Finding the lost siblings of the Sun Cheng Liu Lund Observatory - - PowerPoint PPT Presentation

Finding the lost siblings

• f the Sun

Cheng Liu Lund Observatory Collaborators:

• S. Feltzing, G. Ruchti, T. Bensby, L. Lindegren

Lund Observatory

• A. Brown

Leiden University

• S. Portegies Zwart

University of Amsterdam

Tuesday, February 11, 14

Introduction--I

• Most stars form within embedded clusters
• Stars were born with the Sun in an open cluster

(<100 Myr) called solar siblings

• Properties of parent cluster: three observations
• Gas planets spread out r ~ 30 AU in solar system
• Ordered planetary orbits and excited orbits of Kuiper Belt objects
• Short-lived radioactive isotopes found within meteorite

Lada & Lada 2003, Adams 2010, Portegies Zwart et al. 2009

103Mpc−3 < ρc <105 Mpc−3 4000 < N <105 R ~1− 3pc

Estimated properties of parent cluster:

Tuesday, February 11, 14

Introduction--II

• 10000
• 5000

5000 10000

• 10000
• 5000

5000 10000 y [pc] x [pc] GC

* Birthplace of the Sun is ~2.8 kpc further outer

axisymmetric + velocity dispersion Sun

Portegies Zwart et al. 2009

birth place

* Derive phase-space distribution of solar siblings

Simulations:

About 10-60 siblings within 100 pc from the Sun

Tuesday, February 11, 14

Introduction--II

• 10000
• 5000

5000 10000

• 10000
• 5000

5000 10000 y [pc] x [pc] GC

* Have chance to find solar siblings in solar vicinity

axisymmetric + velocity dispersion Two spiral arms Sun Sun

Mishurov & Acharova 2011

birth place

Simulations:

Tuesday, February 11, 14

Introduction--III

• Solar siblings have the common origin with the Sun:

kinematics, metallicities, elemental abundances, and stellar ages

• Kinematics information could be lost caused by spiral arms,

while chemical abundances are preserved in star’s atmosphere

• Homogeneous chemical signatures of members of parental

cluster allow us to tag the Sun’s siblings (Chemical tagging)

• Determine the birth place of the Sun and better understand

the mechanisms of the radial migration in the Galactic disk

Freeman & Bland-Hawthorn 2002; Mitschang et al. 2013

Tuesday, February 11, 14

Our sample

• The methodology developed in Brown et al. (2010) is used to

select candidates from the Hipparcos Catalogue

Parallax:

̟ ≥ 10 mas

Relative parallax: Proper motion:

∧ σ̟/̟ ≤ 0.1

∧ µ ≤ 6.5 mas yr−1

• selection criteria of phase-space
• exclude very young stars with

colours bluer than (B-V)=0.4

• 57 candidates were selected, while high resolution (>48000) and SNR

(>150) spectra (UVES, FIES, FEROS) of 33 targets were observed

(B−V) M V 0.5 1 1.5 −5 5 10

Hipparcos colour cut

• bserved

Tuesday, February 11, 14

5616 5617 5618 5619 5620 0.0 0.2 0.4 0.6 0.8 1.0

MgH 5615.045 C2 5615.441 MgH 5615.489 MgH 5615.489 MgH 5615.521 Fe 1 5615.644 C2 5615.926 C2 5616.156 C2 5616.172 C2 5616.269 C2 5616.317 MgH 5616.906 C2 5617.198 MgH 5617.358 C2 5617.750 C2 5617.870 G d 1 5617.910 C2 5617.924 C2 5618.061 MgH 5618.393 Fe 1 5618.632 MgH 5618.875 Fe 1 5619.225 MgH 5619.526 Fe 1 5619.595 C2 5619.649 C2 5619.658 C2 5619.808 MgH 5619.995 MgH 5620.279 MgH 5620.654 MgH 5620.673 MgH 5620.799 MgH 5620.923

5616 5617 5618 5619 5620

• 0.3
• 0.2
• 0.1

0.0 Residuals 5616 5617 5618 5619 5620 Wavelength 0.0 0.2 0.4 0.6 0.8 1.0 Intensity 079672_SME

Stellar parameters

• SME (Spectroscopy Made Easy) is used to determine Teff, logg, [Fe/H],

Vsini and elemental abundances

• Initial model atmosphere: Teff and logg were calibrated from photometric

and astrometric data ( [Fe/H]=0 )

• The Sun and other 4 stars, with well-known Teff and logg and high

resolution spectra, are used to construct a homogeneous line list

HIP 79672: Teff=5796, logg=4.34, [Fe/H]=0.058

Spectral analysis

Jofre et al. 2013

Tuesday, February 11, 14

Demarque et al. 2004, Bensby et al. 2011

• Stellar ages are estimated from

Y2 isochrones by maximising probability distribution functions

• 17 stars have solar metallicity, while the age of 4 candidates (HIP 10786, 30344, 40317

and 112584) are consistent with the solar age within one-sigma

• Systematic errors: ▵T ~ 60K, ▵logg ~ -0.07dex, ▵[Fe/H] ~ 0.06dex

3.6 3.7 3.8 1 2 3 4 5 6 7 8 9 (a)

log(T eff) M V

3.6 3.7 3.8 2.5 3 3.5 4 4.5 5 (b)

log( T eff) log g

5.01 Gyr 4.47 Gyr 3.16 Gyr

Tuesday, February 11, 14

Elemental abundances

• Flat trends for most of elements with abundances close to the

solar abundances, while [Al/Fe] and [Ni/Fe] show a transition

• Uncertainty of abundances are between 0.05 and 0.08 dex

[X/Fe]=[X/H]-[Fe/H]

−0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Na/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Mg/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Al/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Si/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Ca/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Ti/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Cr/Fe] −0.3 −0.2 −0.1 0.1 0.2 0.3 −0.2 0.2 [Fe/H] [Ni/Fe]

Tuesday, February 11, 14

Chemical tagging

• Mitschang et al. (2013) defined a metric to quantify

chemical difference between two stars using element abundances (Nc = 9)

• Pc that a probability of two stars belong to the same

cluster responds to

δC =

NC

• C

ωC |Ai

C − Aj C|

NC

δC

• Confidence limit of probability

Plim = 90 percent

Tuesday, February 11, 14

calculate for any one star and the Sun

δC

Process to identify solar siblings

possible cluster stars non-cluster stars Pc < 90 re-calculate for any two stars

δC

cluster stars Pc > 90 cluster detection confidence Pclus

Tuesday, February 11, 14

Results

• Three sibling candidates (HIP 7764, 21158 and 40317)

and the Sun may be from a dissolved cluster based on chemical tagging

• The cluster detection confidence is Pclue = 94 percent
• 4 stars (HIP 10786, 30344, 40317 and 112584) are

consistent with the solar metallicity and age within one- sigma

• Only star -HIP 40317- could be the lost sibling of the

Sun ([Fe/H]=0.04, Age ~ 4.4 Gyrs, and Pc > 90)

Tuesday, February 11, 14

Conclusion

• Chemical tagging is a powerful technical to detect dissolved

cluster

• Very small fraction of candidates (1/33≃0.03) are solar siblings
• The selection criteria developed in Brown et al. (2010) are

not optimal

• Since a smooth and axisymmetric Galactic potential has been

used to model stellar orbits in Brown’s work, a more realistic potential could prove more efficient at finding solar siblings

• If our results are true, then only few solar siblings left because
• f perturbation from the spiral arms or inner bar

Tuesday, February 11, 14

Thanks for listening!

Tuesday, February 11, 14