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

TL;DR

This work tackles the challenge of measuring stellar rotation periods in crowded open clusters with TESS by leveraging FFIs and Gaia-based vetting. The authors extract and quality-filter light curves for NGC 3532, using Lomb–Scargle periodograms across three TESS cycles and correcting for systematics, source confusion, and half-period harmonics. They achieve a robust expansion of the cluster's rotation census, recovering 69% of the literature periods in at least one cycle and producing 885 reliable $P_{ m rot}$ measurements (including 706 new ones), with high agreement (up to 86–87%) for the most confident measurements when compared to Fritzewski et al. 2021. The study demonstrates that, with appropriate caution, TESS data can yield meaningful rotation measurements in crowded fields, enabling expansive gyrochronology studies across ages and metallicities, and lays groundwork for applying this approach to other clusters in future TESS data releases.

Abstract

The Transiting Exoplanet Survey Satellite (TESS) has observed nearly the entire sky, producing full-frame images (FFIs) every 30 min (Cycles 1$-$2), 10 min (Cycles 3$-$4), and now 200 s (Cycle 5+), over 27-day sectors. Light curves extracted from FFIs can be used to measure stellar rotation periods ($P_{\rm rot}$) in nearby open clusters, and are well-suited for studying low-mass stars ($\lesssim$1.2 M$_\odot$) younger than $\approx$1 Gyr, whose $P_{\rm rot}$ are generally still $\leq$15 days. A challenge to exploiting TESS data fully is its 21$''$ pixel size, which can cause strong signals from a source to contaminate the signals of nearby sources in the crowded environments found, e.g., in the more distant and/or richest clusters. We conducted a test with the young ($\approx$350 Myr old), moderately distant (470 pc), and rich open cluster NGC 3532 ($N_\star$ > 3000), which has an extensive $P_{\rm rot}$ catalog from ground-based photometry, to examine the reliability of $P_{\rm rot}$ obtained from TESS data. We recovered 69% of the literature periods from at least one of the three TESS cycles in which NGC 3532 was observed before any quality analysis. We then used all available TESS data for low-mass members of NGC 3532 and, applying a set of quality cuts that combined information from TESS and from Gaia, measured $P_{\rm rot}$ for 885 cluster stars, adding 706 new $P_{\rm rot}$ to the existing catalog. We conclude that, when considered with appropriate caution, TESS data for stars in crowded fields can yield reliable $P_{\rm rot}$ measurements.

Figures (15)

• Figure 1: Comparison of a TESS Sector 10 FFI 40$\times$40 pixel cutout (left) and ground-based DSS image (right) for a representative cluster target. The red plus symbols are the 159 Hunt2023 cluster members in the image, and the blue diamonds are 32 variables included in the Gaia DR3 variability catalog GaiaDR3_Variability_method. Clearly, the TESS pixels are large enough to include contributions from several sources, including background variables.
• Figure 2: Gaia CMD for NGC 3532 based on the Hunt2023 catalog of 3414 members. The absolute magnitudes were calculated using inverted Gaia DR3 parallaxes; we applied a reddening $A_V$ = 0.093 mag cummings2018. The red circles are the 276 stars in the Hunt2023 catalog with fritz2021_periods$P_{\hbox{\scriptsize rot}}$ measurements. The blue circles are the 1358 low-mass, main-sequence cluster members for which we have TESS light curves and that met our selection criteria for making new $P_{\hbox{\scriptsize rot}}$ measurements.
• Figure 3: CPDs for NGC 3532 low-mass members with TESS light curves for each TESS cycle in which they were observed before (left) and after (right) our analysis removing erroneous measurements. As in Figure \ref{['fig:cmd']}, blue circles represent our targets for $P_{\hbox{\scriptsize rot}}$ measurement with TESS; red circles represent stars with a $P_{\hbox{\scriptsize rot}}$ measurement that agrees with that from fritz2021_periods within 15%. In each CPD, we omit stars for which we measured $P_{\hbox{\scriptsize rot}} > 15$ days.
• Figure 4: The impact on the Cycle 1 $P_{\hbox{\scriptsize rot}}$ distribution of requiring a minimum power when validating $P_{\hbox{\scriptsize rot}}$ measurements. The left panel is a histogram of the periodogram powers for the Cycle 1 light curves; the orange bins correspond to powers below the 30$^{\rm th}$ percentile for this dataset (0.05 in this case). The right panel is the full distribution of Cycle 1 $P_{\hbox{\scriptsize rot}}$, with the values drawn from measurements with a power below the 30$^{\rm th}$ percentile indicated with orange crosses. Those low-power measurements include many outliers, such as the K stars with $P_{\hbox{\scriptsize rot}}$$>$ 10 days or $<$1 day, and other problematic measurements, such as the strip of $\approx$3 day $P_{\hbox{\scriptsize rot}}$ across the observed color range.
• Figure 5: TESS Cycle 1 CPD depicting our identification of $P_{\hbox{\scriptsize rot}}$ measurements that are the result of systematic errors. Measurements of $P_{\hbox{\scriptsize rot}}$$<$ 15 days with periodograms powers greater than the 30$^{\rm th}$ percentile power for this cycle are plotted as blue circles. $P_{\hbox{\scriptsize rot}}$ measurements ranging from 3.0 to 3.5 days with periodogram powers $<$0.1 are marked with orange crosses and are likely the result of an (unspecified) systematic error in the Cycle 1 data. Setting a $P_{\hbox{\scriptsize rot}}$ range and raising the minimum power required for consideration allows us to remove a number of unreliable measurements missed with the cut illustrated in Figure \ref{['fig:power_threshold']}.