Fermi Detection of a Luminous Gamma-ray Pulsar in a Globular Cluster

We report the Fermi Large Area Telescope detection of gamma-ray (>100 megaelectronvolts) pulsations from pulsar J1823-3021A in the globular cluster NGC 6624 with high significance (~7 sigma). Its gamma-ray luminosity L_gamma = (8.4 +/- 1.6) x10^34 ergs per second, is the highest observed for any millisecond pulsar (MSP) to date, and it accounts for most of the cluster emission. The non-detection of the cluster in the off-pulse phase implies that its contains < 32 gamma-ray MSPs, not ~100 as previously estimated. The gamma-ray luminosity indicates that the unusually large rate of change of its period is caused by its intrinsic spin-down. This implies that J1823-3021A, has the largest magnetic field and is the youngest MSP ever detected, and that such anomalous objects might be forming at rates comparable to those of the more normal MSPs.

💡 Research Summary

The Fermi Large Area Telescope (LAT) has revolutionized high‑energy astrophysics since its 2008 launch, revealing that many globular clusters (GCs) emit γ‑rays as a collective output of their resident millisecond pulsars (MSPs). However, individual MSPs within GCs have rarely been identified as distinct γ‑ray sources because of limited spatial resolution and the faintness of most cluster pulsars at GeV energies. In this paper the authors report the first robust detection of γ‑ray pulsations from a single GC pulsar, J1823‑3021A, located in the globular cluster NGC 6624.

Using precise radio timing solutions obtained from long‑term observations with the Jodrell Bank, Parkes, and Nançay telescopes, each γ‑ray photon recorded by Fermi‑LAT between 4 August 2008 and 4 October 2010 was assigned a rotational phase. An H‑test applied to photons above 0.1 GeV yields a test statistic of 64, corresponding to a ∼6.8 σ detection. The γ‑ray light curve shows two peaks at phases 0.01 ± 0.01 and 0.64 ± 0.01, aligned with the main radio components, and the spectrum is well described by a power law with index 1.4 ± 0.3 and an exponential cutoff at 1.3 ± 0.6 GeV. The phase‑averaged energy flux is Fγ = (1.1 ± 0.1 ± 0.2) × 10⁻¹¹ erg cm⁻² s⁻¹.

Adopting a distance of 8.4 ± 0.6 kpc for NGC 6624, the inferred γ‑ray luminosity is Lγ = 4πd²fΩFγ ≈ 8.4 × 10³⁴ erg s⁻¹ (with a beaming correction factor fΩ ≈ 0.9). This makes J1823‑3021A the most luminous γ‑ray MSP known to date, exceeding the luminosities of all previously catalogued MSPs.

Crucially, the authors examined the off‑pulse intervals (phases 0.07–0.60 and 0.67–0.90) and found no detectable point source. By assuming a typical MSP spectrum and scaling to the full rotation, they derived a 95 % confidence upper limit on the off‑pulse flux of 5.5 × 10⁻¹² erg cm⁻² s⁻¹. This limit translates into an upper bound of < 32 γ‑ray MSPs in the cluster, far below earlier estimates of ∼100. The result is consistent with independent estimates based on the correlation between γ‑ray luminosity and stellar encounter rate, which predict ∼30 ± 15 MSPs.

The pulsar’s spin period is P = 5.44 ms, typical for MSPs, but its observed period derivative, ˙P_obs = 3.38 × 10⁻¹⁸ s s⁻¹, is one to two orders of magnitude larger than for most MSPs. While line‑of‑sight acceleration within the cluster (a_l) can contribute to the apparent ˙P, the collapsed core of NGC 6624 prevents a reliable mass model, and the pulsar’s projected offset of only 0.018 pc from the cluster centre suggests that a_l cannot fully account for the large ˙P_obs.

Using the γ‑ray efficiency η = Lγ/Ė, where the spin‑down power Ė = 4π²I ˙P/P³ (I ≈ 10⁴⁵ g cm²), the authors infer a minimum intrinsic period derivative ˙P ≈ 3.4 × 10⁻¹⁹ s s⁻¹ (for fΩ = 0.9). Even assuming an unrealistically high efficiency η = 1, this intrinsic ˙P would still be ∼10 % of the observed value, indicating that the majority of ˙P_obs is intrinsic. With η ≈ 0.1, a more realistic value for MSPs, the intrinsic spin‑down power is Ė ≈ 8.3 × 10³⁵ erg s⁻¹, and the corresponding γ‑ray efficiency is consistent with those measured for other γ‑ray MSPs.

From the intrinsic ˙P the surface dipole magnetic field is estimated as B₀ = 3.2 × 10¹⁹ √(P ˙P) ≈ 4.3 × 10⁹ G, considerably higher than the typical 10⁷–10⁹ G fields of recycled MSPs. The authors discuss three possible explanations for the combination of a relatively long spin period (5.44 ms) and a strong magnetic field: (i) the neutron star may have a larger mass and moment of inertia than assumed, (ii) the accretion phase could have proceeded at super‑Eddington rates in a non‑spherical geometry, or (iii) the magnetic field may have been weaker during accretion and subsequently increased, a behavior observed in some young, non‑recycled pulsars but not previously in MSPs.

The characteristic age τ_c = P/(2 ˙P) derived from the intrinsic ˙P is ≈ 2.5 × 10⁷ yr, making J1823‑3021A the youngest MSP known, with only J1824‑2452A in M28 possibly comparable. Because the spin‑down timescale is ∼10² times shorter than that of typical radio‑bright MSPs, such high‑B, young MSPs are expected to be observable for a relatively brief interval. Nonetheless, the detection of J1823‑3021A, together with the similar object in M28, suggests that the formation rate of these anomalous, high‑magnetic‑field MSPs in globular clusters may be comparable to that of the more conventional MSP population.

In summary, the paper demonstrates that (1) a single MSP can dominate the γ‑ray output of its host globular cluster, (2) γ‑ray observations combined with precise radio timing provide a direct probe of intrinsic spin‑down and magnetic field properties, and (3) there exists a sub‑population of young, high‑magnetic‑field MSPs in globular clusters whose formation may be more common than previously thought. These findings place new constraints on MSP evolutionary models, particularly on the mechanisms that reduce magnetic fields during the recycling process and on the accretion histories required to produce such energetic pulsars.