THE $\mu$ TAU ASSOCIATION: A 60 MYR-OLD COEVAL GROUP AT 150 pc FROM THE SUN

Young stellar associations in the Solar neighborhood ($\lesssim$ 200 pc) are valuable laboratories to study stellar evolution and refine our age-dating methods because they contain groups of stars with many different masses that formed coevally from the same molecular cloud (e.g., Zuckerman & Song, 2004; Torres et al., 2008). Their proximity is valuable because their members appear brighter, but it also causes them to be spread over larger areas of the sky, which makes their initial identification less straightforward. Obtaining credible lists of members with low contamination by unrelated field stars is challenging and typically requires measuring the six-dimensional position and space velocity of each member. As these stars formed from a single molecular cloud, they share the same velocities typically within $\simeq$ 2–4 km s^-1, allowing us to distinguish them from most field stars.

Until recently, trigonometric distance measurements were only available for a limited set of bright stars (e.g., Perryman et al. 1997), and radial velocity measurements of stars in the Solar neighborhood were even more limited to small-scale samples (e.g., see Gontcharov, 2006; White et al., 2007). This led to the identification of co-moving and coeval massive stars that represented only the tip of the iceberg of each young association of stars in our neighborhood (Zuckerman & Song, 2004; Torres et al., 2008). Efforts have been made to identify the lower-mass population based on various methods that can assign membership probabilities with missing parts of the 6-dimensional space and velocity, including the convergent point method (Mamajek, 2005; Torres et al., 2006) and various other flavors of selection cuts in space-velocity and/or photometry (Zuckerman & Song, 2004; Kraus et al., 2014; Riedel et al., 2017; Shkolnik et al., 2017) as well as methods based on Bayesian statistics (Malo et al., 2013; Gagné et al., 2014, 2018).

The second data release of the Gaia mission (Gaia DR2 hereafter; )¹¹1 changed this landscape completely in April of 2018 by providing trigonometric distance measurements for $\simeq$ 1.3 billion stars with an unprecedented precision, as well as radial velocities for more than 7.2 million bright stars. This allowed us to complete the 6-dimensional kinematics for a number of stars on a completely new scale, which led to a plethora of scientific discoveries that quickly unveiled the spatial and kinematic structure of the Solar neighborhood as well as the Milky Way in general. Some of these discoveries include many new associations of stars (Oh et al., 2017; Faherty et al., 2018; Gagné et al., 2018a; Kounkel & Covey, 2019; Meingast et al., 2019), a large number of new M-type members of known associations (Gagné et al., 2018c; Gagné & Faherty, 2018; Luhman, 2018; Reino et al., 2018; Zuckerman, 2019; Tang et al., 2019), as well as the discovery of tidal disruption tails around three older, nearby clusters: the Hyades (Röser et al., 2019), Praesepe (Röser & Schilbach, 2019) and Coma Ber (Tang et al., 2019).

This paper presents the discovery and characterization the MUTA association, based on an initial list of massive co-moving and coeval members that had been discovered in historical surveys but never before published. The advent of Gaia DR2 allowed us to complete this list and characterize MUTA such that it will become yet another important laboratory for the investigation of stellar evolution and the grounds for discovery of age-calibrated brown dwarfs and exoplanets. In Section 2, we present the initial list of MUTA members, which we use to build a spatial-kinematic model (Section 3) to search for additional members with the BANYAN $\Sigma$ Bayesian identification tool (Gagné et al., 2018) in Section 4. In Section 5, we present an iterative method to correct interstellar extinction in Gaia DR2 color-magnitude diagrams, required because the photometric bandpasses are wider than usual. We discuss the properties of MUTA as a whole and its individual members in Section 6, including their present-day mass function and stellar activity indicators, and a comparison with the Galactic kinematic structures recently unveiled by Kounkel & Covey (2019). We summarize and conclude this work in Section 7.

The existence of a distinct group of co-moving young stars in the vicinity of the Taurus-Auriga (Kenyon et al., 2008) star-forming region first appeared in a spatial distribution of Cas-Tau OB-type stars assembled by Blaauw (1956). Cas-Tau was identified by Blaauw (1956) as an extended group of co-moving stars with an expansion age of $\simeq$ 50 Myr that seems to be on the way to being dissolved. They noted that Cas-Tau may share a common origin with an extended stream of stars around the $\alpha$ Persei cluster (e.g., see Heckmann & Lübeck, 1958; Lodieu et al., 2019) identified by Rasmuson (1921). An over-density in the Cas-Tau stars seemed to be located at Galactic coordinates $\left(\ell,b\right)=\left(190^{\circ},-10^{\circ}\right)$, and was recovered as part of the de Zeeuw et al. (1999) census of nearby OB associations (see their Fig. 19). This over-density overlaps with subgroup 5 of Cas-Tau defined by Blaauw (1956), with five B-type stars in common (29 Tau, 30 Tau, 35 Eri, $\mu$ Tau, $\mu$ Eri) and one additional star (40 Tau) not in common that seems to be an un related background star. Combining this list of 12 early-type stars assembled by de Zeeuw et al. (1999) to other co-moving B, A and F-type stars in the range $\ell$ from 170° to 205° and $b$ from $-$40° to $-$27° as well as nearby ROSAT entries (Boller et al., 2016) in the same region yielded a total set of 35 stars that appeared to be young and co-moving within 15 $\mathrm{mas}\,\mathrm{yr}^{-1}$ of the average proper motions of the de Zeeuw et al. (1999) list ($\mu_{\alpha}\cos\delta$ = 21.0 $\mathrm{mas}\,\mathrm{yr}^{-1}$, $\mu_{\delta}$ = $-$20.5 $\mathrm{mas}\,\mathrm{yr}^{-1}$). Four of these 37 stars are clear outliers either in $XYZ$ (HD 23110, TYC 657–794–2, and HD 28796) or $UVW$ (HIP 18778) and were excluded from our initial list. The resulting 33 stars are listed in Table 2 with their properties. We tentatively named this group $\mu$ Tau Association (MUTA) after one of its brightest members. We assigned initial members with MUTA identification numbers (from 1 to 30) in order of decreasing $V$-band brightness. We assigned the same MUTA ID to binaries with separations below 15^′′.

In a more recent analysis or the Gaia Data Release 1 (DR1), Oh et al. (2017) recovered about a third of the stars in Table 2 as three broken up groups of co-moving systems, which they named Groups 43 (6 matches), 52 (3 matches) and 60 (4 matches). The overlap between our initial list of MUTA members and the Oh et al. (2017) sample are shown in Figure 1, where part of the Taurus star-forming region can be seen at a similar distance from the Sun (see e.g. Wichmann et al. 2000), and the Hyades cluster (Perryman et al., 1998) also appears in the foreground. The method that Oh et al. (2017) used to identify systems of co-moving stars works directly in proper motion and parallax space, which tends to recover spatially large moving groups only as broken parts, explaining why the spatially extended MUTA was broken up in three groups, similarly to other nearby young moving groups (Faherty et al., 2018).

We cross-matched our initial list of MUTA members with Gaia DR2 data to build a color-magnitude sequence shown in Figure 2 to demonstrate they constitute massive OBA-type stars ($G-G_{\rm RP}<0.2$) and a well-defined sequence of later-type stars ($G-G_{\rm RP}>0.2$), providing further evidence that they are coeval and young.

The earliest-type member in our initial list is 29 Tau (MUTA 5), a B3 V-type star (Beavers & Cook, 1980), which corresponds to a mass of $\simeq$ 5.4 $M_{\odot}$ (Pecaut & Mamajek, 2013). Hohle et al. (2010) and Gullikson et al. (2016) estimated the mass of 29 Tau based on evolutionary tracks and found respective values of $6.0\pm 0.7$ $M_{\odot}$ and $5.4\pm 0.6$ $M_{\odot}$, consistent with the expected mass for a B3 star. Following the evolutionary tracks of Choi et al. (2016), such a star has a main-sequence life of only $\simeq$ 80 Myr, indicating that the MUTA association is likely younger than the Pleiades.

We note that both $\mu$ Tau (MUTA 2) and $\tau^{1}$ Ari are known eclipsing binaries (Avvakumova et al., 2013). While the first is part of our initial list of members, $\tau^{1}$ Ari was identified in an earlier parsing of de Zeeuw et al. (1999) but was not included because of its discrepant $UVW$ motion (it is separated from the other stars by $\simeq$6.3 km s^-1). A further analysis of their respective light curves might be useful for constraining models of stellar structure at young ages.

The BANYAN $\Sigma$ tool (Gagné et al., 2018) makes it possible to identify additional stars with similar Galactic positions $XYZ$ and space velocities $UVW$ compared to our initial list of MUTA members, if we provide it a 6-dimensional multivariate Gaussian model for MUTA in $XYZUVW$ space. One of the main benefits of BANYAN $\Sigma$ is its ability to recover stars with only partial kinematics, often a consequence of missing radial velocity or parallax measurements. The BANYAN $\Sigma$ tool currently includes kinematic models for 29 nearby young associations, which consist of the 27 associations described in Gagné et al. (2018), as well as the recently discovered Volans-Carina (Gagné et al., 2018a) and the Argus associations (Makarov & Urban, 2000), whose census of members was recently revised by Zuckerman (2019).

We compiled literature radial velocity measurements for the stars listed in Table 2 to identify a set of 25 core members with complete kinematics (see Table 2). This list excludes any gravitationally bound companion to avoid artificially giving each system more weight in the kinematic construction of the MUTA model (consistent with the model construction method of Gagné et al. 2018). HD 28715, HD 23990, HD 23538, and HD 27687 (MUTA 11, 13, 14, and 17, respectively) currently do not have radial velocity measurements and were not included in Table 2 although they are likely part of MUTA based on their position in a color-magnitude diagram (see Figure 2) and their common proper motion and parallax compared to the other members.

The methodology described in Gagné et al. (2018; see their Section 5) was used to build a $XYZUVW$ multivariate Gaussian model of the stars listed in Table 2. In summary, a 6-dimensional average vector and covariance matrix in $XYZUVW$ space were built by calculating the average, variance and covariances of the 25 core members with full kinematics listed in Table 2. When calculating the averages, variances and covariances, the individual measurements were weighted proportionally to the squared inverse of their individual error bars to minimize the impact of low quality measurements. The covariance matrix is then regularized to ensure its determinant is finite and positive with a singular value decomposition step. The resulting model is shown in Figure 3.

The average sky position of MUTA members is 04:01:29.54, $+$07:59:33.3 ($60.3731$°, $7.9926$°) with a standard deviation of 5° in both directions. The average galactic coordinates $\left(\ell,b\right)$ are ($182.4658$°, $-31.8645$°) with a standard deviation of ($9$°, $3$°). The total velocity $S_{\rm tot}$ of the members averages 28.3 km s^-1 with a standard deviation of 2.3 km s^-1. The $UVW$ values we find correspond to a convergent point of $103.380$°, $-29.325$° in right ascension and declination (06:53:31, $-$29:19:30).

The kinematic model described in Section 3 was combined with the BANYAN $\Sigma$ tool to identify candidate members of MUTA in Gaia DR2 data. We pre-selected only Gaia DR2 entries with right ascensions in the range 10–150°, declinations in the range $-$20 to $+$40° and trigonometric distances within 300 pc of the Sun. These limits are significantly wider than the ranges of sky positions (47° to 72° and $-$3.5° to 16.5°, respectively) and distances (all in the range 130–220 pc) of the initial list of members. The sky positions, proper motions and parallaxes from Gaia DR2 were used to determine a membership probability, as well as the Gaia DR2 radial velocities when available. We selected only the stars with Bayesian membership probabilities above 90% and a maximum likelihood separation of less than 5 km s^-1 from the core of our MUTA kinematic model in $UVW$ space as new candidate members. The latter criterion avoids selecting stars that would fit all BANYAN $\Sigma$ models poorly, including its model of the local Galactic neighborhood.

These selection criteria resulted in a set of 503 additional candidate members which are listed in Table 3. Their common proper motion is illustrated in Figure 4 and their positions in a Gaia DR2 $G-G_{\rm RP}$ color versus absolute $G$ magnitude are shown in Figure 5. Their sky positions are located in the range 37–74° and $-$4° to $+$29° in right ascension and declination, and their trigonometric distances are in the range 100–220 pc, indicating that our initial filtering of Gaia DR2 entries was likely appropriate to encompass the full distribution of MUTA members.