The Phantom Periodic Radio Burst: How a Pulsar Fooled CHIME Astronomers ✦
Context
Fast radio bursts (FRBs) remain one of the most intriguing mysteries of modern astrophysics: powerful millisecond pulses of extragalactic origin. Detecting periodicity in an FRB signal would be a real breakthrough, pointing to an exotic source—perhaps a neutron star in an extreme environment. That's why the 2022 announcement by the CHIME/FRB collaboration of the first periodic FRB with a 217 ms period grabbed so much attention. However, the subsequent check reminded us how critical calibration accuracy is for radio interferometers: an error in sky position can turn a harmless galactic object into a sensation. This story, from the detection of a mysterious signal to the retraction of the paper, illustrates the fine line between discovery and artifact in the era of big surveys, especially for pulsars. The complex interstellar medium, including cosmic dust, which can distort observational data, also plays a significant role in such misunderstandings.
Methods
The key to the solution came from two independent observations. First, the new CHIME/Slow backend recorded bursts from the known pulsar PSR J0248+6021, whose morphology exactly matched FRB 20191221A. Second, archival data revealed "twin" pulses arriving on the same day with nearly identical characteristics but at different declinations. Comparing daily calibration solutions revealed an anomaly: due to rain, several receiver elements failed, leading to the flagging of most of the upper frequency band (≳600 MHz) on the calibration day. The algorithm, lacking good data, used an outdated calibration, resulting in a phase gradient in the reconstructed beam patterns equivalent to a positioning error of ∼20°. Additionally, analysis of the dispersion measure (spectroscopic characteristics) confirmed that the signal matches a source inside the Milky Way, not an extragalactic object.
Results
Detailed modeling showed that the calibration error affected only a limited time window on December 21, 2019, and manifested as a frequency-dependent shift in the recovered positions of radio bursts. For FRB 20191221A, this resulted in a false localization 20 degrees south of the true position of the pulsar PSR J0248+6021. Remarkably, due to strong scattering in the dense HII (ionized hydrogen) region where the pulsar resides, its dispersion measure turned out anomalously high—four times higher than predicted for that direction—perfectly disguising the source as extragalactic. This pulsar, like all neutron stars, formed in a supernova explosion, which explains its rapid rotation. After applying the correct calibrations, the "duplicate" burst at the false declination vanished, and the true signal matched the pulsar's coordinates flawlessly. A check of all observing days revealed only one additional day (March 11, 2019) with similar anomalies but no registered FRBs; for the rest of the events in the Second CHIME/FRB Catalog, no such problems were found.
Implications
This case forces a reevaluation of the reliability of automated transient classification in large surveys. It shows that even meticulously calibrated instruments are susceptible to rare but catastrophic errors that can spawn false discoveries. For the FRB population, this means a statistical correction: one of the most intriguing periodic source candidates has been removed, narrowing the parameter space for theories predicting strict periodicity (e.g., from neutron stars in binary systems or magnetars). Moreover, the incident spurred the development of more robust calibration methods, particularly diagnostics for secondary peaks in the beam pattern, which now issue real-time warnings of potential glitches.
Future development
The CHIME/FRB experience is already being incorporated into new projects like CHIME/FRB Outriggers and the future Square Kilometre Array (SKA). The developed method of daily calibration monitoring by constructing a beam grid and searching for spurious maxima will become standard practice for survey-mode radio interferometers. In the future, machine learning may predict calibration degradation moments based on weather data and receiver telemetry. In parallel, deeper study of pulsar PSR J0248+6021 in the radio band will shed light on the mechanisms of its giant pulses and scattering in the turbulent environment of the ionized hydrogen region.
Impact
The results will influence the compilation of future fast radio burst catalogs, as well as strategies for identifying periodic signal candidates in experiments like ASKAP and MeerKAT. CHIME's lessons will benefit any multi-wavelength survey where astrometric precision is crucial.
Next steps
Immediate next steps include reanalyzing all archival CHIME/FRB data with the new diagnostic tool, and automatically applying multiple independent calibration solutions to events falling into risk zones. The collaboration also plans to publish updated reliability criteria for periodic sources.
Key open problems
This story ties directly into the classic problem of separating galactic and extragalactic transient populations—specifically, the long-standing question of whether fast radio bursts could have a galactic origin. It also highlights that uncertainties in models of free electron distribution in the Milky Way can mislead even with precise localizations, and that the search for exotic objects demands exceptionally careful handling of systematic effects. The work of Jocelyn Bell Burnell on pulsar discovery reminds us that great findings can also begin with noise, but in the era of terabyte data streams, the cost of error is multiplied many times over.
🎯 The pulsar PSR J0248+6021 is a real record-holder: its dispersion measure reaches nearly 400 pc/cm³, which is anomalously high for a direction far from the galactic plane. This happens because the line of sight passes through a dense region of ionized hydrogen—the very HII region that fatally disguised it as an extragalactic source.
Key numbers
- period: 217 ms
- declination error: ~20 degrees
- calibration failure frequency: ≳600 MHz
- dispersion measure: ~400 pc/cm³
- number of affected days: 2 out of several years