Evidence for Radial Expansion at the Core of the Orion Complex with Gaia EDR3
Cameren Swiggum, Elena D'Onghia, João Alves, Josefa Gro?schedl, Michael Foley, Catherine Zucker, Stefan Meingast, Boquan Chen, Alyssa Goodman
DDraft version January 28, 2021
Typeset using L A TEX twocolumn style in AASTeX63
Evidence for Radial Expansion at the Core of the Orion Complex with
Gaia
EDR3
Cameren Swiggum, Elena D’Onghia, Jo˜ao Alves, Josefa Gro ß schedl, Michael Foley, Catherine Zucker, Stefan Meingast, Boquan Chen, and Alyssa Goodman Department of Astronomy, University of Wisconsin, 475 North Charter Street, Madison, WI 53706, USA University of Vienna, Department of Astrophysics, T¨urkenschanzstraße 17, 1180 Vienna, Austria Center for Astrophysics | Harvard & Smithsonian, 60 Garden St., Cambridge, MA 02138, USA Sydney Institute for Astronomy, The University of Sydney, School of Physics A28, Camperdown, NSW 2006, Australia
ABSTRACTWe present a phase-space study of two stellar groups located at the core of the Orion complex:Brice˜no-1 and Orion Belt Population-near (OBP-near). We identify the groups with the unsupervisedclustering algorithm, Shared Nearest Neighbor (SNN), which previously identified twelve new stellarsubstructures in the Orion complex. For each of the two groups, we derive the 3D space motions ofindividual stars using
Gaia
EDR3 proper motions supplemented by radial velocities from
Gaia
DR2,APOGEE-2, and GALAH DR3. We present evidence for radial expansion of the two groups from acommon center. Unlike previous work, our study suggests that evidence of stellar group expansionis confined only to OBP-near and Brice˜no-1 whereas the rest of the groups in the complex showmore complicated motions. Interestingly, the stars in the two groups lie at the center of a dustshell, as revealed via an extant 3D dust map. The exact mechanism that produces such coherentmotions remains unclear, while the observed radial expansion and dust shell suggest that massivestellar feedback could have influenced the star formation history of these groups.
Keywords:
Star forming regions: individual (Orion Complex) – stars: kinematics; dynamics; ages INTRODUCTIONThe formation and evolution of star clusters is a com-plex process that depends crucially on the formation ofthe individual stars and how feedback mechanisms, in-cluding winds, radiation pressure, and supernovae, affecttheir environment. Young stars are usually grouped inclusters and are located within their natal star-formingregions. In contrast to younger stars, older stars arefound dispersed throughout the Galactic field.The role of gas in the formation and evolution ofbound clusters or associations of stars is still poorlyunderstood (e.g., Krause et al. 2020). A popular sce-nario posits that stars form and temporarily persist indense molecular clouds, held together by the gravita-tional potential of the remaining gas (Lada & Lada2003). The gas is eventually blown-out by feedback fromstellar winds and supernovae events. Stars may dispersedue to this gas expulsion and the associated change ingravitational potential (Hills 1980; Goodwin & Bastian
Corresponding author: Cameren [email protected] ≈
400 pc, Zucker et al. 2020; Großschedl et al.2018), its mass ( > M (cid:12) ; Lada et al. 2010), and thepresence of multi-phase gas and stars at various evolu-tionary stages, Orion represents one of the best placesto observe young stars in their natal environments.Before the Gaia era, the classification of reliable as-sociations and young clusters within the Orion complexwas a challenge. While overdensities of stars were iden-tified (e.g. Blaauw 1964; Brown et al. 1994; Bally 2008),often these overdensities appeared superimposed on theplane of the sky. The lack of accurate distances andproper motions prevented a separation of distinct groups a r X i v : . [ a s t r o - ph . GA ] J a n Swiggum et al. in true 3D space, making it difficult to classify the stellarassociations in the Orion complex.This situation has changed with the advent of the Eu-ropean Space Agency’s (ESA)
Gaia
Mission, which hasprovided distances and proper motions for tens of thou-sands of young stars in the vicinity of Orion (Gaia Col-laboration et al. 2016). These data, supplemented byradial velocities (RVs) from
Gaia and the Sloan Dig-ital Sky Survey IV (SDSS-IV; Majewski et al. 2017)APOGEE-2 (Blanton et al. 2017), and GALAH DR3(Buder et al. 2020), have allowed the use of clusteringalgorithms to identify the stellar associations in phase-space, with implications for understanding how youngclusters form and evolve.A recent study of the proper motions of stellar pop-ulations within the Orion Nebular Complex (ONC) re-veals that this young star-forming region is bound, withlittle evidence of expansion (Kuhn et al. 2019). Otherworks explored the entire Orion complex and identifiednumerous stellar groups of various ages (e.g. Kubiak etal. 2017; Kounkel et al. 2018; Kos et al. 2019; Zari et al.2019). A 6D phase-space analysis of the λ Orion regionhas shown evidence for expansion (Kounkel et al. 2018).Additionally, a gas study of Orion has also shown evi-dence for expansion likely due to strong stellar feedback(Großschedl et al. 2020).Chen et al. (2020) applied an unsupervised clusteringalgorithm to Gaia DR2 data and identified 21 stellargroups in the Orion Complex. While nine of them over-lap with stellar groups previously identified in the Orionregion (Brice˜no et al. 2007; Alves & Bouy 2012; Kounkelet al. 2018), twelve were newly discovered. Recently,Kounkel et al. (2018) cataloged the Orion region groupsusing a different clustering algorithm. Kounkel (2020)further analyzed the groups’ phase space and claimedthat the entire complex is expanding from a central re-gion. Here we show that the expansion is most evidentin two massive stellar groups in Orion, namely OBP-near and Brice˜no-1, which exhibit a “Hubble flow” likeexpansion pattern similar to those found in other star-forming regions (e.g. Wright & Mamajek 2018; Wrightet al. 2019; Kuhn et al. 2019; Cantat-Gaudin et al. 2019;Rom´an-Z´u˜niga et al. 2019). Our analysis uses
Gaia
EDR3, APOGEE-2, and GALAH DR3 to characterizethe stars in the phase space and determine the age ofthe two groups of which OBP-near is newly discoveredin a previous work (Chen et al. 2020).In Section 2, we layout the data and methodology usedto identify the stellar groups and study their dynamicsand ages. Section 3 reports the main results of our study.Section 4 discusses physical scenarios that could lead to our results and compare to previous works. Our resultsare briefly summarized in Section 5. DATA ANALYSIS2.1.
SNN Stellar Groups
Chen et al. (2020) applied two clustering algorithms,namely Shared Nearest Neighbor (SNN; Chen et al.2018) and
EnLink (Sharma & Johnston 2009), to
Gaia
DR2 in the vicinity of Orion, to uncover stellar groupsclustered in RA, DEC, parallaxes ( α, δ, (cid:36) ) and propermotions ( µ α , µ δ ). Advantages of this approach and itsapplied analysis include: 1) the ability to omit stars thatunlikely belong to any group (unclassified stars); 2) theuse of a density threshold criterion to balance the highand low-density regions in the 5D-space; 3) the assign-ment of group stability scores based on how many timesthe star appears in each of the total iterations (7,000 inChen et al. 2020) and group membership probability.Our study focuses on two groups labeled as SNN 1 andSNN 3 in Chen et al. (2020). SNN 1 is named “Brice˜no-1” due to its significant overlap with the 25 Orionis stel-lar population studied by Brice˜no et al. (2007). Thegroup is the most stable of all groups recovered in theoriginal SNN analysis with a stability score of 3393 (outof 7,000). The plane-of-sky region of this stellar popula-tion is notably classified as the Ori OB1a sub-associationand its average distance from the Sun is 350 pc. SNN3 is named “OBP-near” and is newly discovered by theSNN clustering algorithm. It lies in the Ori OB1b sub-association (Blaauw 1964) in the plane of the sky. OBP-near has a stability score of 2847, making it the thirdmost stable group discovered in the SNN analysis. Itsaverage distance from the Sun is 360 pc, establishing it-self at the front of the Orion Belt Population, hence thedistinguishing title ”near”.Compared to
Gaia
DR2, EDR3 shows ∼
30% im-provement in parallax precision and an improvementin proper motions by a factor of ∼ Gaia
EDR3 databaseusing the same region selection for Orion as used inChen et al. (2020), specifically as 75 ◦ < α < ◦ , − ◦ < δ < ◦ , 2 < (cid:36) < − < µ α < − < µ δ < (cid:36)/σ (cid:36) > ,
811 stars. We perform 500 iterations ofSNN, each adopting a range of values for the free pa-rameters n xyz and d pm . The n xyz parameter defines thenumber of nearest neighbors in 3D Cartesian position-space computed for each star. The d pm parameter is The prominent B star 25 Ori is not a member of SNN 1, hencethe differing choice in the group’s name. rion Expansion n xyz and d pm between10–800 pc and 0–0.40 mas/yr, respectively. For the ini-tial group finding, the parameters min samples and eps are fixed to the values in eps = 0 . min samples = 20.This choice selects candidates with at least 20 stars hav-ing more than 50% shared nearest neighbors. We set eps = 0 . min samples , is set to 15 (see Chen et al. (2020) for de-tails). We adopt a probability cut of 2%, meaning a starwas included in a given group for at least 10 out of 500iterations, greatly reducing the number of overlappedmembers. The average probability is 16% after this cut.Using EDR3, we retrieve Brice˜no-1 and OBP-near,along with the other SNN groups. Around 80% of theoriginal Brice˜no-1 stars are part of the group with 306new stars. Around 52% of the original OBP-near starsbelong to the group with an additional 252 new stars.Most of the stars that are no longer members of thegroups are reclassified as field stars. Gaia
EDR3 averageposition and velocities of both groups are provided inTable 1. 2.2.
Radial Velocities
To study the 3D kinematics of both OBP-near andBrice˜no-1, we require radial velocity (RV) measure-ments. We first cross-match the stars from OBP-nearand Brice˜no-1 with the Sloan Digital Sky Survey IV(SDSS-IV; Majewski et al. 2017) APOGEE-2 (Blantonet al. 2017) adopting
SYNTHVELIO_AVG as the RV mea-surement and SYNTHVERR as their corresponding errors.Additionally, we include more RVs by cross-matchingthe two groups to GALAH DR3 (Buder et al. 2020).When a star has RV measurements in multiple surveys,we adopt the RV with the highest S/N, leading to a pref-erence for APOGEE-2 RVs, followed by GALAH DR3,and then
Gaia
DR2. There is no apparent systematicshift between APOGEE-2 and
Gaia
DR2 RVs, howeverthere is a positive shift in GALAH DR3 RVs by ∼ . − when comparing to APOGEE-2. We subtractthis value from GALAH DR3 RVs to be inline with the Gaia and APOGEE-2 RVs.There is a large spread in the RV distribution, likelydue to spectroscopic binaries in the sample. A CMDinspection of the groups’ stars with discrepant RVs re-veals that many are O and B-type stars. According toChini et al. (2012), the binary rate for stars with mass > M (cid:12) is ∼
80% and drops to ∼
20% for stars with An average of the RV measurements from cross-correlating to thebest-fit synthetic spectrum. mass ∼ M (cid:12) . We exclude these stars by restricting theRVs to the range 17 < RV <
26 km s − . Overall, a totalof 221 stars with accurate RVs remain in the sample ofboth groups.2.3. Positions and Kinematics
Knowing the positions and velocities of the twogroups, we can constrain their dynamical evolution. Weconvert parallax to distance using d = 1 /(cid:36) . We adoptthe Astropy python package (Astropy Collaboration etal. 2018) and convert the sky coordinates, distances,proper motions, and radial velocities of both groups’stars to 3D Galactic cartesian coordinates ( X, Y, Z )pc and Heliocentric velocity components to (
U, V, W )km s − . A correction is made to the solar motion by sub-tracting ( U, V, W ) (cid:12) = (11 . , . , .
25) km s − fromthe stars’ velocities (Sch¨onrich et al. 2010) to obtain(u,v,w) velocities with respect to the LSR. We also sub-tract Galactic rotation values at each star’s Galacto-centric radius using the MilkyWay2014 potential fromthe
Galpy python package (Bovy 2015). The differ-ence in galactic rotation across the two groups is modest( ∼ .
006 km s − pc − ). A correction to the proper mo-tions’ perspective contraction is made using Equation13 from van Leeuwen (2009).Errors on the positions and velocities are estimatedusing a Monte Carlo approach. The astrometric andRV errors are assumed to be normally distributed anduncorrelated given the proximity of Orion and the re-gion’s data quality. For a given star, Gaussian distribu-tions are created for each measurement with the inputerrors corresponding to the distributions’ standard de-viations and the observed values corresponding to themean. These distributions are sampled in parallel 10,000times and transformed from spherical to HeliocentricGalactic cartesian coordinates. The standard deviationsof each dimension’s resulting distribution are then cal-culated and stored as their corresponding errors.Finally, we define our study’s reference frame by sub-tracting the median ( u, v, w ) motions from the combinedgroups and adopting the notation ( X, Y, Z, v x , v y , v z ).2.4. Age Estimates
To estimate the ages of Brice˜no-1 and OBP-near,we fit isochrones to the groups’ CMDs using their
Gaia
EDR3 photometry. We use the latest PARSEC isochrones with the updated EDR3 passband definitions(Marigo et al. 2017). Similar to previous we considerboth a stellar population’s age ( t ) and extinction ( A V )as free parameters when fitting the models with an age http://stev.oapd.inaf.it/cgi-bin/cmd Swiggum et al.
Table 1.
Summary of GroupsGroup Name ¯ α ¯ δ ¯ d ¯ µ α µ δ ¯RV N N RV (A-2) N RV (GDR3) N RV ( Gaia )(deg) (deg) (pc) (mas/yr) (mas/yr) (km/s)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)Brice˜no-1 81 . .
9) 2 . .
1) 346 (11) 1 . . − . .
3) 21 . .
5) 892 71 26 5OBP-near 83 . . − . .
9) 357 (14) 1 . . − . .
2) 22 . .
9) 655 105 7 7
Note —The two stellar groups as identified with the SNN clustering algorithm: Brice˜no-1 and OBP-near. In column (1) we give the nameof the group; (2) average/1 σ spread in RA; (3) average/1 σ spread in DEC; (4) average/1 σ spread in distance; (5) average/1 σ spreadin PMRA; (6) average/1 σ spread in PMDEC; (7) average/1 σ spread in heliocentric radial velocity; (8) number of stars in each group;(9) number of APOGEE-2 radial velocities used; (10) we number GALAH DR3 radial velocities used; (11) number of Gaia
DR2 radialvelocities used. range of 1 < t <
20 Myr and a step size of 0.05 Myrand an extinction range of 0 . < A V < . . Z = 0 . m , to come from an isochrone withparameters, θ = ( t, A V ) as:ln ( L ( θ , m )) = n (cid:88) i =1 ln (cid:18) π ) / σ i (cid:19) − χ χ = n (cid:88) i =1 (cid:32) M obs g i − M g i ( θ , m ) σ i (cid:33) (2)A quality cut is applied to the Gaia photometryfollowing phot_g_mean_flux_over_error >
20 and phot_rp_mean_flux_over_error >
20. We interpo-late the isochrones to match the stars’
Gaia G − G RP color. M g i is the absolute magnitude calculated fromthe G band photometry and parallax. M g i ( t, m ) is theabsolute magnitude of the interpolated isochrone pointcorresponding to the same color as the observed star.The best-fit age (hereafter t iso ) is found by maximiz-ing Eq. 1 (minimizing Eq. 2) where the summation isperformed across all CMD points of the group. Age un-certainties are determined by finding the range of ageswith corresponding likelihoods within 1 σ of the maxi-mum likelihood. These statistical uncertainties do notconsider the biases that contribute to isochrone fitting:unresolved binaries, model uncertainties, and differen-tial extinction. An unresolved binary sequence is appar-ent in both groups and is likely biasing the ages towardsyounger values. RESULTS When viewed in three spatial dimensions, the posi-tions and kinematics reveal that the two groups are ex-panding radially away from a common center (Figure 1).Figure 2 shows a correlation between the position andvelocity for both stellar groups. The slopes of the lin-ear fits to the position-velocity profiles are ( κ x , κ y , κ z )= (0 . ± . , . ± . , . ± . − pc − . Hence, stars further away from the groups’ cen-ter are moving faster. A fit to the stars’ speeds as afunction of radius yields a slope of κ r = 0 . ± . − pc − when the center is estimated by-eye to be X = − , Y = − , Z = −
105 pc. This center is cho-sen as opposed to a center calculated using the averagestellar position due to the non-uniform spatial distribu-tion of stars when considering the two groups’ combinedprofile. The stars’ average speed is 2 . ± .
03 km s − and the velocity dispersion of both groups is ∼ . − .These linear profiles show evidence for ballistic expan-sion, where the highest velocity stars have had sufficienttime to travel further outwards from the center. A pro-cess where stars are accelerated outwards due to violentchanges in the region’s gravitational potential may alsoplay a roll in creating the observed expansion profile(Zamora-Avil´es et al. 2019).The correlation is tighter in y − v y and z − v z than in x − v x . A Kendall’s τ correlation test yields τ x − v x = . τ y − v y = .
73, and τ z − v z = .
50, each with a statisticalsignificance of p < × − . We note that among thethree directions, the x-axis is the most aligned with theline-of-sight direction. As illustrated in Figure 2, thisbias results with large uncertainties in the parallaxesand RV estimates for the correlation projected alongthis component.The expansion time estimate can be inferred using thestars’ velocities as a function of position. For each of theposition-velocity slopes found in Figure 2 we calculatethe expansion time, t exp = 1 /γκ , where γ = 1 . rion Expansion OBP-nearOBP-near Briceño-1Briceño-1
Figure 1.
3D kinematics of stars with available RVs in OBP-near and Brice˜no-1 in Heliocentric Galactic cartesian coordinatesafter subtracting the combined average of their kinematics in each dimension. Larger, lighter-colored cones represent faster-moving stars, and smaller, darker-colored cones represent slower-moving stars. The directions of the cones’ apexes represent thedirections of the stars’ motions. The dashed line in each panel marks the spatial division of the two groups from that viewingangle. An interactive version is available here, or in the online version of the published article.
Swiggum et al. pc Myr − km − s. We find t exp,x = 8 .
58 Myr, t exp,y =8 .
00 Myr, t exp,z = 7 .
72 Myr, and t exp,r = 9 .
68 Myr.When tracing the stars’ positions back in time (assumingconstant velocity), we find that the groups were mostcompact around 7 . t iso , and their associated uncertainties. Fromthese isochrones, both groups are found to have an ex-tinction of A V = 0 . ∼ t iso value and itsoffset on the CMD. However, the strong main-sequenceoverlap between the two groups makes this offset subtle.Note again that the t iso values are likely lower limitsdue to the bias from unresolved binaries in the sample.Additionally, Brice˜no-1 is positioned along the LOS, di-rectly in front of the 20 Myr old group, ASCC20 (Koset al. 2019, SNN7 in Chen et al. 2020), possibly biasingthe isochrone fit for Brice˜no-1 towards an older value.However, SNN results do not indicate significant over-lap from ASCC20. An investigation into the infraredcolors of both groups’ stars is presented in Appendix Aand points towards OBP-near indeed having a slightlyyounger age than Brice˜no-1. DISCUSSIONA recent study performed a hierarchical clusteringanalysis on Orion and claimed that the entire complexshows clear radial expansion from a central region dueto a SNe explosion (Kounkel et al. 2018; Kounkel 2020).These studies recovered 190 individual groups, whichwere then recombined into larger groups as Orion A, B,C, D, and λ Orion. Orion D was reported to consistof stars associated with the Orion OB1ab region andmore diffuse populations just outside of it. We applythe SNN clustering algorithm to the region and find that43% (594 stars) of OBP-near and Brice˜no-1 stars canbe cross-matched to Orion D of Kounkel et al. (2018).Other groups recovered in Chen et al. (2020) includ-ing ASCC20 (SNN 7; Kos et al. 2019), L1616 (SNN 8),and Orion X (SNN 20; Bouy & Alves 2015) partiallycross-match to Orion D as well. Additionally, 50% ofstars from OBP-d (SNN 6; Kubiak et al. 2017), OBP-b(SNN 9; Kubiak et al. 2017), ome Ori (SNN 14; Chen etal. 2020), OBP-West (SNN 16; Chen et al. 2020), andOBP-far (SNN 19; Chen et al. 2020) overlap with OrionC.It was claimed that Orion D shows signs of expansionby analyzing proper motions with Gaia DR2 (Kounkelet al. 2018); however, Orion D contains a mix of smallergroups that differ in age and kinematics. Isochrone fits
10 5 0 5 10 15 20 25 x (pc) v x ( k m s ) x = ± 2.55 pc = 0.115±0.009 km s pc
30 20 10 0 10 20 y (pc) v y ( k m s ) y = ± 1.12 pc = 0.122±0.003 km s pc
15 10 5 0 5 10 15 z (pc) v z ( k m s ) z = ± 0.89 pc = 0.126±0.005 km s pc r (pc) Sp ee d ( k m s ) r = ± 3.56 pc = 0.104±0.007 km s pc Figure 2.
Position-velocity profiles of stars in the combinedreference frame of OBP-near (blue points) and Brice˜no-1 (redpoints). Each panel shows the position-velocity profile for agiven Cartesian component. Regression lines (orange) arefit to the combined profiles of OBP-near and Brice˜no-1 inorder to estimate the expansion slope, κ , which is quoted inthe legends. Individual velocity errors are displayed and themean of the position errors is shown in the lower right-handcorner of each panel. to each of the SNN groups in this region indicate anage range between 6 and 20 Myr, with the oldest beingASCC20. These values suggest that Orion D has mul-tiple stellar populations which presumably did not formtogether.Furthermore, previous work placed all of these groupsin a single, common reference frame and proposed thatthe entire complex is currently expanding from a centralregion due to a supernova event that occurred 6 Myrago (Kounkel 2020). As part of the proposal, Orion rion Expansion t = 6.8 +3.62.9 MyrBriceño-1OBP-near t = 9.0 +4.03.4 MyrBriceño-1
G G RP (mag) t = 6.2 +3.52.4 MyrOBP-near M G ( m a g ) Figure 3.
The CMDs of OBP-near (cyan dots) and Brice˜no-1 (red dots) stars: the top panel shows the combined CMD ofboth groups with their combined best-fit isochrone (dottedline); the middle panel shows the CMD of Brice˜no-1 andits best-fit isochrone; the bottom panel shows the CMD ofOBP-near and its best-fit isochrone. The legend in eachpanel quotes the associated age of the isochrone fit and itsuncertainty.
C and D were once part of the same molecular cloud,split by the supernova explosion that separated them into two distinct regions. To test the hypothesis we ex-tend the analysis of OBP-near and Brice˜no-1 to the restof the SNN groups. While most of the original groupsare re-derived in Section 2, here we utilize the origi-nal catalog from Chen et al. (2020) (except OBP-nearand Brice˜no-1) as their study more carefully considersthe entire Orion complex. We update the astrometryof these SNN groups by cross-matching to
Gaia eDR3.SNN 13, 18, 22, and 25 are discarded, as Chen et al.(2020) finds them to strongly overlap with some of thelarger identified groups.We follow the SNN groups’ motion in a common ref-erence frame backwards/forwards in time ± ∼ ∼
20 Myr old groups overlappingwith Orion D are moving through the complex ratherthan expanding from Orion’s core. Hence, signs of ex-pansion are most evident in OBP-near and Brice˜no-1while the other groups show a more complex dynamicalhistory.The radial expansion shown by OBP-near andBrice˜no-1 groups may be related to past SNe explosionsor strong stellar winds which might have occurred at thecore of Orion. Such winds and SNe would be expectedto create cavities in the interstellar medium, manifest-ing as shells and bubbles (see e.g. Smith et al. 2020; Kim& Ostriker 2018). Figure 5 shows the spatial locationof OBP-near and Brice˜no-1 (colored dots), as compared
Swiggum et al. G r o up A g e ( M y r ) OBP-nearOBP-near Briceño-1Briceño-1 Lambda OrionLambda OrionNGC 1980 NGC 1980
Figure 4.
Interactive animation tracing the evolution of the SNN groups (Chen et al. 2020) (colored dots) forwards (+8 Myr)and backwards ( − to a high-resolution 3D spatial map of the nearby in-terstellar medium derived from 3D dust mapping (Leikeet al. 2020). OBP-near and Brice˜no-1 lie at the centerof a dust shell, which would indicate that past feed-back events occurred within the two groups. Evidenceof a dust shell is also seen in a complementary 3D dustmap of the Orion A region, with the nearest part ofthe dust shell coincident with a previously undiscoveredforeground cloud at a distance between ≈ −
345 pc(see Figure 6 in Rezaei Kh. et al. 2020). Past stud-ies have also found evidence for an expanding H I shellthat roughly surrounds the two groups’ plane-of-sky po-sitions but with a center more aligned with OBP-near(Chromey et al. 1989; Ochsendorf et al. 2015).One possible scenario to explain our observations sug-gests that stars which today identify as OBP-near andBrice˜no-1 formed approximately at the same time as onebound cluster at the core. A major feedback event dis-persed the gas in the original molecular cloud after thestars formed. While the violent change of the gravita-tional potential might unbind the system and lead thestars to expand, it is unclear whether this mechanismcan drive the stars into a ballistic expansion with thesymmetry displayed in Figure 1. As stated previously,rapid changes to the potential following gas/dust disper-sal could aid in accelerating the stars (Zamora-Avil´es etal. 2019). Another theory posits that star associations may formin gas compressed layers of the expanding shell fromshock fronts after SN events or other massive stellarfeedback. In this scenario, the grouped stars will moveat the same velocity as the expanding gas shell theyformed in (Elmegreen & Lada 1977). Yet, it is unclearwhether this process will lead to the expansion with theproperties observed in Figure 1. Such a scenario, wherestar formation is triggered sequentially, could explain anintrinsic age difference between OBP-near and Brice˜no-1; if feedback took place closer to one side of the progen-itor cloud, then the onset of star formation could haveoccurred at different times across different locations ofthe expanding gas shell. CONCLUSIONSThe astrometric measurements of the
Gaia mis-sion provided in DR2 and EDR3, supplemented byAPOGEE-2 and GALAH DR3, have enabled a sys-tematic study of the 3D kinematics of OBP-near andBrice˜no-1, two stellar groups at the core of the Orioncomplex. This work shows direct evidence of ballisticexpansion occurring locally at the core of the Orion com-plex. While previous works found some evidence of co-herent motions in the Orion complex, this study showscompelling evidence of stellar expansion in a massivestar-forming region. The stars are currently located atthe center of a dust shell, suggesting that their expan- rion Expansion -100-125-75 Z [ p c ] - - - - Y [pc] X [pc] Y [ p c ] Briceño-1OBP-nearDust-130-160-190-110-150 - - - - X [pc] - - - - Z [ p c ] -100-130-160-70 Figure 5.
The spatial location of OBP-near and Brice˜no-1, overlaid on a 3D spatial map of interstellar dust (Leike et al. 2020).As apparent in the top down X-Y view (middle panel), the OBP and Brice˜no-1 groups lie at the center of a dust shell. Thestars’ sizes indicate their probability to be group members assigned by SNN. An interactive version of this figure is availablehere or in the online version of the published article. sion links to stellar feedback events. However, the pro-cess driving the relatively symmetric radial expansionshown in this study remains unclear. Upcoming work(Foley et al. in prep) will present further evidence forthe connection between this radial expansion and stellarfeedback. Future
Gaia data releases are anticipated tosignificantly improve the astrometry of stars, which willallow for more accurate kinematic analyses of the entireOrion complex. Forthcoming observations and numeri-cal studies of the interstellar medium will also help shedlight on the relationship between feedback events andstellar dynamics Software : Astropy (Astropy Collaboration et al.2018), NumPy (van der Walt et al. 2011), Plotly, Mat-plotlib (Hunter 2007), Glue (Robitaille et al. 2017), andGalpy (Bovy 2015).0
Swiggum et al.
REFERENCES
Alves, J. & Bouy, H. 2012, A&A, 547, A97.doi:10.1051/0004-6361/201220119Alves, J., Zucker, C., Goodman, A. A., et al. 2020, Nature,578, 237Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M.,et al. 2018, AJ, 156, 123. doi:10.3847/1538-3881/aabc4fBally, J. 2008, Handbook of Star Forming Regions, VolumeI, 459Baumgardt, H. & Kroupa, P. 2007, MNRAS, 380, 1589.doi:10.1111/j.1365-2966.2007.12209.xBessell, M. S. & Brett, J. M. 1988, PASP, 100, 1134.doi:10.1086/132281Blanton, M. R., Bershady, M. A., Abolfathi, B., et al. 2017,AJ, 154, 28Blaauw, A. 1964, ARA&A, 2, 213Bouy, H. & Alves, J. 2015, A&A, 584, A26.doi:10.1051/0004-6361/201527058Bovy, J. 2015, ApJS, 216, 29Brice˜no, C., Hartmann, L., Hern´andez, J., et al. 2007, ApJ,661, 1119Brown, A. G. A., de Geus, E. J., & de Zeeuw, P. T. 1994,A&A, 289, 101Buder, S., Sharma, S., Kos, J., et al. 2020, arXiv:2011.02505Cantat-Gaudin, T., Jordi, C., Wright, N. J., et al. 2019,A&A, 626, A17. doi:10.1051/0004-6361/201834957Chen, B., D’Onghia, E., Pardy, S. A., et al. 2018, ApJ, 860,70. doi:10.3847/1538-4357/aac325Chen, B., D’Onghia, E., Alves, J., et al. 2020, A&A, 643,A114. doi:10.1051/0004-6361/201935955Chini, R., Hoffmeister, V. H., Nasseri, A., et al. 2012,MNRAS, 424, 1925. doi:10.1111/j.1365-2966.2012.21317.xChromey, F. R., Elmegreen, B. G., & Elmegreen, D. M.1989, AJ, 98, 2203. doi:10.1086/115289Cropper, M., Katz, D., Sartoretti, P., et al. 2018, A&A,616, A5Cutri, R. M., Wright, E. L., Conrow, T., et al. 2013,Explanatory Supplement to the AllWISE Data ReleaseProducts, by R. M. Cutri et al.Elmegreen, B. G. & Lada, C. J. 1977, ApJ, 214, 725.doi:10.1086/155302Fang, M., Kim, J. S., Pascucci, I., et al. 2017, AJ, 153, 188.doi:10.3847/1538-3881/aa647bGaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al.2016, A&A, 595, A1. doi:10.1051/0004-6361/201629272Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al.2018, A&A, 616, A1Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al.2020, arXiv:2012.01533 Geen, S., Watson, S. K., Rosdahl, J., et al. 2018, MNRAS,481, 2548. doi:10.1093/mnras/sty2439Goodwin, S. P. & Bastian, N. 2006, MNRAS, 373, 752.doi:10.1111/j.1365-2966.2006.11078.xGroßschedl, J. E., Alves, J., Meingast, S., et al. 2018, A&A,619, A106. doi:10.1051/0004-6361/201833901Großschedl, J. E., Alves, J., Teixeira, P. S., et al. 2019,A&A, 622, A149. doi:10.1051/0004-6361/201832577Großschedl, J. E., Alves, J., Meingast, S., et al. 2020,arXiv:2007.07254Hills, J. G. 1980, ApJ, 235, 986. doi:10.1086/157703Hunter, J. D. 2007, Computing in Science and Engineering,9, 90. doi:10.1109/MCSE.2007.55Kim, C.-G. & Ostriker, E. C. 2018, ApJ, 853, 173.doi:10.3847/1538-4357/aaa5ffKoenig, X. P. & Leisawitz, D. T. 2014, ApJ, 791, 131.doi:10.1088/0004-637X/791/2/131Kos, J., Bland-Hawthorn, J., Asplund, M., et al. 2019,A&A, 631, A166. doi:10.1051/0004-6361/201834710Kounkel, M., Covey, K., Su´arez, G., et al. 2018, AJ, 156,84. doi:10.3847/1538-3881/aad1f1Kounkel, M. 2020, ApJ, 902, 122.doi:10.3847/1538-4357/abb6e8Krause, M. G. H., Offner, S. S. R., Charbonnel, C., et al.2020, SSRv, 216, 64. doi:10.1007/s11214-020-00689-4Kubiak, K., Alves, J., Bouy, H., et al. 2017, A&A, 598,A124Kuhn, M. A., Hillenbrand, L. A., Sills, A., et al. 2019, ApJ,870, 32. doi:10.3847/1538-4357/aaef8cKuhn, M. A., Hillenbrand, L. A., Carpenter, J. M., et al.2020, ApJ, 899, 128. doi:10.3847/1538-4357/aba19aLada, C. J. 1987, Star Forming Regions, 115, 1Lada, C. J. & Lada, E. A. 2003, ARA&A, 41, 57.doi:10.1146/annurev.astro.41.011802.094844Lada, C. J., Muench, A. A., Luhman, K. L., et al. 2006,AJ, 131, 1574. doi:10.1086/499808Lada, C. J., Lombardi, M., & Alves, J. F. 2010, ApJ, 724,687. doi:10.1088/0004-637X/724/1/687Leike, R. H., Glatzle, M., & Enßlin, T. A. 2020, A&A, 639,A138. doi:10.1051/0004-6361/202038169Luri, X., Brown, A. G. A., Sarro, L. M., et al. 2018, A&A,616, A9Majewski, S. R., Schiavon, R. P., Frinchaboy, P. M., et al.2017, AJ, 154, 94Marigo, P., Girardi, L., Bressan, A., et al. 2017, ApJ, 835,77Ochsendorf, B. B., Brown, A. G. A., Bally, J., et al. 2015,ApJ, 808, 111. doi:10.1088/0004-637X/808/2/111 rion Expansion Rezaei Kh., S., Bailer-Jones, C. A. L., Soler, J. D., et al.2020, A&A, 643, A151. doi:10.1051/0004-6361/202038708Robitaille, T., Beaumont, C., Qian, P., et al. 2017, ZenodoRom´an-Z´u˜niga, C. G., Roman-Lopes, A., Tapia, M., et al.2019, ApJL, 871, L12. doi:10.3847/2041-8213/aafb06Sch¨onrich, R., Binney, J., & Dehnen, W. 2010, MNRAS,403, 1829Sharma, S. & Johnston, K. V. 2009, ApJ, 703, 1061.doi:10.1088/0004-637X/703/1/1061Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ,131, 1163. doi:10.1086/498708Smith, R. J., Treß, R. G., Sormani, M. C., et al. 2020,MNRAS, 492, 1594. doi:10.1093/mnras/stz3328Ward, J. L. & Kruijssen, J. M. D. 2018, MNRAS, 475,5659. doi:10.1093/mnras/sty117Ward, J. L., Kruijssen, J. M. D., & Rix, H.-W. 2020,MNRAS, 495, 663. doi:10.1093/mnras/staa1056Wright, N. J. & Mamajek, E. E. 2018, MNRAS, 476, 381.doi:10.1093/mnras/sty207 Wright, N. J., Jeffries, R. D., Jackson, R. J., et al. 2019,MNRAS, 486, 2477. doi:10.1093/mnras/stz870van der Walt, S., Colbert, S. C., & Varoquaux, G. 2011,Computing in Science and Engineering, 13, 22.doi:10.1109/MCSE.2011.37van Leeuwen, F. 2009, A&A, 497, 209.doi:10.1051/0004-6361/200811382Zamora-Avil´es, M., Ballesteros-Paredes, J., Hern´andez, J.,et al. 2019, MNRAS, 488, 3406.doi:10.1093/mnras/stz1897Zari, E., Brown, A. G. A., de Bruijne, J., et al. 2017, A&A,608, A148. doi:10.1051/0004-6361/201731309Zari, E., Brown, A. G. A., & de Zeeuw, P. T. 2019, A&A,628, A123Zucker, C., Speagle, J. S., Schlafly, E. F., et al. 2020, A&A,633, A51. doi:10.1051/0004-6361/201936145 Swiggum et al.
APPENDIX A. INFRARED COLORS OF OBP-NEAR AND BRICE ˜NO-1 −0.2 0.0 0.2 0.4 0.6 0.8 1.0 H − K S (mag)−0.20.00.20.40.60.81.01.21.41.6 J − H ( m ag ) A V = 3 mag Bessel main sequenceBessel giant branchBrice ̃no-1OBP-Near 0 1 2 3 4 W W (mag)0.00.20.40.60.81.0 W − W ( m ag ) A V = 5 mag Class III Class IIBrice ̃no-1 (CIII 105, CII 24)OBP-Near (CIII 75, CII 56)
Figure 6.
Infrared color-color diagrams for stellar members in Brice˜no-1 (red) and OBP-near (blue).
Left: A V = 3 mag is shown as black arrow, indicating the direction of reddeningfor extincted sources. The two gray dashed lines show the reddening band, plotted parallel to the extinction vector and abovethe main sequence. Right:
WISE mid-infrared color-color diagram. An extinction vector of length A V = 5 mag is shown asblack arrow. The gray dashed line separates the groups roughly into Class II and Class III YSOs. The number of sources thatfall in each class per group is given in the legend. Figure 6 shows two infrared color-color diagrams to investigate how extinction is influencing the photometry ofthe two groups by using near-infrared photometry, and to investigate their evolutionary status by using mid-infraredphotometry. For this analysis we use near-infrared photometry, J (1 . µ m), H (1 . µ m), and K S (2 . µ m), from theTwo Micron All-Sky Survey (2MASS; Skrutskie et al. 2006), and mid-infrared photometry, W1 (3 . µ m), W2 (4 . µ m),and W3 (12 µ m), from the Wide-field Infrared Survey Explorer (WISE, AllWISE catalog; Cutri et al. 2013), retrievedform https://irsa.ipac.caltech.edu. To combine the 2MASS and WISE catalogs with the Gaia EDR3 astrometry weused a 1” cross-match radius, only keeping the closest nearest neighbor.The left panel in Fig. 6 shows the 2MASS near-infrared color-color diagram, highlighting which sources in the twogroups show significant signs of reddening, caused either by foreground extinction or by a circumstellar disk. Inparticular, sources above the main sequence, located within the reddening band, are thought to be dominated byforeground extinction. Only sources that pass the following quality criteria were used for the diagram:j sig , h sig , k sig < . , j snr , h snr , k snr > . (A1)j sig , h sig , k sig < . , j snr , h snr , k snr > . (A2)With these cuts there are 850 and 642 sources left in Brice˜no-1 and OBP-near, respectively. The majority of thesources in the two groups (more than 95%) show no or very little reddening, with no significant difference between the rion Expansion A V = 3 to 8 mag, while most of these are classified as Class II YSOs based ontheir mid-infrared color (see below), hence the reddening could be caused by circumstellar dust. The source showingthe highest reddening in Fig. 6 (2MASS J05414177-0151456 in OBP-near) is projected on top of the young embeddedcluster NGC 2024 in Orion B and might rather belong to this cluster, since some contamination by other groups in ourSNN selected samples can not be ruled out.The right panel in Fig. 6 shows a WISE mid-infrared color-color diagram. To use near- to mid-infrared photometryto distinguish evolutionary classes within a young stellar population based on the infrared excess determined from thespectral energy distribution (SED) is a proofed tool since the 80s (e.g. Lada 1987; Lada et al. 2006; Großschedl et al.2019). This allows to estimate which of the young stellar objects (YSOs) are still embedded in an envelope (Class I),are pre-main-sequence stars surrounded by a circumstellar disk (Class II), or which are pre-main-sequence stars thathave already dissipated their envelope and disk and are on their way to the main-sequence (Class III). The latter classcan not be strictly separated from main-sequence stars using solely mid-infrared colors, however, since we are dealingin this study with populations younger than about 10 Myr, the classification into Class III for sources with no clearinfrared excess is a feasible approach. We like to note that the most massive stars of the two populations have likelyalready reached the main-sequence while still being young. Regarding Class I protostars, there are likely no protostarsincluded in the Class II samples, since they tend to have redder colors beyond the borders of the displayed axes. Figure6 only includes a rather small fraction (about 17%) of the SNN selected members, due to the inferior resolution andsensitivity of WISE compared to Gaia or 2MASS. To get reliable WISE photometry, several quality criteria have beenapplied to avoid erroneous photometry and contamination by extended emission, to which especially the W3 passbandis susceptible to. The quality criteria are as follows:w > , w < , w3mag 1 − w3mag 6 < . (A3)For more details on the WISE quality criteria see Großschedl et al. (2019). With these cuts there are 129 and 131sources left in Brice˜no-1 and OBP-near, respectively. The Class II YSOs are separated from Class III and main-sequencestars with W1 − W2 > − . · ( W − W − .
3) + 0 ..
3) + 0 .. ..