Carbon-Enhanced Metal-Poor (CEMP) in the Milky Way KASI Seminar, May 27, 2015 Young Sun Lee Chungnam National University

Outline Part I  CEMP Stars  Identification from Metal-Poor (MP) Stars  Origin of CEMP Stars Part 2  Exploring the Milky Way  Constraining the Halo’s Initial Mass Function (IMF)  Halo Dichotomy with CEMP Stars  Search for the Signature of the Nucleosynthesis of First Generation of Stars (Pop III)  Looking forward 2

3 Metal-Poor (MP) Stars  Efforts to searching for Very Metal-Poor (VMP; [Fe/H] < -2.0) stars  HK Survey of Beers and colleagues  HES(Hamburg ESO) Survey of Christlieb and colleagues  These surveys identified several thousand VMP stars  Many tens of VMP stars from  Original Sloan Digital Sky Survey (SDSS)  SEGUE (Sloan Extension for Galactic Understanding and Exploration) I and II  Ongoing SDSS (e.g., BOSS)  LAMOST (Multi-Object fiber Spectroscopic Telescope)

4 Why We Need Large Numbers of MP Stars ?  Extremely MP stars have recorded the heavy element abundances produced in the first generations of stars  Nucleosynthesis in the early universe  Determination of the frequency of C-enhanced or alpha-rich stars  Initial mass function and star formation history  Identification of relatively rare objects(r-process / s-process enhanced stars) amongst MP stars  Site for production of neutron capture elements  Change in the nature of the Metallicity Distribution Function (MDF) as a function of distance may reveal the assembly history of the MW

Known MP Stars – Pre and Post SDSS/SEGUE-1/SEGUE-2 Name Metallicity Pre Post Metal-Poor (MP) [Fe/H] < ,000150,000+ Very Metal-Poor (VMP) [Fe/H] < ,000 30,000+ Extremely Metal-Poor (EMP) [Fe/H] < Ultra Metal-Poor (UMP) [Fe/H] < Hyper Metal-Poor (HMP)[Fe/H] < Mega Metal-Poor (MMP)[Fe/H] <  Until 2013 (Nomenclature; Beers & Christlieb 2005) 5

Known MP Stars – Pre and Post SDSS/SEGUE-1/SEGUE-2 Name Metallicity Pre Post Metal-Poor (MP) [Fe/H] < ,000150,000+ Very Metal-Poor (VMP) [Fe/H] < ,000 30,000+ Extremely Metal-Poor (EMP) [Fe/H] < Ultra Metal-Poor (UMP) [Fe/H] < Hyper Metal-Poor (HMP)[Fe/H] < Mega Metal-Poor (MMP)[Fe/H] < Septa Metal-Poor (SMP)[Fe/H] < Octa Metal-Poor (OMP) [Fe/H] < Giga Metal-Poor (GMP) [Fe/H] < Note that EMP stars probably include additional UMP, HMP, MMP, SMP, OMP, GMP stars, but wont be revealed as such until high-resolution follow-up conducted (contamination due to interstellar CaII K and/or carbon) 6

7 Identification of Carbon-Enhanced Metal-Poor Stars  The HK and HES surveys revealed an unexpectedly large number of VMP stars with anomalously strong CH bands  Both surveys identified about 100 CEMP stars for [Fe/H] < -2.5 CH G band

8 High-Resolution Follow-up  Detailed chemical-abundance analyses of VMP stars confirmed:  Most VMP stars exhibit similar abundance pattern  But, there are peculiar objects with strong enrichments or deficiencies of light elements such as C, N, O, Na, Mg, Si, etc.  Objects with enhanced carbon are the most common variety

Carbon-Enhanced Metal-Poor (CEMP) Stars  CEMP  Carbon-Enhanced Metal-Poor (CEMP)  Not Carbon-Extremely Metal-Poor  CEMP defined by [Fe/H] +0.7 or [C/Fe] > +1.0 (Beers & Chrislieb 2005)  [C/Fe]  Coin a term Carbonicity similar to Metallicity (e.g., Carollo et al. 2012) 9

Identifying CEMP Stars in SDSS/SEGUE 10  Developed a program to estimate [C/Fe] from the SDSS/SEGUE stellar spectra using CH G band (Lee et al. 2013)

CEMP Stars in SDSS/SEGUE

Thousands of CEMP Stars Identified by SDSS/SEGUE Lee et al. (2013) – CEMP stars from SDSS/SEGUE + Literature Sample 12  The largest list of CEMP stars ever made

13 Frequency of CEMP Stars Frequency of CEMP Stars  Interestingly the fraction of CEMP stars increases as the metallicity decreases  This indicates that a large amount of carbon was produced in the early history of the Milky Way Then, a question arises “how?” Lee et al. (2013)

14 Patterns of n-capture Elements  Another interesting aspect of CEMP stars is that they have different enhancement of n-capture element  High-resolution spectroscopy of CEMP stars revealed different level of enhancement of n- capture elements  s-process (e.g., Ba or Sr) rich  r-process (e.g., Eu) rich  r/s-process (Both Eu and Ba or Sr) rich  No n-capture (neither Eu nor Ba)  Indicative of different astrophysical sites of carbon production at early times

15 Subclasses of CEMP Stars  CEMP Stars are further divided into four more groups depending on the enhancement of the s- process element (Ba) and r-process element (Eu)  CEMP-s and CEMP-n accounts for over 95 % Beers & Christlieb ARAA (2005)

Origin of CEMP Stars  CEMP stars in the Galaxy likely have had multiple mechanisms to produce carbon  Each subclass displays distinct properties  CEMP-s  Variation of radial velocity, indicating a binary system  Mostly discovered for [Fe/H] > -3.0  Account for about 80 % of CEMP stars  Origin: AGB binary mass transfer  Carbon and s-process elements are produced during the AGB 16

Origin of CEMP Stars – Con’t  CEMP-no  Occur preferentially at the lowest metallicities ([Fe/H] < -3.0)  No radial velocity variation, so not in a binary  3 of the 4 stars known with [Fe/H] < -4.5  Origin: A few mechanisms -Rapidly rotating massive ( M sun ) MMP stars -Faint SNe of intermediate mass (25-60 M sun ) with mixing and fallback  CEMP-r, CEMP-r/s  Probably, they are originated with supernova, but not well known, and need more sample 17

18 Difference in [Fe/H] Distribution: CEMP-s vs. CEMP-no CEMP-s CEMP-no  Aoki et al. (2007) demonstrated that the CEMP-no stars occur preferentially at lower [Fe/H] than the CEMP-s stars  About 80% of CEMP stars are CEMP-s, 20% are CEMP-no  Global abundance patterns of CEMP-no stars is incompatible with AGB models at low [Fe/H]

Evidence of Faint SN Model: BD+44:493 – A 9 th Magnitude Star  Ito et al. (2009) report on discovery that BD+44 is an [Fe/H] = -3.8, and carbon-enhanced and low nitrogen  Light-element abundance pattern similar to those for CEMP-no stars  No RV variation at levels > 0.5 km/s over past 25 years  Identified as CEMP-no star  Compared the abundance pattern with faint SN model with 25 M sun 19

Abundance Pattern Compared to 25 M Sun Mixing/Fallback Model Low N, compared with some other CEMP-no stars with enhanced N => indication of the origin of faint SNe. Rapidly rotating scenario needs high [N/Fe] 20

Exploring the Milky Way with CEMP Stars  Constraining the Initial Mass Function (IMF) of the Galactic halo  Dual property of the halo with CEMP stars  Search for the signature of the Nucleosynthesis of First Generation of Stars (Pop III) 21

Constraining the Halo’s IMF  It is possible to derive the mass of progenitors of CEMP-s stars by AGB model  CEMP-s stars are produced in the range of M=1-8 M sun  But, efficiently in M=1-3 M sun  CEMP stars are dominated by CEMP-s (over 80 %)  Constraints on the IMF of the Galactic halo  AGB population synthesis model predicts the CEMP frequency  Compare that with the observation (e.g., Suda et al. 2013; Lee et al. 2014) 22

 Lee et al. (2014) compared AGB population synthesis model prediction with that from SDSS/SEGUE data  The transition of the IMF occurred between [Fe/H] = -2.5 and Transition time Constraining the Halo’s IMF

24 Dichotomy of the Galactic Halo  Structural components of the stellar populations  Bulge / Thin Disk / Thick Disk (MWTD) / Halo  New results from SDSS revealed ( Carollo et al. 2007, 2010) InnerOuter Distance (R) < kpc> kpc Rotational Velocity (V φ ) ~0-50 km/s-40 to -70 km/s Distribution Eccentric orbits (oblate shape) More spherical shape [Fe/H] Origin Dissipative collapse Accreted

25  This dichotomy should appear in other chemical elements as in [Fe/H]  [Mg/Fe] is higher in the inner halo by 0.1 index than the outer halo (Roederer 2009)  What about [C/Fe]?  Carollo et al. (2012) more studies the distribution of [C/Fe] in the inner and outer halo Contract in Other Elements?

Spatial Distribution of [C/Fe] It shows that [C/Fe] continuously changes from low to high as |Z| increases (not expected for single halo) Lee et at. in preparation 26

Global CEMP Fraction vs. |Z| Clear increase of f (CEMP) with |Z| (not expected for single halo) Lee et at. in preparation 27 Carollo et al. 2012

Inner/Outer Halo CEMP Fractions f (CEMP) OH ~ 2 x f (CEMP) IH roughly constant IH/OH 28 Carollo et al. (2012)

What Do These Imply ?  The distribution of CEMP stars indicates that there is likely to be more than one source of C production at low metallicity, and that the difference can be associated with assignment to inner/outer halo  It is speculated that the majority of CEMP stars associated with the inner halo will be CEMP-s, while those associated with the outer halo will be CEMP-no 29

Fraction of CEMP-no and CEMP-s in the Inner/Outer Halo  Sample of 183 stars with high-resolution spectroscopy  Include about 50 CEMP stars  Need more sample, it is underway with SDSS/SEGUE Carollo et al Inner halo Outer halo

Connection with Ultra-faint Dwarf Galaxies  Ultra-faint SDSS dwarf galaxies possess lots of CEMP stars, some of which have low n-capture abundances  Building block of the outer halo? 31  SEGUE1 (Frebel et al. 2014)  Three of the 7 giants that have [Fe/H] < -3.5 are CEMP stars  All three are CEMP-no stars  Concluded that disrupted UF dSph galaxies could account for the CEMP-no stars found in the outer halo of the MW

Search for Nucleosynthesis Production of First Generation Stars (Pop III)  Different initial mass produces different chemical abundance pattern  Chemical abundance pattern leads to the mass of the progenitor  Model versus Observation 32

Expected Chemical Signatures 33 Number per Mass Bin M CEMP-s CEMP-no Enhanced Light elements Carbon-normal Normal light elements Tumlinson 2002

As if Right on Cue …  Nature – March 2014 A single low-energy, iron-poor supernova as the source of metals in the star SMSS J S. C. Keller, M. S. Bessell, A. Frebel, A. R. Casey, M. Asplund, H. R. Jacobson, K. Lind, J. E. Norris, D. Yong, A. Heger, Z. Magic, G. S. Da Costa, B. P. Schmidt, & P. Tisserand 34

Discovery of a Star with [Fe/H] = -7.1  Announcement of the discovery of a star with metallicity [Fe/H] < -7.1  More than 10,000,000 times lower than the Sun  And of course, it is a CEMP-no star, with the same light element abundance pattern as other CEMP-no stars 35

High-resolution Spectrum The high-resolution spectrum from Magellan shows lack of detectable Fe lines ( Keller et al., Nature 2014) Stellar parameters: T eff = 5100 log g = 2.3 [Fe/H] < -7.1 [C/Fe] ~ +4.5 UMP HMP 36

Close-up of Ca II K Region A comparison of the spectrum of SMSS with other UMP ([Fe/H] < -4) and HMP ([Fe/H] < -5) stars in the regions of CaII K. 37

Observed Elemental Abundance Pattern Note singular detections of C, Mg, and Ca (Keller et al. 2014) 38 ● SMSS Faint SN model of 60 M sun star Pollution from a single SN

Evidence for Something Missing  Not all stars with [Fe/H] < -2.5 are carbon-enhanced, in particular for the abundance range -3.5 < [Fe/H] < -2.5  Including at least one star, with [Fe/H] ~ -5.0, and a number with [Fe/H] ~-4.0, without the detection of the chemical signature of CEMP-no stars  Where did they come from? 39

Aoki et al. (Science, Aug. 22, 2014) A chemical signature of first-generation very massive stars W. Aoki, N. Tominaga, T. C. Beers, S. Honda, Y. S. Lee Abstract: Numerical simulations of structure formation in the early universe predict the formation of some fraction of stars with several hundred solar masses. No clear evidence of supernovae from such very massive stars has, however, yet been found in the chemical compositions of Milky Way stars. We report on an analysis of a very metal-poor star SDSS J , which possesses elemental-abundance ratios that differ significantly from any previously known star. This star exhibits low [alpha-element Fe] ratios and large contrasts between the abundances of odd and even element pairs, such as scandium/titanium and cobalt/nickel. Such features have been predicted by nucleosynthesis models for supernovae of stars more than 140 times as massive as the Sun, suggesting that the mass distribution of first-generation stars might extend to 100 solar masses or larger. 40

Abundance Patterns Like No Other  SDSS J  Cool MS star with [Fe/H] = -2.5, NOT carbon-enhanced, and with elemental-abundance ratios unlike any previously studied very low-metallicity star.  Abundance ratios between adjacent odd- and even-element pairs are very low: [Na/Mg] = -0.56, [Sc/Ti] < -0.99, [Co/Ni] =  n-capture elements are quite low compared to other VMP stars: [Sr/Fe] < -1.8, [Ba/Fe] <

42 Comparison to Standard Supernova Model 炭素 ● SDSS J ▲ Comparison star (G39-36, [Fe/H]=-2.1) Core collapse supernova model Does not reproduce C, Na, Mg, Al, Si, and Co abundances

43 Comparison to Other Supernova Models ● SDSS J Pair Instability Supernova (PISN) with M=130 M sun Core-collapse very-massive star model, M=1000 M sun Mass distribution of first-generation stars might extend to 100 M sun or larger

The Path Forward  Expansion of numbers of identified CEMP stars, in particular with [Fe/H] < -2.5, which include both CEMP-s and CEMP-no stars, both from HK/HES, SDSS/SEGUE, APOGEE, and the ~ 8 million medium-res spectra coming from LAMOST  For detailed dual property of the Galactic halo  CEMP Fraction as a function of [Fe/H]  Connection with ultra-faint dwarf galaxies 44

The Path Forward – Con’t  Detailed chemical abundance analysis from high- resolution follow-up of VMP ([Fe/H] < -3.0) stars  Establishment of the frequency of such objects as SDSS J or SMSS J0313, based on high-resolution spectroscopic surveys of the many thousands of stars known with [Fe/H] < -2.5  Refinement of various SN models (PISN or very massive core collapse) with observed abundance pattern  Gemini Korean time is a good opportunity for the high- resolution follow-up of EMP stars  Lots of faint targets in SDSS/SEGUE/BOSS  Mostly too faint (g > 17) for 8-10m class telescope  Really good targets for GMT 45

Expected Signature  Carbon-rich stars (CEMP-no stars)  Enhanced with other light elements (e.g., N, O, Mg, etc.) and lack of over-abundances of neutron-capture elements  Origin  Associated with production by “faint SNe” – progenitors with mass on the order of M sun undergoing mixing and fallback  Or rapidly rotating mega metal-poor (MMP; [Fe/H] < -6.0) stars  Both eject large amount of CNO, but little heavy metals Low-mass stars formed with the help of C, O cooling 46

Expected Signature – Con’t  Carbon-poor (or normal) metal-poor stars  Enhanced with light-element abundances, apparently formed with cooling other than C, O -Perhaps by dust (e.g. silicates)? We don’t know yet  One star found ([Fe/H] ~ -5.0; Caffau et al. 2012)  Origin  Possibly associated with production by first-generation objects of very high mass ( M sun ), which produce large amounts of heavy metals, but little carbon 47

48 Dichotomy of the Galactic Halo  Structural components of the stellar populations  Bulge / Thin Disk / Thick Disk (MWTD) / Halo  New results from SDSS revealed ( Carollo et al. 2007, 2010)  Inner Halo -Dominant at R < kpc, highly eccentric (slightly prograde) orbits, metallicity peak at [Fe/H] = -1.6, likely associated with major/major collision of massive components  Outer Halo -Dominant at R > kpc, uniform distribution of eccentricity, (including highly retrograde) orbits, metallicity peak around [Fe/H] = -2.2, likely associated with accretion from dwarf-like galaxies over an extended period, up to present

Relation to Classical CH or Ba Stars  Classical CH stars (or Ba stars)  Recognized by Bidelman and Keenan  G or K red giants  Overabundance of s-process element such as Ba II  Enhanced features of CH, CN, C2 etc.  Radial velocity variation, indicating a binary system  Result of Asymptotic Giant Branch (AGB) mass transfer  Metal-rich ([Fe/H] = -1.0) counterpart of CEMP-s stars 49

Global CEMP Fraction and vs [Fe/H] Global variation shows smooth increase of f (CEMP) vs. [Fe/H]. Clear increase of vs. [Fe/H] 50 Carollo et al. 2012