Redeveloping a CLEAN Deconvolution Algorithm for Scatter-broadened Radio Pulsar Signals

Abstract

Abstract Broadband radio waves emitted from pulsars are distorted and delayed as they propagate toward the Earth due to interactions with the free electrons that compose the interstellar medium (ISM), with lower radio frequencies being more impacted than higher frequencies. Multipath propagation in the ISM results in both later times of arrival for the lower frequencies and causes the observed pulse to arrive with a broadened tail described via the pulse broadening function. We employ the CLEAN deconvolution technique to recover the pulse broadening timescale and by proxy the intrinsic pulse shape. This work expands upon previous descriptions of CLEAN deconvolution used in pulse broadening analyses by parameterizing the efficacy on simulated data and developing a suite of tests to establish which of a set of figures of merit leads to an automatic and consistent determination of the scattering timescale and its uncertainty. We compare our algorithm to the cyclic spectroscopy method of estimating the scattering timescale, specifically to the simulations performed in Dolch et al. (2021). We test our improved algorithm on the highly scattered millisecond pulsar J1903+0327, showing the scattering timescale to change over years, consistent with estimates of the refractive timescale of the pulsar.