1Formation of the Milky Way[CO fig 24.6]+ [CO Tbl 24.1]+ dark matter halo> 230 kpc radius~ 200x1010MsunStellar Halo• 11-13 Gyr old- horizontal branch• very low Z (metallicity)• elongated orbits•0.3x1010MsunNuclear Bulge• 0.2 - 10 Gyr old-age-flatnesscorrelation•high Z (metallicity)• elongated orbits- but much smaller than for halo stars•1x1010MsunThick Disk• ~10 Gyr old• moderately low Z (metallicity)• elongated orbits•0.3x1010MsunThin Disk•~8 Gyr•solar Z• circular orbits•6x1010MsunGas Disk•< 10 Gyr old• above solar Z• circular orbits• 0.5 x1010MsunELS, 1962, ApJ, 136, 7482Measuring metallicity from colorsFor main-sequence stars:• More metals Î higher absorption coefficient.• Line-blanketing vs. back-warming.Line blocking as a function of λEffect on color-color diagramEggen, Lynden-Bell & Sandage (ELS)Angular Momentum vs. δ(U-B)• Rule out star formation before collapse• If gas cloud self-supporting on galaxy-sized scale, also self-supporting on star-sized scale.• Ang. Momentum of metal-poor stars same as for stars in circular orbits at R ~ 5 kpc, but halo stars have much higher energy orbits Î galaxy not in dynamical equilibrium when halo stars formed.Î halo stars formed from infalling gas.Timescale & extent of collapse:• Large observed e Î collapse faster than rotation period of ~108yrs. Î all halo GCs formed within 108yrs• Collapse in Z direction is factor ~25• From max Z of halo stars vs. thickness of gas disk• Collapse in R direction is factor ~ 10• Angular momentum of individual gas elements stays constant as gas goes from collapse to disk.• Î compare Rmaxof halo stars to R of disk stars with same angular momentum. Orbits of 221 stars selected by high velocities relative to Sun.δ(U-B) Î metallicity; δ(U-B) ~ 0 Î solar metallicity; δ(U-B) large Î metal poor.3Problems for ELS Model• Halo stars have angular momentum ~ 0 – ~ ½ of all halo stars are in retrograde orbits• Globular cluster age spread– 3 billion yr spread not consistent with freefall timescale tff~ 6x108yrs.• Range of globular cluster chemical abundances – Near galactic center Î metal rich, but (perhaps) older.– Outer halo Î wider range in [Fe/H], but on average younger.– Spheroid vs old disk G.C. distributions:• Multi-component disk with different ages.• Evolution of chemical abundances– G dwarf problemGlobular Cluster Distribution on skyMetal poorMetal richGlobular Cluster Ages and MetallicitiesAge-metallicity degeneracy in fits • Must use several indices.• Easy to confuse age effect with metallicity effect.• Hard to get absolute calibration of metallicity. Worthey et al. 1994Trager et al. 20004Closed Box Models(and friends and relatives)Gas Î stars Î enriched gasS = mass of starsM = mass of metals (heavy elements) in ISMG = total mass of gas in ISMAssume instantaneous recycling from massive stars.From a new generation of stars:dS = mass of low mass stars added to Sp dS = mass of heavy elements added to M frommassive stars in this generation.where p = yield.dM = p dS – Z dS= -p dG + Z dG since dG = -dSBut dZ = d(M/G) = (1/G) dM – (M/G2) dG= -p (dG/G)Z(t) = -p ln [G(t)/G(0)] or G(t)=G(0) e-Z(t)/pMetallicityZ = M/GZ~ 0.02G dwarf problemS[Z<Z(t)] = S(t) = G(0) – G(t)= G(0) { 1 – e-Z(t)/p}Z(t) = gas metallicity at time tCompare to case when gas had some arbitrary fraction αof that metallicity:S[Z<αZ(t)] 1-XαS[Z< Z(t)] 1-Xwhere X = ~ 0.1 – 0.2Predicts broad distribution in metallicity of stars.Î S[Z<1/4 Z] = 0.4 S[Z<Z]Very different than what is observed in solar neighborhood:S[Z<1/4 Z] = 0.02 S[Z<Z]Also… Leaky box (gas driven out by stars),Accreting box models.=G(t )G(0)Bottom-up (Hierarchical Merger) Formation of Milky WayobservedClosed box, fully processedSee alsoZinn 1980, 1985Searle & Zinn (1978)– Motivated by large range in metal abundances of halo globular clusters in outer halo• No radial abundance gradient Î Not slow contraction• Consistent with either – formation during freefall» ELBS model –or mergers.– Claim that age is an additional parameter:• Tightly bound clusters – small age spread– inner halo collapsed in < 109yrs • Loosely bound clusters – large age spread– Outer halo tcollapse> 109 yrs.– Distribution of [Fe/H] Î all gas used up • (if a closed box model)• No gas left over for later formation of disk• Alternate explanation: many stripped subsystems– MW assembled from ~ 108 Msun“proto-galactic fragments” which had already formed stars and undergone chemical evolution.• Halo formed from different fragments than disk, so different angular momentum is OK.• Dense central region evolved rapidly Î high [Fe/H] bulge of today• Halo formed from inside to outside (bulge formed first).r (kpc)[Fe/H]Z/ZsunÎNumber Î5Embellishments to Bottom-Up Model• Thick disk – formed during slow dissipative collapse of remaining gas, which had angular momentum• Heating, metal enrichment through SNe over ~400 million yrs.•T ~ 106K Î scale height ~ 1.6 kpc• Collapse to old thin disk– Star formation in thin disk over next 5 billion yrs– Further gradual collapse to thinner gas disk we see today.• Metallicity vs. orbital eccentricity of halo starsdata used by ELBS (1962) Tim Beers’ bulletin board (2000)Milky Way Mergers• Recent/current dwarf galaxy mergers – Sagittarius– Monoceros– Canis-Majoris• Show up as star streams in halo• Magellanic clouds– Magellanic stream (H I gas)– But recent result casts doubtSDSSObserved by 2MASS6MW Formation: Bottom-Up or Top-Down?• Favoring bottom-up– ΛCDM cosmology says so!– Small galaxies currently merging with MW– Halo has two major components • Distinct metallicities and kinematics (Carrollo, Beers et al. 2007)• Favoring top-down– Disk clearly formed from gas, not from stars pre-formed in smaller sub-units.– ΛCDM predicts 100s of low mass DM halos still orbiting MW• Only 10-15 are seen.• But SDSS is starting to find more• Top-down apologia– Thick disk may be stars stirred up from thin disk by accretion of dwarf galaxies.– Bulge stars may be formed from gas falling in from halo and disk. The issue is still unclear…May be a combination of both, or bottom-up may do it