J1641+3627H (M13H) Millisecond Pulsar in M13

• The pulsar is discovered using sensitive Fourier-domain acceleration searches applied to FAST archival data over seven years.
• It was detected during rare scintillation-bright states, achieving approximately 8.2σ significance in two brief observations.
• The binary nature is inferred from significant line-of-sight accelerations, though limited data prevent constraining its orbital parameters.

J1641+3627H, commonly designated M13H, is a millisecond pulsar in the globular cluster M13 discovered with the Five-hundred-meter Aperture Spherical radio Telescope (FAST) in the study "Illuminating Hidden Pulsars: Scintillation-Enhanced Discovery of Two Binary Millisecond Pulsars in M13 with FAST" (Yin et al., 8 Aug 2025). It was identified through a sensitive Fourier-domain acceleration search applied to 84 FAST observations spanning seven years. The source was detected in only 2 of 84 observations, each during scintillation-brightened states and with significant apparent line-of-sight acceleration, which is consistent with a binary system. At present, however, the extremely low detection rate prevents a phase-connected timing solution and leaves orbital parameters and system classification unconstrained (Yin et al., 8 Aug 2025).

1. Discovery setting and search methodology

M13H emerged from a reanalysis of a FAST archival set comprising 84 observations of M13 spanning 2018 to 2025 February 25 (Yin et al., 8 Aug 2025). All observations used the central beam of the FAST 19-beam receiver in L band from 1.0 to 1.5 GHz, with 4096 channels of width 0.122 MHz, a sampling time of 49.152 $\mu$s, and $n_p = 2$ polarizations. Typical full-length integrations were approximately 4000–5000 s, while the 2021 October 1 SnapShot mode consisted of $4\times 2000$ s pointings; M13H was detected in the initial 2000 s segment.

The search pipeline was explicitly optimized for faint and accelerated systems. Radio-frequency interference masking was performed with rfifind in PRESTO, dedispersion with prepsubband over $\mathrm{DM} = 28$–$32~\mathrm{cm^{-3}\,pc}$ in steps of $0.05~\mathrm{cm^{-3}\,pc}$, Fourier transformation with realfft, and low-frequency noise mitigation with rednoise. Candidate generation used accelsearch with harmonic summing up to 32, a deliberately low detection threshold of $2\,\sigma$, and candidate sifting with JinglePulsar; diagnostic folding and initial ephemerides were obtained with prepfold. Searches were performed on both full-length data and coherently segmented sub-integrations, generally 1500–2500 s long.

The acceleration-search formalism was presented through the Fourier-bin drift parameter $z_{\rm max}$,

$$z_{\rm max} \approx \frac{a_{\rm l}\, h\, f\, T_{\rm obs}^2}{c},$$

where $a_{\rm l}$ is the constant line-of-sight acceleration, $n_p = 2$0 the harmonic number, $n_p = 2$1 the spin frequency, $n_p = 2$2 the observation duration, and $n_p = 2$3 the speed of light. Search configurations included $n_p = 2$4 for isolated-pulsar full-length searches, $n_p = 2$5 for segmented accelerated searches, and $n_p = 2$6 for full-length accelerated searches. The paper gives an illustrative sensitivity scale: for a 200 Hz signal, a 2000 s observation with $n_p = 2$7 reaches sensitivity up to $n_p = 2$8. Parallel jerk searches on segmented data, using $n_p = 2$9 and $4\times 2000$0 with four-harmonic summing, yielded no additional detections. In context, the segmented acceleration searches were important because they reduced orbital smearing and aligned the search with short scintillation-bright intervals.

2. Detection record and measured observables

M13H was detected in exactly two observations, corresponding to a detection rate of $4\times 2000$1 (Yin et al., 8 Aug 2025). The two successful epochs were MJD 59488, the discovery epoch, and MJD 59966, the confirmation epoch. In both cases the pulsar was found in segmented data rather than through a more significant recovery in the full-length integrations.

Epoch	Measured parameters	Detection context
MJD 59488	$4\times 2000$2; $4\times 2000$3 ms; $4\times 2000$4	Discovery in initial 2000 s SnapShot segment
MJD 59966	$4\times 2000$5; $4\times 2000$6 ms; $4\times 2000$7	Confirmation in $4\times 2000$8 s segment

The detections were weak but statistically compelling in the folding diagnostics. prepfold yielded reduced $4\times 2000$9–$\mathrm{DM} = 28$0 and noise probabilities below $\mathrm{DM} = 28$1 and $\mathrm{DM} = 28$2, corresponding to approximately $\mathrm{DM} = 28$3 Gaussian significance in both epochs. By comparison, full-length acceleration searches recovered M13H only at slightly reduced significance, with values of $\mathrm{DM} = 28$4 and $\mathrm{DM} = 28$5 in the reported tests with higher $\mathrm{DM} = 28$6.

The pulse profile is reported as single-peaked. Potential substructure could not be assessed because of the limited signal-to-noise ratio, and no duty cycle or detailed morphology parameters were given. The estimated L-band flux density in the two detections spans 2.12–3.48 $\mathrm{DM} = 28$7Jy over 1.05–1.45 GHz. A median flux density was not reported.

3. Binary interpretation and orbital non-determination

The discovery paper presents M13H as a millisecond pulsar in a binary, with the binary interpretation resting on the significant apparent accelerations measured in both detections (Yin et al., 8 Aug 2025). Both accelerations are negative line-of-sight values, $\mathrm{DM} = 28$8 and $\mathrm{DM} = 28$9, and the change in apparent barycentric spin period between the two epochs is consistent with the source having been observed at different orbital phases rather than as an isolated rotator.

The same paper is equally explicit about the current limits of inference. It states that the "extremely low detection rate currently prevents constraints on orbital parameters or classification." No robust constraints are provided for the orbital period $32~\mathrm{cm^{-3}\,pc}$0, projected semi-major axis $32~\mathrm{cm^{-3}\,pc}$1 or $32~\mathrm{cm^{-3}\,pc}$2, or eccentricity $32~\mathrm{cm^{-3}\,pc}$3. An appendix gives a temporary ephemeris in the BT model solely for folding, but the paper emphasizes that many solutions are possible, so these values are not reported as measurements.

A qualitative estimate is noted under the acceleration-search assumption $32~\mathrm{cm^{-3}\,pc}$4: the orbital period is estimated to be at least approximately 0.5 days. This is not a measured parameter and remains unconstrained. Likewise, the standard pulsar mass function,

$32~\mathrm{cm^{-3}\,pc}$5

is given only for context. For M13H, $32~\mathrm{cm^{-3}\,pc}$6, companion mass, and categorical labels such as black widow or redback are not provided because no timing solution exists. A plausible implication is that M13H may belong to a compact and substantially accelerated binary configuration, but the available detections do not permit a formal classification.

4. Scintillation as the enabling detection mechanism

The paper attributes the detectability of M13H primarily to fortuitous flux enhancement by interstellar scintillation (Yin et al., 8 Aug 2025). Both new pulsars in the study were discovered during scintillation-brightened states, and for M13H specifically the source would otherwise have remained undetected given its weak observed flux density and low recovery rate.

No scintillation bandwidths or timescales are reported for M13H itself. The paper instead situates M13H within the broader scintillation phenomenology of the M13 line of sight. M13 pulsars show strong time–frequency modulation due to diffractive interstellar scintillation, and non-detections among previously known M13 pulsars are often attributed to scintillation. It also notes that a screen-like scattering medium along the M13 line of sight has been detected, at a distance of $32~\mathrm{cm^{-3}\,pc}$7 kpc and a height of $32~\mathrm{cm^{-3}\,pc}$8 kpc above the Galactic plane, and that M13A shows scintillation timescales of $32~\mathrm{cm^{-3}\,pc}$9 to $0.05~\mathrm{cm^{-3}\,pc}$0 minutes. These values are contextual rather than measurements for M13H.

The observational interpretation is therefore twofold. First, scintillation raised M13H above the survey threshold on the two successful epochs. Second, segmentation of the data improved sensitivity both by reducing orbital smearing and by matching the analysis window to intervals of enhanced flux. The paper further notes that during the 5000 s confirmation observation no significant scintillation was observed, and the reduced significance of the full-length search was more likely due to orbital acceleration; nevertheless, the overall inference is that scintillation facilitated the two successful detections.

5. Timing status, astrometric information, and measurement limits

No phase-connected timing solution is currently available for M13H (Yin et al., 8 Aug 2025). The low detection rate and weak flux prevent the derivation of a coherent timing model, and consequently there are no reported timing residuals, time-of-arrival counts, or parameter uncertainties from a fitted timing solution.

Astrometric information is correspondingly limited. Right ascension, declination, and quantitative cluster-centric offsets are not provided. The only location statement is qualitative: M13H was detected only in the central beam during the 2021 October 1 SnapShot observation covering a 4.5 arcmin region, which suggests that the pulsar is located in the core region of M13.

The flux-density estimates were derived using the paper’s radiometer-equation framework with adopted parameters $0.05~\mathrm{cm^{-3}\,pc}$1, $0.05~\mathrm{cm^{-3}\,pc}$2 K, $0.05~\mathrm{cm^{-3}\,pc}$3, $0.05~\mathrm{cm^{-3}\,pc}$4, and $0.05~\mathrm{cm^{-3}\,pc}$5 MHz over 1.05–1.45 GHz. Even with FAST sensitivity, the resulting 2.12–3.48 $0.05~\mathrm{cm^{-3}\,pc}$6Jy detections remain near the effective survey limit. This explains why profile details such as substructure and duty cycle could not be measured robustly, and why the source currently remains a detection-level object rather than a fully timed binary MSP.

6. Relation to the M13 pulsar population and prospective characterization

Within the same FAST study, M13H is contrasted most clearly with J1641+3627G (M13G), whose characterization is far more complete (Yin et al., 8 Aug 2025). M13G is a 4.32 ms pulsar with $0.05~\mathrm{cm^{-3}\,pc}$7, a phase-connected timing solution spanning 6.4 years, a circular 0.12 d orbit, minimum and median companion masses of $0.05~\mathrm{cm^{-3}\,pc}$8 and $0.05~\mathrm{cm^{-3}\,pc}$9, classification as a black widow, no eclipses, and 57 detections used in timing. M13H, by contrast, is an 11.21 ms pulsar with only two detections, significant apparent accelerations, no orbital solution, undetermined classification, and a single-peaked profile seen only in scintillation-favorable states.

This contrast is significant for interpreting the M13 pulsar census. Previously known M13 pulsars already exhibited large flux variations and occasional non-detections due to scintillation, so M13H is not an isolated anomaly within the cluster. Rather, it exemplifies a population of systems that may remain hidden unless search strategies simultaneously address low flux density, severe orbital smearing, and intermittent scintillation gain. This suggests that some cluster MSPs can be discovery-limited by propagation effects and search configuration as much as by telescope sensitivity.

The paper identifies the observational requirements for advancing beyond the current state of knowledge. Additional epochs are needed, preferably coincident with scintillation-brightened states, to secure repeated detections across orbital phases. Those detections would enable more robust acceleration and jerk characterization, determination of $2\,\sigma$0, $2\,\sigma$1, and $2\,\sigma$2, and eventual system classification. Continued use of segmented, overlapping coherent acceleration searches is specifically motivated to mitigate orbital smearing and exploit scintillation windows, while more computationally intensive methods such as template-bank or phase-modulation searches are suggested as possible routes to uncovering highly accelerated systems.