ASKAP J1935+2148: Ultra-Long-Period Radio Transient

• ASKAP J1935+2148 is a long-period radio transient that exhibits three distinct emission states with highly polarized bright and weak pulses and null intervals.
• Observations with ASKAP and MeerKAT enabled high-time-resolution polarimetry and phase-connected timing, confirming a stable 54-minute rotation period.
• Its unusual placement in the pulsar death valley and state-dependent polarization challenge standard neutron star models, suggesting magnetically powered reconnection processes in ultra-long-period magnetars.

ASKAP J1935+2148, formally ASKAP J193505.1+214841.0, is a long-period, emission-state-switching radio transient discovered with the Australian Square Kilometre Array Pathfinder and followed up with MeerKAT. Its defining observational property is a stable rotation period of approximately $54$ minutes accompanied by three distinct radio emission states: bright, tens-of-seconds pulses that are highly linearly polarized; weak, sub-second pulses that are highly circularly polarized; and intervals with no detectable emission. The discovery paper concluded that its radio coherence, phase stability, and compactness constraints exclude a magnetic white dwarf origin and instead favor a neutron star, likely an ultra-long-period magnetar, although the source also challenges standard neutron-star radio-emission phenomenology by lying in the pulsar “death valley” while remaining radio loud (Caleb et al., 2024).

1. Discovery, designation, and observational basis

ASKAP J1935+2148 was discovered serendipitously on 2022-10-15 during a target-of-opportunity ASKAP observation of GRB 221009A. MeerKAT localization gives the position as RA (J2000) $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$, Dec (J2000) $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$. The source lies within $5.6''$ of the magnetar SGR J1935+2154 and on the edge of the supernova remnant in which SGR J1935+2154 is centered, but no secure association with either object is claimed (Caleb et al., 2024).

The initial ASKAP observations used a central frequency of $887.5$ MHz, bandwidth $288$ MHz, and $10$ s integration time, with ASKAP operating in square_6×6 PAF configuration and a field of view of approximately $30\ {\rm deg}^2$. A fast pulse-detection pipeline subtracted visibilities of neighboring $10$ s integrations to remove quiescent field emission, then imaged the residuals to identify bright pulses; subsequent refinement used model subtraction, phase shifting to the source, dynamic spectra extraction, and baselines $>200$ m. Polarization was characterized through RM synthesis and Stokes $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$0 imaging. MeerKAT follow-up used a central frequency of $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$1 MHz and bandwidth $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$2 MHz, with $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$3 s imaging and beamformed full-Stokes recording with PTUSE at $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$4 sampling, APSUSE total intensity at $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$5, and real-time searches with TUSE/MeerTRAP at $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$6. Calibration used PKS J1939−6342 for flux and bandpass and PKS J2011−0644 for phase, while baseline variation was mitigated using an off-source APSUSE beam (Caleb et al., 2024).

These observing setups are central to the source’s phenomenology. ASKAP’s $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$7 s imaging cadence was naturally sensitive to the broad bright-state envelopes, whereas MeerKAT beamforming revealed the narrow weak-state pulses and enabled polarization-resolved timing. A plausible implication is that the apparent diversity of pulse morphologies partly reflects instrumental access to different temporal scales, although the discovery paper also emphasizes that the emission states represent genuine physical mode changes rather than merely sampling effects.

2. Timing solution and rotational phenomenology

A phase-connected timing solution gives a period of $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$8 s, equivalent to approximately $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$9 min. Although a parameter table lists $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$0 s, the timing analysis adopts $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$1 s. Times of arrival were derived from ASKAP imaging and MeerKAT beamformed detections, then fit barycentrically with tempo2 using the JPL DE436 ephemeris. In the timing analysis, the position was held fixed to the interferometric localization, the dispersion measure was fixed to the value that optimized the signal-to-noise ratio of individual MeerKAT pulses, and the ToA uncertainties were taken to reflect pulse width and jitter (Caleb et al., 2024).

The phase stability is unusually stringent for such an extreme-period source. MeerKAT pulses arrived within $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$2 ms of their predicted times, corresponding to approximately $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$3 of a rotation period, which confirms phase coherence across instruments and emission modes. The inferred period derivative remains only an upper limit, $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$4 at $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$5, and no significant torque variations were detected. The light-cylinder radius is correspondingly large, $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$6 (Caleb et al., 2024).

In the $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$7–$= +21^\circ\ 48'\ 41.504'' \pm 0.6''$8 plane, ASKAP J1935+2148 lies within the pulsar “death valley,” below standard death lines derived from curvature-radiation vacuum-gap and space-charge-limited-flow models, yet it is radio loud. That combination is one of the source’s principal theoretical difficulties. The discovery paper treats this as evidence that current neutron-star emission models are incomplete for ultra-long-period objects, while a later theoretical study uses the same placement to argue that conventional rotation-powered gap discharge is unlikely to sustain the source’s radio emission continuously (Yang et al., 27 Aug 2025).

3. Emission states and long-term evolution

Three distinct emission states were observed over approximately eight months. The bright pulse state, seen with ASKAP in 2022 October–November, consists of pulses with widths of $= +21^\circ\ 48'\ 41.504'' \pm 0.6''$9–$5.6''$0 s, linear polarization fraction $5.6''$1, and low circular polarization $5.6''$2. Its spectral index across ASKAP’s $5.6''$3 MHz band centered at $5.6''$4 MHz is $5.6''$5. Peak flux densities reach $5.6''$6 mJy in $5.6''$7 s images, with several pulses between roughly $5.6''$8 and $5.6''$9 mJy, and the implied duty cycle is approximately $887.5$0 for the $887.5$1 min period. Fifteen such pulses were detected across the early ASKAP epochs (Caleb et al., 2024).

The weak pulse state, seen with MeerKAT, is markedly different. The pulse widths are approximately $887.5$2 ms, about $887.5$3 narrower than ASKAP’s brightest envelope. These pulses are highly circularly polarized with $887.5$4, retain significant linear polarization with $887.5$5, and show a flat position angle across the main pulse when the signal-to-noise ratio permits. Their spectral index across the $887.5$6 MHz MeerKAT band centered at $887.5$7 MHz is $887.5$8. Measured flux densities are $887.5$9 mJy in a $288$0 s image average on 2023 February 3, with pre- and post-cursor subpulses, and $288$1 mJy on 2023 May 8. No broader envelope was detected in PTUSE full-Stokes beamformed data despite sub-millisecond sampling. The duty cycle is approximately $288$2, and only two pulses were detected, interspersed with quiescent epochs (Caleb et al., 2024).

The third state is a quenched or null state with no detectable pulses in multiple ASKAP and MeerKAT sessions. Its fractional occurrence is approximately $288$3–$288$4 across epochs $288$5–$288$6, consistent with intermittency, nulling, or large flux variability. Bright ASKAP pulses ceased after 2022-11-05, and only weak MeerKAT pulses were detected in 2023, indicating an emission-mode change. The discovery paper further suggests that the weak mode likely persists between nulls and remains below ASKAP’s sensitivity and time resolution (Caleb et al., 2024).

This state switching is the source’s most distinctive observational feature. Unlike other long-period sources, ASKAP J1935+2148 was identified as the first to exhibit drastic variations in emission modes reminiscent of neutron stars, specifically the transition from broad, highly linearly polarized pulses to narrow, highly circularly polarized pulses. A subsequent ULPP-oriented interpretation took this combination of burstiness, small duty cycle, and intermittency as qualitatively consistent with localized reconnection-driven activity, but that remains a model-dependent reading rather than an observationally established mechanism (Yang et al., 27 Aug 2025).

4. Propagation diagnostics, spectra, and coherent radio emission

The source has a measured dispersion measure of $288$7, determined from MeerKAT real-time detections; ASKAP could not constrain the DM because of its $288$8 s integrations. The rotation measure is consistent across instruments: $288$9 and $10$0, although a table lists $10$1. These RM values are consistent with the smoothed Galactic foreground and nearby pulsars, providing no evidence for an additional, large source-local RM (Caleb et al., 2024).

Faraday conversion was examined as a possible explanation for the large circular polarization in the weak state. Modeling found no significant frequency dependence of circular polarization across the MeerKAT band, which disfavors simple Faraday conversion scenarios. The large $10$2 is therefore likely intrinsic or produced by more complex propagation, such as partially coherent addition of modes. The pulses are detected across the full corrected bandpasses in both ASKAP and MeerKAT. The state-dependent spectra are also striking: the bright state is mildly inverted, while the weak state is steep, a behavior described as reminiscent of state-dependent spectral changes in pulsars and magnetars (Caleb et al., 2024).

The source is clearly a coherent emitter. Using the Rayleigh–Jeans brightness temperature,

$10$3

with $10$4 and $10$5, the discovery analysis adopted $10$6 kpc, $10$7–$10$8 s, $10$9 GHz, $30\ {\rm deg}^2$0 m, and $30\ {\rm deg}^2$1 Jy for the bright envelopes, giving $30\ {\rm deg}^2$2–$30\ {\rm deg}^2$3 K, consistent with an inferred range of $30\ {\rm deg}^2$4–$30\ {\rm deg}^2$5 K. The weak pulses, with $30\ {\rm deg}^2$6 s, $30\ {\rm deg}^2$7 GHz, and $30\ {\rm deg}^2$8 mJy averaged over $30\ {\rm deg}^2$9 s, likewise yield $10$0 K. These values are well above the incoherent synchrotron limit and point to coherent radio-emission mechanisms (Caleb et al., 2024).

The inferred monochromatic radio luminosities are approximately $10$1 in the ASKAP bright state and $10$2 in the MeerKAT weak state. A later theoretical paper instead used an isotropic radio-luminosity estimator and argued that the duty-cycle-averaged luminosity is of order $10$3–$10$4, still exceeding the spin-down power implied by the current $10$5 upper limit; that energetic interpretation underpins its reconnection-powered ULPP model (Yang et al., 27 Aug 2025).

5. Distance, local environment, and multiwavelength constraints

The DM-inferred distance is approximately $10$6 kpc in YMW16 and $10$7 kpc in NE2001, giving an average distance of approximately $10$8 kpc. On that basis, the measured DM suggests that ASKAP J1935+2148 is in the foreground relative to SGR J1935+2154. Its location near the Galactic plane and close to SGR J1935+2154 in SNR G57.2+0.8 is regarded as broadly consistent with a neutron-star origin, but not as evidence of a direct physical association (Caleb et al., 2024).

No persistent radio continuum counterpart has been detected. A stacked ASKAP deep image at $10$9 MHz shows no $>200$0 source, with rms approximately $>200$1, although confusion from the SNR edge complicates background subtraction. In X-rays, neither Chandra nor Swift identified a counterpart despite deep, multi-epoch coverage. Chandra, with a combined $>200$2 ks exposure in the $>200$3–$>200$4 keV band, gives a $>200$5 count-rate limit of $>200$6, corresponding to unabsorbed flux limits $>200$7 for a $>200$8 keV blackbody and $>200$9 for a $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$00 power law, or $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$01. Swift/XRT, with $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$02 ks total exposure, provides less constraining limits and shows no flaring behavior over 2010–2022 (Caleb et al., 2024).

Near-infrared and optical data likewise provide only limits or unrelated foreground sources. VLT/HAWK-I imaging from 2015 April 2 detected a source within $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$03 of the radio position, PSO J293.7711+21.8119, with $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$04, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$05, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$06 in Vega magnitudes, and PS1 photometry $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$07, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$08. A Keck/LRIS spectrum showed a featureless red continuum consistent with an L/T dwarf, and the distance estimates imply that it is a foreground star at $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$09 kpc and is unlikely to be associated with the radio transient. If no counterpart is assumed, the $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$10 limits at the radio position are $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$11, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$12, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$13 mag, with PS1 stack limits $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$14, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$15, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$16, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$17, and $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$18 (Caleb et al., 2024).

These non-detections are important because they constrain proposed source classes. In particular, they leave no clear accretion signature, no persistent synchrotron nebula, and no luminous X-ray counterpart. A plausible implication is that ASKAP J1935+2148 belongs to a low-radiative-efficiency compact-object population whose dominant observable output can remain confined to sporadic coherent radio bursts.

6. Source classification, comparative context, and theoretical interpretations

The strongest exclusion in the discovery analysis is against a magnetic white dwarf origin. Under pair-production-based coherent-emission models, the compactness required to sustain pair cascades leads to a lower bound on the emitting object’s radius,

$= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$19

where $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$20 and $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$21 near the polar cap. With $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$22 s, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$23 G, and $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$24, this gives $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$25 cm $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$26, much larger than the typical white-dwarf radius of approximately $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$27. Even more conservative choices of $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$28 strengthen the exclusion. Accreting cataclysmic variables are also disfavored because their radio luminosities, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$29–$= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$30, are far below that of ASKAP J1935+2148, and their ECMI origin in M-dwarf coronae does not match the observed behavior (Caleb et al., 2024).

The neutron-star interpretation follows from both phenomenology and energetics. Using the dipole estimate

$= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$31

the $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$32 upper limit implies $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$33 G and a spin-down luminosity $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$34 for canonical $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$35. Because the radio luminosity exceeds $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$36, the emission is interpreted as magnetically powered rather than spin-down powered, consistent with magnetar physics involving twisted or multipolar fields, reconnection, or untwisting. The source’s very low X-ray luminosity and long period are therefore described as compatible with an older population of ultra-long-period magnetars having low quiescent X-ray output but retaining coherent radio emission (Caleb et al., 2024).

ASKAP J1935+2148 also occupies a distinctive place within the broader long-period transient and ULPP population. Relative to GLEAM-X J162759.5−523504 at approximately $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$37 min and GPM J1839−10 at approximately $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$38 min, its period is roughly $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$39 longer than GLEAM-X and roughly $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$40 longer than GPM J1839−10. GLEAM-X was active for approximately three months, GPM J1839−10 for more than three decades, whereas ASKAP J1935+2148 showed bright pulses for weeks before transitioning to weak pulses and quenching over months. In period and mode variability it has been described as bridging GCRT J1745−3009, which showed $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$41-minute pulses with $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$42-minute periodicity and later narrower $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$43-minute pulses with variable circular polarization, and the newer long-period sources, but with more extreme state-dependent polarization reversals and spectral changes (Caleb et al., 2024).

A later theoretical study places ASKAP J1935+2148 within a set of eight radio-band ultra-long-period pulsars and argues for a distinct ULPP phenomenology. In that framework, a Crab-like young pulsar can evolve to $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$44 min on a timescale of approximately $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$45–$= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$46 Myr through magnetic-dipole-plus-wind braking, and the radio emission is powered predominantly by magnetic reconnection in localized, superstrong, multipolar fields near the polar cap rather than by rotational energy. The same paper emphasizes that five of the eight ULPPs, including ASKAP J1935+2148, have $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$47, and proposes a dichotomy in which normal pulsars are powered by rotation-driven gap discharges whereas ULPPs are powered by reconnection-induced continual radio bursts (Yang et al., 27 Aug 2025). This interpretation is not a consensus statement for the field; it is a specific explanatory model built on the observed period, $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$48 limit, death-valley placement, and radio energetics.

Several issues remain unresolved. The discovery study explicitly identifies open questions concerning the driver of the state switching, the origin of radio loudness below standard death lines, the cause of circular-polarization dominance in the weak mode, and the need for longer-baseline phase-connected timing to measure $= 19^{\rm h}\ 35^{\rm m}\ 05.175^{\rm s} \pm 0.3''$49, search for torque noise, and test for additional periodicities such as precession or binary orbital modulation. It also calls for broadband, high-time-resolution polarimetry, deeper X-ray observations, harder high-energy searches, more sensitive IR/optical campaigns, continued radio imaging for faint persistent emission, and systematic searches for similar long-period, mode-switching transients (Caleb et al., 2024). Taken together, these priorities define ASKAP J1935+2148 not merely as an unusual source, but as a diagnostic object for the physics of coherent radio emission, magnetospheric state changes, and compact-object evolution at the longest known neutron-star-like periods.