The AMBRE Project - ast.cam.ac.uk · 5th December 2012 AMBRE Presentation 3 Galactic Archaeology...

The AMBRE Project

Institute of Astronomy / Observatoire de la Cte d'Azur

5th September 2012, IoA

Galactic Archaeology with the ESO stellar archive

C. Clare Worley

5th December 2012 AMBRE Presentation 2

Presentation Summary

Galactic Archaeology & Stellar Classification

The AMBRE Project

Galactic Archaeology with the FEROS archived spectra


Galactic Archaeology

The exploration of large spectral datasets for underlying structures and

populations within the Galaxy in order to create a chemical and kinematic

chart with which to test theories of galactic formation & evolution.

Tracing the fossil record within the Milky Way

Key Characteristics:Position, Direction, Velocity, Age, Temperature, Gravity, Chemical Composition


The Gaia-ESO Survey

A census of a thousand million stars in our GalaxyThe Gaia Mission

300,000 stars sampling all major components of the Milky Way First homogeneous overview of the distributions

of kinematics & elemental abundance

Revolutionise our understanding of Galactic and Stellar Evolution


Stellar Parametrisation

CLASSICAL DETERMINATION:

Measurement of individual spectral line strengths Atomic/Molecular linelists, Stellar Atmospheric models,

Radiative transfer code Curve of Growth, Spectrum Synthesis Too time consuming for large samples

47 Tuc NGC 6388 NGC 362

50 stars in 3 globular clusters

Worley et al. 2008, 2009, 2010

25 million stars in the Milky Way

?

Effective Temperature: Teff Surface Gravity: log g Metallicity: [Fe/H] or [M/H] (global) Chemical Abundances: [X/Fe] e.g. [/Fe]: =Mg,Si,Ca,Ti...


Automated Stellar ParametrisationOptimisation Methods:

Minimum distances, Genetic algorithms, Gauss-Newton methodPattern Recognition Methods: Artificial Neural Networks, Support Vector Machines, Decision Trees

Projection on set Vectors: Principal Component Analysis

MATISSE: MATrix Inversion for Spectrum SynthEsisRecio-Blanco, Bijaoui, de Laverny (2006); Bijaoui, Recio-Blanco, de Laverny (2008)

Developed at Observatoire de la Cote d'Azur primarily for the Gaia mission Local multi-linear regression method B vectors reflect sensitivity of key parameters ( = Teff; log g; [M/H]; [/Fe]) Based on High Resolution Optical Synthetic Spectra Grid (de Laverny et al. 2012)


The AMBRE Project

Worley et al. 2012, A&A, 542, 48,


The AMBRE ProjectArchologie avec Matisse: aBondances dans les aRchives de l'Eso

ESO Spectrograph Resolving Power Spectral Domain Approximate No. archived spectra

FEROS 48,000 350nm - 920nm 20,000HARPS 115,000 378nm - 691nm 125,000UVES 40,000 to 110,000 300nm - 1100nm 100,000

Flames/GIRAFFE 5,600 to 46,000 370nm - 900nm >100,000Total Sample >350,000

This extensive dataset covers a large range of wavelengths and resolutions, including the wavelength domain and resolutions of Gaia RVS.

This is a unique opportunity to carry out comprehensive testing of MATISSE which is essential to its preparation for use in the Gaia Mission (GSP-spec).

ESO/OCA project: 2009 - 2012

Collaborators: P. de Laverny, A. Recio-Blanco, V. Hill,C. Ordenovic, F. Guitton, J.C. Gazzano, M De Pascale and A. Bijaoui


Key Objectives for the AMBRE ProjectESO: Advanced data products for ESO archive

Homogeneous determination of stellar parameters for archived spectra Key information for each spectra in the ESO database Available to the entire astronomical community ESO Archive

OCA: Essential testing of MATISSE for GSP-spec & Gaia Gaia RVS wavelengths and resolutions Large scale dataset of real stellar spectra Extensive error analysis prior to Gaia's launch

Galactic Archaeology: Studies of the Milky Way Creation of a galactic chemical chart Selection of specific stellar sample, i.e. halo, thick disk, thin disk... Kinematic & chemical analysis of homogeneous samples


The AMBRE:FEROS Analysis

FEROS Targets in Galactic Coordinates

Majority of Southern Sky is sampled

FEROS: Instrument Resolution ~ 48,000

Wavelengths 350nm to 920nm

No. Archived Spectra 21551

No. Stars ~ 6285

MATISSE SET-UP: 17 Wavelength Regions: Mg triplet (~5160 ), H (4861 )

1500 from 4000 to 7000 Synthetic Grid & B Functions at Resolution ~15,000


Comparison to literature values (SIMBAD, PASTEL, S4N...)

Radial Velocity determinationVrad, VsiniError estimatePreferred-mask parameters

Spectral Processing AWavelength selectionNormalisationCosmic ray removalSignal-to-noiseConstant Spectral FWHM

Spectral Processing BSpectral Processing ARadial Velocity correctionMeasured Spectral FWHM

MATISSEStellar parameters (Teff, log g, [Fe/H], [/Fe])Synthetic spectra

Spectral Processing C

Modified Spectral Processing BNormalisation using synthetic spectra

Select alternate: Radial VelocityNormalisation

Final Stellar parametersFinal Radial Velocity

Final Normalised SpectraFinal Synthetic Spectra

YES NO

AMBRE:FEROS Analysis Pipeline

Tests for:Convergence?Goodness of fit? (2)

MATISSEStellar parametersTests


FEROS Validation: Homogeneous Reference Samples


Radial Velocity Comparison

S4N Gaia RVS Standards

Vrad : Cross-correlation of observed spectrum with binary masks(Private communication, C. Melo)


PASTEL Analysis: Effective Temperature


PASTEL Analysis: Surface Gravity


PASTEL Analysis: Metallicity


S4N and Bensby et al 2003: [/Fe]


Stellar Parameters for the ESO: FEROS archived spectraNo. of FEROS archived spectra = 21551No. of spectra analysed in SPC = 19181No. of spectra rejected = 15043No. of spectra accepted = 6508 (~3087 stars)

Rejection Criteria I: 2370 spectra prior to SPC

Extreme Emission Features (nova), Extremely Noise Spectra, Poor Normalisation, Missing orders

Rejection Criteria II: 12673 spectra after SPC

Radial Velocity Determination Errors and spectral FWHM

Mainly hot/fast rotating stars: outside limits of Synthetic Grid (FGK stars)


FEROS Spectra: The Parameters6508 spectra 3087 stars 30.2 % of archived sample


Galactic Archaeology with the ESO:FEROS Archived Spectra


FEROS Stars: Cross-matching Catalogues Proper Motions & Apparent Magnitudes (V, B, J, K ) (PPMXL)

Distances

1 Hipparcos new reduction (van Leeuwen 2007)

2 Absolute Magnitudes using Isochrones: MV, MB, MJ, MK

Kordopatis et al, 2011a&b

& Interstellar reddening from Theoretical Colours

Gazzano et al. 2010, 2012; Gonzalez et al, 2009

D = 100.2(m M +5) ZGC = Z + Dsinb

Distance from Sun & above Galactic Plane


PPMXL: V, B, J, K, Proper Motions Hipparcos: V, B-V, Parallaxes, Proper Motions

FEROS Stars: Photometry

PPMXL

Hipparcos


FEROS Stars: Metallicity Distribution


Disk Stars: Metallicity Gradients?

FEROS PPMXL SampleRGC, ZGC ~ 1.5 kpc

Dwarfs Not useful for gradients Useful as Volume-limited,

Local Solar Neighbourhood

Giants Dispersed to higher Z More metal-poor

(pc) (pc)

(pc) (pc)


Chemically & Kinematically Characterised Stellar Sample in the Local Solar Neighbourhood

Stellar Populations: Thick & Thin Disk

Total (radial,vertical) kinetic energyvs

rotational energyThin Disk: < ~50kms-1Thick Disk: ~50kms-1 ~150kms-1Halo: > ~150kms-1 (Bensby 2009)


Open Clusters: Metallicity Gradient107 stars (147 spectra) over 26 open clusters.

1. Star Coords < 0.5o of cluster centre,

2. Star Vrad < 2 kms-1 of cluster Vrad.

Other Studies:Pedicelli et al. 2009Haywood 2008Chen et al. 2008


Galactic Archaeology: SN Pollution in Milky Way

SN Type II

SN Type Ia

Comparison to Bensby et al. (2003)[/Fe] to [M/H]with log g > 3.4 dex.

Dwarfs, Sub-Giants

Supernova Type II:-element production

Supernova Type Ia:Fe production

Change in regime from SN Type II to SN Type Ia as pollution source.

21 Thick Disk45 Thin Disk


SummaryGalactic Archaeology is a rapidly developing field of research

Wealth of data from large scale surveys Advances in automated analysis Development of crucial calibration datasets

The AMBRE Project takes advantage of the pre-existing large spectral datasets in the ESO Archive

Accepted parameters for 30.2% of FEROS spectra Recycling of archive spectra UVES, HARPS analyses almost complete Archive-based galactic chemical chart


Acknowledgements

Observatoire de la Cte d'Azur European Southern Observatory Centre National d'Etudes Spatiales

11th September 2012 AMBRE Presentation 31

MATISSE: MATrix Inversion for Spectral SynthEsis

Stellar parameters ( = Teff, log g, [M/H], individual chemical abundances) are derived by the projection of an input observed spectrum O() on a vector function B().

The B() function is an optimal linear combination of theoretical spectra S() calculated from a synthetic spectra grid (the learning phase).

Recio-Blanco, Bijaoui, de Laverny (2006); Bijaoui, Recio-Blanco, de Laverny (2008)

B()Oi()

The B() function reflects the sensitivity of the particular spectral regions to the parameter in question, .

Local multi-linear regression method

i Si()

i


The Grid of High Resolution Synthetic Spectra

~16000 synthetic spectra Optical domain: 3000 - 12000

MARCS model atmospheres (Gustafsson et al., 2010) Teff : 2500 8000 K log g : 0.0 +5.0 dex [M/H] : -5.0 +1.0 dex [/Fe] : -0.4 +0.8 dex

Molecular line lists from Bertrand Plez

Atomic line lists : depending on the analysis grid

MATISSE TRAINING PHASEB functions are generated corresponding to each point in the spectral grid for each parameter

Computed using OCA Mesocentre ~ 50,000hrs computing time

de Laverny et al. 2012, A&A, 544, 126


MATISSE: Using the B functions

B0: Teff(0), log g(0), [M/H](0)

Bi: Teff(i), log g(i), [M/H](i)

BBi i B Bi -1i -1

Final: Teff, log g, [M/H]

Teff

log g

[M/H]

3D Representation of Synthetic Grid


MATISSE: Java InterfaceInputs Comparison of Synthetic and Observed Spectra

Quality flag


The FEROS Archived Spectra: Normalisation

SPB

SPB Final

SPCi=0


FEROS Spectra: The RejectionsWell-defined CCF vrad 10 km/s

FWHM CCF 50 km/s

SPC

Spectral FWHM 0.8 m Strong line FWHM 8 m

C. Clare Worley Automated Stellar Classification & Galactic Archaeology Preparing MATISSE for the ESA Gaia MissionAutomated Stellar Classification & Galactic ArchaeologyGlobular Clusters & Large Scale SurveysSlide 4Slide 5Slide 6Slide 7AMBRE Archologie avec Matisse: aBondances dans les aRchives de l'EsoKey Objectives for Gaia, ESO and Galactic ArchaeologySlide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28Slide 29Slide 30Slide 31Slide 32Slide 33Slide 34Slide 35Slide 36

The AMBRE Project - ast.cam.ac.uk · 5th December 2012 AMBRE Presentation 3 Galactic Archaeology...

Documents

Transcript of The AMBRE Project - ast.cam.ac.uk · 5th December 2012 AMBRE Presentation 3 Galactic Archaeology...