An emission-state-switching radio transient with a 54-minute period

The study investigates the nature of a newly discovered long-period radio transient, ASKAP J1935+2148, exhibiting a 53.8-minute period and distinct emission states. Long-period radio transients are rare and poorly understood; proposed origins include magnetic white dwarfs and magnetars, but no consensus exists. The source was first detected serendipitously during ASKAP observations of GRB 221009A near the magnetar SGR 1935+2154. The research aims to characterize the source’s emission properties (periodicity, pulse morphology, polarization, spectrum), assess intermittency and mode switching, measure dispersion and rotation measures, and evaluate physical models (white dwarf versus neutron star/magnetar) capable of producing the observed coherent radio emission. The work is significant because the source lies in the pulsar “death valley,” challenging standard radio emission theories, and because it displays emission-state switching reminiscent of neutron-star pulsars, potentially illuminating the physics of ultralong-period magnetars and coherent emission mechanisms.

The paper situates ASKAP J1935+2148 among known long-period radio transients such as GLEAM-X J162759.5−523504.3 and GPM J1839−10, and the Galactic Centre Radio Transient GCRT J1745−3009. GLEAM-X J1627 showed activity for only three months, whereas GPM J1839−10 has persisted for decades. GCRT J1745−3009 exhibited 77-minute periodicity with later episodes of narrower, weaker pulses and variable circular polarization. These comparisons underscore that long-period sources can show diverse and evolving emission states. Prior models invoke magnetic white dwarfs (isolated or in interacting binaries) and magnetars. However, the radio luminosities of cataclysmic variables (10^21–10^25 erg s−1) are far below those observed here, and coherent radio emission from isolated magnetic white dwarfs has not been detected to date. Population studies suggest only a limited number of neutron-star long-period radio emitters are expected, while a larger white-dwarf population is possible, yet explaining coherent emission remains challenging in both scenarios. The work also references pulsar mode-switching behavior (e.g., PSR J1107−5907, PSR B0823+26, PSR B2111+46) as an analogue for the observed emission-state switching.

Observations and detection:

• Discovery with ASKAP during a target-of-opportunity observation (12–23 October 2022) near SGR 1935+2154. A fast-imaging pulse-detection pipeline subtracted visibilities between adjacent 10 s integrations, imaged residuals for transient detection across 36 beams, and flagged peaks above noise. Follow-up used deeper model subtraction, phase shifting to the source, and extraction of dynamic spectra (>200 m baselines) to improve sensitivity and analyze polarization.
• Source localization: ASKAP initial localization (~1.5″). Precise MeerKAT localization (RA 19h35m05.175s ± 0.3″, Dec +21°48′41.504″ ± 0.6″).

MeerKAT follow-up (L band, c856M4k configuration):

• Simultaneous beamformed and imaging observations at 1,284 MHz with 856 MHz bandwidth; 2 s imaging time resolution and beamformed data with 38.28 μs sampling via PTUSE; real-time searches with MeerTRAP/TUSE at 306.24 μs sample and up to 140 ms boxcar widths across DM 0–5118 pc cm−3.
• Calibration: PKS J1939–6342 (flux/bandpass), PKS J2011–0644 (phase). FBFUSE formed incoherent and up to 780 coherent beams tiled within the primary beam. Beamformed total-intensity data recorded; full polarization on-source recorded with PTUSE (psrfits).
• Post-processing: Used APSUSE off-source beam for baseline subtraction of PTUSE on-source data; performed single-pulse search with TransientX over DM 120–160 pc cm−3 and max width 1 s, yielding two pulses. Extracted polarization profiles and dynamic spectra; corrected for RM and DM.

Timing and period estimation:

• Constructed times-of-arrival (ToAs) from ASKAP (midpoint of 10 s integrations) and MeerKAT (PSRCHIVE PAT on beamformed peaks). TOA uncertainties accounted for pulse width (HWHM) and 10 s non-detections bracketing bursts. Used TEMPO2 with JPL DE436 ephemeris to fit a phase-connected solution for period P and constrain Ṗ without fitting position or DM (well determined by imaging and S/N optimization).

Multiwavelength and archival searches:

• Radio: VLITE (340 MHz) datasets (2017–2023; A and B configurations) formed 4–10 s snapshot time series after continuum subtraction and phase shifting; no pulses detected. VLA P-band (∼325 MHz) in 2013–2014 and GMRT/VLA L-band near SGR 1935+2154 (2020–2021) with 5–5.4 s snapshots; no pulses detected.
• Persistent radio imaging: Stacked ASKAP deep image at 887.5 MHz showed no unpulsed source above 3σ (RMS 25 μJy/beam); background subtraction uncertain due to proximity to an SNR edge.
• X-ray: Chandra (five observations totaling 157.7 ks; CIAO 4.15.1, CALDB 4.10.2) reprocessing, reprojection, and stacking; 3σ upper limit of 6.3×10^−5 counts s^−1 (0.3–10 keV). Using NH ≈ 4×10^21 cm^−2 and spectral models (kT=0.3 keV blackbody or Γ=2 power law), unabsorbed flux limits F_BB < 1.3×10^−15 and F_PL < 1.7×10^−15 erg s^−1 cm^−2, implying L_X ≈ 4×10^30 erg s^−1 at 4.85 kpc. Swift/XRT (291 visits; 302.4 ks total) showed no flares; 3σ count-rate limit 1.7×10^−4 s^−1 and less constraining flux limits.
• Optical/NIR: VLT HAWK-I JHKs imaging (300 s each) found a source within 1.5″ (J=18.4±0.1, H=17.3±0.1, Ks=17.1±0.1); Keck/LRIS 3200–10,000 Å spectrum showed featureless red continuum consistent with an L/T dwarf; likely foreground (chance alignment ~5%). If unassociated, derived limits J>21.4, H>20.5, Ks>19.8. PS1 DR2: z=22.0±0.3, y=20.4±0.1; no g,r,i detections.

Polarization and propagation analyses:

• ASKAP pulses: >90% linear polarization, <3% circular; RM ≈ +159.3±0.3 rad m^−2 via RM synthesis; consistent with Galactic foreground and nearby pulsars; no large local RM contribution.
• MeerKAT pulses: >70% circular, ~40% linear; RM consistent (+159.8±0.3 rad m^−2). Tested Faraday conversion using phenomenological Poincaré-sphere rotations; found no frequency dependence indicative of conversion, suggesting intrinsic circular polarization or complex mode addition.

Model constraints:

• Estimated distance from DM (NE2001 and YMW16): 4.3–5.4 kpc; adopted average 4.85 kpc for luminosity estimates.
• Assessed white-dwarf and neutron-star scenarios using period, Ṗ constraints, radio luminosity, and radius constraints derived from death-line physics and pair-cascade requirements. Evaluated cataclysmic-variable radio luminosities and ruled them out for this source’s luminosity.
• Compared location in P–Ṗ diagram with death lines and other long-period sources; ASKAP J1935+2148 lies in the pulsar death valley.

• Discovery of ASKAP J1935+2148, a long-period radio transient with P ≈ 3,225.31 s (≈53.8 min) and a phase-connected timing solution; upper limit on period derivative Ṗ ≤ (1.2±1.5)×10^−11 s s^−1.
• Three emission states observed: • Strong pulse mode (ASKAP, 887.5 MHz): 15 bright pulses with widths 10–50 s, >90% linear polarization, <3% circular, in-band spectral index α=+0.4±0.3; brightest peak flux density 234.7 mJy; duty cycle ~1.5% in imaging. • Weak pulse mode (MeerKAT, 1,284 MHz): two pulses ~370 ms wide with >70% circular and ~40% linear polarization, spectral index α=−1.2±0.1; 2 s image peak ~9 mJy (≈26× fainter than brightest ASKAP pulse); duty cycle ~0.01%; pre- and post-cursor components detected in one epoch. • Quenched/null state: non-detections in multiple epochs; source estimated quenched ~40–50% of the time across ASKAP and MeerKAT campaigns.
• RM consistently measured at ≈ +159.3 to +159.8 rad m^−2; consistent with Galactic foreground, indicating no strong local RM contribution.
• DM = 145.8±3.5 pc cm^−3 (MeerKAT), implying a distance of 4.3–5.4 kpc (YMW16/NE2001), average ~4.85 kpc.
• Intermittency and pulse arrivals: MeerKAT pulses arrive within ~319 ms of predicted phase (in phase with ASKAP pulses), confirming a common periodic origin.
• X-ray non-detections: Chandra stacked limit F_BB < 1.3×10^−15 and F_PL < 1.7×10^−15 erg s^−1 cm^−2 (0.3–10 keV), corresponding to L_X ≈ 4×10^30 erg s^−1 at 4.85 kpc; Swift/XRT shows no flares across 2010–2022.
• No persistent radio continuum detected in ASKAP deep stack (RMS 25 μJy/beam), though background subtraction is uncertain due to proximity to an SNR edge.
• Archival radio searches (VLITE 340 MHz, VLA P-band 325 MHz, GMRT/VLA L-band): no pulses detected.
• Model implications: Based on period and compactness constraints from pair-production death-line physics, an isolated magnetic white-dwarf origin is ruled out (minimum radius estimates incompatible with white-dwarf sizes even for conservative curvature parameters). Cataclysmic variables are too radio-faint by many orders of magnitude. A neutron-star origin (likely an ultralong-period magnetar or related neutron star) is favored; radio luminosity exceeds spin-down power, implying alternative (magnetically powered) emission processes.
• Source lies in the pulsar death valley on the P–Ṗ diagram, challenging standard spin-down powered radio emission models and indicating unusual magnetospheric conditions and strong beaming (~1% duty cycle).

The findings demonstrate that ASKAP J1935+2148 is a periodically emitting, highly polarized coherent radio source with distinct emission states that evolve over months and occur at consistent rotational phases. The strong linearly polarized tens-of-seconds pulses contrasted with the weak, highly circularly polarized ~370 ms pulses point to different coherent emission regimes or magnetospheric states. The consistent RM and phase alignment across telescopes confirm a single source producing both states. The source’s location in the pulsar death valley and radio luminosity exceeding spin-down power argue against canonical pulsar emission via pair cascades sustained by rotational energy alone. Radius constraints for coherent emission, given the observed period, disfavor an isolated magnetic white dwarf; cataclysmic variables lack sufficient radio luminosity. In contrast, a neutron-star scenario—potentially an ultralong-period magnetar—can accommodate the coherent emission and extreme beaming, with radio power drawn from magnetic energy (e.g., reconnection, twisted fields). The mode-switching behavior, polarization changes, and intermittency closely resemble those seen in intermittent/mode-switching pulsars (e.g., PSR J1107−5907), supporting a neutron-star-like magnetospheric origin. The absence of bright X-ray emission is consistent with some long-period radio transients and lower-luminosity or older magnetar subclasses. Overall, the results address the research question by identifying the likely neutron-star nature of the source, constraining physical models, and highlighting the need for non-standard emission mechanisms.

The paper reports the discovery and characterization of ASKAP J1935+2148, a 54-minute period radio transient exhibiting three distinct emission states with markedly different polarization and temporal properties. A phase-connected timing solution, polarization measurements, spectral indices, and multiwavelength limits collectively indicate coherent emission and disfavour an isolated magnetic white-dwarf origin based on compactness constraints. The data support a neutron-star origin, likely an ultralong-period magnetar or related object with magnetically powered radio emission. The source’s mode switching, intermittency, and position in the pulsar death valley challenge conventional pulsar emission theories and expand the parameter space of neutron-star radio emitters. Future work should include continued monitoring to search for additional periodicities, long-term evolution of emission states, and signs of a possible companion. Deeper, simultaneous multi-frequency and multiwavelength campaigns (including sensitive X-ray observations) could further constrain emission mechanisms and environment, while population studies will refine expectations for the prevalence of such ultralong-period neutron stars.

• ASKAP imaging time resolution (10 s) prevents resolving subsecond structures within the wide pulses and precludes direct DM estimation from ASKAP data.
• MeerKAT beamformed data are polarization-calibrated but not flux-calibrated; flux densities are taken from 2 s imaging, potentially missing peak brightness on shorter timescales.
• The source lies on the edge of a supernova remnant; uncertain background subtraction limits the ability to confirm or rule out faint persistent radio emission.
• Intermittency and nulling lead to non-detections in a substantial fraction (~40–50%) of observing epochs, limiting continuous coverage of emission-state evolution.
• X-ray non-detections set stringent limits but do not completely rule out low-luminosity magnetar scenarios; Swift limits are less constraining than Chandra.
• Archival low-frequency radio searches (VLITE/VLA P-band) have higher noise levels and primary-beam attenuation at large offsets, possibly limiting detectability of weak pulses.