, but not as an additive mass loss on the AGB.
I presume the smaller timesteps on the AGB are bringing us closer to the desired, converged solution, right?