Red Clump Stars, Stellar Twins, and the Prospects of Chemical Cartography with Gaia
Keith Hawkins Columbia Univ. (Simons Fellow)
25 April 2017 [email protected]
Ignace Gaston Pardies, Star Map Plate 2: Cetus, Aquarius,
Andromeda. etc, Paris, 1693
A Vision of Chemical Cartography
Ingredients : chemical abundances, Distances, sky positions, ages?
Adapted from Greg Stinson and Maria Bergemann.[Mg/Fe] as a function of spatial position in Gaia-ESO
What can cartography tell us
• chemical history across the galaxy
• chemical substructure
• Radial migration
Hayden+2015
Note: This (and many Galactic structure) work(s) made use of Red Clump Stars
The Era of Large Surveys
N ~ 100,000N ~ 1,000,000 N ~ 10,000,00025+ elements15+ elements 18+? elements2012—
2014— 2021-2026?
N ~ 100,000
2008—N ~ 483,0008+ elements
2003-2013
15 (20*+) elements
Chemical ‘Labeling’ Galactic components ‘cluster’ in chemical space! (cartography is the next step —> need distances)
R. Wyse Talk, Hawkins+ 2015b; see also Hogg+2016, Blanco-Carisma+2016, Lambert+2017, Ness+2017; and many others
Ingredient 2: Distances
Red clump stars (Stanek+1998, …)
stellar twins (Jofre+2016, …)
Distances: Red Clump
Red clump
Use Gaia parallaxes to assess:
(1) How good of a standard candle is the RC? (2) Update magnitude of RC in J, H, Ks, W1, W2, W3, W4, G, NUV*
Red Clump Datasets
• Parallax cut at 30% (prior v.s. data dominated posterior)
• N ~ 970 (30%)
• N ~ 180 (10%)
APOKASC (Elsworth+2016), APOGEE-RC (Bovy+2014), Laney+2012, APO1m (Feuillet+2016)
G-K
MK
-1.62 (Laney+2012)
Distances: Red Clump The Model
Mλ
N
Distances: Red Clump The Model
RC
Mλ
N
Distances: Red Clump The Model
RC
Mλ
N
Contaminates
Distances: Red Clump The Model
RC
Mλ
N
Contaminates
M, σM — RC Mc=M, σMc — cont.
fout — fraction of contamination
Distances: Red Clump The Model
RC
Mλ
N
Contaminates
M, σM — RC Mc=M, σMc — cont.
fout — fraction of contaminationOther important parameter(s):
L — scale-length of exponentially decreasing space density prior (Bailer-Jones+ 2015;
Astraatmadja+2016)
Distances: Red Clump The Model
RC
Mλ
N
Contaminates
M, σM — RC Mc=M, σMc — cont.
fout — fraction of contaminationOther important parameter(s):
L — scale-length of exponentially decreasing space density prior (Bailer-Jones+ 2015;
Astraatmadja+2016)‘Nuisance parameters’:
- distance - Extinction
Distances: Red Clump The Model
RC
Mλ
N
Contaminates
M, σM — RC Mc=M, σMc — cont.
fout — fraction of contaminationOther important parameter(s):
L — scale-length of exponentially decreasing space density prior (Bailer-Jones+ 2015;
Astraatmadja+2016)‘Nuisance parameters’:
- distance - Extinction
Use STAN statistical software to sample posterior in parameters
M = �1.62+0.01�0.01
0.12
0.16
0.20
0.24
�M
�M = 0.17+0.02�0.02
0.6
0.8
1.0
1.2
�M
,out
�M,out = 0.71+0.10�0.08
208
216
224
232
240
L(p
c)
L (pc) = 222.67+4.39�4.31
�1.65
0
�1.62
5
�1.60
0
�1.57
5
M
0.08
0.16
0.24
0.32
0.40
f out
0.12
0.16
0.20
0.24
�M
0.6
0.8
1.0
1.2
�M,out20
821
622
423
224
0
L (pc)0.08
0.16
0.24
0.32
0.40
fout
fout = 0.17+0.05�0.05
Red Clump K(JHG+)-MagnitudeM = abs mag
σM = dispersion of
abs mag
L = scale-length of exponential
decreasing space density
prior fout =
contamination fraction
Mk = -1.62 +/- 0.013 ; σMK = 0.17+/- 0.02; fout ~ 17%
σM,out = dispersion of abs mag of
outlier
Hawkins+ 2017 in prep
M = �1.62+0.01�0.01
0.12
0.16
0.20
0.24
�M
�M = 0.17+0.02�0.02
0.6
0.8
1.0
1.2
�M
,out
�M,out = 0.71+0.10�0.08
208
216
224
232
240
L(p
c)
L (pc) = 222.67+4.39�4.31
�1.65
0
�1.62
5
�1.60
0
�1.57
5
M
0.08
0.16
0.24
0.32
0.40
f out
0.12
0.16
0.20
0.24
�M
0.6
0.8
1.0
1.2
�M,out20
821
622
423
224
0
L (pc)0.08
0.16
0.24
0.32
0.40
fout
fout = 0.17+0.05�0.05
Red Clump K(JHG+)-MagnitudeM = abs mag
σM = dispersion of
abs mag
L = scale-length of exponential
decreasing space density
prior fout =
contamination fraction
Mk = -1.62 +/- 0.013 ; σMK = 0.17+/- 0.02; fout ~ 17%
σM,out = dispersion of abs mag of
outlier(1) Consistent with literature in Ks
(2) dispersion sets floor of ~10% distance uncertainty
(fair standard candle)
Hawkins+ 2017 in prep
Red Clump K(JHG+)-Magnitude
Hawkins+ 2017, in prep
NewGood Fair
Agreement with literature:
Poor
RC in Galex NUV M = 8.95+0.05
�0.04
0.60
0.75
0.90
1.05
�M
�M = 0.83+0.06�0.07
1.2
1.8
2.4
3.0
3.6
�M
,out
�M,out = 1.48+0.47�0.32
195
210
225
L(p
c)
L (pc) = 211.39+6.23�6.00
8.80
8.88
8.96
9.04
9.12
M
0.15
0.30
0.45
0.60
f out
0.60
0.75
0.90
1.05
�M
1.2
1.8
2.4
3.0
3.6
�M,out
195
210
225
L (pc)0.15
0.30
0.45
0.60
fout
fout = 0.10+0.19�0.07
σM = 0.83 +/- 0.07!!
New
RC in Galex NUV
Mohammed+ (with KH), 2017 in prep
New
RC in Galex NUV
Mohammed+ (with KH), 2017 in prep
Takeaway: 1. RC is a good standard candle; but the bluer the
band the more ‘population effects’ (e.g. [Fe/H]) should be accounted for
2. Photometric Metallicity index using NUV
New
Stellar Twins
credit: IoA, Cambridge
Interests: Use twins to map the bulge, obtain distances to stars at larger distances than Gaia can reach
Distances: Stellar Twins
Jofre et al. 2016
Distances: Stellar Twins
Mädler et al. (incl. KH) 2016 Figure taken from Gaia collab. et al. 2016Twins used to help ‘settle’ Pleiades debate
Stellar Twins in APOGEEAlternative ways to find stellar twins: χ2
~600 APOGEE-TGAS stars with quality spectra+ASPCAP params
Stellar Twins in APOGEEAlternative ways to find stellar twins: χ2
5000 10000 15000 20000 25000 30000 35000 400000.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
<(�
$/$
TG
AS)>
5000 10000 15000 20000 25000 30000 35000 40000
�2
0.4
0.6
0.8
1.0
1.2
�(�
$/$
TG
AS)
Stellar Twins in APOGEEAlternative ways to find stellar twins: χ2
Stellar Twins in APOGEEAlternative ways to find stellar twins: χ2
Stellar Twins in APOGEEAlternative ways to find stellar twins: χ2
What are the Prospects for chemical cartography?
Adapted from Greg Stinson and Maria Bergemann.[Mg/Fe] as a function of spatial position in Gaia-ESO
What are the Prospects for chemical cartography?
Great! with RC stars, Gaia, stellar twins and a ton of large surveys for chemistry.
Adapted from Greg Stinson and Maria Bergemann.[Mg/Fe] as a function of spatial position in Gaia-ESO