Investigating time variability of X-ray absorption in Swift GRBs

TL;DR

This study performs a time-resolved analysis of 199 Swift GRBs to investigate changes in X-ray absorption, finding most absorption likely occurs on large scales and that early spectral analysis remains reliable.

Contribution

It introduces a Bayesian time-resolved spectral analysis of GRBs to assess absorption variability, providing new insights into the origin and stability of X-ray absorption.

Findings

Seven GRBs show decreasing column density, but alternative models are possible.

Most excess absorption originates on large scales of host galaxies or intergalactic medium.

Early Swift data reliably estimate total absorption without bias from variability.

Abstract

The existence of excess absorption in the X-ray spectra of GRBs is well known, but the primary location of the absorbing material is still uncertain. To gain more knowledge about this, we have performed a time-resolved analysis of the X-ray spectra of 199 GRBs observed by the \textit{Swift} X-ray telescope, searching for evidence of a decreasing column density ($N_{\mathrm{H,intr}}$) that would indicate that the GRBs are ionizing matter in their surroundings. We structured the analysis as Bayesian inference and used an absorbed power-law as our baseline model. We also explored alternative spectral models in cases where decreasing absorption was inferred. The analysis reveals seven GRBs that show signs of a decrease in $N_{\mathrm{H,intr}}$, but we note that alternative models for the spectral evolution cannot be ruled out. We conclude that the excess absorption in the vast majority of GRBs must originate on large scales of the host galaxies and/or in the intergalactic medium. Our results also imply that an evolving column density is unlikely to affect the spectral analysis of the early X-ray spectra of GRBs. In line with this, we show that estimating the total $N_{\mathrm{H,intr}}$ from early {\it Swift} data in Window Timing mode reveals the same increasing trend with redshift as previous results based on data taken at later times, but with tighter constraints.