Using Stellar Rotation to Identify Tidally Stripped Members of the Praesepe Open Cluster

The precise astrometry from the Gaia mission has enabled the discovery of extended stellar structures in the Milky Way, including both new structures and tidal features associated with known star clusters (e.g., Kounkel & Covey 2019; Meingast & Alves 2019; Röser et al. 2019). Using Gaia DR2, Röser & Schilbach (2019) identify 389 stars that may have been tidally stripped from the benchmark Praesepe open cluster. Astrometric data alone, however, cannot confirm whether these tidal tails contain true siblings of the remaining cluster core stars.

Because the period distributions of open clusters evolve with age, we can use stellar rotation periods (P_rot) to support or rule out candidate tidal tail members. Stellar rotation has been successfully employed to confirm the existence of and age-date a variety of young stellar structures including stellar streams (e.g., Pisces–Eridanus, Theia 456: Curtis et al. 2019; Andrews et al. 2022) and diffuse coronae surrounding open clusters (e.g., NGC 2516: Bouma et al. 2021). Recently-formed star clusters show a roughly uniform distribution of P_rot between 1 and 10 days for stars with masses 0.1M_⊙ ≲ M_* ≲ 1.3M_⊙ (e.g., NGC 6530 and the Orion Nebula Cluster; Bouvier et al. 2014). This initial distribution evolves so that at a given age, the higher-mass stars in this range tend to follow a single-valued mass–period relationship, while lower-mass stars still have a range of P_rot values. The overall distribution also evolves to slower periods over time. Therefore, if candidate tidal tail members follow the same period distribution as in the cluster core, we can assume that they are the same age as and formed along with the core stars.

To measure rotation periods for possible tidally stripped members of Praesepe, we turn to NASA's Transiting Exoplanet Survey Satellite (TESS) mission. We select 96 candidate tidal tail members from Röser & Schilbach (2019), observed in TESS sectors 7, 20, 21, and 34. Many of our targets were observed in two sectors. We use a Python GUI ⁷ to extract and detrend light curves from the full-frame images using two methods: Causal Pixel Modeling (CPM) and Simple Aperture Photometry (SAP). We use the unpopular package (Hattori et al. 2021) to produce the CPM light curves, and photutils (Bradley et al. 2022) to produce the SAP light curves.

We measure rotation periods using a Lomb–Scargle periodogram before visually inspecting each light curve. One author inspected all 96 targets twice, selecting the best sector(s) as appropriate and assigning a quality score each time. Ambiguous results were then resolved through discussion with a co-author. This left us with 64 light curves that were too noisy or systematics-dominated to measure a period, or which appeared flat with no variability. Thirty-two stars show clear periodic variability, and can be assessed for cluster membership.

Figure 1 shows a color–period diagram comparing our 32 new P_rot to existing P_rot for cluster core stars. The primary sources of these core P_rot are K2 data analyzed by Douglas et al. (2017) and Rampalli et al. (2021). We use the same core membership catalog as Rampalli et al. (2021), which is restricted to stars with membership probabilities P_mem ≥ 70%. Our new TESS P_rot, along with Gaia colors and object identifiers, are provided as "data behind the figure." Our ability to measure longer periods with TESS is limited by the relatively short sector duration, midway data gaps, and the calibration process that seems to suppress longer signals, but overall when we measure a P_rot for a star, it lines up very well with the core distribution. Twenty-eight stars clearly follow the slow-rotator sequence (for (G_BP − G_RP) ≲ 2) or the broad distribution of periods for M dwarfs (G_BP − G_RP ≳ 2). All stars shown are also consistent with the cluster main sequence in a Gaia M_G versus (G_BP − G_RP) diagram. Therefore, we find strong evidence that 28 of our 32 new rotators are truly former members of Praesepe.

Zoom In Zoom Out Reset image size

There are four notable outliers among our TESS targets: three targets with (G_BP − G_RP) ≈ 1.5 and P_rot ≈ 6 days, and a very rapid rotator with (G_BP − G_RP) ≈ 1.4. For the three outliers with P_rot ≈ 6 days, the TESS sector length and mid-sector data downlink prevent us from confidently identifying a P_rot value. Their light curves are also consistent with P_rot ≈ 12 days and a double-dip feature—in this case, the measured P_rot ≈ 6 days would represent a 1/2 harmonic of the true period. The mid-sector data downlink perfectly interrupts the signal every 12 days, meaning we cannot say for certain whether the true P_rot for these stars is 12 days. If our detections represent a half-period harmonic, however, then these stars would be consistent with the slow rotator sequence in the cluster core.

Finally, the star with (G_BP − G_RP) ≈ 1.4 and P_rot ≈ 1 days (Gaia DR2/EDR3 5753188583680600320) is likely a binary star. We measure a P_rot that would be unusually fast for a single star with the same color, which would seem to rule it out as a tidally stripped member. It has a Gaia EDR3 RUWE value of 1.386, however, suggesting that it is an unresolved binary system. The star is also overluminous in the Gaia M_G versus (G_BP − G_RP) diagram. The very rapid rotation suggests it is either a short-period, equal-mass, tidally synchronized binary system, or a hierarchical triple system; in either case, this makes it an interesting target for radial velocity (RV) follow-up. We can neither confirm nor rule out this star as a former Praesepe member without RVs, so for now we simply say that it could be consistent with prior membership in the cluster.

Our study suggests that Röser & Schilbach (2019) were successful in identifying real associates of the Praesepe cluster, and demonstrates the utility of P_rot as supporting information for identifying former cluster members. The stars for which we could measure P_rot are all consistent with the cluster; while P_rot would let us identify some older field stars, we do not find any, indicating that the underlying catalog is reasonably accurate.

To conclude, we present 32 new rotation periods for candidate tidal tail members of the Praesepe open cluster. Twenty-eight of these stars have P_rot completely consistent with the cluster core, and we consider them truly former members. Four other stars are technically outliers: three of them are consistent with the cluster sequence if we double their measured periods, and one is likely a tidally synchronized binary. We cannot rule out these stars as former cluster members at this time, but they could be confirmed with additional data. Based on how many of our targets with P_rot are consistent with the Praesepe cluster core, this supports the idea that these stars are part of the same structure as Praesepe, likely tidal tails (or an extended filament of star formation).

The code used to produce Figure 1 and the "data-behind-the-figure" may be found at https://github.com/jessmcdivitt/Praesepe-Tails and Zenodo (McDivitt & Douglas 2022).

J.M. acknowledges support from the Lafayette College EXCEL Scholars Program. S.T.D. acknowledges support from the National Aeronautics and Space Administration (NASA) under grant No. 80NSSC19K0377 issued through the TESS Guest Investigator program. J.L.C. is supported by NSF AST-2009840 and NASA TESS GI grant No. 80NSSC22K0299 (G04217).

Facilities: TESS - , Gaia. -