Discovery of the linear energy dependence of the spectral lag of X-ray bursts from SGR J1935+2154

Spectral lag of the low-energy photons with respect to the high-energy ones is a common astrophysical phenomenon (such as gamma-ray bursts and the Crab Pulsar) and may serve as a key probe to the underlying radiation mechanism. However, spectral lag in keV range of the magnetar bursts has not been systematically studied yet. In this work, we perform a detailed spectral lag analysis with the Li et al.'s Cross-Correlation Function (Li-CCF) method for SGR J1935+2154 bursts observed by Insight-Hard X-ray Modulation Telescope (HXMT), Gravitational Wave High-energy Electromagnetic Counterpart All-sky Monitor (GECAM), and Fermi/Gamma-ray Burst Monitor (GBM) from 2014 July to 2022 January. We discover that the spectral lags of about 61 per cent (non-zero significance >1 sigma) bursts from SGR J1935+2154 are linearly dependent on the photon energy (E) with t(lag)(E) = alpha(E/keV) + C, which may be explained by a linear change of the temperature of the blackbody-emitting plasma with time. The distribution of the slope (alpha) approximately follows a Gaussian function with mean and standard deviation of 0.02 ms keV(-1) (i.e. high-energy photons arrive earlier) and 0.02 ms keV(-1), respectively. We also find that the distribution can be well fitted with three Gaussians with mean values of similar to-10.009, 0.013, and 0.039 ms keV(-1), which may correspond to different origins of the bursts. These spectral lag features may have important implications on the magnetar bursts.

Automated summary

The spectral lag of low-energy photons compared to high-energy photons is a common astrophysical phenomenon that can provide insights into the underlying radiation mechanism. In this study, the spectral lag of magnetar bursts from SGR J1935+2154 observed between July 2014 and January 2022 was systematically analyzed using the Li-CCF method. It was found that approximately 61% of the bursts showed a linear relationship between the spectral lag and photon energy, which can be attributed to a linear change in the temperature of the emitting plasma. The distribution of the slope followed a Gaussian function, with a mean value of 0.02 ms keV(-1) (indicating earlier arrival of high-energy photons) and a standard deviation of 0.02 ms keV(-1). Additionally, the distribution was well-fitted with three Gaussian functions, suggesting different origins of the bursts. These spectral lag features have important implications for understanding magnetar bursts.