BANYAN. XII. New Members of Nearby Young Associations from GAIA–Tycho Data

Young associations and moving groups in the solar neighborhood are valuable laboratories in which to study the properties of age-calibrated stars in detail (Zuckerman & Song 2004; Torres et al. 2008). Their proximity makes it possible to gather high-quality data more easily, perform high-angular resolution imaging of exoplanets, and bring this characterization to the lowest-mass and faintest substellar objects with a calibrated age (e.g., see Marois et al. 2008; Delorme et al. 2013; Liu et al. 2013; Naud et al. 2014; Gagné et al. 2015a, 2017a; Macintosh et al. 2015; Faherty et al. 2016). However, the main difficulty in studying such associations is also a consequence of their proximity: their members are distributed over large areas of the sky (e.g., see Figure 1 of Gagné et al. 2018c), making it difficult to identify members without measuring their full 6D kinematics, consisting of the XYZ Galactic coordinates and UVW space velocities. These require measuring the trigonometric parallax and heliocentric radial velocities of stars, which is challenging to perform on all stars in a large area of the sky.

Until recently, the discovery and kinematic characterization of most young associations and moving groups in the solar neighborhood relied on parallaxes obtained by the Hipparcos mission (Perryman et al. 1997), allowing us to discover only the most massive and brightest members of these young associations, in general within ∼100 pc. The stars bright enough to have a Hipparcos parallax measurement only represent ∼10% of the total stellar population within 30 pc, and 2% within 100 pc (Perryman et al. 1997; Pecaut & Mamajek 2013), which is a consequence of the initial mass function peaking at mass of ∼0.25 M_⊙ (Bochanski et al. 2010) corresponding to the M spectral class. Recent studies focused on various statistical methods to obtain the full kinematics of only the most likely members of these associations, based on their sky position, proper motion, and photometry, as well as radial velocities and parallaxes when available (e.g., see Mamajek 2005; Rodriguez et al. 2011; Malo et al. 2013, 2014; Gagné et al. 2014; Kraus et al. 2014; Gagné et al. 2015b; Murphy & Lawson 2015; Riedel et al. 2017; Shkolnik et al. 2017), and have started to uncover and confirm the kinematics of a fraction of their low-mass stars and substellar objects. The Gaia mission (Gaia Collaboration et al. 2016b) has already started to benefit this field of research with the Data Release 1 (Gaia–DR1; Gaia Collaboration et al. 2016a), which provided 2 million parallaxes for the Tycho catalog (Høg et al. 2000).⁶ This will be even more true of Gaia DR2, which will provide a billion parallaxes that will allow to complete the population of all young associations within 150 pc down to ∼0.12 M_⊙ (Smart et al. 2017). This completion of the stellar population of young associations in the solar neighborhood will have many direct applications, such as comparing their initial mass functions (e.g., see Chabrier 2005; Luhman 2007; Bochanski et al. 2011; Jeffries 2012; Gagné et al. 2017b), providing strategic targets for the direct imaging of exoplanets, understanding the stellar formation history of the solar neighborhood, and will provide important benchmark populations to characterize the fundamental properties and chemical abundances of coeval stars (e.g., see King et al. 2000; Schuler et al. 2006).

The latest membership classification tool, BANYAN Σ (Gagné et al. 2018c), benefited from Gaia–DR1 data to refine its kinematic models of the 27 well-characterized young associations within 150 pc of the Sun. The ages of these associations are in the range ∼1–850 Myr, and their general characteristics are listed in Table 1, which is a summarized version of Table 1 from Gagné et al. (2018c). The BANYAN Σ tool uses Bayesian inference to determine the membership probability that a star belongs to any of these 27 associations, based on its sky position, proper motion, and, optionally, its radial velocity and distance. When radial velocity and/or distance are not known, these observables are marginalized and a membership probability is still calculated. BANYAN Σ includes more associations and generates less contaminants at a fixed recovery rate compared to all previous tools available in the literature (e.g., Gagné et al. 2014; Malo et al. 2014; Riedel et al. 2017). Furthermore, its analytical solving of marginalization integrals makes it less computationally intensive, and amenable to analyze much more easily large data sets such as Gaia–DR1 as well as the upcoming data releases.

Table 1.  Young Associations Included in This Study

Group	$\langle \varpi \rangle$ a	$\langle \nu \rangle$ b	${S}_{\mathrm{spa}}$ c	${S}_{\mathrm{kin}}$ d	Age	References
Name	(pc)	(km s−1)	(pc)	(km s−1)	(Myr)
118TAU	100 ± 10	14 ± 2	3.4	2.1	∼10	1
ABDMG	${30}_{-10}^{+20}$	${10}_{-20}^{+10}$	19.0	1.4	${149}_{-19}^{+51}$	2
βPMG	${30}_{-10}^{+20}$	10 ± 10	14.8	1.4	24 ± 3	2
CAR	60 ± 20	20 ± 2	11.8	0.8	${45}_{-7}^{+11}$	2
CARN	30 ± 20	${15}_{-10}^{+7}$	14.0	2.1	∼200	3
CBER	${85}_{-5}^{+4}$	−0.1 ± 0.8	3.6	0.5	${562}_{-84}^{+98}$	4
COL	50 ± 20	${21}_{-8}^{+3}$	15.8	0.9	${42}_{-4}^{+6}$	2
CRA	139 ± 4	−1 ± 1	1.5	1.7	4–5	5
EPSC	102 ± 4	14 ± 3	2.8	1.8	${3.7}_{-1.4}^{+4.6}$	6
ETAC	95 ± 1	20 ± 3	0.6	2.0	11 ± 3	2
HYA	42 ± 7	${39}_{-4}^{+3}$	4.5	1.2	750 ± 100	7
IC2391	149 ± 6	15 ± 3	2.2	1.4	50 ± 5	8
IC2602	146 ± 5	17 ± 3	1.8	1.1	${46}_{-5}^{+6}$	9
LCC	110 ± 10	14 ± 5	11.6	2.2	15 ± 3	10
OCT	${130}_{-20}^{+30}$	${8}_{-9}^{+8}$	22.4	1.3	35 ± 5	11
PL8	130 ± 10	22 ± 2	5.0	1.1	∼60	12
PLE	134 ± 9	6 ± 2	4.1	1.4	112 ± 5	13
ROPH	131 ± 1	−6.3 ± 0.2	0.7	1.6	<2	14
TAU	120 ± 10	16 ± 3	10.7	3.6	1–2	15
THA	${46}_{-6}^{+8}$	${9}_{-6}^{+5}$	9.1	0.8	45 ± 4	2
THOR	96 ± 2	19 ± 3	3.9	2.1	${22}_{-3}^{+4}$	2
TWA	60 ± 10	10 ± 3	6.6	1.5	10 ± 3	2
UCL	130 ± 20	5 ± 5	17.4	2.5	16 ± 2	10
UCRA	147 ± 7	−1 ± 3	4.5	1.8	∼10	16
UMA	${25.4}_{-0.7}^{+0.8}$	−12 ± 3	1.2	1.3	414 ± 23	17
USCO	130 ± 20	−5 ± 4	9.9	2.8	10 ± 3	10
XFOR	100 ± 6	19 ± 2	2.6	1.3	∼500	18

Notes. The full names of young associations are: 118 Tau (118TAU), AB Doradus (ABDMG), β Pictoris (βPMG), Carina (CAR), Carina-Near (CARN), Coma Berenices (CBER), Columba (COL), Corona Australis (CRA),  Chamaeleontis (EPSC), η Chamaeleontis (ETAC), the Hyades cluster (HYA), Lower Centaurus Crux (LCC), Octans (OCT), Platais 8 (PL8), the Pleiades cluster (PLE), ρ Ophiuci (ROPH), the Tucana-Horologium association (THA), 32 Orionis (THOR), TW Hya (TWA), Upper Centaurus Lupus (UCL), Upper CrA (UCRA), the core of the Ursa Major cluster (UMA), Upper Scorpius (USCO), Taurus (TAU), and χ¹ For (XFOR).

^aPeak of distance distribution and ±1σ range. ^bPeak of radial velocity distribution and ±1σ range. ^cCharacteristic spatial scale in XYZ space. ^dCharacteristic kinematic scale in UVW space.

References. (1) Mamajek (2016), (2) Bell et al. (2015), (3) Zuckerman et al. (2006), (4) Silaj & Landstreet (2014), (5) Gennaro et al. (2012), (6) Murphy et al. (2013), (7) Brandt & Huang (2015), (8) Barrado y Navascués et al. (2004), (9) Dobbie et al. (2010), (10) Pecaut & Mamajek (2016), (11) Murphy & Lawson (2015), (12) Platais et al. (1998), (13) Dahm (2015), (14) Wilking et al. (2008), (15) Kenyon & Hartmann (1995), (16) Gagné et al. (2018c), (17) Jones et al. (2015), and (18) Pöhnl & Paunzen (2010).

Download table as:  ASCIITypeset image

In this paper, we apply the BANYAN Σ tool to the Tycho catalog stars that benefit from parallaxes in the Tycho–Gaia Astrometric Solution (TGAS hereafter) to identify 32 new bona fide members and 219 new candidate members of the 27 nearest young associations. In Section 2, we build Gaia–2MASS M_G versus G−J color-magnitude sequences for field stars and young associations of different age categories that will complement the kinematic analysis of BANYAN Σ in our determination of Bayesian membership probabilities. Our method for selecting new candidate members is described in Section 3, and we investigate their signs of youth such as UV, X-ray emission and model isochrones in Section 4. The conclusion of this work is presented in Section 5.

In this section, color-magnitude sequences are built for field stars and members of young moving groups at different ages. For the sequence to be useful for the most stars and across a large range of spectral types regardless of age, it is preferable to use a color combination that spans a wide wavelength window. The Gaia G-band⁷ and 2MASS (Skrutskie et al. 2006) J-band magnitudes respect this criterion and are both readily available for a large number of stars. In Figure 1, we show the G − J color as a function of spectral type for all bona fide members of young associations compiled by Gagné et al. (2018c), for which TGAS and 2MASS cross-matches were already provided. The members are grouped by ages to demonstrate that the spread in G − J color is not significant in any age category. This figure demonstrates how the G − J color provides a good proxy for spectral type in the A0–L0 range.

Zoom In Zoom Out Reset image size

The full data set of TGAS was used to build a color-magnitude sequence in absolute M_G band as a function of G − J color. The TGAS catalog entries were cross-matched with the nearest 2MASS entries using a search radius of 4'' to achieve this. To build a monotonic sequence that can be described as a function in this color-magnitude space, giant stars were filtered out from the TGAS data. A simple color-magnitude region cut was used to achieve this (displayed in Figure 2); all stars redder or brighter than a color-magnitude region described by the three straight lines connecting these four (G − J, M_G) coordinates were rejected:

$$\begin{eqnarray}{p}_{1} & = & (1.02,-\infty ),\\ {p}_{2} & = & (1.02,2.40),\\ {p}_{3} & = & (1.14,4.00),\\ {p}_{4} & = & (3.00,8.00).\end{eqnarray} \tag{ 1 }$$

Zoom In Zoom Out Reset image size

This rejection of giants from the field color-magnitude sequence means that a search for young stars that is not using parallax information may be susceptible to giant stars contaminating a sample of young stellar candidate members. This will, however, be mitigated by the kinematic analysis of BANYAN Σ that will reject low-proper motion objects unless they happen to match those of a young moving group by chance. In the current analysis, all entries of TGAS that fall in the region described in Equation (1) are ignored.

All stars in TGAS were first split in G − J color bins of 0.05 mag in the range −0..5 mag to build the field sequence. A first sequence was built by measuring the median absolute M_G-band magnitude in each G − J color bin. A cumulative distribution function of the absolute M_G-band magnitudes of all stars within each bin was then built, and used to determine the positive and negative error bars in absolute M_G-band magnitude that would each encompass 34% of the population on both sides of the median value, such that both error bars contain 68% of all data and correspond to a ±1σ range. All three resulting median and ±1σ color-magnitude sequences were then smoothed with a 2-cells wide running average over all color bins. The resulting color-magnitude sequence of field stars is presented in Figure 2.

Similar color-magnitude sequences were then built for the bona fide members of young associations compiled by Gagné et al. (2018c) (although the giants rejection criterion was not applied, as we have detailed literature information on these stars confirming that none of them are giants). All members were first assigned to one of three age categories, which were found to assemble the stars that follow similar color-magnitude sequences: (1) younger than 20 Myr (i.e., members of 118TAU, CRA, EPSC, ETAC, LCC, ROPH, TAU, TWA, UCL, UCRA, and USCO); (2) 20–100 Myr (i.e., members of βPMG, CAR, COL, IC2602, IC2391, OCT, PL8, THA, and THOR); and (3) 100–800 Myr (i.e., members of ABDMG, CARN, CBER, HYA, PLE, UMA, and XFOR). Known unresolved binaries were not included here, which will make the current search less sensitive in discovering binary systems in young associations.⁸ The resulting color-magnitude diagrams are displayed in Figure 3. The 100–800 Myr sequence is fainter than the field sequence at colors bluer than G − J ≈ 1.0–1.2, and merges with it at redder colors. This can be attributed to field stars more massive than 0.96–1.00 M_⊙ (corresponding to G − J ∼ 1–1.2; Pecaut & Mamajek 2013 ⁹ ) that start to depart from the main sequence onto the giant branch after ∼8–9 Gyr (Choi et al. 2016). At younger ages of 20–100 Myr, stars redder than G − J ≈ 1.0–1.2 are brighter than the field sequence because their radii are still inflated from their young age (e.g., see Soderblom et al. 2014). This effect is more dramatic for stars younger than 20 Myr, to the point where their sequence merges with that of the field at colors bluer than G − J ≈ 1.0–1.2. This illustrates how stars coming into and departing from the main sequence are hard to distinguish using isochrones alone.

Zoom In Zoom Out Reset image size

The BANYAN Σ software uses sky position, proper motion and optionally radial velocity and distance to assign a Bayesian probability that a star is a member of a known young association within 150 pc of the Sun. TGAS includes parallax measurements, sky position, proper motion, and G-band magnitudes for 2,057,050 stars, but it does not contain radial velocity measurements. We used the same method described in Section 2 to cross-match TGAS entries with 2MASS to obtain the J-band magnitudes of all stars. Giant stars were filtered out from this sample with the criterion described in Section 2 (only for stars with ${M}_{G}\lt 3$ to avoid rejecting low-mass stars younger than ∼20 Myr.), resulting in a sample of 1,338,580 stars. All stars with a trigonometric distance measurement at a statistical significance of less than 2σ (i.e., measurement over error less than 2) or with a missing 2MASS J-band measurement were rejected, further refining the sample to 1,190,699 stars.

All color-magnitude sequences described in Section 2 were used to assign a Bayesian probability in each young association directly from a comparison of their G − J color and absolute G-band magnitude, assuming that the vertical spread around each sequence is Gaussian. These probabilities are calculated by BANYAN Σ through the constraint_dist_per_hyp and constraint_edist_per_hyp keywords, and are subsequently multiplied to the Bayesian prior probabilities in determining membership probabilities (see Gagné et al. 2018a, 2018b, 2018c for more details). The same color-magnitude sequences could be used to constrain the acceptable distances of any star without a parallax measurement, by using the same keyword in the BANYAN Σ software; it then automatically determines which method to adopt depending on whether trigonometric distances are available.

Only stars with a Bayesian membership probability above 90% for any young association were selected for further consideration.¹⁰ This probability threshold was designed to produce similar recovery rates of 50% for all associations in BANYAN Σ (Gagné et al. 2018c). BANYAN Σ assigns each candidate member of a young association with an optimal radial velocity that maximizes its membership probability (see Gagné et al. 2018c for more details). These optimal radial velocities were combined with measured kinematics to derive optimal space velocities UVW that correspond to the best-matching values that can be expected for the candidate member. The candidate members with an optimal UVW located at more than 5 km s⁻¹ or 5σ from the core of its best-matching association kinematic model were also rejected, as this is larger than the typical spread of young association members (1–4 km s⁻¹; Gagné et al. 2018c). This cut rejected 491 objects, which are heavily skewed toward large distances (313 of them have distances above 150 pc). Such candidate members might correspond to field interlopers with peculiar velocities or members of yet unknown young associations, as well as more distant associations not yet included in BANYAN Σ (e.g., Lupus). These selection criteria generated a total of 1560 candidate members, which were cross-matched with the compilation of bona fide and candidate members compiled by Gagné et al. (2018c). Only the 830 stars that were not already listed in this compilation will be considered here.

The 830 candidate members were cross-matched with SIMBAD (Ochsenbein et al. 2000), the RAVE data release 5 catalog (Kunder et al. 2017) and the Kharchenko et al. (2007) catalog to assign them radial velocity measurements. These cross-matches retrieved radial velocity with measurement errors below 10 km s⁻¹ for 291 stars. 67 stars were found to have a radial velocity measurement in both RAVE and other catalogs; the RAVE measurements were preferred in these cases. Only six of them had radial velocity measurements that differed by more than 2σ; we found no evidence that these stars are binaries in the literature. All stars with radial velocity measurements were re-analyzed with BANYAN Σ, and the same probability and optimal UVW selection criteria were used to reject an additional 122 candidate members. This resulted in a total of 539 candidate members without a radial velocity measurement, and 169 candidate members with radial velocity measurements, for a total of 708 candidates which are listed in Table 4.

In Figure 4, all 708 candidate members identified here are compared with the field and young color-magnitude diagrams built in Section 2. There are a few notable cases where stars fall well outside of the young color-magnitude sequences, despite their having a Bayesian membership probability above 90%. This indicates that they were a poor fit to even the field color-magnitude sequence. An investigation of the six such cases that have a radial velocity measurement (HD 145501, ROXs 43, CD–33 10685, MZ Lup, DR Tau, and TYC 8881–551–1) reveals that they have circumstellar disks or infrared excesses that hint at a possible disk (Evans et al. 2003; Cieza et al. 2009; Kraus et al. 2012; Gáspár et al. 2016; McDonald et al. 2017), which likely explains their peculiar position in the color-magnitude diagram.

Zoom In Zoom Out Reset image size

The Galactic coordinates XY of all 708 candidates are compared with the BANYAN Σ spatial models in Figure 5. Ten candidates (DR Tau and nine others without radial velocity measurements in Table 4) are located at distances further than 200 pc and could be members of associations not included in BANYAN Σ. In Table 2, we list the 60 targets that have ambiguous membership in more than one young association.

Zoom In Zoom Out Reset image size

A literature search was performed to determine whether the candidates identified in Section 3 were already known as candidate members of a young association, or whether they display signs of youth. A total of 254 candidates were never identified as candidate members of young associations in the literature, 38 of which now have complete kinematics. These stars are listed in Table 3. There are three stars in Table 3 (TYC 8881–551–1, TYC 8098–597–1 and CD–31 11053) that were identified as spectral binaries by (Torres et al. 2006). This makes them uncertain members because their orbital motion could affect the measured radial velocities, and we therefore do not draw a firm conclusion about their membership here. DR Tau has complete kinematics consistent with TAU, but it is located at a distance of ∼207 pc, significantly larger than its other members (∼120 pc). It is possible that TAU is spatially much larger than previously thought, but we do not draw any firm conclusions on the membership of DR Tau here. Similarly, CD–33 10685 (at a distance of ∼141 pc) has complete kinematics consistent with UCL, but Merín et al. (2008) classify it as a member of Lupus, which is too distant to have been included in the BANYAN Σ models. We will therefore wait for the future inclusion of Lupus in the BANYAN Σ models before drawing a conclusion on its membership.

Table 3.  Candidate Members with Full Kinematics

Name	Spectral	Isochronal	G − J	NUV − G	X-Ray	Li EW	Signs of	Bona fide	Referencesd
	Typea	Ageb (Myr)	(mag)	(mag)	HR1	(mÅ)	Youthc	Member
ABDMG
CD–26 1643	F9V	⋯	0.9	5.0	−0.15	⋯	X	Y	1, ⋯
HD 61518	F5V	⋯	0.8	⋯	−0.23	⋯	IR	Y	2, ⋯
HD 147512	G8/K0V	⋯	1.0	6.3	⋯	⋯	IR	Y	3, ⋯
HD 221239	K2.5V	⋯	1.3	7.4	−1.00	⋯	UV	Y	4, ⋯
BPMG
TYC 8098–597–1e	K3V	⋯	2.0	⋯	0.16	25	X	?	5, 6
HD 207043	G5V	⋯	1.0	5.9	⋯	⋯	IR	Y	7, ⋯
CAR
HD 37402	F6V	⋯	0.8	4.8	−0.17	110	Li, IR	Y	8, 9
CARN
S1* 329	K7V(ke)	⋯	2.1	8.5	0.47	⋯	X, UV	Y	7, ⋯
L 106–104	M3	⋯	2.2	⋯	⋯	⋯	IR, Ca	Y	10, ⋯
COL
HD 29329	F7V	⋯	0.8	⋯	−0.01	88	X, IR	Y	11, 11
TYC 8881–551–1e	K0IV/V	⋯	1.8	6.1	−0.22	⋯	UV, Sp	?	12, ⋯
V* AI Lep	G2V	${25}_{-5}^{+3}$	1.1	⋯	0.00	213	X, Li, IS	Y	11, 11
LCC
TYC 8649–1758–1	(K2)	${13}_{-4}^{+5}$	1.5	⋯	⋯	⋯	IR, Ca, IS	Y	⋯, ⋯
TYC 8653–1049–1	(G8)	${28}_{-6}^{+3}$	1.3	⋯	⋯	⋯	IR, Ca, IS	Y	⋯, ⋯
OCT
TYC 7053–832–1	(G6)	⋯	1.2	5.8	⋯	⋯	UV, IR, Ca	Y	⋯, ⋯
HD 35212	F5V	⋯	0.8	⋯	−0.27	⋯	IR	Y	2, ⋯
HD 275012	G5	⋯	1.2	⋯	−0.01	⋯	X, IR, Ca	Y	13, ⋯
CD–35 2433	(G2)	⋯	0.9	⋯	−0.02	⋯	X, IR	Y	⋯, ⋯
HD 42122	F7/G0V	${18}_{-4}^{+2}$	0.8	4.7	⋯	⋯	IR, IS	Y	8, ⋯
TYC 8534–1810–1	(G7)	${45}_{-9}^{+70}$	1.2	⋯	0.11	⋯	X, IR, Ca, IS	Y	⋯, ⋯
TYC 8895–112–1	(G4)	⋯	1.0	⋯	−0.32	⋯	IR, Ca	Y	⋯, ⋯
TYC 8104–898–1	(G1)	⋯	0.9	⋯	⋯	⋯	⋯	N	⋯, ⋯
TYC 9178–1390–1	(K1)	40 ± 10	1.4	⋯	⋯	⋯	IR, IS	Y	⋯, ⋯
TYC 9341–1233–1	(G9)	⋯	1.3	8.0	⋯	⋯	⋯	N	⋯, ⋯
TAU
HD 284659	A2	${8}_{-2}^{+1}$	0.2	⋯	⋯	⋯	IS	?	13, ⋯
THA
HD 10863	F0IV	$\leqslant 1000$	0.6	5.5	⋯	⋯	IS, Sp	Y	14, ⋯
UCL
CD–31 11053e	K3Ve	5 ± 1	1.8	⋯	0.68	470	X, Li, IR, IS	?	6, 6
TYC 7295–853–1	(G7)	${16}_{-3}^{+2}$	1.2	⋯	⋯	⋯	IR, Ca, IS	Y	⋯, ⋯
TYC 7296–1194–1	(K0)	${22}_{-5}^{+3}$	1.3	⋯	⋯	⋯	Ca, IS	Y	⋯, ⋯
TYC 7353–768–1	G8V	${18}_{-5}^{+2}$	1.1	⋯	1.00	270	X, Li, IR, IS	Y	6, 6
HD 321958	G9V	${14}_{-3}^{+2}$	1.2	⋯	0.51	275	X, Li, IR, IS	Y	6, 6
TYC 8332–2024–1	K5Ve	⋯	1.7	⋯	0.03	480	X, Li	Y	6, 6
V* V991 Sco	G6/8	9 ± 1	1.1	⋯	−0.05	⋯	X, IR, IS	Y	15, ⋯
CD–23 13197	K0IV(e)	${13}_{-4}^{+2}$	1.3	⋯	0.52	360	X, Li, IR, IS, Sp	Y	6, 6
HD 317637	K2V	${13}_{-3}^{+2}$	1.4	⋯	0.03	390	X, Li, IR, IS	Y	6, 6
BD–18 4557	K2IV(e)	${9}_{-2}^{+1}$	1.5	⋯	0.18	420	X, Li, IR, IS, Sp	Y	6, 6
CD–43 11887	G9V	${9}_{-3}^{+2}$	1.3	⋯	0.13	350	X, Li, IR, IS	Y	6, 6
USCO
CD–25 11942	K0IV	${6.3}_{-0.7}^{+0.8}$	1.4	⋯	0.01	310	X, Li, IS, Sp	Y	6, 6

Notes.

^aSpectral types in parentheses were estimated using the G − J color with the spectral type–color relations of Pecaut & Mamajek (2013); see also http://www.pas.rochester.edu/~emamajek/EEM_dwarf_UBVIJHK_colors_Teff.txt. ^bPre-main-sequence ages derived from MIST isochrones in absolute G versus G − J. See Section 4 for more detail. The color-magnitude positions of HD 284659, HD 42122, and TYC 8534–1810–1 are also consistent with respective post-main-sequence ages of ${560}_{-100}^{+70}$ Myr, ${4.5}_{-0.9}^{+0.5}$ Gyr and ${10}_{-6}^{+4}$ Gyr. ^cSigns of youth compiled from the literature. See Section 4 for more details. X: X-ray emission with HR1 ≥−0.15; UV: Galex NUV − G versus G − J consitent with youth; Li: Lithium absorption above 100 mÅ; E: Mid-infrared excess; Lm: Luminosity class consistent with youth; Is: Young isochronal age consistent with; Ca: Ca ii infrared triplet age consistent with proposed association. ^dReferences for: (1) spectral type and (2) Li absorption. ^eSpectral binary (Torres et al. 2006).

References. (1) Jaschek et al. (1964), (2) Houk (1978), (3) Houk & Swift (1999), (4) Gray et al. (2003), (5) Kharchenko (2001), (6) Torres et al. (2006), (7) Gray et al. (2006), (8) Houk & Cowley (1975), (9) Moór et al. (2013), (10) Rousseau et al. (1996), (11) White et al. (2007), (12) Nordström et al. (2004), (13) Nesterov et al. (1995), (14) Pribulla et al. (2014), and (15) Houk (1982).

Download table as:  ASCIITypeset image

We found that 53 objects were already known as bona fide members with full kinematics and were either overlooked or excluded in the Gagné et al. (2018c) compilation of bona fide members. For example, some members of HYA identified by Gaia Collaboration et al. (2017) were not clearly identified as cluster or stream members, and some ABDMG members identified by (Zuckerman & Song 2004) had a "questionable membership flag." We consider that their high membership probability calculated in this work warrants considering them as bona fide members, as it demonstrates that they have kinematics consistent with other unambiguous members. A histogram of the associations in which new candidate members were identified here is displayed in Figure 6. In Figure 7, we display the fractional number of new candidate members identified here compared with the number of currently known bona fide members. Some associations such as PL8 and OCT were not extensively studied in the literature, and as a consequence our sample will make a significant contribution to their number of known members upon full confirmation of their kinematics.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Seventy-five others objects were already known as candidate members, and the addition of a TGAS parallaxe or a newly compiled radial velocity confirms their membership in Sixty six cases, or revises it to a different association in nine cases. One other star (EX Cet) that we identify as a candidate member of βPMG was classified as a member of the Hercules Lyra (HLY) association by López-Santiago et al. (2006), but this association was not included in BANYAN Σ because it is likely composed of non-coeval stars (Mamajek 2015). Given that EX Cet is located at 4 km s⁻¹ from the core of the BANYAN Σ kinematic model of βPMG, we do not draw any conclusion on its membership here.

The G − J versus NUV − G colors of the stars with an entry in the data release 5 of GALEX (Martin et al. 2005) were compared with the field and young sequences and are displayed in Figure 8. A total of 133 stars in our sample have an entry in GALEX; the main reason for this incompleteness is the poor coverage of GALEX at Galactic latitudes south of l = 30°. This figure is similar to Figure 2 of Rodriguez et al. (2011), but uses Gaia colors instead of those based on Tycho V. The field sequence was built from a combination of the field sample of Rodriguez et al. (2011), complemented with all stars in the sample of Oh et al. (2017), from which all known young stars were removed, as well as any star with a BANYAN Σ probability above 1% of belonging to a known young association. Any groups of stars in the Oh et al. (2017) sample that display signs of youth as a population (see Faherty et al. 2018) were also excluded from our field sample. The sample of young associations compiled by Gagné et al. (2018c) are compared with the field NUV sequence in Figure 8, which allowed us to derive a simple criterion that delimitates regions dominated by young stars:

$$\begin{eqnarray}{NUV}-G & \lt & 6.5(G-J)-1.09\ \mathrm{if}\ G-J\lt 1.476,\\ {NUV}-G & \lt & 1.5(G-J)-6.30\ \mathrm{if}\ G-J\geqslant 1.476,\end{eqnarray} \tag{ 2 }$$

Zoom In Zoom Out Reset image size

This criterion is especially efficient at identifying the NUV excess of G-type stars or later (G – J > 1). For earlier-type stars, the young and field sequences become gradually harder to distinguish, and this criterion will fail to identify most young stars.

The candidate members identified here were similarly cross-matched with the second ROSAT all-sky survey catalog (Boller et al. 2016) and the XMM-Newton slew survey Source Catalog v2.0 (XMM-SSC 2018). A total of 222 stars in our sample have an entry in either one of these catalogs. The main limitation of this cross-match is the limited ROSAT sensitivity, which does not allow it to detect the furthest and/or lowest-mass stars in our sample. All stars with a HR1 hardness ratio above −0.15, consistent with the observed distribution of young βPMG, TWA, and THA members in Kastner et al. (2003), were flagged as likely young. The distribution of HR1 values are displayed in Figure 9. Stars with a mid-infrared excess detected at a $\geqslant$ 1.5σ statistical significance in the McDonald et al. (2017) catalog were also flagged as likely young. McDonald et al. (2017) analyzed most of the TGAS sample, but some of the stars could not be cross-matched to older catalogs because they ignored their proper motion, and they used a more conservative cut on the distance precision (they required a statistical significance above 2.4σ versus our 2σ requirement), which resulted in only 656 stars in our sample having such a mid-infrared excess measurement. These measurements are displayed in Figure 10. Žerjal et al. (2017) provide age constraints with a precision of ∼50% for RAVE survey stars based on the Ca ii infrared triplet chromospheric activity indicator. There are 17 stars in our sample that have such activity-based ages consistent with their respective association. Lithium equivalent widths were also reported by Torres et al. (2006), White et al. (2007), da Silva et al. (2009), Moór et al. (2013), and Pecaut & Mamajek (2016) for 114 stars in our sample. Measurements of lithium equivalent widths above 100 mÅ were adopted as a sign of youth (e.g., see Moór et al. 2013). A weaker lithium line is not necessarily inconsistent with youth, but it does not provide a strong indication of youth. All youth indicators based on NUV, X-ray, lithium, infrared excess, or the Ca ii infrared triplet are reported in Tables 3 and 4.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 4.  Candidate Members Recovered in This Work

Name	Assoc.	Spectral	R.A.	Decl.	${\mu }_{\alpha }\cos \delta$	μδ	Distance	Rad. Vel.	HR1	NUV	IRex	Known	Youthb	Referencesc
		Typea	(hh:mm:ss)	(dd:mm:ss)	(mas yr−1)	(mas yr−1)	(pc)	(km s−1)		(mag)	$N\sigma$	Group
New Bona Fide Members
CD–26 1643	ABDMG	F9V	04:21:33.518	−25:43:10.14	42.74 ± 0.31	−55.46 ± 0.33	54.84 ± 0.85	23 ± 2	−0.15	13.099 ± 0.004	0.0	⋯	X	1, 2, ⋯, ⋯
HD 61518	ABDMG	F5V	07:37:02.743	−52:21:56.94	−31.93 ± 0.05	6.13 ± 0.06	60.95 ± 0.84	31 ± 1	−0.23	⋯	2.7	⋯	E	3, 4, ⋯, ⋯
New Young Candidate Members
HD 236717	ABDMG	K0IV/V	01:24:27.324	+57:51:04.71	112.66 ± 0.50	−96.16 ± 0.48	38.50 ± 0.35	⋯	⋯	⋯	⋯	⋯	Lm	23, ⋯, ⋯, ⋯
BD+49 646	ABDMG	K0	02:22:34.008	+50:33:35.42	75.56 ± 0.90	−99.52 ± 0.34	51.90 ± 0.61	⋯	−0.12	⋯	⋯	⋯	X	24, ⋯, ⋯, ⋯
New Candidate Members with No Known Signs of Youth
G 133–40	ABDMG	(M0)	01:47:29.691	+34:13:06.28	119.9 ± 1.9	−195.28 ± 0.60	29.72 ± 0.42	⋯	⋯	21.4 ± 0.3	⋯	⋯	⋯	⋯, ⋯, ⋯, ⋯
CD–30 658	ABDMG	(K2)	01:54:17.741	−29:23:24.20	74.40 ± 0.85	−64.5 ± 1.9	48.2 ± 2.0	⋯	⋯	18.94 ± 0.06	0.0	⋯	⋯	⋯, ⋯, ⋯, ⋯
Reclassified Candidate Members
2M0346+1709	ABDMG	K4/5	03:46:14.875	+17:09:08.33	47.23 ± 0.60	−113.02 ± 0.30	53.02 ± 0.84	⋯	0.16	19.21 ± 0.07	⋯	Nonmember	X,UV	22, 20, ⋯, 42
EX Cet	BPMG	G9Vk	01:37:35.640	−06:45:39.00	171.90 ± 0.05	−98.17 ± 0.04	23.96 ± 0.17	11.706 ± 0.002	−0.30	13.960 ± 0.004	0.0	HLY	⋯	43, ⋯, ⋯, 44
Confirmed Candidate Members
CD–41 2076	ABDMG	K5Ve	05:48:30.422	−41:27:30.24	13.5 ± 1.5	−6.5 ± 2.3	55.5 ± 1.4	31 ± 1	−0.06	⋯	2.2	ABDMG	X, E	10, 10, 10, 10
V1841 Ori	BPMG	K2IV	05:00:49.303	+15:26:59.82	17.3 ± 2.3	−58.2 ± 1.5	52.35 ± 0.84	18.1 ± 0.9	−0.05	⋯	1.0	BPMG	X, Li, Lm	10, 65, 10, 65
Known Bona Fide Members
PW And	ABDMG	K0Ve	00:18:21.057	+30:57:19.56	141.45 ± 0.47	−172.94 ± 0.62	28.22 ± 0.32	−11.0 ± 0.1	−0.09	⋯	⋯	ABDMG	X, Li	40, 94, 39, 42
HD 15407	ABDMG	F5V	02:30:50.800	+55:32:52.83	81.32 ± 0.04	−95.56 ± 0.03	49.44 ± 0.74	−13 ± 1	⋯	⋯	4.2	ABDMG	E	95, 12, ⋯, 95
Known Candidate Members
HW Cet	ABDMG	(K2)	03:12:34.343	+09:44:55.75	61.2 ± 1.6	−91.36 ± 0.83	55.19 ± 0.89	⋯	⋯	⋯	⋯	ABDMG	Li	11, ⋯, ⋯, 115
HD 24681	ABDMG	G5V	03:55:20.447	−01:43:46.57	42.09 ± 0.90	−91.43 ± 0.59	55.24 ± 0.76	⋯	−0.04	⋯	0.0	ABDMG	X, Li	⋯, ⋯, 39, 116
Rejected Candidate Members
TYC 8104–898–1	OCT	(G1)	06:38:02.156	−45:13:48.30	−15.31 ± 0.73	15.26 ± 0.77	161.1 ± 6.6	10 ± 2	⋯	⋯	1.2	⋯	⋯	⋯, 7, ⋯, ⋯
TYC 9341–1233–1	OCT	(G9)	23:09:11.244	−72:21:12.78	27.8 ± 1.0	4.77 ± 0.92	115.3 ± 3.6	4.8 ± 0.9	⋯	19.32 ± 0.06	0.9	⋯	⋯	⋯, 7, ⋯, ⋯

Notes.

^aSpectral types in parentheses were estimated using the G − J color with the spectral type–color relations of Pecaut & Mamajek (2013); see also http://www.pas.rochester.edu/~emamajek/EEM_dwarf_UBVIJHK_colors_Teff.txt. ^bSigns of youth compiled from the literature. See Section 4 for more detail. X: X-ray emission with HR1 ≥−0.15; UV: Galex NUV − G versus G − J consitent with youth; Li: Lithium absorption above 100 mÅ; E: Mid-infrared excess; Lm: Luminosity class consistent with youth; Is: Young isochronal age consistent with; Ca: Ca ii infrared triplet age consistent with proposed association. ^cReferences for: (1) spectral type, (2) radial velocity, (3) lithium detection, and (4) membership candidacy in a young association.

References. (1) Pribulla et al. (2014), (2) Holmberg et al. (2007), (3) Jaschek et al. (1964), (4) Kharchenko et al. (2007), (5) White et al. (2007), (6) Nordström et al. (2004), (7) Kunder et al. (2017), (8) Houk (1978), (9) Kharchenko (2001), (10) Torres et al. (2006), (11) Houk & Cowley (1975), (12) Gontcharov (2006), (13) Moór et al. (2013), (14) Nesterov et al. (1995), (15) Gray et al. (2006), (16) Kordopatis et al. (2013), (17) Houk & Swift (1999), (18) Valenti & Fischer (2005), (19) Houk (1982), (20) Soubiran et al. (2013), (21) Rousseau et al. (1996), (22) Gray et al. (2003), (23) Yoss (1961), (24) Hill & Schilt (1952), (25) Wright et al. (2003), (26) Roeser & Bastian (1988), (27) Cannon & Mayall (1949), (28) Abt (1981), (29) Houk & Smith-Moore (1988), (30) Wilson (1953), (31) Westerlund et al. (1981), (32) Upgren et al. (1972), (33) Fabricius et al. (2002), (34) Loth & Bidelman (1998), (35) Jackson & Stoy (1955), (36) Goedicke (1945), (37) da Silva et al. (2009), (38) Skiff (2014), (39) Wehinger & Hidajat (1973), (40) Christian et al. (2001), (41) Gaidos et al. (2014), (42) López-Santiago et al. (2006), (43) Stephenson (1986b), (44) Schlieder et al. (2010), (45) Esplin et al. (2014), (46) Kohlschütter (1920), (47) Cannon & Pickering (1993), (48) Alcalá et al. (1996), (49) Röser et al. (2011), (50) Elliott et al. (2016), (51) Messina et al. (2010), (52) Makarov & Urban (2000), (53) Pecaut & Mamajek (2016), (54) Slawson et al. (1992), (55) Hoogerwerf (2000), (56) Krautter et al. (1997), (57) Hughes et al. (1993), (58) Glaspey (1972), (59) Patterer et al. (1993), (60) Cheetham et al. (2015), (61) Viana Almeida et al. (2009), (62) Kharchenko et al. (2004), (63) Merín et al. (2008), (64) Rizzuto et al. (2011), (65) Kraus et al. (2014), (66) Johnson & Mitchell (1958), (67) Cayrel de Strobel et al. (2001), (68) Mermilliod et al. (2009), (69) Hartman et al. (2010), (70) Kraft (1967), (71) Stauffer et al. (2007), (72) Mendoza (1956), (73) Magazzù et al. (1997), (74) Tokovinin & Smekhov (2002), (75) Wilson (1962), (76) Schlieder et al. (2012), (77) Hoffleit et al. (1970), (78) Evans (1967), (79) Chen et al. (2011), (80) Upgren (1962), (81) Mermilliod et al. (2008), (82) Song et al. (2012), (83) Mamajek et al. (2002), (84) Dahm et al. (2012), (85) Köhler et al. (2000), (86) Luhman & Mamajek (2012), (87) Preibisch & Zinnecker (1999), (88) Slettebak (1963), (89) Schlaufman & Casey (2014), (90) Desidera et al. (2015), (91) Adams et al. (1935), (92) Galli et al. (2017), (93) Hoogerwerf & Aguilar (1999), (94) Montes et al. (2001), (95) Melis et al. (2010), (96) van Leeuwen et al. (1986), (97) Wichmann et al. (2000), (98) Osawa (1959), (99) Walter & Boyd (1991), (100) Abt & Morrell (1995), (101) Zuckerman & Song (2012), (102) de Zeeuw et al. (1999), (103) Gahm et al. (1989), (104) Bertiau (1958), (105) Walter et al. (1994), (106) Shkolnik et al. (2017), (107) Gaia Collaboration et al. (2017), (108) Stephenson (1986a), (109) Endl et al. (2006), (110) Chubak & Marcy (2011), (111) Ammler-von Eiff & Guenther (2009), (112) Keenan & McNeil (1989), (113) King et al. (2003), (114) Zuckerman & Song (2004), (115) Zuckerman et al. (2001), (116) McCarthy & White (2012), (117) Binnendijk (1946), (118) Li & Hu (1998), (119) Stauffer et al. (1991), (120) Gratton (1939), (121) Bouy et al. (2015), (122) Soderblom et al. (1993), (123) Bourgés et al. (2014), (124) Feigelson et al. (1987), (125) Eggen (1969), (126) Rebull et al. (2011), (127) Randich et al. (1995), (128) Lodén (1969), (129) Kraus & Hillenbrand (2007), (130) Casewell et al. (2006), (131) Odenkirchen et al. (1998), (132) Garrison & Gray (1994), (133) Lodén (1980), (134) Spencer Jones & Jackson (1939), (135) Stephenson & Sanduleak (1975), (136) Murphy et al. (2015), (137) Preibisch et al. (1998), (138) Rizzuto et al. (2015), (139) Garrison (1967), (140) Struve & Straka (1962), (141) Carpenter et al. (2006), (142) Lindblad (1922), (143) Joy (1949), (144) Nguyen et al. (2012), (145) Luhman et al. (2009), (146) Bell et al. (2017), (147) Silaj & Landstreet (2014), (148) Mamajek et al. (2000).

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

We used the MIST solar-metallicity model tracks of Choi et al. (2016) to determine isochronal ages for the 38 new candidate members with complete kinematics. To do so, we calculated the minimum $N\sigma$ distance of the star to each isochrone in absolue Gaia M_G versus Gaia–2MASS G − J to build a probability density function as a function of age. The probability density functions were then visually inspected, and were classified in one of three categories to determine whether (1) they do not provide a significant age constraint, (2) they provide a unimodal age constraint, or (3) they provide a bimodal age constraint (i.e., a pre-main sequence age and a post-main sequence age). The resulting pre-main sequence ages are reported in Table 3, and corroborate membership in all cases but one: the pre-main sequence isochronal age of the A2 TAU candidate HD 284659 is ${8}_{-2}^{+1}$ Myr, which is significantly older than the estimated age of TAU (1–2 Myr; Kenyon & Hartmann 1995). It remains unclear whether HD 284659 is a 1–2 Myr member of TAU (in which case the MIST tracks systematically overestimate the age in this very young regime), or if it is part of an older sub-group with similar kinematics to TAU that is not considered in BANYAN Σ. HD 284659 is located spatially within the distribution of TAU members (at 5.8 pc, or 0.6σ from the core of the BANYAN Σ spatial model), and at 3.9 km s⁻¹ (or 1σ) from the core of the BANYAN Σ kinematic model. Finding more 6–9 Myr objects in the vicinity of HD 284659, or calibrating the MIST tracks for 1–2 Myr A-type stars using more empirical data, would help in clarifying the membership of HD 284659. Alternatively, it is possible that HD 284659 is instead a ${560}_{-100}^{+70}$ Myr old star that is starting to depart from the main sequence and that happens to share the spatial position and kinematics of TAU by pure chance. Figure 11 displays the MIST solar-metallicity tracks compared with stars of Table 3 and empirical color-magnitude sequences derived in Section 2.

Zoom In Zoom Out Reset image size

A total of 32 candidate members with full kinematics have at least one sign of youth that is consistent with their kinematic membership, and we therefore propose them as new bona fide members of their respective young associations. The OCT candidate member TYC 8104–898–1 has no literature information that allows us to put a constraint on its age, and its position in a M_G versus G − J color-magnitude diagram is slightly underluminous compared with all solar-metallicity MIST tracks. The RAVE data release 5 catalog (Kunder et al. 2017) provides a sub-solar-metallicity measurement of [m/H] = −0.26 ± 0.09 dex for this star, which may explain its peculiar absolute magnitude. We therefore reject TYC 8104–898–1 as a candidate member of OCT. TYC 9341–1233–1 is also rejected because its GALEX–Gaia NUV − G color is consistent with field stars, and no conclusion is drawn on the TAU membership of HD 284659 because of its inconsistent (but likely young) isochronal age.

Seven stars in our sample have signs of youth but are rejected as candidate members of young associations. These stars could be slightly older and have formed in groups that are now completely dissolved, scattered members of known associations, or members of associations not yet discovered. Similarly, several young brown dwarfs without a clear origin were identified in previous work (e.g., Gagné et al. 2015a; Faherty et al. 2016). The upcoming of Gaia DR2 will help understanding the origin of these objects.

We used the BANYAN Σ tool in combination with TGAS to identify 32 new F0–M3 bona fide members of nearby young associations and 219 additional new candidate members. These new candidate members include HD 121191, a new A-type candidate member of UCL which is known to have a significant infrared excess (Melis et al. 2013; McDonald et al. 2017), and three other infrared excess candidate members of associations older than ∼10 Myr (TYC 8602–718–1 in βPMG; CD–60 2373 in PL8; and TYC 8602–718–1 in CAR). The discovery of accretion disks around these slightly older stars would be valuable to understand their lifetimes, as only few such examples are currently known (e.g., Boucher et al. 2016; Silverberg et al. 2016; Murphy et al. 2018). Five new candidate members of CARN and ABDMG are located within 30 pc of the Sun (the nearest one, HD 19819, is a candidate member of CARN at ∼22 pc), and 26 objects in our sample are new A-type candidate members of young associations. These stars will be particularly valuable for direct-imaging searches of exoplanets, as proximity makes it possible to detect companions at smaller separations, and massive stars are known to have a larger occurrence of giant exoplanet companions which are easier to detect (Lannier et al. 2017). Furthermore, 26 new candidate members (21 with signs of youth) of OCT, TAU, UCL, and USCO are located at spatial distances above 150 pc, making them $\gt 5$σ outliers to the spatial model of their respective association, while displaying consistent kinematics. This hints that these associations may extend to larger distances not yet explored, due to the paucity of parallax measurements previously available beyond 150 pc. The new candidates presented here have the potential to almost double the number of members in associations that were not extensively studied in the literature such as PL8 and OCT.

This work hints that the upcoming Gaia DR2 will allow us to uncover many new young stars, complete their kinematic picture and investigate the initial mass function of several young associations down to the regime of low-mass stars. The identification of new young stars and their assignment of accurate ages will be useful to build standard stellar populations, and will provide valuable targets for direct-imaging searches of exoplanets.

We thank the anonymous referee for useful comments. We thank Dustin Lang for providing the TGAS–2MASS cross-match data, David Rodriguez from providing part of the data used to build the field sequence in Figure 8, and Eric E. Mamajek for useful comments. This research made use of the SIMBAD database and VizieR catalog access tool, operated at the Centre de Données astronomiques de Strasbourg, France (Ochsenbein et al. 2000); data products from the Two Micron All-Sky Survey (2MASS; Skrutskie et al. 2006), which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center (IPAC)/California Institute of Technology (Caltech), funded by the National Aeronautics and Space Administration (NASA) and the National Science Foundation (Skrutskie et al. 2006); and data products from the Wide-field Infrared Survey Explorer (WISE; and Wright et al. 2010), which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory (JPL)/Caltech, funded by NASA. This project was developed in part at the 2017 Heidelberg Gaia Sprint, hosted by the Max-Planck-Institut für Astronomie, Heidelberg. This work made use of data from the European Space Agency (ESA) mission Gaia (http://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC; http://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. We note that 46 of the 254 new candidate members presented in this paper were also independently uncovered in Faherty et al. (2018).

J.G. wrote the codes, manuscript, generated figures, and led all analyses; O.R.L. performed parts of the literature cross-matches and interpretation, and helped generate lists of new candidate members; J.K.F. helped parsing young association literature data and provided general comments; R.D. shared comments and supervized O.R.L.; and L.M. helped with the construction of color-magnitude sequences.

Software: BANYAN Σ (Gagné et al. 2018c).