🔭 astrophysics

Measuring Rotation Periods in Crowded Star Clusters with TESS: A Proof-of-Concept with NGC 3532

This study demonstrates that despite challenges from TESS's large pixel size in crowded fields, applying quality cuts to data from the open cluster NGC 3532 successfully yields reliable rotation periods for 885 low-mass stars, including 706 new measurements that significantly expand existing catalogs.

Original authors: Matthew S. Stafford, Jason L. Curtis, Marcel A. Agüeros

Published 2026-01-27

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Matthew S. Stafford, Jason L. Curtis, Marcel A. Agüeros

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Listening to a Noisy Crowd

Imagine you are at a massive, crowded music festival (the star cluster NGC 3532). You want to listen to the specific rhythm of one singer (a star) to figure out how fast they are spinning.

You have a high-tech microphone (TESS, the space telescope) that can hear almost the entire festival. However, there's a catch: your microphone has a very wide "ear." Instead of hearing just one singer, it picks up the voices of everyone standing within a large circle around them. In a crowded festival, this means your recording is a messy mix of the singer you want and all their neighbors.

This paper is a "proof-of-concept" test. The authors asked: Can we still figure out the true rhythm of a specific singer in this noisy, crowded crowd, or is the data too messy to use?

The Problem: The "Big Pixel" Blunder

Previous telescopes (like Kepler) had tiny, precise ears that could isolate a single singer. TESS, however, has "big ears" (large pixels).

The Analogy: If you try to take a photo of a single flower in a dense garden with a camera that has a very wide lens, you end up capturing the flower plus the bushes, the fence, and the people walking behind it.
The Result: In a crowded star cluster, the light from a bright, spinning neighbor can "bleed" into the data of the star you are studying, making it look like the wrong star is spinning at the wrong speed.

The Experiment: A Test Run in NGC 3532

The authors chose NGC 3532 as their test case because it is:

Crowded: It has over 3,000 stars packed together.
Distant: It's far enough away that the "bleeding" effect is strong.
Well-Known: Scientists already had a "gold standard" list of rotation speeds for these stars from ground-based telescopes. This gave the authors a cheat sheet to check their work.

They used TESS data from three different time periods (Cycles 1, 3, and 5) to create light curves (records of how bright the stars get and dim over time).

The Solution: The "Noise-Canceling" Filter

The authors didn't just accept the messy data. They built a sophisticated set of filters to clean it up, acting like a detective sorting through clues:

The "Volume" Check: They threw out any rhythm that was too quiet (low signal power), assuming it was just random noise.
The "System Error" Hunt: They noticed some stars all seemed to be spinning at exactly 3 days or 5 days, regardless of what they actually were. They realized this was a glitch in the data processing, not real stars, and removed them.
The "Neighbor" Check (Source Confusion): This was the most important step. They looked at the neighborhood of every star.
- The Analogy: If you hear a loud drumbeat in your recording, but you see a drummer standing right next to the singer you are trying to record, you know the drumbeat belongs to the neighbor, not the singer.
- They cross-referenced their data with a massive catalog of known variable stars (Gaia). If a bright, fast-spinning neighbor was nearby, they flagged the measurement as "contaminated" and discarded it.
The "Harmonic" Fix: Sometimes, a spinning star creates a rhythm that looks like it's spinning twice as fast as it actually is (like a fan blade creating a flicker). They used math to spot these "half-speed" tricks and corrected them.

The Results: A New, Bigger List

After all this cleaning, the results were impressive:

Recovery Rate: Before cleaning, they matched the known "gold standard" rhythms 69% of the time. After cleaning, the agreement jumped to 77%. For the most reliable stars, they matched 86–87% of the time.
Expansion: The previous list had rotation speeds for 279 stars. The authors added 706 new stars to the list, bringing the total to 885.
The Takeaway: Even though TESS has "big ears" and the crowd is noisy, if you use the right filters and check your neighbors, you can still get a reliable rhythm for the stars.

The Conclusion

The paper concludes that TESS is a powerful tool for studying crowded star clusters, but you can't just take the raw data at face value. You have to be a careful editor, removing the "noise" from neighbors and fixing the "echoes" of harmonics.

By doing this, they successfully created one of the largest rotation catalogs for an open cluster, proving that we can study how stars spin even in the busiest parts of the galaxy, provided we are careful about who is standing next to whom.

Problem Statement
The Transiting Exoplanet Survey Satellite (TESS) offers the unique capability to measure stellar rotation periods ( $P_{\text{rot}}$ ) for low-mass stars across nearly the entire sky. However, TESS's large pixel scale ( $21''$ ) poses a significant challenge in crowded stellar fields, such as distant or rich open clusters. In these environments, photometric contamination from neighboring variable sources can severely compromise the reliability of rotation period measurements extracted from Full-Frame Images (FFIs). While TESS data has been successfully used for isolated fields, its application to crowded clusters requires robust methods to distinguish true stellar rotation signals from blended variability.

Methodology
The authors conducted a proof-of-concept study using the young ( $\approx 350$ Myr), moderately distant ($470$ pc), and rich open cluster NGC 3532 ( $N_\star > 3000$ ). The study leveraged TESS data from Cycles 1, 3, and 5, which provided observations with cadences of 30 min, 10 min, and 200 s, respectively.

Data Assembly: Light curves were constructed for 1,358 low-mass members of NGC 3532 (selected from the E. L. Hunt & S. Reffert 2023 membership catalog) using 40 $\times$ 40 pixel FFI cutouts. The authors employed Causal Pixel Modeling (CPM) via the unpopular Python package to remove time-varying systematics.
Period Detection: Lomb-Scargle (LS) periodograms were generated for each target across all available cycles. Both one-term and two-term LS periodograms were calculated to identify and correct for harmonic aliases (e.g., half-period harmonics).
Quality Control and Filtering: A multi-step quality assessment was applied to remove spurious measurements:
- Power Thresholds: A minimum periodogram power threshold (30th percentile per cycle) was applied to eliminate weak signals.
- Systematic Error Removal: Specific systematic artifacts (e.g., a cluster of $\approx 3$ -day periods in Cycle 1 and 5–7 day periods in Cycle 3) were identified and removed based on period and power criteria.
- Source Confusion Mitigation: The authors cross-matched their targets with the Gaia DR3 Variability Catalog (GVC) and constructed light curves for potential contaminating cluster members. Any period measurement matching a nearby variable source (within 2') was flagged and removed if the contaminant's signal was stronger or if the period was inconsistent across cycles.
- Harmonic Correction: Measurements identified as half-period harmonics were corrected using two-term LS periodograms and cross-cycle consistency checks.
Validation: The resulting catalog was compared against the existing ground-based photometric catalog by D. J. Fritzewski et al. (2021a), which contained 279 rotation periods for the cluster.

Key Contributions and Results

Recovery Rate: Before applying any quality cuts, the authors recovered 69% of the literature periods (from Fritzewski et al. 2021a) within 15% in at least one TESS cycle.
Expanded Catalog: After applying rigorous quality filters, the authors produced a new rotation catalog for NGC 3532 containing 885 stars. This represents an addition of 706 new rotation periods to the existing literature.
Quality Classification: The new catalog classifies periods into three quality levels based on the number of TESS cycles confirming the measurement:
- Quality 3: Confirmed in all three cycles (261 stars).
- Quality 2: Confirmed in two cycles (309 stars).
- Quality 1: Measured in only one cycle (315 stars).
Agreement with Literature:
- The overall agreement between the final TESS-based catalog and the Fritzewski et al. (2021a) catalog is 77%.
- Agreement is highest for the most confident literature measurements (Class 1), reaching 86%.
- For the authors' own highest-confidence measurements (Quality 3), the agreement with Fritzewski et al. (2021a) is 87%.
Diagnostic Capabilities: The study successfully identified and removed measurements contaminated by background variables (e.g., a Cepheid contaminating nearby cluster members) and corrected harmonic aliases. The analysis revealed that while the method is robust, it struggles with faint, slow rotators where contamination and low signal-to-noise ratios persist.

Significance
The paper concludes that TESS data, when processed with appropriate caution and specific quality filters combining TESS and Gaia data, can yield reliable rotation period measurements even in crowded stellar fields. The authors emphasize that their approach does not rely on pre-existing rotation catalogs, making it applicable to clusters without prior period measurements. By tripling the sample size of rotation periods for NGC 3532, this work provides a substantial dataset for calibrating gyrochronology models for low-mass main-sequence stars. The authors note that while some inaccuracies remain (particularly for slow rotators and faint stars), the methodology offers a viable path forward for expanding rotation catalogs to other distant and crowded open clusters observed by TESS.