The FAST Discovery of a binary millisecond pulsar PSR~J1647-0156B (M12B) with a candidate cross matching algorithm

This paper introduces a candidate cross-matching algorithm based on spin period and dispersion measure consistency to identify recurring pulsar signals, which was successfully applied to FAST data to discover the binary millisecond pulsar PSR J1647-0156B (M12B) in the globular cluster M12.

The Gist

Imagine the universe is a giant, noisy radio station. Hidden among the static are thousands of "cosmic lighthouses" called pulsars. These are dead stars that spin incredibly fast, beaming radio waves at us like a lighthouse beam. Finding them is like trying to hear a single person whispering in a crowded, screaming stadium.

For a long time, astronomers used powerful telescopes (like the FAST telescope in China) to scan the sky. They had a computer program that acted like a sieve, catching potential "whispers" (candidates). But here's the problem: the telescope is so sensitive that it catches too much noise. It also catches the same whisper multiple times if you look at the same spot on different days.

The Problem: The "Needle in a Haystack" that Keeps Moving

When astronomers looked at Globular Clusters (dense groups of stars, like cosmic cities), they found a mess.

• The Haystack: Millions of radio signals.
• The Needles: Real pulsars.
• The Fake Needles: Radio interference from Earth (like your microwave or a satellite) that looks like a pulsar.

Usually, astronomers would look at each day's data separately. If a signal was too faint or looked a little weird, they might ignore it. It's like trying to find a friend in a crowd by looking at a single photo; if they are wearing a hat in one photo and no hat in another, you might think they are two different people.

The Solution: The "Matchmaker" Algorithm

The authors of this paper invented a new tool called the Cross-Matching Algorithm (CMA). Think of this as a super-smart detective or a matchmaker.

Instead of looking at each day's data alone, the CMA takes all the lists of potential signals from different days and asks:

"Hey, did we see the same person twice?"

It uses two specific clues to decide if two signals are the same pulsar:

1. The Spin Speed (Period): Does it spin at almost the exact same speed? (Within 1% difference).
2. The Distance Clue (Dispersion Measure): As radio waves travel through space, they get slowed down by gas. This "slowing" tells us how far away the signal is. If two signals have the same "distance signature" (within 10% difference), they are likely from the same place.

The Analogy: Imagine you are trying to find a specific friend in a city.

• Old Method: You ask, "Did anyone see a tall guy in a red shirt today?" and "Did anyone see a tall guy in a red shirt tomorrow?" separately. You might miss him if he wore a blue shirt on Tuesday.
• New Method (CMA): You gather all the reports. "Wait, the guy seen on Monday and the guy seen on Wednesday both have the same unique scar and walk with the same limp." Bingo! It's the same person.

The Big Discovery: M12B

Using this new "Matchmaker," the team looked at old data from the FAST telescope and found something amazing: PSR J1647-0156B, also known as M12B.

• Why was it missed before? This pulsar is a "binary" system, meaning it's orbiting another star. It has a very short orbit (about half a day). Because it moves so fast around its partner, its signal gets "Doppler shifted" (stretched or squeezed) depending on where it is in its orbit. In a long observation, the signal gets smeared out and looks too faint to be real. It's like trying to take a photo of a race car; if you use a slow shutter speed, the car looks like a blur.
• The Breakthrough: The CMA found that this "blur" appeared in 9 different observations. Even though each individual sighting was weak and looked suspicious, the algorithm saw the pattern: Same spin speed, same distance, appearing 9 times.
• The Result: They confirmed it's a real pulsar! It spins 360 times every second (2.76 milliseconds), has a triple-peaked pulse profile (like a lighthouse with three beams), and is dancing in a tight orbit with a companion star.

Why This Matters

This paper isn't just about finding one new pulsar. It's about changing the rules of the game.

• Efficiency: It saves astronomers hours of staring at computer screens, filtering out the "fake" noise automatically.
• Recovery: It rescues "faint" or "weird" pulsars that were previously thrown away because they didn't look perfect in a single snapshot.
• Future: The authors suggest that if we apply this "Matchmaker" to all the old data we have, we might find hundreds more hidden pulsars that are currently hiding in plain sight.

In short: They built a smart filter that connects the dots between different days of observation, allowing them to hear the faintest whispers of the universe that were previously drowned out by the noise.

Technical Summary

Here is a detailed technical summary of the paper "The FAST Discovery of a binary millisecond pulsar PSR J1647-0156B (M12B) with a candidate cross matching algorithm."

1. Problem Statement

Large-scale radio pulsar surveys, particularly those conducted with highly sensitive telescopes like the Five-hundred-meter Aperture Spherical radio Telescope (FAST), generate massive volumes of data. Standard search pipelines (using FFT, single-pulse searches, or FFA) analyze each observation epoch independently.

• The Challenge: This independent analysis leads to a dramatic expansion in the number of candidates. Weak pulsar signals or those with unusual pulse profiles (e.g., due to scintillation or orbital motion) often fall below the Signal-to-Noise Ratio (SNR) thresholds in individual observations and are discarded during visual inspection.
• The Specific Issue: In Globular Cluster (GC) surveys, the same sky location is often observed dozens of times. Existing methods fail to correlate faint, repeated detections across different epochs, leading to missed discoveries. Furthermore, distinguishing true pulsar candidates from Radio Frequency Interference (RFI) remains difficult when signals are weak.

2. Methodology: The Candidate Cross Matching Algorithm (CMA)

The authors developed a post-processing algorithm designed to sift candidates from multiple observations of the same source.

• Core Logic: The CMA aggregates candidate lists from different observational epochs (generated by PRESTO's ACCEL_sift.py). It compares every candidate against all others to identify pairs that likely correspond to the same physical source.
• Matching Criteria: A candidate pair is considered a match if:
  1. Spin Period ($P$): The relative difference is $\le 1\%$ ($\Delta P/P \le 0.01$).
  2. Dispersion Measure ($DM$): The relative difference is $\le 10\%$ ($\Delta DM/DM \le 0.10$).
• Implementation:
  • Candidates are grouped using a hash mapping dictionary (Python collections.defaultdict).
  • A diagnostic scatter plot of $\Delta DM/DM$ vs. $\Delta P/P$ is generated. Genuine pulsars cluster tightly around the origin (0%, 0%), whereas RFI shows significant scatter in the parameter space.
• Threshold Justification:
  • The 1% period threshold accounts for orbital motion in binary systems (tested against 438 known binary pulsars).
  • The 10% DM threshold balances the need to capture interstellar medium fluctuations and noise while excluding RFI (which often has erratic DM values).

3. Key Contributions

1. Algorithm Development: The creation of a robust CMA that leverages multi-epoch data to recover faint candidates missed by single-epoch analysis.
2. Validation: The method was tested on archival FAST data covering five Globular Clusters (M2, M3, M12, M15, NGC 6517), successfully re-identifying known pulsars that had previously been missed or only detected once.
3. New Discovery: The application of CMA led to the discovery of a new binary millisecond pulsar, PSR J1647-0156B (M12B).

4. Results

• Discovery of PSR J1647-0156B (M12B):
  • Location: Globular Cluster M12 (NGC 6218).
  • Parameters:
    • Spin Period ($P$): $2.76$ ms.
    • Dispersion Measure ($DM$): $42.70 \pm 0.05$cm$^{-3}$ pc.
    • The DM is nearly identical to the known pulsar M12A ($42.50$cm$^{-3}$ pc), confirming M12B's membership in M12.
  • Characteristics: The pulse profile exhibits three narrow peaks and shows scintillation.
  • Binary Nature: The pulsar is in a binary system with an orbital period ($P_b$) of approximately 0.53 days and a nearly circular orbit ($e \approx 0$). The projected semi-major axis is $x \approx 0.73$ light-seconds.
  • Companion Mass: Estimated minimum and median companion masses are $0.15$and$0.18 M_{\odot}$, respectively (assuming a$1.4 M_{\odot}$ pulsar), consistent with Low-Mass X-ray Binary (LMXB) evolution.
• Why it was missed previously: M12B has a short orbital period (~0.5 days). In standard searches using full datasets (lasting >2 hours), the pulsar's signal was smeared out due to orbital acceleration, making it too faint to pass SNR thresholds. It was only detectable in segmented searches (0.5–1 hour), where it appeared too faint to be confirmed individually. The CMA successfully aggregated these 9 faint, segmented detections.
• Performance:
  • The CMA reduced the candidate volume for visual inspection significantly.
  • It successfully recovered 6/10 known pulsars in M2 and 10/15 in M15 that were previously missed or only detected once.
  • It identified M12B, which appeared in candidate lists 9 times but was ignored in prior analyses.

5. Significance

• Efficiency in Pulsar Searches: The CMA provides a scalable solution for handling the "big data" challenge in modern radio astronomy. It transforms the problem of "too many candidates" into an advantage by using multiplicity as a filter for authenticity.
• Recovery of Faint Sources: The method is particularly effective for detecting faint, scintillating, or highly accelerated binary pulsars that standard pipelines discard.
• Future Applications: The authors suggest that applying this algorithm to other archival GC observations could significantly increase the pulsar census in dense stellar environments, potentially revealing more binary systems and testing stellar evolution models.
• Technical Validation: The study confirms that the 1% period and 10% DM tolerance thresholds are optimal for distinguishing real pulsars from RFI in binary systems, offering a standardized approach for future surveys.