Formation pathway of Neutron Star - Helium White Dwarf binaries in latest dev.

[yuzhesong]

Describe the bug
With the recent changes to the code, the behaviour of NS-HeWD binaries have seemingly changed.

Previously, after experiencing nuclear timescale mass transfer, the MS companion of the NS becomes a Hertsprung Gap star, then first giant branch, then a HeWD. It stays as HG star and FGB star both for a substantial amount of time.

Right now, the MS companion becomes a HG star after experiencing nuclear timescale MT, then very quickly becomes a HeWD, skipping the FGB phase in between. Evolution of NS mass, companion mass, separation and stellar types of this binary is shown in the plot below:

There are two problems caused by this.

1. The code is now under-producing millisecond pulsars, as shown in the P-Pdot diagram below. You can re-produce this results with the current MSPDev branch code and the MSPDev branch code with commit 5c90129, by changing the --random-seed to 0 and --number-of-systems to 500,000

2. We previously already didn't produce NS-HeWD binaries with orbital period less than 10 days. With this change, we are now not producing any NS-HeWD binaries with orbital period less than 100 days. Plots below:

Versioning (please complete the following information):

• COMPAS v03.22.01 (shown in the plot as Current Dev)
• COMPAS in the current MSPDev branch (shown in the plot as Current MSPDev)
• COMPAS in MSPDev branch with commit 5c90129 (shown in the plot as Old MSPDev)

To Reproduce
To reproduce the system generated by V03.22.01, use this command:
COMPAS --evolve-pulsars TRUE --number-of-systems 1 --random-seed 72973 --pulsar-birth-magnetic-field-distribution-min 11.0 --pulsar-birth-magnetic-field-distribution-max 13.0 --pulsar-birth-spin-period-distribution-mean 75.0 --pulsar-birth-spin-period-distribution-sigma 25.0 --pulsar-magnetic-field-decay-timescale 1.0 --pulsar-magnetic-field-decay-massscale 0.02 --pulsar-minimum-magnetic-field 7.7 --pulsar-birth-magnetic-field-distribution UNIFORM --pulsar-birth-spin-period-distribution NORMAL --initial-mass-min 5.0 --initial-mass-max 30.0 --minimum-secondary-mass 0.1 --maximum-neutron-star-mass 2.3 --detailed-output True --eddington-accretion-factor 0.2

To reproduce the system generated by the MSPDev branch, use this command:
COMPAS --MSPs-from-AIC TRUE --evolve-pulsars TRUE --number-of-systems 1 --random-seed 72973 --msp-birth-magnetic-field-distribution-mean 8.5 --msp-birth-magnetic-field-distribution-sigma 0.25 --msp-birth-spin-period-distribution-mean 2.0 --msp-birth-spin-period-distribution-sigma 0.5 --pulsar-birth-magnetic-field-distribution-min 11.0 --pulsar-birth-magnetic-field-distribution-max 13.0 --pulsar-birth-spin-period-distribution-mean 75.0 --pulsar-birth-spin-period-distribution-sigma 25.0 --pulsar-magnetic-field-decay-timescale 1.0 --pulsar-magnetic-field-decay-massscale 0.02 --pulsar-minimum-magnetic-field 7.7 --msp-birth-magnetic-field-distribution LOGNORMAL --msp-birth-spin-period-distribution NORMAL --pulsar-birth-magnetic-field-distribution UNIFORM --pulsar-birth-spin-period-distribution NORMAL --logfile-definitions logfile_def.txt --output-path $pwd --initial-mass-min 5.0 --initial-mass-max 30.0 --minimum-secondary-mass 0.1 --maximum-neutron-star-mass 2.3 --neutron-star-accretion-efficiency-parameter 0.2 --detailed-output True --eddington-accretion-factor 0.2

Label the issue:

urgency_moderate - This is a moderately urgent issue

severity_moderate - This is a moderately severe bug

[ilyamandel]

Hi @yuzhesong , @SimonStevenson , all:

The failure to accurately reproduce observed populations is not, in itself, a code problem.

It may be a physics problem: maybe it points to our models of some physical processes not being right -- but not necessarily in the place suggested, e.g., it might be about supernova physics rather than mass transfer physics.

Or it may be pointing to a different formation channel: maybe this is telling you that all short-period NS+HeWD binaries form through triple evolution rather than isolated binary evolution. In that case, it would be a mistake to try to "fix" the problem away, since the mismatch between binary evolution models and observations is actually telling us something very important.

In any case, I don't see an evidence of a clear "bug" in this report, in the sense of code not doing what it's intended to do, and I wouldn't know what to fix. I'm therefore marking this as "will not fix" for now, and will close it after two weeks unless a code issue is clarified.

[ilyamandel]

@yuzhesong and @SimonStevenson -- 2 weeks have passed without comment, so closing this issue. Please feel free to re-open if there is a clear bug in the sense that the code is not doing what we intended it to do.