IFMR_DWDs

Scripts for fitting an Initial-Final Mass-Relation (IFMR) for a set of double white dwarfs (DWDs) with measured Teffs and loggs. Based on the framework first introduced by Andrews et al. (2015), but extended to include analysis of outliers and estimation of systematic Teff/logg uncertainties

Components

• fit_IFMR.py: Fit a multi-segment piecewise linear IFMR to a set of DWDs. Hyper-parameters are the fraction of outlier systems, the extra dtau variance for outliers, and Teff/logg systematic terms.
• fit_IFMR_single.py: Fit a multi-segment piecewise linear IFMR to one DWD. Hyper-parameters are the two initial masses for the system.
• IFMR_config.py: Contains all required variables needed to fit the IFMR. Edit this file to fit the IFMR of your choice.
• plot_IFMR.py: plot outputs from an MCMC run. Use with -h or --help to list all available options.
• Mi_from_Mf.py: Given a white dwarf mass and it's uncertainty, calculate the initial mass and pre-WD lifetime and their uncertainties. Uses a Bayesian approach and takes into account uncertainty in the IFMR itself. Use with -h or --help to see all available options.
• data/DWDs_Teffs_loggs.pkl: A pickled dictionary for each of the 72 fitted DWDs in our sample. The keys are the obversation/system names. The values are an array of Teff/logg pairs, their covariance matrix, and pre-computed data to convert Teffs/loggs to masses and cooling ages (in particular the Jacobians to convert uncertainties), using data from Bedard et al. (2020). Note that WDJ1336-1620AB appears three times as "runA", "runD", and "runAD" for an averaged measurement. If you have your own observations, you can create your own pickled dictionary file, or provide the information directly to the routines/classes in DWDs.py.

Installation

Simply clone this repository to get all available scripts and data.

Dependencies

• python version >=3.9
• emcee (for running MCMC fits)
• corner (for plotting MCMC results)
• numpy, numba, matplotlib for maths and plotting

Data

Shortened versions (last 100 steps) of our MCMC chains are included in the repository. The full chains (10000 steps, up to 2GB each) can be provided upon reasonable request.