# Measuring dust in core-collapse supernovae with a Bayesian approach to line profile modelling

## TL;DR

This paper introduces a Bayesian MCMC method to model dust-affected line profiles in supernovae, enabling precise estimation of dust properties and improving upon previous manual modeling approaches.

## Contribution

It applies an affine invariant MCMC sampler to the DAMOCLES code for rigorous multi-parameter exploration of supernova dust models, demonstrating enhanced parameter constraints.

## Key findings

- Most parameters are tightly constrained using the Bayesian approach.
- A strong dependence between grain size and dust mass is quantified.
- The method successfully re-analyzes SN 1987A data, improving parameter estimation.

## Abstract

Optical and near-IR (NIR) line profiles of many ageing core-collapse supernovae (CCSNe) exhibit an apparently asymmetric bluewards shift often attributed to greater extinction by internal dust of redshifted radiation emitted from the receding regions of the SN ejecta. The DAMOCLES Monte Carlo line radiative transfer code models the extent and shape of these dust-affected line profiles to determine the dust mass that has condensed, in addition to other properties of the dusty ejecta. I present here the application of an affine invariant Markov Chain Monte Carlo (MCMC) ensemble sampler (emcee) to the DAMOCLES code in order to investigate the multi-dimensional parameter space rigorously and characterise the posterior probability distribution. A likelihood function is formulated that handles both Monte Carlo and observational uncertainties. This Bayesian approach is applied to four simulated line profiles in order to test the method and investigate its efficacy. The majority of parameters can be tightly constrained using this method, and a strong (predictable) dependence between the grain size and the dust mass is quantified. The new approach is also applied to the H$\alpha$ line and [O I]6300,6363 \AA\ doublet of SN 1987A at 714d post-outburst, re-examining a previous 5-dimensional smooth model and also investigating a new, more complex, 10-dimensional model that treats both features simultaneously. The dust mass, dust grain size and a range of other parameters can be well constrained using this technique, representing a significant improvement over the previous manual approach.