A KELT-TESS Eclipsing Binary in a Young Triple System Associated with a "Stellar String" Theia 301
Joni-Marie Clark Cunningham, Dax L. Felix, Don M. Dixon, Keivan G. Stassun, Robert J. Siverd, George Zhou, Thiam-Guan tan, David James, Rudolf B. Kuhn, Marina Kounkel
aa r X i v : . [ a s t r o - ph . S R ] S e p Draft version September 9, 2020
Typeset using L A TEX default style in AASTeX63
A KELT-
TESS
Eclipsing Binary in a Young, Triple System Associated with the Local “StellarString” Theia 301
Joni-Marie C. Cunningham, Dax L. Feliz, Don M. Dixon, Joshua Pepper, Keivan G. Stassun, Robert J. Siverd, George Zhou, Daniel Bayliss, Thiam-Guan Tan, Phillip Cargile, David James, Rudolf B. Kuhn,
6, 7 and Marina Kounkel Vanderbilt University, Department of Physics & Astronomy, 6301 Stevenson Center Ln., Nashville, TN 37235, USA Department of Physics, Lehigh University, Bethlehem, PA 18015, USA Observatoire de Gen`eve Perth Exoplanet Survey Telescope Cerro Tololo Inter-American Observatory, Colina El Pino, S/N, La Serena, Chile South African Astronomical Observatory, P.O. Box 9, Observatory 7935, Cape Town, South Africa Southern African Large Telescope, P.O. Box 9, Observatory 7935, Cape Town, South Africa Western Washington University
AbstractHD 54236 is a nearby, wide common-proper-motion visual pair that has been previously identifiedas likely being very young by virtue of strong X-ray emission and lithium absorption. Here we reportthe discovery that the brighter member of the wide pair, HD 54236A, is itself an eclipsing binary (EB),comprising two near-equal solar-mass stars on a 2.4-d orbit. It represents a potentially valuable oppor-tunity to expand the number of benchmark-grade EBs at young stellar ages. Using new observations ofCa II H&K emission and lithium absorption in the wide K-dwarf companion, HD 54236B, we obtain arobust age estimate of 225 ±
50 Myr for the system. This age estimate and
Gaia proper motions showHD 54236 is associated with Theia 301, a newly discovered local “stellar string”, which itself may berelated to the AB Dor moving group through shared stellar members. Applying this age estimate toAB Dor itself alleviates reported tension between observation and theory that arises for the luminosityof AB Dor C when younger age estimates are used.
Keywords: stars: low-mass — stars: eclipsing binaries — stars: fundamental parameters — stars:stellar associations INTRODUCTIONEclipsing binary (EB) stars are foundational to stellar astrophysics by virtue of providing all of the fundamentalphysical properties of stars to high accuracy, with which to test and benchmark theoretical stellar evolution models,and to probe the ages of interesting Galactic populations. In a review of the field, Torres et al. (2012) identifiedapproximately 100 EBs whose parameters have been determined with sufficient accuracy to serve as stringent modelbenchmarks. Only one of these benchmark-grade EBs includes stars still in the pre–main-sequence (PMS) phase ofevolution (V1174 Ori; Stassun et al. 2004). In addition, only about 25 of these EBs have measured metallicities, andan even smaller number have independent age constraints (as provided by, e.g., membership in clusters or movinggroups), limiting the ability of even these benchmark-grade EBs to serve as the most stringent tests of stellar models.Additional benchmarks for early stellar evolution are critical in building our understanding of stellar formation.A number of ground and space based exoplanet transit surveys are operating that find EBs in large numbers, eitheras unintentional false positives or as intentional targets. The KELT survey (Pepper et al. 2007, 2012), has so faridentified four bright transiting exoplanets (Siverd et al. 2012; Beatty et al. 2012; Pepper et al. 2012; Collins et al.2014), in conjunction with a very large number of EBs (e.g. Pepper et al. 2008).In the course of commissioning the KELT-South telescope we identified the bright, X-ray source HD 54236A as anEB. It is listed in the SIMBAD database as a PMS star and listed in the SACY catalog as possessing Li in its spectrum,a diagnostic of stellar youth. It also possesses a wide common-proper-motion (CPM) companion, HD 54236B, whichin turn provides the opportunity for independent checks on key parameters such as system age. Most recently, the
Cunningham et al.
Transiting Exoplanet Survey Satellite (
TESS ) observed the HD 54236 system (HD 54236A has TIC ID 238162238 inthe TESS Input Catalog; Stassun et al. 2019).At the same time, the advent of
Gaia has revolutionized our understanding of stellar multiples (e.g.,Oelkers & Stassun 2018), stellar associations (e.g., Oh et al. 2017), and the 3D structure of the solar neighborhood(e.g., Kounkel & Covey 2019). Indeed, very recent investigations of the full phase-space structure of stars in the solarneighborhood reveal previously unrecognized “stellar strings”, long filamentary groups of young stars that appear toencode the large spiral-arm structures from which they formed (Kounkel et al. 2020). Thus, HD 54236 represents avaluable opportunity to expand the number of known benchmark-grade EBs at young stellar ages, and to test recentsuggestions of young “stellar strings” in the solar neighborhood.In § § § II H&K activity. Finally, in § § THE HD 54236 SYSTEMHD 54236A is a V =9.26 object in the constellation Puppis. It was identified as a member of a common proper motionbinary in the first CCDM catalog (Soderblom et al. 1993). It is also listed as a visual binary in the Washington VisualDouble Stars catalog (Mason et al. 2001). It was identified as a possible pre-main sequence (PMS) object from theSACY survey (Torres et al. 2006): first through its detection as a strong X-ray source, and then through Li observedin its spectrum. Torres et al. (2006) classified HD 54236A as spectral type G0V, and HD 54236B as spectral typeK7V.The total proper motion is 17 . ± .
22 mas yr − for both HD 54236 A and B (values from the Gaia archiveGaia Collaboration et al. 2018), and the equivalent width (EW) of the Li line is given in the SACY catalog as 120 m˚Afor A and 40 m˚A for B. No uncertainties are provided in the SACY catalog for the Li EW values. Below we reportupdated precise measurements of the Li EW and abundances for the stars.HD 54236A is in a fairly crowded part of the sky, and thus all our photometry to date is blended with nearby starsat some level. We will pay special attention in § ∼
21 AU. Surrounding, unrelated fieldstars are also identified. This observational ecosystem includes: • HD 54236A ( V =9.26), spectral type ∼ G0, which is the principal object of study in this paper and which weidentify as an eclipsing binary. • HD 54236B, which is 6.5 arcseconds south of HD 54236A and is V =13.2, spectral type ∼ K7. It is a commonproper-motion companion of HD 54236A, and we confirm below that it shares the systemic radial velocity ofHD 52236A also. • The bright star HD 54262, which is 54 arcseconds east of HD 54236A and is V =9.31. Based on its reportedproper motion of µ α = − . ± .
06 mas yr − and µ δ = − . ± .
06 mas yr − , compared to the proper motionof the HD 54236 system of µ α = − . ± .
06 mas yr − and µ δ = − . ± .
05 mas yr − ( Gaia ), it appearscertain that HD 54262 is not associated with the HD 54236 system. • A very distant ( ∼ V = 14 .
6) that is 7.4 arcseconds to thenorth of HD 54236A (TIC 767642478) is not physically associated with the HD 54236 system.Thus, to our knowledge, the HD 54236 system physically comprises three stars: HD 54236A which we identify belowas an eclipsing binary (two stars, unresolved) and HD 54236B as a wide CPM tertiary companion that is visuallyresolved with a separation from HD 54236A of 6.5 arcseconds. oung EB Triple System in Stellar String Theia 301
30” 2.065’ x 1.679'
E N
Unrelated Field StarsHigh Proper MotionUnrelated Star Unrelated StarK7V neighbor G07 EBCommon-proper motion of HD54236 system
Figure 1.
DSS image of the HD 54236 tertiary system and its common-proper-motion is highlighted in fushcia. The tertiarysystem includes HD 54236A (the eclipsing binary), and HD 54236B, the CPM tertiary member ∼ DATA3.1.
Photometric Observations
KELT-South Photometry
The Kilodegree Extremely Little Telescope-South (KELT-South) is a dedicated exoplanet transit telescope locatedat the Sutherland observing station of the South African Astronomical Observatory (SAAO). The telescope, hardware,and the primary exoplanet survey is described in Pepper et al. (2012). The KELT observations are taken with a KodakWratten R -band.KELT-South conducted a commissioning run from Jan 4, 2010, to Feb 19 2010. During that run, KELT-Southcontinuously observed a field centered at α = 08:16:00, δ = -54:00:00, with repeated 30 s exposures. After removal ofpoor-quality images, we have a total of 3,041 images across 29 individual nights. We generated relative photometryfrom flat-fielded images using a modified version of the ISIS image subtraction package (see also Alard & Lupton 1998;Alard 2000; Hartman et al. 2011), in combination with point-spread function fitting using the stand-alone DAOPHOTII (Stetson 1987, 1990) package, and the SExtractor program (Bertin & Arnouts 1996). A more complete explanationof our pipeline procedures, and the algorithms, are provided in Siverd et al. (2012).One of the transit candidates was HD 54236A. Although the depth and out-of-eclipse variation of the lightcurveindicated that the target was unlikely to be a transiting planet, the prospect that it represented a bright PMS eclipsingbinary prompted us to gather additional data of this target. The KELT light curve of HD 54236A is shown in Fig. 2. Cunningham et al. R e l a t i v e M a g n i t u d e Figure 2.
Phase-folded lightcurve of HD 54236A from KELT. Small symbols represent individual measurements, red linesrepresent the binned light curve.
Because of the very wide-field nature of the KELT observing setup, with the accompanying large pixel scale (23arcseconds per pixel), the KELT lightcurve of HD 54236A (Fig. 2) is blended with all of the other stars identifiedin Figure 1. The largest component of blended flux comes from HD 54262, which is about 0.33 mag fainter thanHD 54236A. Therefore, the eclipse depths seen by KELT are at least 40% shallower than the true depths.An initial periodicity analysis yielded a most likely orbital period of ∼ TESS light curve (see below). 3.1.2.
PEST Photometry
One of the partners of the KELT-South survey is the Perth Exoplanet Survey Telescope (PEST), operated by T.G.Tan. PEST consists of a 12” Meade LX200 SCT f/10 telescope, with a focal reducer yielding f/5. The camera is anSBIG ST-8XME, with 1.2 arcsec/pixel.PEST observed HD 54236A on the night of April 25, 2012, for 119 minutes around the predicted time of primarytransit in the Cousins R -band, although the ephemeris was poor given the elapsed time between the original observa-tions and the PEST observations (over 2 years). The observations caught the last hour of the primary eclipse egressplus an hour after the eclipse. This light curve is more pure than the discovery KELT light curve in that it excludeslight contamination from HD 54262. The PEST light curve does however still include light contamination from thefaint visual companion HD 54236B. The raw photometry of the combined HD 54236A/B system was extracted usingthe C-Munipack software package (written by David Motl) to perform aperture photometry. Relative photometrywas derived using a set of three nearby comparison stars via a custom-written program.The PEST light curve is shown in Figure 3, together with the KELT data scaled by a factor of 3 in flux to accountfor the additional dilution of the KELT data by the high proper motion bright star HD 54262. The addition of thePEST photometry allowed us to refine the orbital period from the KELT data because of the expanded time baseline.3.1.3. TESS Photometry
TESS observed HD 54236 (TIC 238162238) in Sectors 6, 7, and 8 from December 15th, 2018, to February 27th, 2019.The 2-minute cadence light curves were processed using the Lightkurve Python module (Lightkurve Collaboration et al. http://c-munipack.sourceforge.net/ oung EB Triple System in Stellar String Theia 301 Phase −0.04−0.020.000.020.040.06 R e l a t i v e M a g n i t u d e Figure 3.
Phase-folded lightcurve of HD 54236A from KELT ( black points ) with the follow-up data from PEST over-plotted( red points ). Since the KELT data includes the flux from the unrelated high proper motion star HD 54262, the KELT data aremultiplied by a factor of 3 to match the egress slope from the PEST data.
TIC 238162238 Target Pixel File cutouts
Figure 4.
TESS
Target Pixel File cutouts of HD 54236 for Sectors 6, 7, and 8. The red aperture mask is the optimal aperturefrom the SPOC pipeline, chosen to avoid the distant faint star to the north, TIC 767642477, from blending the photometry.
To estimate the background, we required a brightness threshold such that the pixels in the Target Pixel File that are0.1% times the standard deviation below the overall median flux of pixels in the selected aperture. We then subtractedthe background from our source signal and normalized by the median flux.To remove long-term trends and to assist with normalization of the flux baselines between sectors and satellitedownlink times, we utilized a Biweighted Midcorrelation filter (via the Wotan package; M. et al. 2019). We chose awindow size of 3 times the ∼ TESS light curve with the trend line that is fit to the flux baseline in red. Sincethe Biweighted Midcorrelation filter is median-based, it is less sensitive to outliers and ignores the transit events. InFigure 8, we show the final phase-folded
TESS light curve on the orbital period estimated above from the
TESS lightcurve. We report a final system ephemeris combining all of the available light curve data in Section 4.1.
Cunningham et al. N o r m a li z e d R e l a t i v e F l u x N o r m a li z e d R e l a t i v e F l u x TIC 238162238 TESS Sectors 6, 7 and 8 Photometry
Figure 5.
Left: TESS light curve of HD 54236 in Sectors 6, 7, and 8.
Right:
Phase-folded
TESS light curve. B L S P o w e r TIC 238162238 BLS analysis: zoom in on 2.43 +/- 0.1 days
BLS top peak period: 2.4377 daysOut of Eclipse Variation: 2.3782 days
Figure 6.
BLS periodogram, showing the most prominent periodic features, corresponding to the eclipse period but alsoshowing a similar but different period of 2.3782 d which we attribute to the stellar rotation period (see the text).
Spectroscopic Observations
Spectroscopic Observations of the HD 54236 were acquired from the Wide Field Spectrograph (WiFeS) and echellespectrograph on the ANU 2.3 m telescope from the Australian National University , and, the Goodman High Through-put Spectrograph (Clemens et al. 2004), installed on the f/16.6 Nasmyth platform of the 4.1 m SOuthern AstrophysicalResearch (SOAR) telescope. In what follows, each instrument and the analysis of its spectroscopic observations aredescribed. 3.2.1. ANU 2.3m Spectroscopy
We obtained a single 100-sec exposure of HD 54236A using WiFeS (Dopita et al. 2007) on the ANU 2.3 m telescopeat Siding Spring Observatory, Australia. WiFeS is an image slicer integral field spectrograph, where the B3000 gratingand the RT560 dichroic were selected for our observations, giving the spectral coverage of 3500–6000˚A at a resolution of λ/ ∆ λ = 3000. The object spectrum was extracted and reduced with the IRAF packages CCDPROC and KPNOSLIT,with the wavelength solution provided by a Ne-Ar arc lamp exposure taken on the same night. The spectrum is flux http://rsaa.anu.edu.au/observatories/siding-spring-observatory IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research inAstronomy (AURA) under cooperative agreement with the National Science Foundation. oung EB Triple System in Stellar String Theia 301 N o r m a li z e d R e l a t i v e F l u x TIC 238162238 First 10 days of data
Figure 7.
The first 10 days of the
TESS light curve spanning Sectors 6 to 8, shown in
TESS
Julian Days (TJD). The blackpoints are the
TESS data and the red line is the Biweighted Midcorrelation filter applied to the flux baseline. We used a windowsize of 6 hours which equals to 3 times the transit duration reported by BLS. N o r m a li z e d R e l a t i v e F l u x N o r m a li z e d R e l a t i v e F l u x TIC 238162238 TESS Sectors 6, 7 and 8 Photometry
Figure 8.
The phase folded
TESS light curve that was smoothed using a Biweighted Midcorrelation filter as described in § calibrated against exposures of spectrophotometric standard stars from Hamuy et al. (1994), employing techniquesdescribed in Bessell (1999). The resulting flux-calibrated spectrum had signal-to-noise of S/N = 300 per resolutionelement.The reduced object spectrum was fitted to a grid of synthetic spectra from Munari et al. (2005), spaced at intervalsof 100 K in T eff , 0.5 dex in log g , 0.5 dex in [Fe/H], and 0.02 magnitudes in E ( B − V ). Discrepancies between theobject spectrum and the synthetic template are largely due to differences in the flux calibration between them butprevent no analytical obstacle as spectral matching is done based on spectral features, rather than its overall shape.We weighted regions sensitive to log g variations preferentially, including the Balmer jump, the MgH feature at 4800˚A,and the Mg b triplet at 5170˚A. Details of the data reduction and spectral fitting process can be found in § T eff = 6350 ±
100 K, log g = 5 . ± .
3, and [Fe/H] = − . ± . § T eff = 4400 ± g = 4 . ± .
35, [Fe/H] = − . ± .
44 (Bayliss et al. 2013). In addition, from this spectrum we have verified the Li EW of HD 54236A reportedby SACY (120 m˚A). We obtain an EW of 140 ±
20 m˚A for the Li line at 6708 ˚A, representing the blended EW of bothstars in the HD 54236 EB.
Cunningham et al. L o gg KST1C02704 b0043.fits 6353 5.0 -0.5 RMS=0.0378 R M S N o r m a li s e d f l u x data template L o gg KST1C02704S b0043S.fits 4388 4.8 -0.5 RMS=0.0734 R M S N o r m a li s e d f l u x data template Figure 9.
Top:
Spectral analysis of the stellar parameters of HD 54236A, the eclipsing binary.
Bottom:
Spectral analysisof the stellar parameters of the tertiary member, HD 54236B. The observed spectra for both HD 54236A and HD 54236B areshown in blue, while their respective templates modeled in green. Discrepancies between the target spectra and its template arelargely due to differences in flux calibration, however, spectral matching is done based on spectral features rather than overallshape.
The WiFeS spectrograph was also used to measure the radial velocities of the eclipsing stars. One spectrum wasobtained on HJD 2456104.838 using WiFeS at medium resolution ( λ/ ∆ λ = 7000), where both components of thespectroscopic binary were resolved. The observation and data analysis process follows Penev et al. (2013). Thevelocities were measured via cross correlation as described in the next section, and are reported along with the full setof velocities described below in Table 1.3.2.2. High-resolution Spectroscopy: Radial velocities
Nine high-resolution observations of HD 54236A were obtained using the ANU 2.3 m echelle spectrograph, at aresolution of λ/ ∆ λ ≈ − pixel − , in the spectral range 4200–6700˚A, over 20 echelleorders. The data was reduced with the IRAF package CCDPROC, extracted and normalized using ONEDSPEC. Thewavelength solution was provided by Th-Ar arc lamp exposures that bracketed each science exposure. A standard950 sec exposure of the target yields a signal-to-noise of S/N = 40 per resolution element. The instrument setup,observations, and data reduction process are detailed in Zhou et al. (2014).We cross-correlated the object spectra against a series of radial velocity standard star spectra taken during twilighteach night. The cross-correlation peaks for the two spectroscopic components of the object were resolved in theobservations, we were therefore able to extract radial velocities for both components in each observation by fittingGaussians to the two cross-correlation peaks. We take the stronger peak to represent the (more massive) primarycomponent. An example cross-correlation function (CCF) for the observation on HJD 2456057.93945 is shown inFigure 10.The radial velocity is calculated as the mean over all orders not affected by telluric contamination, weighted bytheir respective signal-to-noise ratio; the scatter in the measurements over the orders gives the error in the radialvelocity measurement for that epoch. The complete set of primary and secondary radial velocities determined fromthese observations and are summarized in Table 1. oung EB Triple System in Stellar String Theia 301 Figure 10.
Cross-correlation function for the spectrum taken on HJD 2456057.93945226, with the ANU 2.3 m echelle at R = 23000. The CCF shown is the average CCF of the cross correlations from all echelle orders, excluding those with very lowcross-correlation heights. The cross-correlations are performed against RV standard star exposures taken on the same night. Forclarity, the highest peak (presumed to correspond to higher mass, primary component in the eclipsing binary, has been shiftedin this figure to be at 0 km s − . Table 1.
RV Observations of HD 54236 AHJD Primary RV Secondary RV Source(UTC) (km s − ) (km s − )2456057.93945 − . ± . . ± . − . ± . . ± . . ± . − . ± . . ± . − . ± . . ± . − . ± . . ± . − . ± . − . ± . . ± . − . ± . . ± . . ± . − . ± . . ± . − . ± . − . ± . . ± . SOAR Spectroscopy
HD 54236 A & B were observed on UT 20120605 and UT 20120607 using the Goodman High Throughput Spectro-graph (Clemens et al. 2004), installed on the f/16.6 Nasmyth platform of the 4.1m SOuthern Astrophysical Research(SOAR) telescope. The spectrograph used the 1200 l/mm grating in M5 mode, a 0.46-arcsecond wide slit, a GG-495blue-blocking filter, imaged onto a 4096 × µ m pixels (0.15 arcsecs/pixel).This setup yields a central wavelength of ≃ ≃ ≃ − . Using this setup, HD 54236 A and B were separately observed for exposuretimes of 300- and 900-seconds, respectively, on UT 20120605, and 300- and 600-seconds, respectively, on UT 20120607,resulting in spectra having maximum S/N near to H α of ≃
350 for HD 54236A and 100 for HD 54236B.For each target, we measured heliocentric radial velocities and Li I Cunningham et al. performed using standard IRAF procedures. Heliocentric radial velocities were determined relative to the InternationalAstronomical Union standard stars HR 6468, HR 6859, HD 120223 and HD 126053. Cross-correlation of the velocitystandards against one other shows that the zero-point relative to the standard system is accurate to ∼ . − . TheCCF is the average from all echelle orders, excluding those with very low cross-correlation heights and are performedagainst RV standard star exposures taken on the same night. The radial velocity standard stars on these nights showsthat the system is stable to ≃ − over the course of the night.The radial velocities for the eclipsing components of HD 54236A are reported with the other radial velocity mea-surements in Table 1. For the single star HD 54236B, we obtain a systemic velocity of − . ± . − . As shownbelow, this velocity is consistent with the systemic velocity of the HD 54236A eclipsing binary system.In order to measure the EW of the Li I ± ± Figure 11.
Region of the Li line from the SOAR spectrum.4.
RESULTSWith the above data in hand for the HD 54236 system, including photometric light curves and spectroscopic mea-surements of the component radial velocities and individual stellar spectra, in this section we determine the orbitalparameters as well as constraints on the system age from consideration of the Li absorption and Ca HK emission. Toremind the reader, the system comprises three stars, two eclipsing components in HD 54236A and a single CPM starin HD 54236B. 4.1.
System Ephemeris
Analysis of all of the available light curve and radial-velocity data together (see §§ TESS light curves, spanning a total time baseline of more than 10 yr,leads to the final orbital ephemeris determination of P orb = 2 . ± . = 2455199 . ± . Orbit Solution
We performed a simultaneous Keplerian orbit fit to the radial velocity measurements of the two eclipsing componentsof HD 54236A. We fit the ANU and SOAR radial velocities simultaneously, without attempting to introduce anypossible systematic offset between the two. Thus, in total we fit 11 primary radial velocities and 11 secondary radialvelocities which are listed in Table 1. The resulting orbit solution is shown in Figure 12 and summarized in Table 2.The orbital eccentricity is consistent with zero, which is not surprising given the very short orbital period, hence forthe remainder of our analysis we assume a circular orbit ( e ≡ oung EB Triple System in Stellar String Theia 301 Table 2.
Orbital Solution of HD 54236 A q e v γ a sin i . ± .
012 0 . ± . − . ± . − . ± .
080 R ⊙ -0.4 -0.2 0.0 0.2 0.4 Phase (P=2.43740d)-100-50050100 R V ( k m / s ) Figure 12.
RV solution for HD 54236A: The primary is modeled with the solid line, and the secondary with the dashed line.
PHOEBE Light Curve Model
The PHysics of Eclipsing BinariEs, (Prsa et al. 2011; Prˇsa et al. 2016) commonly referred to as PHOEBE is an EBmodeling code which enhances upon the Wilson-Devinney (Wilson & Devinney 1971) method with improved modelfidelity. Functions are available through PHOEBE that provide generation of highly characterized synthetic lightcurves given orbital and stellar parameters. We invoke PHOEBE to produce synthetic light curves for model fittingto our detrended TESS light curve, adopting the orbital parameters determined from the orbit solution (Section 4.2),such that the principal fitting parameters are the ratio of effective temperatures ( T ratio ), the sum of the stellar radii( R sum ), and the orbital inclination ( i ). This is appropriate given the circular orbit, where to first order T ratio is set bythe relative eclipse depths and R sum is set by the eclipse durations, moderated by i .We use a Genetic Algorithm (GA) to converge to a model light curve solution. Previous work (see, e.g., Metcalfe1999) has shown GA techniques to be robust for model fitting eclipsing binary light curves and thus inspired a modifiedversion we utilize here. For computational efficiency we downsample the TESS 2-min cadence light curve by a factorof 15 and run PHOEBE using the Legacy backend, which is markedly faster. The fitting algorithm starts with thegeneration of an initial random population of 500 model parameter sets [ T ratio , R sum , i ] in our parameter space, eachof which is encoded as a single object representation referred to as a gene. For each model we set the effectivetemperatures by apportioning the spectroscopically determined average (6350 K ) according to the flux-weighted sumof the two stars and the current iteration of T ratio . For each gene a corresponding model light curve is computed inPHOEBE, and the χ between the observed and model fluxes is calculated.To thoroughly explore our parameter space we propagate this population over the course of 20 generations resultingin a total sample of 10,000 models. Iteration of genes from one generation (parent) to the next (child) is done usingsingle point crossover. The members of each parent generation used in the crossover are chosen via roulette wheeluntil re-population is achieved. The probability of each selection is weighted by the reciprocal of the χ goodness-of-fitmetric. To help maintain diversity between generations (i.e., to avoid local minima) we introduce a 10% chance forrandom mutation within the constraints for the parameters at inception.After sampling is complete we discard the first 15 generations in order to prevent the lack of convergence in previousgenerations from contaminating our result. The fitted parameter distributions of the cleaned sample of 2500 modelsis illustrated in the corner plot (Fig. 13). The distributions are gaussian-like in shape with well determined modalvalues. Thus, we determined the converged solutions as the mean values of these distributions and the uncertainties2 Cunningham et al. as the standard deviations. The final converged model is shown over our TESS light curve in Figure 14 as well as theresiduals between the two. There is evidence of some minor systematics in the residuals of the primary eclipse withan amplitude of ∼ .
1% of the eclipse depth of ∼ T ratio together with the spectroscopically determined flux ratio of 3 . / .
5, and solve Stefan-Boltzmann’s Law to obtain R ratio , which with the fitted R sum yields the final individual radii. Similarly, the fitted T ratio together with the final R ratio and the flux ratio yields the final individual temperatures. Finally, using the final i we obtain the stellar masses via the orbit-solution determined a sin i and q . The final stellar parameters are reportedin Table 3. Tratio = 0.95 +0.02−0.02 .
47 8 .
87 9 .
27 9 .
68 0 . I n c l Incl = 79.48 +0.07−0.10 . . . . Tratio . . . . . R s u m . . . . . Incl . . . . . Rsum
Rsum = 2.13 +0.03−0.00
Figure 13.
Corner plot of fitted parameters for cleaned sample of 2500 model light curves. Values atop distribution are themedian values and the 0.16 & 0.84 quantiles.
Table 3.
PHOEBE Fit Parameters and Solutions of HD 54236A T eff , ± KT eff , ± KR . ± . R ⊙ R . ± . R ⊙ i . ± . ◦ M . ± .
029 M ⊙ M . ± .
027 M ⊙ oung EB Triple System in Stellar String Theia 301 −0.2 0.0 0.2 0.4 0.6Phase0.960.970.980.991.001.01 N o r m a li z e d F l u x −0.2 0.0 0.2 0.4 0.6Phase−0.0100−0.0075−0.0050−0.00250.00000.00250.00500.0075 R e s i d u a l s Figure 14.
Light curve fitting for HD 54236A using GA converged model solution. There remain some small systematics inthe residuals in eclipse, the amplitude of which are of order 1% of the eclipse depths. The final stellar parameters from thissolution are listed in Table 3.
SED Analysis
The radii and effective temperature solutions from the PHOEBE analysis (Table 2 were used in Spectral EnergyDistribution (SED) modeling, fitting only for extinction and distance. This is done to ensure that our T eff solutionscorrectly reproduce the observed shape of the SED, and that the inferred distance from this analysis corresponds tothe known Gaia distance. From Figure 15 the SED fits for HD 54236A (top) and HD 54236B (bottom) show a verygood agreement with our T eff determinations. The Gaia distance to HD 54236 is 131.8 ± ± Gaia distance was adopted in fitting the SED to solve for the stellar radius.The stellar radius combined with the log g yields an estimate for the mass. For the radius we find 0 . ± .
06 R ⊙ anda mass estimate of 0 . ± .
11 M ⊙ . These are fully consistent with the empirical relations of Torres et al. (2012) whogive a similar but more precise mass estimate of 0 . ± .
04 M ⊙ .4.5. System Age
In order to establish an age estimate for the HD 54236 system, we consider two commonly used stellar chronometers,namely the abundance of Li and the strength of chromospheric activity in the Ca II H&K lines. While the Ca II H&K activity indicator will be enhanced in the eclipsing pair HD 54236A because of their short orbital period andtherefore likely enhanced rotation, we utilize the existence of the wide tertiary companion star, HD 54236B, to obtaina clean estimate of the Ca II H&K activity age for the system. This analysis is carried out through the comparison ofCa II H&K spectral features against those of similar effective temperatures from three different open clusters spanningthe range of age estimates covered in 4.5.1. Finally, in an attempt to establish association of the system with othernearby stars of similar age, we consider the space motion of the HD 54236 system. For context, in Fig. 16 we showthe components of the HD 54236A system in the T eff –Radius plane compared to the Yonsei-Yale stellar evolutionary4 Cunningham et al. λ ( µ m)-13-12-11-10-9-8 l og λ F λ ( e r g s - c m - ) λ ( µ m)-13-12-11-10 l og λ F λ ( e r g s - c m - ) Figure 15.
Top:
SED fit of HD 54236A
Bottom:
HD 54236B SED fit. Red symbols are observed broadband fluxes, curves arethe best-fit model atmospheres, and blue symbols are the model fluxes for comparison to the observed fluxes. tracks (Yi et al. 2001) for the spectroscopically determined [Fe/H] of − .
5. We also represent all three stars—theeclipsing components of HD 54236A and the tertiary HD 54236B—in the H-R diagram plane compared to the PMSevolutionary models of Baraffe et al. (1998). All of which suggest an age near the ZAMS or slightly younger, consistentwith the other age diagnostics discussed above. 4.5.1.
Lithium
The observed Li EWs for the three stellar components in the HD 54236 system are indicative of a relatively youngsystem age intermediate between the youngest pre-main-sequence clusters and older main sequence clusters. Forexample, from the analysis of Li EW in young stars of various ages, Aarnio et al. (2008) finds that young stars withK spectral type and ages of ∼
50 Myr have Li EW of 100–200 m˚A, somewhat larger than what we observe in theHD 54236B star (spectral type late K). Therefore it is likely that the HD 54236 system has an age somewhat olderthan 50 Myr. To determine the Li inferred age of the system more precisely, we analyzed the observed Li EWs for thethree stars in HD 54236 in the context of other young clusters.First, we corrected the observed EWs for the two eclipsing stars to account for the mutual dilution of the two stars’spectra in our combined light observations. Based on the stars’ final T eff values from our PHOEBE modeling and SEDfits, we converted the Li EW into Li abundances, A ( Li ), using standard curve-of-growth techniques (Soderblom et al.1993) and applying a NLTE correction (Carlsson & Stein 1994).The sample from Sestito & Randich (2005) contains 22 clusters spanning a range of ages from ∼ ∼ A ( Li ) for several of these open clusters ranging in age from 35 Myr to 1.2 Gyr andfollow the same curve-of-growth and NLTE correction in our abundance calculation method for the HD 54236 system.The stellar censuses for our comparison clusters shown in 17 are as follows, in the young clusters with ages less than ∼
100 Myr: NGC 2547 contains 7 stars, IC 2391 has 6, IC 2602 contains 31 and α Per contains 39 stars. In the ZAMSclusters, with ages ∼
100 Myr are the Pleiades with its seven sisters, Blanco 1 which consists of 17 stars, and M35 witha total of 27. The intermediate comparison clusters have ages of ∼ oung EB Triple System in Stellar String Theia 301 R a d i u s [ R s un ] R a d i u s [ R s un ] l og L / L s un Figure 16.
Top: T eff -Radius diagram for the primary star of the HD 54236A system. Middle: T eff -Radius diagram for thesecondary star of the HD 54236A system. Both are in the T eff vs Radius plane and show the Yonsei-Yale stellar evolutionarytracks (Yi et al. 2001), which suggest an age for HD 54236A near the ZAMS (which occurs at an age of 300 Myr in these models)and roughly consistent with an age of 225 ±
50 Myr as determined from other age diagnostics. (Bottom:) H-R diagram showingall three components of the HD 54236 system (black symbols) compared to the PMS evolutionary models of Baraffe et al.(1998); shown are isochrones at ages of 1, 3, 10, 30, 100, and 300 Myr (red dashed lines) and tracks for masses of 0.6–1.2 M ⊙ in increments of 0.1 M ⊙ (blue curves). Cunningham et al.
M34 and M7 or NGC 6457 which have 22, 38, and 24 stars respectively. The oldest clusters used for comparison wereof ages greater than ∼
400 Myr, these included the UMa Group with 14 stellar members, NGC 6633 with 22 stars,Coma Ber with roughly 40 stars forming its distinctive V shape, Hyades containing 14 stars, and NGC 752, the oldestcomparison, having 5 members. The data set used in Sestito & Randich (2005) (and used for comparison in this work)are a collection of independent results from a variety of authors using a degree of models to investigate the evolutionof Lithium abundances and their associated timescales. Figure 17.
Lithium abundance analysis for a range of ages of clusters: Comparison of EW of lithium abundances from theHD 54236 system to several different clusters ranging in age from 35 Myr to 1.2 Gyr (sample clusters and their parameters arefrom Sestito & Randich 2005).From these comparisons we can see that the stellar components in the HD 54236 system (shown with red and blue squares forthe primary and secondary stars in the eclipsing binary primary HD 54236A) are indicative of a relatively young age: betweenPMS clusters and older main sequence clusters.
From visual inspection of Fig. 17, it is clear that the three stars’ A ( Li ) generally follow the trend with T eff observedin young clusters with ages of 100–400 Myr. To constrain the likely age more precisely, we calculated the χ probabilityof the three stars’ A ( Li ) agreement with each of the comparison clusters shown in the figure. We obtain the followingprobabilities for the HD 54236 system Li abundances: P(age <
100 Myr) = 0.03, P(age ∼
100 Myr)= 0.09, P(age200–300 Myr)= 0.40, P(age >
300 Myr) < A ( Li ) with T eff observed forthe three stars in HD 54236 is most consistent with an age of 200–300 Myr. However, a younger age of ∼
100 Myrcannot be entirely ruled out on the basis of the Li abundances alone.4.5.2.
Ca H & K oung EB Triple System in Stellar String Theia 301 II near 3950˚A are a commonly used tracer of chromosphericactivity and thus, through the empirical rotation-activity relationship of FGK dwarfs, can be used as an age indicator.Given the short orbital period of the eclipsing pair in HD 54236A, it is a good assumption that the two stars havetidally interacted and therefore the stars are likely to have their current rotation governed not by the rotation-activity-age relation, but rather by tidal synchronization with the orbit. Indeed, we observe a roughly sinusoidal variation inthe eclipsing system light curve at very nearly the orbital period (see § R ′ HK relationship) are based on single-star angular momentum evolution, so this rules out using the HD 54236Asystem to determine the system age using chromospheric activity measures.Fortunately, the single K-dwarf star in the system (HD 54236B) can be used for this purpose. We utilize the highS/N spectrum obtained with the ANU WiFeS spectrograph. Both because our spectrum was not absolutely fluxcalibrated and because the late spectral type of HD 54236B places it beyond the calibration of existing age-activityrelationships (e.g. Wright et al. 2011; Mamajek 2008),we have instead simply compared our observed spectrum withsuitable late-type standards from several young clusters.In Figure 18, we plot the WiFeS spectrum of the Ca II H&K spectral region for HD 54236B against stars with similareffective temperatures (4200 K ≤ T eff ≤ K ) from three different open clusters spanning the same range of agesfound in our analysis of the Li abundances in the previous section 4.5.1.The open cluster spectra were taken from the Keck HIRES and VLT UVES public archives and convolved down tomatch the lower resolution of the WiFeS observation. We performed a manual examination of the available spectra toconfirm that the open cluster stars plotted are typical in their level of activity for their respective ages.The agreement of the comparison open cluster spectra to the HD 54236B spectrum outside of the Ca II H&K linesis good evidence that the selected open cluster stars are of appropriately similar spectral type. This is importantbecause the change in level of activity in this age range is a fairly sensitive function of spectral type. For instance, atPleiades age the Ca II H&K emission cores transition from very strongly present for T eff . T eff & II emission (even at the WiFeS resolution) at levels significantly higher than observed in HD 54236B. In contrast, thereis also agreement between the observed mild emission in HD 54236B and the M7 comparison star (Bottom panel, 18).Together, these comparisons strongly imply an age for HD 54236B (and therefore of the HD 54236 system) that isolder than the Pleiades ( ∼
130 Myr) and younger than M7 ( ∼
300 Myr).4.5.3.
Isochrone fitting
Ancillary age estimates for HD 54236 B are fitted using photometry and MIST stellar evolutionary models by
MINESweeper (for extensive details on the fitting procedure, including priors used in the inference, see Cargile et al.2019). Using a Bayesian approach,
MINESweeper estimates stellar parameters from MIST stellar evolutionary models(Choi et al. 2016).
MINESweeper employs multi-nested ellipsoid sampling (due to its efficiency in sampling multi-modal likelihood surfaces like stellar isochrones; Feroz et al. 2019) and utilizes optimized interpolation following Dotter(2016). Fitting
Gaia
DR2, 2MASS, and UNWISE photometry for HD 54236B,
MINESweeper and the latest MISTstellar evolutionary models yield an age estimate of 149 +173 − -Myr for HD54236 B, in agreement with our other ageestimators outlined above. Being that HD 54236B is a single star, binarity is not an added complication to fittingand, having lower mass than its EB companion is also less evolved than the HD 54236A EB, therefore, HD 54236B’sposition on in HR diagram space is more sensitive to the system’s age.4.5.4. Summary of Age Estimates
Our observation of the Ca II H&K mission in the HD 54236B K-dwarf star implies an age for HD 54236B (andtherefore of the entire HD 54236 system) that is clearly older than the Pleiades ( ∼
130 Myr) and clearly younger thanM7 ( ∼
300 Myr). This result is in excellent accord with that inferred from the Li abundances of all three stars in theHD 542436 system, for which we obtained a most likely age of 200–300 Myr. As the Li results more clearly rule outan age as young as 130 Myr and the Ca II H&K results more clearly rule out an age as old as 300 Myr, we adopt amost likely system age of 225 ±
50 Myr on the basis jointly of the Li abundances of all three stars and the Ca II H&Kemission of the HD 54236B component. This adopted most likely system age of 225 ±
50 Myr is also in agreementwith the latest MIST models projected age estimate of 225 ±
50 Myr. DISCUSSION: MEMBERSHIP, STELLAR STRINGS, AND AB DOR8
Cunningham et al.
Figure 18. Ca II H&K analysis of the WiFeS spectrum for HD 54236B against stars with similar T eff from three different openclusters spanning the range of ages estimated from the Li abundances in § II H&K analysis for HD 54236B(black line).
Top: comparison with IC 2391 (age ∼
50 Myr).
Middle: comparison with Pleiades (age ∼
130 Myr with a rangeof T eff from 4200 K - 5420 K). Bottom:
Comparison with M7 (age ∼
300 Myr). The open cluster spectra were taken from KeckHIRES and VLT UVES public archives and convolved to match the resolution of the WiFeS observation of HD 54236B. Takentogether, these comparisons strongly imply an age of HD 54236B older than the Pleiades ( ∼
130 Myr) and clearly younger thanM7 ( ∼
300 Myr).
Gaia
DR2 parallax estimates (Bailer-Jones et al. 2018) of HD 54236A puts it at a distance of 133 pc. Thegalactic space velocity is calculated using the
Gaia distance, proper motion and radial velocity estimates follow-ing Johnson & Soderblom (1987). The
U V W velocities for the HD 54236 system are calculated using astrolibRChakraborty & Feigelson (1995) and are listed in Table 4.
Table 4.
Galactic Motions of HD 54236AParallax
U V W . ± .
024 9 . ± .
61 km s − . ± .
36 km s − − . ± . s − No membership to any population has been established HD 54236 to date. Indeed, the BANYAN Σ MultivariateBayesian Algorithm (Gagn´e et al. 2018) reports 0% probability of this source being associated with any of the 27 oung EB Triple System in Stellar String Theia 301
Gaia
DR2 distance and proper motion estimates(Gaia Collaboration et al. 2018; Bailer-Jones et al. 2018).Nonetheless as the number of known co-moving groups has grown significantly in the recent years, we search for otherpossible candidate populations from which HD 54236A may have originated. In particular, we perform a cross-matchagainst Kounkel & Covey (2019); Kounkel et al. (2020) catalog, searching for the nearest moving group of which thissource may be the most likely member. Examining the sources located closer than π > − of the source, only one group appears to be close, Theia 301, makingit the most likely progenitor. Figure 19.
Phase space of HD 54236A relative to the phase space of Theia 301.
Examining the full phase space correlation between the source and the population as a whole, HD 54236A hasgood agreement with the 3-dimensional spatial distribution of Theia 301, as well as with µ α (Figure 19). Greaterdiscrepancy is found in µ δ , where the source is offset by almost ∼
10 mas yr − (6 km s − ). Such an offset could bedue to multiplicity of HD 54236A affecting the astrometric solution, or due to a dynamical evolution of the populationover time, gaining sufficient speed over the initial conditions to no longer be recoverable in phase-space clustering.Recent kinematic analyses have led to new associations being made for co-moving groups of Ursa Major (e.g.,Gagn´e et al. 2020) and 32 Ori (e.g., Stauffer et al. 2020). Similarly, we find that a possible ancestral link exists betweenthe AB Dor moving group and Theia 301. Six listed members of the AB Dor association also have membership toTheia 301 (Messina et al. 2010). Connection between the stellar censuses of Theia-301 and AB Dor are tenuous as0 Cunningham et al. so few members of AB Dor are known at distances farther than 80 pc and clustering from Kounkel & Covey (2019)struggles to identify members of moving groups within 100 pc. The observational knowledge gaps in established AB-Dormembers beyond 80 pc and clustering within 100 pc converge where we would expect to find shared stellar membersbetween these associations. Even so, there have been 6 stars from Messina et al. (2010) that are listed as members ofAB Dor are also members of Theia 301, serving as a bridge between the two populations. The two populations areindependently estimated to have roughly the same ages. The age of Theia 301 obtained from the population isochronefitting is 8.29 ± § § ∼
50 Myr. That sameyear, Luhman et al. (2005) reported a conservative age range of AB Dor C as 75–150 Myr. When the Close et al.(2005) more youthful age estimate of ∼
50 Myr is used, the luminosity of AB Dor C appears highly discrepant withstellar evolution models predictions. However, when an older age like those suggested by Luhman et al. (2005) andBarenfeld et al. (2013) is used, much of the luminosity discrepancy disappears. Assuming that HD 54236 is indeed amember of Theia 301, and further that Theia 301 is linked to the AB Dor association, our robust age estimates ofHD 54236 through lithium abundances (see § II H&K ( § ∼
225 Myr being applied to AB Dor C mitigates its luminosityproblem. SUMMARY AND CONCLUSIONSWe have shown that HD 54236 is a hierarchical triple system comprising HD 54236A as a previously unknown EBwith stellar masses of 1.18 M ⊙ and 1.07 M ⊙ , and the late-K type star HD 54236B as a wide CPM tertiary, withestimated mass 0.5–0.6 M ⊙ . We have furthermore shown that the system is young, with a most likely system ageof 225 ±
50 Myr from joint consideration of the lithium abundances measured in the three stars and the strength ofthe Ca II H&K chromospheric activity index. The system age is determined independently from the lithium abun-dances and Ca II H&K activity to be 225 ±
50 Myr. Using a number of photometric observations for tertiary memberHD54236 B,
MINESweeper and the latest MIST stellar evolutionary models, yield an age estimate of 149 +173 − Myr forHD 54236B providing additional credence to our reported age estimate. At this age, the solar-mass eclipsing com-ponent stars (HD 54236A) are very nearly zero-age main sequence stars, and the wide low-mass tertiary companion(HD 54236B) is near the end of its PMS evolution. All solutions from our analysis are collected in 5.
Table 5.
HD 54236 System Solutions T eff , ± KT eff , ± KR . ± . R ⊙ R . ± . R ⊙ M . ± .
029 M ⊙ M . ± .
027 M ⊙ i . ± . ◦ q . ± . e . ± . v γ − . ± . − a sin i . ± .
080 R ⊙ Parallax 7 . ± .
024 mas U . ± .
61 km s − V . ± .
61 km s − W − . ± .
23 km s − The HD 54236 system is also interesting from the standpoint of its dynamics, by virtue of representing a tight EB withorbital period of 2.4 d, with a wide tertiary companion that remains unaffected by the ”fountain of youth” which mayexist in EBs close enough to enhance their stellar rotation. The wide tertiary companion also represents a statistically oung EB Triple System in Stellar String Theia 301
Gaia
DR2 distance estimates and propermotion information (Bailer-Jones et al. 2018), HD 54236 was found to have 99% probability of being a field star.That is, every known (as of 2018) stellar association within 150 pc has a nearly 0% probability of having HD 54236Aamong its members. Additionally, its
Gaia radial velocity estimates were not similar to radial velocity estimates of thesystems considered by BANYAN Σ. However, the Kounkel & Covey (2019); Kounkel et al. (2020) catalog presentednewly discovered of stellar associations, including “stellar strings”, and found a likely membership to the stellar stringTheia 301, which has an estimated kinematic age that is also in very good agreement with our independent age estimatefor HD 54236.Finally, the AB Dor association (and thus AB Dor itself) appears to be linked both kinematically and through anumber of shared stellar members to this newly discovered “stellar string” Theia 301. Therefore, the age of ∼
225 Myrthat we have found for HD 54236 (and Theia 301) would imply that the AB Dor system also shares this age. Indeed,applying our age estimate to AB Dor itself alleviates reported tension between observation and theory that arises forthe luminosity of AB Dor C, when younger age estimates ( ∼
50 Myr) are used.
Software:
C-Munipack (http://c-munipack.sourceforge.net/), ISIS image subtraction package (Alard & Lupton1998; Alard 2000; Hartman et al. 2004), SExtractor (Bertin & Arnouts 1996), DAOPHOT II (Stetson, Davis, & Crabtree1990), IRAF (Tody 1986, Tody 1993), MINESweeper (Cargyle et al. 2019), MIST (Choi et al. 2019), Lightkurve; s Pythonpackage for Kepler and
TESS data analysis (Lightkurve Collaboration, 2018), astrolibR; used to convert
Gaia distance,proper motion and radial velocity to galactic coordinates (Chakraborty, Feigelson, & Babu 2014)ACKNOWLEDGMENTSWe thank the reviewer for a constructive report and helpful feedback. J=MC would like to thank Erika McIntire,Nicole Megetarian & S. J. Clark for their continued empowerment of a mother in STEM. K.G.S. acknowledges NSFgrant AST-1009810. We thank A. Aarnio for independently checking our calculation of the U V W space motion ofthe system. We also gratefully acknowledge Jaci Cloete and Piet Fourie for their technical assistance in the operationof the KELT-South telescope at the South African Astronomical Observatory. This research has made use of theSIMBAD database, operated at CDS, Strasbourg, France.2
Cunningham et al.
REFERENCES
Aarnio, A.-N., Weinberger, A.-J., Stassun, K.-G., Mamajek,E.-E., & James, D.-J. 2008, AJ, 136:6, 2483 ,doi: 10.1088/0004-6256/136/6/2483Alard, C. 2000, Astronomy and Astrophysics Supplement,144, 363 , doi: 10.1051/aas:2000214Alard, C., & Lupton, R.-H. 1998, ApJ, 503:1, 325 ,doi: 10.1086/305984Bailer-Jones, C.-A.-L., Rybizki, J., Fouesneau, M.,Mantelet, G., & Andrae, R. 2018, AJ, 152:58, 11,doi: 10.3847/1538-3881/aacb21Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H.1998, A&A, 337, 403.https://arxiv.org/abs/astro-ph/9805009Barenfeld, S.-A., Bubar, E.-J., Mamajek, E.-E., & Young,P.-A. 2013, ApJ, 766:6, 7,doi: 10.1088/0004-637X/766/1/6Bayliss, D., Zhou, G., Penev, K., et al. 2013, AJ, 146:113,11, doi: 10.1088/0004-6256/146/5/113Beatty, T.-G., Pepper, J., Siverd, R.-J., et al. 2012, TheAstrophysical Journal Letters, 756:L39, 6,doi: 10.1088/2041-8205/756/2/L39Bell, C.-P.-M., Mamajek, E.-E., & Naylor, T. 2015,MNRAS, 454, 593 , doi: 10.1093/mnras/stv1981Bertin, E., & Arnouts, S. 1996, Astronomy andAstrophysics Supplemental Series, 117, 393 ,doi: 10.1051/aas:1996164Bessell, M.-S. 1999, Publications of the AstronomicalSociety of the Pacific, 111:765, 1426, doi: 10.1086/316454Cargile, P. A., Conroy, C., Johnson, B. D., et al. 2019,arXiv e-prints, arXiv:1907.07690.https://arxiv.org/abs/1907.07690Carlsson, M., & Stein, R.-F. 1994, ApJ, 440:L29, 4,doi: 10.1086/187753Chakraborty, A., & Feigelson, E.-D. 1995, Astronomy UsersLibrary: Package astrolibRChoi, J., Dotter, A., Conroy, C., et al. 2016, ApJ, 823, 102,doi: 10.3847/0004-637X/823/2/102Clemens, C.-J., Crain, A. J., & Anderson, R. 2004,Ground-based instrumentation for Astronomy,Proceedings, 5492, doi: 10.1117/12.550069Close, L. M., Lenzen, R., Guirado, J. C., et al. 2005,Nature, 433, 286, doi: 10.1038/nature03225Collins, K. A., Eastman, J. D., Beatty, T. G., et al. 2014,AJ, 142:2, 18, doi: 10.1088/0004-6256/147/2/39Dopita, M., Hart, J., McGregor, P., et al. 2007, Ap&SS,310, 255, doi: 10.1007/s10509-007-9510-zDotter, A. 2016, ApJS, 222, 8,doi: 10.3847/0067-0049/222/1/8 Feroz, F., Hobson, M. P., Cameron, E., & Pettitt, A. N.2019, The Open Journal of Astrophysics, 2, 10,doi: 10.21105/astro.1306.2144Gagn´e, J., Faherty, J. K., & Popinchalk, M. 2020, ResearchNotes of the American Astronomical Society, 4, 92,doi: 10.3847/2515-5172/ab9e79Gagn´e, J., Mamajek, E. E., Malo, L., et al. 2018, ApJ, 856,23, doi: 10.3847/1538-4357/aaae09Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al.2018, A&A, 616, A1, doi: 10.1051/0004-6361/201833051Hamuy, M., Suntzeff, N. B., Heathcote, S. R., et al. 1994,PASP, 106, 566, doi: 10.1086/133417Hartman, J. D., Bakos, G. ´A., Noyes, R. W., et al. 2011,AJ, 141, 166, doi: 10.1088/0004-6256/141/5/166Johnson, D. R. H., & Soderblom, D. R. 1987, AJ, 93, 864,doi: 10.1086/114370Kounkel, M., & Covey, K. 2019, AJ, 158, 122,doi: 10.3847/1538-3881/ab339aKounkel, M., Covey, K., & G., S. K. 2020, arXiv e-prints.https://arxiv.org/abs/2004.07261Lightkurve Collaboration, Cardoso, J. V. d. M., Hedges, C.,et al. 2018, Lightkurve: Kepler and TESS time seriesanalysis in Python, Astrophysics Source Code Library.http://ascl.net/1812.013Luhman, K.-L., Stauffer, J.-R., & Mamajek, E.-E. 2005,ApJ, 628, L69, doi: 10.1086/432617M., H., T.-J., D., G.-D., M., & R., H. 2019, AJ, 158, 143,doi: 10.3847/1538-3881/ab3984Mamajek, E. E. 2008, Astronomische Nachrichten, 329, 10,doi: 10.1002/asna.200710827Mason, B. D., Wycoff, G. L., Hartkopf, W. I., Douglass,G. G., & Worley, C. E. 2001, AJ, 122, 3466,doi: 10.1086/323920Messina, S., Desidera, S., Turatto, M., Lanzafame, A. C., &Guinan, E. F. 2010, A&A, 520, A15,doi: 10.1051/0004-6361/200913644Metcalfe, T. S. 1999, AJ, 117, 2503, doi: 10.1086/300833Munari, U., Sordo, R., Castelli, F., & Zwitter, T. 2005,A&A, 442, 1127, doi: 10.1051/0004-6361:20042490Oelkers, R. J., & Stassun, K. G. 2018, AJ, 156, 132,doi: 10.3847/1538-3881/aad68eOh, S., Price-Whelan, A.-M., Hogg, D.-W., Morton, T.-D.,& Spergel, D.-N. 2017, AJ, 153, 257,doi: 10.3847/1538-3881/aa6ffdPenev, K., Bakos, G. ´A., Bayliss, D., et al. 2013, AJ, 145, 5,doi: 10.1088/0004-6256/145/1/5Pepper, J., Kuhn, R. B., Siverd, R., James, D., & Stassun,K. 2012, PASP, 124, 230, doi: 10.1086/665044 oung EB Triple System in Stellar String Theia 30123