Comoving stars in Gaia DR1

AMNH seminar talk

Transcript

1. CO-MOVING STARS Semyeong Oh (Princeton) 2017-02-28 AMNH Gaia DR1 in
2. Outline What are co-moving stars? Why care about co-moving stars? Why now? How do we find co-moving stars? What next?
3. I acknowledge help from Adrian Price-Whelan @Princeton David Hogg @NYU Tim Morton @Princeton David Spergel @Princeton
4. What are co-moving stars?
5. Co-moving stars are any number of stars roughly co-located and on a scale smaller than galactic scale sharing very similar space velocity with dispersions smaller than that of field stars
6. Co-moving stars are any number of stars roughly co-located and on a scale smaller than galactic scale sharing very similar space velocity with dispersions smaller than that of field stars
7. Examples
8. Example 1: wide binaries triples, multiples ?
9. Example 2. Moving Groups • clumps of stars with the same velocity • anomaly from smooth velocity distribution origin still under debate • remnants of clusters? • transient phenomena? Bovy et al. 2009
10. Example 3. Open Clusters • hundreds to thousands stars • loosely gravitationally bound • σ < 1 km/s M44 (Beehive), credit: Bob Franke M45 (Pleiades), credit: Robert Gendler
11. So what? Why care about co-moving stars?
12. Co-moving stars are tracers of Galactic tidal field
13. Some binaries are more valuable for calibrating/testing stellar models • F/G/K - M binaries provides anchors for calibrating M dwarf properties. • Giant-Main Sequence pairs are useful for testing consistency between stellar atmospheric models. Wide separation ensures no interaction since birth.
14. Young moving groups are interesting places to search planets and brown dwarfs Credit: ESO/A.-M. Lagrange et al. debris disk (1996) β Pic b (2008) β Pic Age:12Myr
15. But then, … there are many interesting problems in astronomy Why now?
16. Gaia opens a new era for studying the Milky Way from M. Perryman’s Spitzer lecture (2013)
17. Gaia maps a billion stars • astrometry for a billion stars • 1% of the Milky Way • down to G~20 mag • 25 µas accuracy at 15 mag
18. Many spectroscopic surveys will complement the Gaia data Name Number of sources Product Status RAVE 483k stars RV + Teff, logg, Z complete (2003-2013) DR5 Sep 2016 GALAH 1M stars RV + 15 abundances late 2013- DR1 Sep 2016 Gaia-ESO ≥100k stars RV + ≥12 abundances Dec 2011- DR2 July 2015 APOGEE (-2) 100k giants (300k stars) RV + 15 abundances 1: 2011-2014 (2: 2014-2020)
19. Gaia first data release on Sep 14, 2016
20. Tycho-Gaia Astrometric Solution (TGAS) provides proper motions for 2M stars Credit: Michalik & Lindegren
21. Tycho-Gaia Astrometric Solution (TGAS) provides proper motions for 2M stars Credit: Michalik & Lindegren
22. Tycho-Gaia Astrometric Solution (TGAS) provides proper motions for 2M stars Credit: Michalik & Lindegren
23. Tycho-Gaia Astrometric Solution (TGAS) provides proper motions for 2M stars Credit: Michalik & Lindegren
24. How do we find co-moving stars?
25. We pose the discovery as a model selection problem. Case 1 Case 2
26. We build a probabilistic model for the data on assumptions of space velocities. Case 1 Case 2
27. parallax [mas] position [deg] proper motion [mas/yr] We build a probabilistic model for the data on assumptions of space velocities.
28. We build a probabilistic model for the data on assumptions of space velocities. with uncertainties
29. We build a probabilistic model for the data on assumptions of space velocities.
30. L(~ µ |~ v, r) p(~ v) p(r | $) We build a probabilistic model for the data on assumptions of space velocities.
31. Case 1 Case 2 We build a probabilistic model for the data on assumptions of space velocities.
32. We build a probabilistic model for the data on assumptions of space velocities.
33. We marginalize over unknown distances and 3D velocities L1 L2 = R L(~ µ1 |~ v, r1) L(~ µ2 |~ v, r2) p(~ v) p(r1 | $1) p(r2 | $2)drd3~ v ⇧i=1,2 R L(~ µi |~ v, r) p(~ v) p(r | $i) dr d3~ v • We use the ratio of fully marginalized likelihoods (“Bayes factor”) as the quantity to select co-moving stars.
34. We examine 271,232 pairs with [S/N]ϖ > 8, |∆x| < 10 pc, |∆vt | < 10 km/s
35. ln L1/L2 > 6 We select pairs with as candidate co-moving pairs random pairs
36. Of 271,232 pairs in the initial sample, we find 13,058 candidate co-moving pairs among 10,606 stars. resulting in 4,555 connected components. Some pairs are connected.
37. Of 271,232 pairs in the initial sample, we find 13,058 candidate co-moving pairs among 10,606 stars. resulting in 4,555 connected components. Some pairs are connected.
39. Of 271,232 pairs in the initial sample, we find 13,058 candidate co-moving pairs among 10,606 stars, resulting in 4,555 connected components. Some pairs are connected.
40. Size of connected components varies from 2 to 151.
41. Largest connected component: 151 stars, 3465 edges
42. Largest connected component: 151 stars, 3465 edges … a.k.a. the Pleiades!
45. NCC > 5
46. Melotte 111 (Coma) Melotte 25 (Hyades) Alessi 13 NGC 2632
 beehive Explore yourself at smoh.space/vis/gaia-comoving-stars
48. The number of very wide separation co-moving pairs increases at > 1pc 10 2 10 1 100 101 separation [pc] 10 1 100 101 102 103 104 count all size=2 (mutually exclusive) size>2
49. Jiang & Tremaine (2010) predicted these wide separation co-moving pairs Jacobi radius rJ = 1.7 pc Jiang & Tremaine (2010)
52. 0.0 0.5 1.0 1.5 2.0 2.5 G J [mag] 2 0 2 4 6 8 10 G + 5(log ˆ d + 1) [mag] separation < 1 pc (N=297) 0.0 0.5 1.0 1.5 2.0 2.5 G J [mag] 2 0 2 4 6 8 10 separation > 1 pc (N=3939) mutually exclusive pairs
53. What next?
54. What information is in statistics of co-moving pairs? • rJ = 1.7 pc Jiang & Tremaine (2010)
55. Are very wide separation pairs different from “the field”? 0.0 0.5 1.0 1.5 2.0 2.5 G J [mag] 2 0 2 4 6 8 10 G + 5(log ˆ d + 1) [mag] separation < 1 pc (N=297) 0.0 0.5 1.0 1.5 2.0 2.5 G J [mag] 2 0 2 4 6 8 10 separation > 1 pc (N=3939) mutually exclusive pairs
56. Can we constrain the 3D velocities of co-moving pairs? • Proper motions of a group of co-moving stars are projections of the same 3D velocity vector. (“astrometric radial velocity”) • Dravins, Lindegren, Madsen published a series of work:
 a few km/s accuracies for clusters using HIPPARCOS
57. Summary Co-moving stars include wide binaries, moving groups, open clusters. They are interesting tools to study Galactic dynamics and constrain stellar models. Gaia makes now an opportune time. We build a statistical model to select co-moving pairs. Simulating wide binary evolution, Comparing color-magnitude diagram with the field …
58. Melotte 111 (Coma) Melotte 25 (Hyades) Alessi 13 NGC 2632
 beehive