---
title: Ultra-long Period Pulsars (ULPs)
url: https://www.emergentmind.com/topics/ultra-long-period-pulsars-ulps
type: topic
---

# Ultra-long Period Pulsars (ULPs)

Searching arXiv for recent papers on ultra-long period pulsars and related formation/emission models.
Ultra-long period pulsars (ULPs), and in neutron-star-specific usage ultra-long period pulsars or ultra-long period radio pulsars (ULPPs), are coherent radio transients or pulsars with periodicities far longer than the historical radio-pulsar interval of $\sim 0.002$–$12\,\mathrm{s}$. Current usage is not uniform. Some papers use “ULPs” for coherent radio-pulsating transients with periods of order minutes to hours, whereas some formation models define isolated neutron-star ULPs as systems with $P\gtrsim10^3\,\mathrm{s}$. The topic is astrophysically significant because these objects intersect several populations—isolated neutron stars, magnetars, and compact binaries—and because standard rotation-powered pulsar models generally predict pair cascades only for $P\lesssim10\,\mathrm{s}$ [2201.11704] [2406.12352] [2507.10682].

## 1. Nomenclature and class boundaries

The modern ULP literature combines at least two partially overlapping categories. One consists of slowly rotating neutron stars, exemplified by PSR J0901–4046 with $P=75.88554711\,\mathrm{s}$, whose timing and polarimetry identify it as a radio-emitting neutron star [2206.01346]. The other consists of ultra-long period radio transients with pulse periods in the $\sim10^3$–$10^4\,\mathrm{s}$ range, some of which are now confirmed compact binaries rather than isolated neutron stars [2505.06125].

Representative sources illustrate this heterogeneity:

| Source | Period | Classification or note |
|---|---:|---|
| PSR J0901–4046 | $75.88554711\,\mathrm{s}$ | radio-emitting neutron star |
| CHIME J0630+25 | $421.4\,\mathrm{s}$ | ultra-long period object |
| GLEAM-X J1627 | $1091.2\,\mathrm{s}$ | long-period radio transient |
| GPM J1839–10 | $1318.2\,\mathrm{s}$ | long-period radio transient |
| ILT J1101+5521 | $7531.2\,\mathrm{s}$ | confirmed WD–M-dwarf binary |
| GLEAM-X J0704–37 | $10496.6\,\mathrm{s}$ | confirmed WD–M-dwarf binary |
| ASKAP J1839–0756 | $2.32217\times10^{4}\,\mathrm{s}$ | $6.45\,\mathrm{h}$ coherent radio transient |

The confirmed binary identifications are especially important. ILT J1101+5521 and GLEAM-X J0704–37 are confirmed white-dwarf–M-dwarf binaries, with primary masses $M_1\approx0.87\pm0.3\,M_\odot$ and companion masses $M_2\approx0.19$–$0.32\,M_\odot$ from optical spectroscopy [2505.06125]. This establishes that “ultra-long period object” is not synonymous with “isolated neutron star.” A plausible implication is that the observational class is heterogeneous, with only a subset representing genuine neutron-star ULP pulsars.

## 2. Observational phenomenology

The best-studied neutron-star member is PSR J0901–4046. Its discovery paper reported $P=75.88554711\,\mathrm{s}$ and $\dot P=2.25\times10^{-13}\,\mathrm{s\,s^{-1}}$, implying $\tau\simeq5.3\,\mathrm{Myr}$, $B_s\approx1.3\times10^{14}\,\mathrm{G}$, and $\dot E\approx2.0\times10^{28}\,\mathrm{erg\,s^{-1}}$. The average pulse has $W_{50}\simeq300\,\mathrm{ms}$ at both $1.3\,\mathrm{GHz}$ and $0.8\,\mathrm{GHz}$, corresponding to a duty cycle of about $1\%$. Single pulses were grouped into seven morphological classes: “normal,” “spiky,” “double-peaked,” “split-peak,” “triple-peaked,” “quasi-periodic,” and “partially nulling.” Bright pulses show quasi-periodic sub-pulse structure with $P_2$ from $\simeq9.6\,\mathrm{ms}$ to $\simeq338\,\mathrm{ms}$, most commonly $\sim76\,\mathrm{ms}$, and some rotations show partial nulling in which $\sim40\%$ of the underlying envelope energy is intermittently dropped out [2206.01346].

Longer-baseline follow-up sharpened this picture. A coherent timing model over $2.6\,\mathrm{yr}$ gave $P=75.88554698(3)\,\mathrm{s}$ and $\dot P=(2.44\pm0.01)\times10^{-13}\,\mathrm{s/s}$ with RMS timing residual $\simeq7.6\,\mathrm{ms}$, or $\approx10^{-4}P$, demonstrating exceptional rotational stability. High-time-resolution MeerKAT data showed two distinct quasi-periodic microstructure timescales: $\tau_{\rm peak}\simeq60$–$90\,\mathrm{ms}$ with mean $\simeq72.9\,\mathrm{ms}$, and $\tau_{\rm peak}\simeq13$–$33\,\mathrm{ms}$ with mean $\simeq20.9\,\mathrm{ms}$. The source was not detected below $\sim500\,\mathrm{MHz}$, suggesting a low-frequency turnover, and $W_{50}$ remained nearly constant from $544$ to $4032\,\mathrm{MHz}$, consistent with zero radius-to-frequency mapping [2505.04430].

At the extreme long-period end, ASKAP J1839–0756 has $P=2.32217\times10^{4}\,\mathrm{s}=6.45\,\mathrm{hr}$, with an interpulse separated by $177.8^\circ\pm3.0^\circ$, a main-pulse $W_{50}=320$–$710\,\mathrm{s}$, and duty cycle $\delta\approx1.4\%$–$3.1\%$. Its main pulses have $L/I=60\%$–$90\%$ and $V/I=30\%$–$60\%$, interpulses have $L/I\approx90\%$ and $V/I\lesssim10\%$, and the polarization-position-angle swings are consistent with antipodal-pole emission. The spectral index lies in the interval $-2.99\le\alpha\le-2.27$, with a turnover at $300$–$600\,\mathrm{MHz}$ implied by MWA non-detection [2501.09133].

The broader phenomenology includes rare or non-repeating bursts. ASKAP J175534.9$-$252749.1 produced a single $\sim110\,\mathrm{s}$ coherent burst with peak flux density $\sim200\,\mathrm{mJy}$, linear polarization up to $\sim60\%$, circular polarization up to $\sim40\%$, $\mathrm{RM}=961\pm45\,\mathrm{rad\,m^{-2}}$, and spectral index $\alpha=-3.1\pm0.1$. In one 2024 census, all seven known ULPs, plus ASKAP J1755, lay at $|b|\lesssim3^\circ$, suggesting a thin-disk concentration [2406.12352].

## 3. Tension with standard spin-down and radio-emission theory

The basic theoretical difficulty is twofold: producing very long spin periods and sustaining coherent radio emission at such periods. In standard pulsar phenomenology, the inferred dipolar surface field is commonly written as
$$
B_s = 3.2\times10^{19}\sqrt{P\,\dot P}\ {\rm G},
$$
while the spin-down luminosity is
$$
\dot E = 4\pi^2 I\,\frac{\dot P}{P^3}.
$$
For PSR J0901–4046, these relations place the source in the magnetar-field regime while leaving it with a very small $\dot E$ [2206.01346].

Standard vacuum-gap or slot-gap curvature-radiation pictures generally predict pair cascades only when $P\lesssim10\,\mathrm{s}$, producing the traditional “death line” tension for minute-to-hour objects [2406.12352]. In the J0901–4046 discovery analysis, the source lies beyond the RS75/CR93 death lines but above the SCLF line, suggesting that non-dipolar surface fields or alternative particle-acceleration regimes may still permit pair cascades [2206.01346]. This is not a general solution for the whole class, because the longer-period sources occupy even more extreme regions of $P$–$\dot P$ space.

Population-synthesis calculations under pure dipole spin-down sharpen this tension. In the neutron-star scenario, even extreme assumptions—constant high $B$, $20\%$ duty cycle, fallback-driven slow-$P_0$ tails, or bimodal high-$B$ birth distributions—yield zero detectable neutron stars with $P\gtrsim10^3\,\mathrm{s}$ and $\dot E$ above the reference radio-emission threshold. The synthetic neutron-star period distributions pile up around $10$–$100\,\mathrm{s}$ and then drop steeply. In the white-dwarf scenario, by contrast, magnetic white dwarfs naturally populate the $10^2$–$10^5\,\mathrm{s}$ interval in large numbers, but classical pair-cascade emission remains difficult there as well [2307.10351].

This suggests that the ULP problem is not reducible to spin evolution alone. Any successful neutron-star model must explain both why some objects reach $P\gg10\,\mathrm{s}$ and why radio emission persists beyond the conventional rotational death valley.

## 4. Proposed origin channels

Several distinct formation channels have been advanced, and they imply different birth environments, magnetic-field requirements, and event rates.

| Channel | Key conditions | Characteristic outcome |
|---|---|---|
| Supernova fallback disk | $B>10^{14}\,\mathrm{G}$ and initial fallback accretion rates $\sim10^{22-27}\,\mathrm{g\,s^{-1}}$ | spin periods $>100\,\mathrm{s}$; in unstable-disk models $P\sim10^4$–$10^5\,\mathrm{s}$ |
| Wide-binary wind-fed accretion | $P_{\rm orb}\sim10^3$–$10^4\,\mathrm{d}$, second supernova disruption in $>99\%$ of systems | isolated ULPPs with $P_s\sim10^1$–$10^4\,\mathrm{s}$ and tail to $>10^6$–$10^8\,\mathrm{s}$ |
| Shock-inflated companion disk | close binary, $a_0\simeq20\,R_\odot$, $M_c\simeq4\,M_\odot$, $B_0\gtrsim10^{14}\,\mathrm{G}$ | bimodal population with canonical pulsars at $P\lesssim10\,\mathrm{s}$ and ULPs at $10^3\lesssim P<10^5\,\mathrm{s}$ |
| Late-blooming magnetar | core-threading currents and delayed Hall-dominated evolution | radio-loud awakening after $\sim0.1\,\mathrm{Myr}$ and periods $P\sim10^3$–$10^4\,\mathrm{s}$ by $\sim1\,\mathrm{Myr}$ |

In the supernova fallback picture, a newborn neutron star interacts with fallback matter through an accretion disk. A 2022 parameter study found that very long spin periods $>100\,\mathrm{s}$ can be reached in the presence of strong, magnetar-like magnetic fields $>10^{14}\,\mathrm{G}$ and moderate initial fallback accretion rates $\sim10^{22-27}\,\mathrm{g\,s^{-1}}$. The same study treated PSR J0901–4046 and GLEAM-X J162759.5-523504.3 as case studies and concluded that the fallback scenario could represent a viable channel for long-period isolated pulsars [2201.11704].

A later extension introduced thermal-viscous instability into the fallback disk. In that model, the disk evolves self-similarly until neutralization or self-gravity truncates it, while magnetar spin, magnetic field, and inclination evolve under accretion, magnetic-disk, and dipole torques. Thermal-viscous fronts modulate $\dot M$, and in the Model II-2 case the disk remains active up to $\sim10^5\,\mathrm{yr}$, allowing $P$ to climb into the ULPP regime with $P\gtrsim10^4\,\mathrm{s}$. The same simulations found a pronounced U-shaped magnetic-inclination distribution, with $\sim20$–$50\%$ of ULPPs becoming nearly aligned or nearly orthogonal rotators; the authors argued that extra mechanisms are still required to explain radio emission [2410.05944].

A distinct binary-origin channel places the first-born neutron star in a wide high-mass X-ray binary. The spin evolution is governed by $I\,d\Omega/dt=N$, with transitions among ejector, propeller, and accretor phases determined by the ordering of $R_m$, $R_{\rm co}$, and $R_{\rm lc}$. In this picture, wind-fed accretion from the massive companion spins the neutron star down from $P_0\simeq0.2\,\mathrm{s}$ to ultra-long periods before the secondary explodes and disrupts the system. Grid studies give final $P_s\sim10^1$–$10^4\,\mathrm{s}$, while Monte Carlo binary population synthesis yields a tail extending to $P_s>10^6$–$10^8\,\mathrm{s}$, with an estimated Milky Way ULPP birthrate of $\simeq1.4\times10^{-6}\,\mathrm{yr^{-1}}$ at solar metallicity [2507.00946].

Another binary-assisted pathway invokes a close pre-supernova binary in which the newborn neutron star traverses the companion’s shock-inflated envelope. Hydrodynamic simulations for $a_0\simeq20\,R_\odot$ and $M_c\simeq4\,M_\odot$ yield a disk-formation fraction of $\sim10\%$ among unbound neutron stars, with captured disk masses $M_{d,0}\sim10^{-7}$–$10^{-2}\,M_\odot$. Disk interaction then produces a short-lived propeller phase, and for $B_0\gtrsim10^{14}\,\mathrm{G}$ the star is spun down from $P_0\sim0.01$–$0.1\,\mathrm{s}$ to $P\sim10^3$–$10^5\,\mathrm{s}$ on timescales of $10$–$10^3\,\mathrm{yr}$. The implied Milky Way formation rate is $\sim10^{-4}\,\mathrm{yr^{-1}}$ [2507.10682].

Finally, the “late-blooming” magnetar scenario does not require external accretion. Magnetothermal calculations show that if electric currents thread the fluid core at crust freezing, the star can remain multiband silent for an initial period of approximately $0.1\,\mathrm{Myr}$ while cooling passively. Once the crust temperature falls enough for Hall evolution to dominate, crustal failures inject magnetospheric twist and amplify spin-down torque, yielding $P\sim10^3$–$10^4\,\mathrm{s}$ by $t\sim1\,\mathrm{Myr}$ [2505.05373].

The coexistence of these models suggests that there may be more than one physical route to the ULP regime. Binary channels naturally explain some observed long-period radio transients, while fallback and magnetothermal channels address isolated neutron-star candidates.

## 5. Radio-emission mechanisms beyond the classical death line

The emission problem is often treated separately from the formation problem. One recent approach considers local magnetospheric twists driven either by crustal plastic motion or by thermoelectric action from crustal temperature gradients. In that framework, a twisted open-field bundle requires a parallel current
$$
j_{\rm twist}\approx \frac{c}{4\pi R_s}\,\Delta\Psi\,B,
$$
and a gap develops once $j_{\rm twist}>j_{\rm GJ}$. The corresponding critical twist is
$$
\Psi_{\rm crit}\approx10^{-5}\,R_{s,6}\,P_3^{-1}\sin^{-1}\theta_{{\rm fp},-1}.
$$
Pair cascades then proceed via resonant inverse-Compton scattering or curvature radiation, but only for magnetar-like field strengths $B\gtrsim10^{14}\,\mathrm{G}$ and long periods. The model gives
$$
P_{\rm RICS}\gtrsim120\,(T/10^{6.5}\,{\rm K})^{-5}\ {\rm s},
$$
and
$$
P_{\rm curv}\gtrsim150\,(v_{\rm pl}/10^3\,{\rm cm\,yr^{-1}})^{-7/6}\ {\rm s}.
$$
For a fiducial radio efficiency $\eta_r=10^{-4}$, the predicted radio luminosities are $L_r\sim10^{30}$–$10^{33}\,\mathrm{erg\,s^{-1}}$, consistent with GLEAM-X J1627 and GPM J1839–10, and the model predicts simultaneous thermal or non-thermal X-ray/UV counterparts [2406.04135].

The late-blooming magnetar picture gives a related, but more explicitly evolutionary, emission mechanism. During the initial passive-cooling phase, the quiescent X-ray luminosity can fall below $\sim10^{31}\,\mathrm{erg\,s^{-1}}$ within $\sim0.1\,\mathrm{Myr}$. When the Hall parameter reaches the nonlinear regime, crustal failures begin injecting twist into the external field. In core-threading models, waiting times between such failures peak at $\sim1\,\mathrm{yr}$, plastic motion can last months, and the resulting duty cycles are $D\sim1\%$–$10\%$, before beaming corrections. These features were proposed to reproduce sparse and highly variable radio-loud windows in Galactic ULPs [2505.05373].

A more explicitly non-rotational picture has been developed for ASKAP J1935+2148 and related systems. In that model, local “spot” fields of order $B_i\sim10^{16}\,\mathrm{G}$ store magnetic energy
$$
E_B\sim1.7\times10^{47}\Bigl(\frac{B}{10^{16}\,{\rm G}}\Bigr)^2{\rm erg},
$$
and Ohmic/Hall decay powers an X-ray luminosity of order $L_X\sim5\times10^{33}\,\mathrm{erg\,s^{-1}}$. If a fraction $\eta\sim10^{-4}$–$10^{-5}$ emerges in radio, then $L_R\sim10^{-4}L_X\sim5\times10^{29}\,\mathrm{erg\,s^{-1}}$, consistent with observed radio powers. The same work argues that five of eight ULPPs have $L_R>\dot E$, supporting a non-rotational energy reservoir for at least part of the class [2508.19801].

Not all well-observed neutron-star candidates fit a simple magnetar template. PSR J0901–4046 has a very short duty cycle of $\sim1.4^\circ$, no magnetar-like outbursts or timing glitches, and highly stable timing, making it more similar in some respects to radio pulsars with periods $>5\,\mathrm{s}$ than to radio-loud magnetars [2505.04430]. This suggests that the radio-loud ULP population may include multiple magnetospheric states rather than a single canonical emission regime.

## 6. Discriminants, selection effects, and future tests

Because the observational class is heterogeneous, the central empirical task is to distinguish isolated neutron stars from compact binaries and from rarer alternatives. Space-based gravitational-wave interferometers provide one such discriminator when the radio period is the orbital period. For the known ultra-long period objects treated as nearly monochromatic binaries, the predicted one-month and four-year LISA signal-to-noise ratios are exceptionally large for some systems: CHIME J0630+25 has SNR $667$ $(>10^3)$ in one month and $>10^3$ $(>10^4)$ in four years, while GLEAM-X J1627 has SNR $15.9$ $(48.6)$ in one month and $110$ $(337)$ in four years. A detection would confirm binary nature and measure the chirp mass; a non-detection for nearby, short-period systems would favor isolated magnetar interpretations [2505.06125].

A more exotic alternative is self-lensing in an edge-on pulsar–black-hole binary, where lensing once per orbit makes a fast pulsar appear as a minute-to-hour transient. The model predicts apparent periods $T_{\rm app}\simeq P_{\rm orb}$, burst durations set by the lensing timescale, and a Galactic yield of order a few systems for $10\,\mathrm{mJy}$ sensitivity. However, when applied to GLEAM-X J1627, PSR J0901–4046, and GPM J1839–10, the required black-hole masses, short coalescence times, large negative $\dot P_b$, and implied merger rates are inconsistent with observations, so this explanation is disfavored for those specific sources [2401.12494].

Search strategy is itself a major bottleneck. The discovery paper of J0901–4046 emphasized that very long periods, narrow duty cycles, and low harmonic content cause such signals to be filtered out or flagged as interference in standard Fourier pipelines; low dispersion measures can intensify that confusion [2206.01346]. This is consistent with the view that current samples are strongly incomplete. A 2025 detection study introduced FITrig, a GPU-accelerated image-based search method with both image-domain and image-frequency-domain branches. On $50{\rm K}\times50{\rm K}$ images, FITrig ran in $7.2\,\mathrm{s}$ versus $31.0\,\mathrm{s}$ for SOFIA 2 alone, a $4.3\times$ speedup, and reduced false positives by up to $858.8\times$ at $6\sigma$ significance while retaining the ability to detect pulsars $20$ times fainter than surrounding steady features; PSR J0901–4046 was recovered at $f=0.013\,\mathrm{Hz}$ with $\sim1^{\prime\prime}$ localization accuracy [2509.21754].

Future observational tests are correspondingly diverse. Binary-origin models predict that long radio surveys with CHIME, LOFAR, SKA, and FAST should uncover additional ULPPs, especially systems with $P_s=10^3$–$10^6\,\mathrm{s}$ and low persistent X-ray luminosities [2507.00946]. Shock-inflated-companion models predict active remnant disks with mid-infrared excesses of order $L_{\rm IR}\sim10^{32}\,\mathrm{erg\,s^{-1}}$ and $T_{\rm IR}\sim200$–$500\,\mathrm{K}$ during the first few $\times10^4\,\mathrm{yr}$ [2507.10682]. Magnetothermal models predict very low quiescent X-ray fluxes, intermittent outburst windows, and polarimetric signatures linked to crustal-failure geometry [2505.05373]. This suggests that progress on ULPs will come less from a single decisive observable than from combining radio timing, broadband polarimetry, deep X-ray and infrared follow-up, and, for binary candidates, gravitational-wave constraints.

Source: https://www.emergentmind.com/topics/ultra-long-period-pulsars-ulps