GAIA Composition, Formation and Evolution of our Galaxy thanks to Michael Perryman, the GAIA Science Team (present and past) and 200+ others

M83 image (with Sun marked) M83 (AAO, D. Malin) ‘the Sun’

GAIA: Key science objectives Structure and kinematics of our Galaxy: –shape and rotation of bulge, disk and halo –internal motions of star forming regions, clusters, etc –nature of spiral arms and the stellar warp –space motions of all Galactic satellite systems Stellar populations: –physical characteristics of all Galactic components –initial mass function, binaries, chemical evolution –star formation histories Tests of galaxy formation: –dynamical determination of dark matter distribution –reconstruction of merger and accretion history  Origin, Formation and Evolution of the Galaxy

Overview 1.Measurement principles 2.The GAIA satellite and mission 3.Some science examples 4.Schedule, organisation of work 5.Current situation 6.Summary

Some history… 1725: Stellar aberration (Bradley), confirming: –Earth’s motion through space –finite velocity of light –immensity of stellar distances 125 B.C.: Precession of the equinoxes (Hipparchus) 1717: First proper motions (Halley) 1783: Sun’s motion through space (Herschel) : First parallaxes (Bessel/Henderson/Struve) 610 B.C.: Obliquity of the ecliptic (Anaximander) 1761/9: Transits of Venus across the Sun (various) – solar parallax

Principle of parallax measurement

Astrometry from antiquity to GAIA

Hipparcos ESA mission 1989 – 1993 (catalogue in 1997) stars in 2 filters Limiting magnitude: V=12.4 Complete to V=7.3 to 9.0 Median precisions (H p <9.0): 0.7 mas positional 0.7 mas/yr 1.0 parallax Distance <10%: Distance <20%: Tycho (star mapper): stars in 2 filters 7 mas (H p <9.0)

Measurement Principle

Measurement principle: b

Global astrometry basic angle

Measurement principle: c

GAIA: complete, faint, accurate

Scientific design considerations Astrometry (V < 20): –completeness  on-board detection –accuracies: 10  as at 15 mag (science) –continuously scanning satellite, two viewing directions  global accuracy, optimal with respect to observing time Radial velocity (V < 17-18): –third component of space motion –account for perspective acceleration (nearby, fast stars) Photometry (V < 20): –astrophysical diagnostics (4-band + 11-band) + chromatic correction  extinction,  T eff ~ 200 K, [Fe/H] to 0.2 dex

Satellite Deployable: solar array/sun-shield Size: 8.5m diameter (4.2m stowed) 2.9m height (2.1m for payload) Mass: 3100 kg (800 kg payload) Power: 2600 W Launch: dual Ariane 5 Orbit: Sun-Earth L2 (Lissajous) Data rate (phased array): 1 Mbs -1 sustained 3 Mbs -1 downlink (1 ground station) Launch date: Attitude control: FEEP thrusters Design lifetime: 5 years ESA only mission

Payload overview Two astrometric instruments: field of view = 0.6 o  0.6 o separation = 106 o Monolithic mirrors: 1.7 m  0.7 m Non-deployable, 3-mirror, SiC optics Astrometric focal planes: TDI CCDs Radial velocity/photometry telescope Survey principles: revolving scanning onboard detection complete and unbiased sample

Payload configuration

Astrometric instrument: Light path

Astrometric focal plane Sky mapper: - detects all objects to 20 mag - rejects cosmic-ray hits - mag + x,y to main field Main field: - area: 0.66 x 0.56 deg = 0.37 deg 2 - size: 60  70 cm 2 - number of CCDs: 17 x 8=136 - CCDs: 2780 x 2150 pixels - 120´´/s  0.86s per CCD Pixels: - size: 9 x 27  m 2 (37 x 111 mas) - read in TDI mode Broad-band photometry: - 4 band (chromatic correction)

On-board source detection Requirements and constraints: –unbiased sky sampling (mag, colour, resolution, etc.) –no all-sky catalogue at GAIA resolution (0.1 arcsec) to V~20 –cannot transmit entire sky at 0.1 arcsec resolution (telemetry limitations) Solution: on-board detection and sampling –no input catalogue or observing programme (big effort for Hipparcos) –good detection efficiency to V~21 mag –maximum star density: ~ 3 million stars/deg 2 (Baade’s Window) –reduces data rate from several Gbps to a few Mbps Will therefore also detect: –supernovae: 10 5 expected –microlensing events: ~1000 photometric –variable stars (eclipsing binaries, Cepheids, etc) –Solar System objects, including near-Earth asteroids and KBOs

Sky scanning principle Spin axis: 55 o to Sun Scan rate: 120 ´´/s (3 hours) Precession rate: 0.20 ´´/s (76 days) Continuously observing Full sky coverage Each position observed 67 times (on average) per astrometric instrument

Scanning law Observations over 5 months Ecliptic co-ordinates

Astrometric accuracies Derived from detailed analysis: image formation (polychromatic PSF) evaluation vs. spectral type/reddening comprehensive detector signal model sky background and image saturation attitude rate errors and sky scanning on-board detection probability on-ground location estimation (centroid to pixel in hardest case) error margin of 20 per cent included results folded with Galaxy model Fraction of stars with given relative parallax error vs. magnitude (towards Galactic poles) 5-year accuracies in  as

Accuracy example: Stars at 15 mag with   /  0.02

CCD centroiding tests Astrium contract (Sep 2000) ‘GAIA-mode’ operation EEV CCD  m pixels Illumination: 240,000 e - Frequencies: TDI: 2.43 kHz Readout: 90 kHz Differential centroid errors: rms = pixels (  1.2  theoretical limit)

Survey accuracies compared

1  as is a very small angle... the Earth seen from the Pleiades (100 pc) Baden-Württemberg seen from  Centauri (1.3 pc) a human hair observed in Kabul from Heidelberg (5000 km) Earth-Sun system seen from 1 Mpc (by definition) a grain of rice seen on the Moon ( km)

Design requirements Need to monitor basic angle variations to 1  as: –25  K thermal gradients –16 pm peak-to-peak variation over 0.85m  onboard interferometer High thermal and mechanical stability: –no moving parts (electronically steerable phased array) –L2 orbit (dynamical stability, eclipse avoidance, low radiation)

Astrometric reduction Approximate single epoch astrometry (<<1´´) from star mapper and GAIA orbit/attitude Great circle scans  1D positions in two fields separated by a well-known basic angle (monitored) Object matching in different scans (e.g. GSC II method) Full sphere reduction to determine 5 astrometric parameters per star by a global iterative method (>100 measures) Binary models fitted to systems with large residuals GAIA observations of quasars (known and new) put the astrometry on a quasi-inertial reference system

Object matching Sky scan (highest accuracy along scan)

GAIA spectrophotometry and radial velocities High resolution spectra for: - 3 rd component of space motion - perspective acceleration - stellar abundances, rotation velocities Medium band photometer for: - classification of all objects - physical parametrization of stars T eff, log g, [Fe/H], [  /H], A( )

Radial Velocity Measurement Concept F3 giant S/N = 7 (single measurement) S/N = 130 (summed over mission)

Radial Velocity and Photometric Instrument Mounted on same toroidal support Observes same scanning circles Independent star mappers Photometry for all stars (to 20 mag) Radial velocities to ~ 18 mag 1-10 km/s accuracy 0.5´´ spatial resolution

Spectral Sequences around Ca II Effect of temperature: A to M starsEffect of metal abundance in G stars

Photometric system and accuracies 11 medium band (in Spectro, 100 obs.) 4 broad band filters (in Astro, 2x67 obs.) SNR: at V=15 ( at V=20) (end of mission) F33B45 B63B82

GAIA and our Galaxy 10  as = 10% distances at 10 kpc 10  as/yr = 1 km/sec at 20 kpc

GAIA: Why a survey to 20 mag?

GAIA capabilities Distances: <0.1% for stars <1% for 21 million <10% for 220 million Transverse motions: <0.5% km/s for 44 million <3 km/s for 210 million <10 km/s for 440 million Radial velocities to a few km/s complete to V= band photometry ( nm) at ~100 epochs over 4 years Complete survey of the sky to V=20, observing 10 9 objects: 10 8 binary star systems (detected astrometrically; 10 5 orbits) disk white dwarfs brown dwarfs planetary systems resolved galaxies 10 5 quasars 10 5 extragalactic supernovae Solar System objects ( presently known)

Stellar astrophysics Comprehensive luminosity calibration, for example: –distances to 1% for 18 million stars to 2.5 kpc –distances to 10% for 150 million stars to 25 kpc –rare stellar types and rapid evolutionary phases in large numbers –parallax calibration of all distance indicators e.g. Cepheids and RR Lyrae to LMC/SMC Physical properties, for example: –clean Hertzsprung-Russell sequences throughout the Galaxy –solar neighbourhood mass function and luminosity function e.g. white dwarfs (~200,000) and brown dwarfs (~50,000) –initial mass and luminosity functions in star forming regions –luminosity function for pre main-sequence stars –detection and dating of the oldest (disk and halo) white dwarfs

Binary and multiple stars Constraints on star formation theories Orbits for > 100,000 resolved binaries (separation > 20 mas) Masses to 1% for > 10,000 objects throughout HR diagram Full range of separations and mass ratios Interacting systems, brown dwarf and planetary companions Photocentric motions: ~10 8 binaries Photometry: >10 6 eclipsing binaries Radial velocities: >10 6 spectroscopic binaries

Brown dwarfs Isolated systems (Haywood, Les Houches 2001): –very red  very few objects (photometry limited by CCD physics) –detections at G=20 mag sensitive to models (slope of IMF):  = 1: ~ objects (also from 2MASS constraints) –sampling biased to very young and massive BDs (<100 pc): 10% with M<0.05 Ms; ~50% with age < 1 Gyr Binary systems (Quist, A&A 2001, 370, 672): –many to 100pc with P= yr will be detected –orbital solutions for P < 15 yr –from 280,000 primaries to 100pc, expect ~3000

Exosolar planets: Detection domains No sin i ambiguity in mass determination from astrometry

Known planetary orbits

Planets: astrometric signatures ….  in arcsec if a in AU and d in pc

Sampled orbits simulated for GAIA 47 UMa (Lattanzi et al. 2000)

Multiple systems seen with GAIA:  And (Sozzetti et al. astro-ph/ ) - relative inclination important for stability - V rad does not give i or    And B (P=4.6 d) undetected C (241 d) and D (1200 d) detected For large  and well sampled P (<5 years) Mp and i are found as if single, hence: coplanarity to a few degrees

Expected astrometric planetary discoveries Monitoring of hundreds of thousands of stars to 200 pc for 1M J planets with P < 10 years: –complete census of all stellar types (P=2-9 years) –actual masses, not just lower limits (m sin i) –20,000-30,000 planets expected to pc –e.g. 47 UMa: astrometric displacement 360  as Orbits for many (~ 5000) systems Masses down to 10 M Earth to 10 pc

Solar System objects: Detection principles ASM1 ASM3 AF1 AF2 AF3 AF17 …….. Star NEO Detection to V = 20 mag and SAA o At 300 arcsec/s moves by 1.2 arcsec in 15 s Sky Mapper Astrometric Field Detection Confirmation 2.2  1.3 arcsec  0.89 arcsec 2

The Inner Solar System Jupiter Trojans (610) numbered periodic comets other comets minor planets (numbered/multiple apparition) perihelia < 1.3 AU

Deep and uniform detection of all moving objects: complete to 20 mag discovery of ~ new objects (cf. 65,000 presently) taxonomy and mineralogical composition versus heliocentric distance diameters for ~1000 asteroids masses for ~100 objects orbits: 30 times better than present, even after 100 years Trojan companions of Mars, Earth and Venus Edgeworth-Kuiper Belt objects: ~300 to 20 mag + binarity + Plutinos Near-Earth Objects: –~1600 Earth-crossing asteroids > 1 km (100 currently known) –GAIA detection: m at 1 AU, depending on albedo GAIA: Studies of the Solar System

Solar eclipse (image) General Relativity (light bending)

Light Bending at L2 by solar system bodies Klioner (2002) de Bruijne (2002)

Gravitational light deflection de Bruijne (2002)

General Relativity Parametrized Post-Newtonian (PPN) formulation –  = 1.0 for General Relativity (GR) –alternative scalar-tensor theories deviate by GAIA will measure  to 5  from positional displacement at large angles from the Sun –  currently known to –GAIA tests GR at times lower mass than presently –effect of Sun: 4 mas at 90 o ; Jovian limb: 17 mas; Earth: 40  as Microlensing: photometric (~1000) and astrometric (few) events

Galaxies, quasars, and the reference frame Parallax distances, orbits, and internal dynamics of nearby galaxies Galaxy survey ( resolved at 0.1´´ in four bands, 0.5´´ in 11 bands) ~500,000 quasars: kinematic and photometric detection ~100,000 supernovae Galactocentric acceleration: 0.2 nm/s 2   (aberration) = 4  as/yr Globally accurate reference frame to ~0.4  as/yr

Schedule Acceptance Technology Development Design, Build, Test Launch Observations Analysis Catalogue

Future schedule : Phase A technical preparatory study : –study of critical items identified during concept study (focal plane, CCDs, SiC mirrors, data handling, etc) –objective: confidence in technology, cost, schedule –end 2003: scientific and technical report : Phase B detailed design : Phase C/D construction and launch mid-2012: launch (formally) –ESA target date: mid-2011 –community target: mid-2010

Cost at completion

High performance, small pixel (9  m) CCDs Focal plane assembly and verification; on-board detection Large silicon-carbide (SiC) mirrors (1.7  0.7 m 2 ) Ultra-stable large size SiC structures for payload optical bench Large deployable solar array/sunshield assembly Phased-array antenna for high data rate (1 Mbps) in far orbit (L2) Data processing, object classification and physical parametrization Technology developments and challenges

Organisation of scientific work Working groups: about 150 European ‘core’ and ‘associate’ members

GAIA Science Team (GST) Frederic Arenou (Meudon) Coryn Bailer-Jones (MPIA, Heidelberg) Ulrich Bastian (ARI, Heidelberg) Erik Hoeg (Copenhagen) Andrew Holland (Leicester) Carme Jordi (Barcelona) David Katz (Meudon) Maria Lattanzi (Torino) Lennart Lindegren (Lund) Xavier Luri (Barcelona) Francois Mignard (Nice) Michael Perryman (Project Scientist, ESA)

Current GAIA activities ESA/industrial: Phased array antenna (Alcatel) High stability optical bench (Astrium) Telemetry budget and compression algorithms (ESA) FEEPs: will be based on SMART-2 activities (ESA + industry) Proof of concept data reduction system (Barcelona + GMV) New important industrial study contracts (CCD and focal plane study, mirrors) are on hold due to internal ESA SPC/IPC conflict Scientific community (working groups): On-board detection algorithms Detailed specification of radial velocity instrument Definition of optimal photometric system Simulations of data stream Development of object classification and physical parametrization methods

Data Analysis: Concept and requirements Capacity: ~200 Terabytes (1 Tb = 1000 Gb) Overall system: centralised global iterative approach Peculiarity: data mixed-up in time, position and object Processing requirements: entire task is ~10 19 flop Data base structure: e.g. Objectivity Results: time-critical results available early (NEO, supernovae etc) Prototype: Hipparcos global astrometry re-reduced during concept study

Data processing comparison GAIA: 1 Mbps, 200 TB SDSS: 20TB will be collected; final catalogue ~ 1 TB Planck: data rate 100kbps, maps ~ 2 Gb VLT: 80 TB over 6 years GSC2: 6 TB at end 1999 CERN: LHC (~ Petabytes) SLAC: BaBar (~1TB/day) Human Genome Database: ~25 GB

Data processing demonstration (GDAAS) under development since mid-2000 (GMV/UB/CESCA) sky divided into hierarchical triangular mesh (level 6) presently: 8 nodes, 4 processors per node, 0.5 Tbyte disk telemetry ingestion, object matching, sphere iteration iterative processing for 1 million stars (final results April 2002) platform for further development/experimentation Processing 1 N Special Processing Master Database 1 N

Object classification/physical parametrization classification as star, galaxy, quasar, supernovae, solar system objects etc. determination of physical parameters: - T eff, logg, [Fe/H], [  /H], A( ), V rot, V rad, activity etc. combination with parallax to determine stellar: - luminosity, radius, (mass, age) use all available data (photometric, spectroscopic, astrometric) must be able to cope with: - unresolved binaries (help from astrometry) - photometric variability (can exploit, e.g. Cepheids, RR Lyrae) - missing and censored data (unbiased: not a ‘pre-cleaned’ data set) multidimensional iterative methods: - cluster analysis, k-nn, neural networks, interpolation methods required for astrometric reduction (identification of quasars, variables etc.)  produce detailed classification catalogue of all 10 9 objects

Top level classification system

Recent developments Nov. 2001:Ministerial meeting reduces science budget by 435 MEuro over (20% reduction) Nov. 2001:ESA+Astrium start to identify cost savings Oct. 2000:GAIA approved by SPC for launch by 2012 Dec. 2001:SPC requests review of whole ESA science program April 2002:Astrium presents revised GAIA design to GST May 2002:ESA/Astrium deliver final report to SPC June 2002:SPC decide on future science program April 2002:AWG subgroup (Rix, Aerts, Ward) reports on GAIA to AWG; (AWG to SSAC to SPC)

GAIA cost saving measures (ongoing) Substantial savings possible if released from ESA constraints: 1.re-use of satellite subsystem designs from other missions 2.cheaper contract competition mechanism 3.Soyuz-Fregat launch instead of Ariane 5 (110  40 MEuro) -smaller mirrors in along-scan direction, accommodated by increased across-scan dimension, field-of-view size... Other possible cost savings from: smaller solar array/sunshield by reducing solar aspect angle (accommodate for lost astrometric precision in optics) C-SiC optical bench unification of pixel scales field superposition (as in Hipparcos) smaller ESA project team Cost reduction from 570 to MEuro possible without any reduction in accuracies or scientific goals

Main performances and capabilities Accuracies: –4  as at V=10 10  as at V=  as at V=20 –radial velocities to few km/s complete to V=17-18 –10  as  distances to 1% at 1 kpc or 10% at 10 kpc (V=15) –10  as/yr at 20 kpc  1 km/s Capabilities: –sky survey in four bands at ~0.1 arcsec spatial resolution to V=20 –15 band multi-epoch photometry to V=20 –dense quasar link to inertial reference frame  every star observed in the Galaxy and Local Group will be seen to move  GAIA will quantify a 6-D phase space for over 300 million stars and a 5-D phase-space for over 10 9 stars

GAIA and our Galaxy 10  as = 10% distances at 10 kpc 10  as/yr = 1 km/sec at 20 kpc

Hyades (animation) Hyades 60,000 years Aldebaran (non-member) core radius  3 pc direction of motion Perryman et al. (1997)

Gravitational light deflection de Bruijne (2002)

GAIA stands for... Global Astrometric Interferometer for Astrophysics Galactic Astrophysics through Imaging and Astrometry General Astrometric Instrument for Astronomy Great Advances In Astrophysics Great Accuracy In Astrometry