Speaker
Description
Type-I X-ray bursts are thermonuclear explosions on the surfaces of accreting neutron stars (NSs). During these bursts, the rapid proton-capture process (rp-process) is thought to synthesize heavier nuclei and leave observable imprints on the burst light curves. While most X-ray bursters exhibit irregular burst recurrence times and profiles, only a few sources, known as “clocked” bursters, display highly regular bursting activity. These systems provide a unique opportunity to constrain the properties of accreting NSs and accreted material through detailed comparisons between observations and theoretical models.
We report the discovery of an anomalous bursting regime in the prototypical clocked burster GS 1826−238 using observations with the CubeSat X-ray observatory NinjaSat and NuSTAR. In May 2025, GS 1826−238 underwent a soft-to-hard state transition for the first time in a decade. During the subsequent hard state, NinjaSat observations revealed a new clocked-bursting phase with an unprecedentedly short recurrence time of $t_{\rm rec} = 1.603 ± 0.040\ {\rm hr}\ (1\sigma)$. Previous observations showed that the burst recurrence time decreases with increasing mass accretion rate, reaches a minimum value of ~3 hr, and then increases again. The observed recurrence time of 1.6 hr is therefore nearly a factor of two shorter than the previously observed minimum. In addition, the bursts exhibited reduced blackbody normalization and fluence compared with earlier clocked-bursting epochs. NuSTAR observations in February 2026 independently identified another short-recurrence phase with $t_{\rm rec} = 1.91±0.06\ \rm hr$, confirming that this behavior persists over multiple epochs.
The observed recurrence times are difficult to reconcile with the conventional picture in which bursts ignite uniformly over the NS surface. We propose instead that fuel accumulated over only a fraction of the stellar surface, resulting in an enhanced local accretion rate. This scenario naturally explains the unusually short recurrence times together with the reduced emitting area and burst fluence. Our results demonstrate that even for GS 1826−238, the benchmark source for observation–model comparisons of thermonuclear bursts, more realistic models incorporating multidimensional accretion may be required.
| Category | Observation |
|---|