Runchao Chen · Jun Yang · Yihan Yin · Binbin Zhang
Nanjing University · Zhengzhou University · University of Hong Kong
Published in The Astrophysical Journal
ApJ 998, 289 (2026)
The research team of Professor Binbin Zhang at Nanjing University, in collaboration with researchers from Zhengzhou University and the University of Hong Kong, conducted a systematic search for millisecond-timescale periodic signals in a sample of 532 short gamma-ray bursts (sGRBs) recorded by the BATSE detector onboard the NASA Compton Gamma Ray Observatory (CGRO). In this "needle-in-a-haystack" systematic survey, the team discovered the second statistically significant millisecond-timescale periodic signal—a coherent oscillation at approximately 1100 Hz (corresponding to a period of ~0.9 milliseconds) in short gamma-ray burst GRB 960616, which occurred in 1996. Its physical properties are highly consistent with the millisecond-scale periodic signal first discovered in GRB 230307A, further supporting the physical picture that the central engines of short gamma-ray bursts harbor rapidly spinning millisecond magnetars.
The Most Violent Explosions in the Universe and Their Mysterious "Engine"
Gamma-ray bursts (GRBs) are the most violent high-energy explosive phenomena known in the universe. Within just a few seconds, a single gamma-ray burst can release energy exceeding hundreds or even thousands of times the total energy radiated by the Sun over its entire lifetime. Among them, "short" gamma-ray bursts lasting less than 2 seconds are generally believed to originate from the merger of two neutron stars—a catastrophic collision that is also an important source of gravitational waves, as manifested by the GW170817 event in 2017.
But what remains after two neutron stars collide? Such products serve as "central engines" of GRBs and may have two fates: either collapsing into a black hole, driving the explosion via an accretion disk jet; or surviving as an extremely rapidly rotating, ultra-strongly magnetized neutron star—a millisecond magnetar. A magnetar can spin thousands of times per second, with a magnetic field intensity 100 billion times stronger than Earth's magnetic field, making it one of the most extreme physical laboratories in the universe. If it truly exists, its rotation acts like a precise "cosmic clock," pulsating at millisecond intervals, and continuously channeling rotational energy into the jet via its intense magnetic field.
Can the beats of this "cosmic clock" traverse billions of light-years in the universe and leave imprints on human detectors? This is the central question of this study.
From a Single Case to a Sample: The Necessity of Systematic Search
In 2025, Professor Binbin Zhang's team for the first time detected a millisecond-scale periodic signal at approximately 909 Hz lasting about 160 milliseconds in GRB 230307A—the second brightest gamma-ray burst ever observed (Chen et al. 2025, Nature Astronomy). This was interpreted as a spin imprint left by a newborn millisecond magnetar in the jet, constituting the first direct observational evidence of a magnetar-driven gamma-ray burst, attracting widespread attention in the scientific community.
However, the discovery of a single event, however remarkable, always faces the statistical challenge of "coincidence." Is this an accident, or a pattern? If the signal truly originates from the spin of a millisecond magnetar, should similar millisecond-scale periodic signals also be found in a larger sample of short gamma-ray bursts? To answer this question, the research team turned to the short gamma-ray burst archive accumulated over nearly a decade—BATSE.
"Needle in a Haystack": Systematic Search in the BATSE Archive
BATSE (Burst and Transient Source Experiment) was mounted on NASA's Compton Gamma Ray Observatory (CGRO, operational 1991–2000), consisting of eight large-area detectors that accumulated high-quality data from thousands of gamma-ray bursts over nearly a decade. Its "Time-Tagged Event" (TTE) mode records the arrival time of every photon with an ultra-high time resolution of approximately 2 microseconds, providing ideal conditions for searching millisecond-scale periodic signals. To resolve kilohertz-level periodic signals, data time resolution better than 0.5 milliseconds is required, and BATSE's temporal precision far surpasses this threshold.
The research team extracted the public BATSE TTE data from 532 short gamma-ray bursts from the archive, conducting a systematic search while simultaneously considering the effects of both temporal and spectral dimensions: dividing each burst's data into sliding time windows of approximately 100 ms width and 50 ms step, scanning in the frequency range of 500 to 2500 Hz using Rayleigh tests, generating a total of 186,327 independent candidate frequency–power pairs.
"This is a genuine needle-in-a-haystack search. A blind search of high-time-resolution data at this scale is statistically extremely challenging—fluctuations of random noise can themselves produce seemingly significant peaks. Therefore, we must start from the overall statistical distribution of all trial results to reliably identify signals that truly exceed the random background, rather than being misled by statistical 'phantoms.'" — Runchao Chen (First Author, Ph.D. student, Nanjing University) |
A Pearl in the Ocean: The 1100 Hz Millisecond Pulsation in GRB 960616
After systematically surveying all 532 samples, the research team ultimately identified the only statistically significant millisecond-scale periodic signal, originating from the short gamma-ray burst GRB 960616 (BATSE trigger number 5502), which triggered BATSE on June 16, 1996. The main burst lasted approximately 30 ms, and throughout the entire main pulse, the light curve continuously displayed coherent oscillations at approximately 1100 Hz (period ~0.9 ms). This signal was most prominent in BATSE's highest energy channel (energy >325 keV), with a fractional modulation amplitude of approximately 47%—meaning that the gamma-ray radiation intensity varies periodically by nearly "half" its amplitude over time.
Statistically, the probability of this signal being spurious is extremely low. After correcting for all 186,327 independent frequency trials, the trial-corrected False Alarm Probability (FAP) is approximately 5.54×10−4; Monte Carlo simulations yield a chance probability of approximately 2.69×10−6, corresponding to a one-sided Gaussian significance of approximately 4.55σ, which constitutes a high-confidence detection in astronomy.
Figure 1: Weighted Wavelet Z-transform (WWZ) time-frequency power spectrum of GRB 960616 across four energy channels of the two BATSE detectors with the smallest photon incident angles (LAD0 and LAD1). The red-boxed region shows the ~1100 Hz periodic signal. The signal is temporally coincident with the main burst phase and significantly strengthens with increasing photon energy—reaching peak significance in the highest energy channel (>325 keV), clearly exhibiting an energy-dependent characteristic.
Multiple Verification: Making a Signal Indisputable
To ensure this signal is absolutely not a statistical "phantom," the research team verified it from four mutually independent angles, forming a rigorous "cross-validation network":
● Time-frequency consistency (WWZ dynamic power spectrum): The ~1100 Hz oscillation appears independently across multiple energy channels of multiple independent detectors, concentrated during the main burst phase, ruling out the possibility of instrumental noise from a single detector.
● Bayesian power spectrum analysis (Lorentzian QPO model): The best-fit central frequency from Bayesian power spectrum analysis (Lorentzian QPO model) is approximately 1096 Hz, with a quality factor Q ≈ 29.4, indicating the signal's frequency bandwidth is only about 1/29 of the central frequency, presenting a highly coherent narrow-peak feature. The logarithmic Bayes factor is approximately 64, strongly supporting the introduction of an oscillating component beyond the noise component.
● Phase folding (period-folded light curve): Folding all photons during the main burst at the 1100 Hz period yields a clear sinusoidal phase profile with fitted fractional amplitude A = 0.47 ± 0.04, highly consistent with the search results.
● Gaussian process regression (data-driven Monte Carlo): Joint modeling was performed under a Gaussian process framework assuming a periodic component, and the fitted aperiodic model was used to generate 186,000 simulated light curves without any periodic component but with intensity consistent with the observations. The results show that the observed Lomb–Scargle power cannot be reproduced in any of the simulations, indicating that a purely aperiodic process is extremely unlikely to produce the periodic signal observed in GRB 960616.
Figure 2: Phase-folded light curve of GRB 960616 at 1100 Hz. The black step line shows the folded photon count phase distribution; the red curve is the sinusoidal model fit (red shading: 1σ confidence interval). The folded result shows a standard sinusoidal modulation waveform with fractional amplitude A ≈ 47%, indicating a highly coherent periodic oscillation.
Physical Picture: How the Magnetar's "Spin Code" Is Imprinted in Gamma Rays
Why can the rotational beats of a millisecond magnetar leave imprints in the distant gamma-ray emission? The team's earlier work on GRB 230307A proposed a corresponding theoretical model that predicted the frequency scale, coherence, and energy dependence of this periodic modulation. The periodic signal observed in GRB 960616 is highly consistent with these model predictions across all key features.
In this framework, GRB 960616 was driven by a rapidly spinning newborn millisecond magnetar. An angle exists between its spin axis and magnetic axis (analogous to the offset between Earth's magnetic axis and rotation axis), giving the magnetic field driving the jet a directional character that varies regularly with the rotation period. The magnetic energy in the jet is dissipated through internal magnetic reconnection, randomly generating numerous "mini-jets" in various directions at different locations in the emission region—like small projectiles scattered in all directions within the jet interior, with those pointing toward Earth forming "hot spots" that are the primary sources of radiation we receive.
When the number of effective mini-jets is large enough, the contributions from various directions add up and are nearly symmetric, "averaging out" the spin modulation so no periodic signal emerges. However, if during the entire main pulse of a burst the effective mini-jets are intrinsically fewer and the distributional asymmetry persists throughout, the spin modulation can continuously appear with high amplitude—this is precisely the situation observed in GRB 960616. This also explains why the signal is strongest at high energies (>325 keV): high-energy photons are dominated by a very small number of mini-jets whose orientations most closely align with the line of sight, further weakening the "averaging" caused by the random distribution of mini-jets; low-energy photons aggregate contributions from a wider range of directions and are more easily "averaged out."
From Coincidence to Pattern: The Scientific Significance of the Second Discovery
Only one significant signal was identified among 532 BATSE short gamma-ray bursts, which itself conveys important physical information: millisecond-timescale periodicity is not a universal feature of short gamma-ray bursts; its detectability is constrained by multiple stringent conditions—including the relatively long post-merger survival and high spin rate of the magnetar, a magnetically dominated jet dissipation mechanism, and sufficient photon statistics.
Nevertheless, when GRB 960616 and GRB 230307A are examined side by side, the remarkable physical consistency between the two signals is far more profound than mere statistical counting:
● Extremely high coherence: Both signals present a nearly strictly periodic narrow-peak structure, with quality factors far exceeding the expectation of broadband random noise.
● Consistent timescales: The signal periods are both approximately 0.9–1.1 milliseconds, falling precisely within the theoretically expected range for the spin period of a newborn millisecond magnetar.
● Identical energy dependence: The modulation amplitude is most prominent at high energies and diminishes at low energies in both cases, fully consistent with the theoretical predictions of the ICMART jet model.
● Self-consistent physical picture: Both cases can be explained within the magnetically dominated jet framework; the difference lies only in the geometric details of the jet emission region.
This finding strongly suggests that in some bursts driven by double neutron star mergers, the rotational information of the central engine can indeed traverse the complex medium of a relativistic jet and leave observable modulation imprints in the gamma-ray light curve. The appearance of the second case pushes this possibility from an "interesting isolated event" toward "a physically backed statistical trend," and also provides a clear search target for future high-time-resolution gamma-ray observations.
Behind the Paper
"Finding the second millisecond periodic signal was one of the goals we most urgently hoped to achieve after discovering the signal in GRB 230307A. The discovery of a single event always faces the question of 'is it accidental,' while repeated verification in independent samples can fundamentally change our understanding of the physical nature of this phenomenon. BATSE provided one of the richest short gamma-ray burst archives in human history, and its microsecond-level time-resolution data brought us a rare historical opportunity. In the future, we will continue searching for similar signals in more luminous gamma-ray bursts, and deepen our theoretical understanding of how magnetar spin information is transmitted to jet radiation through numerical simulations." — Professor Binbin Zhang (Corresponding Author, Nanjing University) |
"The signal in GRB 960616 persisted throughout the entire main burst phase, unlike the situation in GRB 230307A where it appeared only within a brief window—this difference itself provides richer clues for the physics of jet radiation." — Runchao Chen (First Author, Ph.D. student,, Nanjing University) |
Funding Support
This research was supported by the National Natural Science Foundation of China (12573046, 12121003), the Ministry of Science and Technology Key R&D Program (2022YFF0711404, 2022SKA0130102), the National SKA Telescope Program (2022SKA0130100), the China Manned Space Engineering Program (CMS-CSST-2021-B11), the Fundamental Research Funds for the Central Universities, and the Jiangsu Province Innovation and Entrepreneurship Program.
References
1. R.-C. Chen et al. (2025). Evidence for a brief appearance of gamma-ray periodicity after a compact star merger. Nature Astronomy. https://doi.org/10.1038/s41550-025-02649-w
2. C. Chirenti et al. (2023). Kilohertz quasiperiodic oscillations in short gamma-ray bursts. Nature, 613, 253–256.
3. B. Zhang & H. Yan (2011). The ICMART model of gamma-ray bursts. The Astrophysical Journal, 726, 90.
4. W. S. Paciesas et al. (1999). The Fourth BATSE Gamma-Ray Burst Catalog. The Astrophysical Journal Supplement, 122, 465–495.
5. H. Sun et al. (2025). Magnetar emergence in a peculiar gamma-ray burst from a compact star merger. National Science Review, 12, nwae401.
Paper link: https://doi.org/10.3847/1538-4357/ae3a9 6