Recovering Intrinsic Pulsar Profiles and Scattering Parameters with a CLEAN-Based Algorithm for High-Precision Timing

TL;DR

This work evaluates a CLEAN-based deconvolution pipeline (CBADeS) for recovering intrinsic pulsar profiles and ISM scattering parameters from simulated PSRFITS data. By modeling the observed profile as a convolution of the intrinsic shape with a pulse broadening function and applying iterative CLEAN decomposition, the authors extract $\tau_{sc}$, $\alpha$, DM, and ToA residuals, validating performance across single and multi-component profiles. They compare multiple pulse broadening functions, implement FoMs to select optimal scattering times, and demonstrate how de-scattering improves timing precision, especially at higher S/N, while highlighting limitations at low S/N and ensuring robust PBF selection via Bayesian analysis. The study outlines potential extensions to FRBs, RRATs, and magnetar bursts, and discusses future improvements in PBF diversity, computational efficiency, and application to real datasets. Overall, CBADeS provides a model-aware, data-driven approach to mitigate scattering in pulsar timing and ISM studies, with implications for improving PTA timing precision and ISM turbulence diagnostics.

Abstract

In high precision pulsar timing, the accurate recovery of intrinsic pulsar profiles and their associated scattering parameters is of paramount importance. In this paper, we present a comprehensive study focused on the retrieval of intrinsic pulsar profiles through the utilization of a CLEAN-based algorithm as described in Bhat et al. (2003). The primary objective of this study is to elucidate the capabilities of our pipeline in the context of recovering the intrinsic profiles and associated parameters, such as dispersion measure, frequency scaling index, scattering time, pulse broadening function, and time of arrival residuals. We use simulated profiles to rigorously test and validate the efficiency of our recovery pipeline. These simulated profiles encompass single and multi-component Gaussians, designed to emulate the diverse nature of pulsar profiles. By comparing the recovered profiles and parameters to their injected values, as derived from simulations, we provide a robust evaluation of the pipeline's performance along with its drawbacks and limitations.

Figures (13)

• Figure 1: Plot of scattered, reconstructed and intrinsic profile along with summed up CCs. The arrangement of the CCs give the shape of the final reconstructed profile. Note that the injected intrinsic profile is shifted to the left for more visibility
• Figure 2: A plot of all FoMs. The red dashed line shows the best $\tau_{sc}$.
• Figure 3: Violin plot of the difference between injected and chosen $\tau_{sc}$ by each FoM. The units in the y axis are in terms of the step size of the Trial $\tau_{sc}$ as explained in Section \ref{['choosetau']}.
• Figure 4: The scattered, reconstructed and injected intrinsic profiles are plotted for 6 profile shapes. Note: The injected intrinsic profile (red) is shifted to the left to prevent overlap with the reconstructed profile.
• Figure 5: Top Panel:Obtained Alpha values of 120 archive files. Bottom Panel: obtained $\tau_{sc}$ values at the central frequency (400MHz).