Embed Size (px)
Citation preview
MNRAS 445 4241ndash4246 (2014) doi101093mnrasstu1991
The Gaia-ESO Survey the most metal-poor stars in the Galactic bulge
L M Howes1lsaquo M Asplund1 A R Casey2 S C Keller1 D Yong1 G Gilmore2
K Lind23 C Worley2 M S Bessell1 L Casagrande1 A F Marino1 D M Nataf1
C I Owen1 G S Da Costa1 B P Schmidt1 P Tisserand1 S Randich4 S Feltzing5
A Vallenari6 C Allende Prieto78 T Bensby5 E Flaccomio9 A J Korn3
E Pancino1011 A Recio-Blanco12 R Smiljanic13 M Bergemann2 M T Costado14
F Damiani9 U Heiter3 V Hill12 A Hourihane2 P Jofre2 C Lardo10
P de Laverny12 L Magrini4 E Maiorca4 T Masseron2 L Morbidelli4
G G Sacco4 D Minniti151617 and M Zoccali1718
1Research School of Astronomy and Astrophysics Australian National University ACT 2601 Australia2Institute of Astronomy University of Cambridge Madingley Road Cambridge CB3 0HA UK3Department of Physics and Astronomy Division of Astronomy and Space Physics Uppsala University Box 516 SE-751 20 Uppsala Sweden4INAF - Osservatorio Astrofisico di Arcetri Largo E Fermi 5 I-50125 Florence Italy5Lund Observatory Department of Astronomy and Theoretical Physics Box 43 SE-221 00 Lund Sweden6INAF - Padova Observatory Vicolo dellrsquoOsservatorio 5 I-35122 Padova Italy7Instituto de Astrofisica de Canarias E-38205 La Laguna Tenerife Spain8Departamento de Astrofisica Universidad de La Laguna E-38206 La Laguna Tenerife Spain9INAF - Osservatorio Astronomico di Palermo Piazza del Parlamento 1 I-90134 Palermo Italy10INAF - Osservatorio Astronomico di Bologna via Ranzani 1 I-40123 Bologna Italy11ASI Science Data Center Via del Politecnico SNC I-00133 Roma Italy12Laboratoire Lagrange (UMR7293) Universite de Nice Sophia Antipolis CNRS Observatoire de la Cote drsquoAzur CS 34229 F-06304 Nice cedex 4 France13Department for Astrophysics Nicolaus Copernicus Astronomical Center ul Rabianska 8 PL-87-100 Torun Poland14Instituto de Astrofısica de Andalucıa-CSIC Apdo 3004 E-18080 Granada Spain15Departamento de Ciencias Fisicas Universidad Andres Bello Republica 220 Santiago Chile16Vatican Observatory V00120 Vatican City State Italy17Millenium Institute of Astrophysics Av Vicuna Mackenna 4680 Macul Santiago Chile18Instituto de Astrofısica Facultad de Fısica Pontificia Universidad Catolica de Chile Av Vicuna Mackenna Santiago Chile
Accepted 2014 September 23 Received 2014 September 20 in original form 2014 July 14
ABSTRACTWe present the first results of the EMBLA survey (Extremely Metal-poor BuLge stars withAAOmega) aimed at finding metal-poor stars in the Milky Way bulge where the oldeststars should now preferentially reside EMBLA utilizes SkyMapper photometry to pre-selectmetal-poor candidates which are subsequently confirmed using AAOmega spectroscopy Wedescribe the discovery and analysis of four bulge giants with minus272 le [FeH] le minus248the lowest metallicity bulge stars studied with high-resolution spectroscopy to date UsingFLAMESUVES spectra through the Gaia-ESO Survey we have derived abundances of twelveelements Given the uncertainties we find a chemical similarity between these bulge stars andhalo stars of the same metallicity although the abundance scatter may be larger with some ofthe stars showing unusual [αFe] ratios
Key words stars abundances ndash stars Population II ndash Galaxy bulge ndash Galaxy evolution
1 IN T RO D U C T I O N
The first stars in the Universe (referred to as Population III stars)have been extensively searched for both in the local Universe
E-mail louisehowesanueduau
(Frebel amp Norris 2013 and references therein) and at high red-shift (eg Bromm amp Loeb 2006 Cooke Pettini amp Murphy 2012)but despite massive efforts no true Population III star has yet beenfound There is an argument that no such stars should remain to-day models of their formation indicate that they would have beenmassive and short-lived (Nakamura amp Umemura 2001 Abel Bryanamp Norman 2002) Recent simulations however have suggested that
Ccopy 2014 The AuthorsPublished by Oxford University Press on behalf of the Royal Astronomical Society
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4242 L M Howes et al
disc fragmentation could have produced smaller mass stars someof which may have survived to the present day (Clark et al 2011)
Surveys focusing on the discovery of these old and metal-poor stars have almost exclusively targeted the Galactic halo (egChristlieb et al 2008) although some more recent studies havelooked at dwarf galaxies of the Local Group (eg Frebel Kirby ampSimon 2010) The halo is known to be on average more metal-poorthan other Galactic components and some of these halo stars passthrough the solar neighbourhood making them relatively uncom-plicated to observe The number of metal-poor halo stars discoveredhas been growing and there are now chemical abundances for gt400metal-poor stars with [FeH] lt minus25 (Norris et al 2013)
It is not obvious though that the halo is the ideal place to lookfor the first stars Using cold dark matter simulations DiemandMadau amp Moore (2005) predicted that if any were to survive to thepresent day 30ndash60 per cent of them would reside within the inner3 kpc of the Galaxy a population density of first stars that would be1000 times greater than that of the solar neighbourhood Tumlinson(2010) has modelled the current spatial distribution of all stars with[FeH] lt minus35 that formed prior to z = 15 and shown that becauseof the inside-out construction of dark-matter haloes the oldest aswell as the most metal-poor stars should be more frequent in thecentral regions of the Galaxy In other words even if these stars mayhave originated elsewhere they are now most likely to be locatedwithin the central regions of the Galaxy
Few if any dedicated attempts have been made to search theGalactic bulge for extremely metal-poor stars The bulge is knownto be metal rich with a metallicity distribution function (MDF)peaking at [FeH] sim +03 (eg Gonzalez et al 2013 Ness et al2013a) Furthermore the huge number of stars in the bulge the dis-tance to the bulge and the high degree of extinction in the Galacticplane make it practically very difficult to find metal-poor stars thereThe ARGOS survey (Ness et al 2013a) spectroscopically studied28 000 stars in the bulge at R asymp 10 000 identifying only 16 starswith minus28 le [FeH] le minus20 outlining the extent of the problemof finding metal-poor stars in a metallicity unbiased survey of thebulge Similar results were found in the BRAVA survey (Kunderet al 2012) The APOGEE survey found five new metal-poor stars(Garcıa Perez et al 2013) with minus21 le [FeH] le minus16 from 2400observed bulge stars To date no bulge star with [FeH] lt minus21has been exposed to a high-resolution abundance analysis
This Letter is the first in a series of papers exploring the results ofthe EMBLA1 (Extremely Metal-poor BuLge stars with AAOmega)survey which aims to find the most metal-poor stars in the bulgeHere we present the results of our initial observations from whichwe have analysed four bulge stars with [FeH] lt minus2
2 O BSERVATIONS
Our observations are conducted in three stages We first acquireuvgriz photometry from the SkyMapper telescope identifyingmetal-poor candidates that we then confirm spectroscopically withthe AAOmega multi-object spectrograph Finally we analyse themost interesting stars using high-resolution spectra obtained with8 m class telescopes
The SkyMapper telescope is a 13 m telescope capable of imag-ing in six bandpasses with a 57-square-degree field of view (Kelleret al 2007) The filters have been designed to optimize both stellar
1 In Nordic mythology Embla was the first woman born in the middle ofthe world from the remains of giants
Figure 1 Two-colour plot using the g v and i bands of SkyMapper todemonstrate the metallicity dependence on the (v minus g) minus 2(g minus i) colourThe coloured circles are data taken from both EMBLA and the ARGOSsurvey (Ness et al 2013a) with [FeH] determined spectroscopically (fieldat (lb)=(0minus10)) The red trapezium shows our selection criteria for metal-poor candidates The arrow represents the mean reddening vector in thisfield E(B minus V) = 017 (Schlegel Finkbeiner amp Davis 1998)
and extragalactic astronomy in particular the narrow v-band fil-ter centred on the Ca II K line provides a useful stellar metallicityindicator We have obtained SkyMapper photometry taken duringcommissioning for more than 100 deg2 of the bulge with each fieldcontaining of the order of 106 stars ranging from 12th to 18th mag-nitude From (v minus g) minus 2(g minus i) (g minus i) two-colour diagrams we areable to select the most metal-poor candidates (Fig 1) Fig 1 showsour selection lsquoboxrsquo which accounts for the effects of reddening
The AAOmega spectrograph on the AAT provides spectra ofasymp350 stars (plus asymp50 sky and guide fibres) simultaneously over a2 degree diameter field-of-view (Sharp et al 2006) Approximately8500 bulge stars were observed in 2012 and 2013 All of theseobservations were taken using the 1700D grating for the red armand the 580V grating for the blue arm providing a resolving powerof about 10 000 over the 845ndash900 nm and of 1300 over the 370ndash580 nm region The data were reduced using 2DFDR and analysedwith the SICK PIPELINE (Casey 2014b) to measure radial velocitiesand to determine the stellar parameters (Teff log g [FeH] and[αFe])
Fig 2 shows the raw MDF of the EMBLA survey uncorrected forselection biases In comparison to the relatively unbiased ARGOSMDF (Ness et al 2013a) the average metallicity is approximately08 dex lower with a significant tail of stars down into the ex-tremely metal-poor regime More than 300 stars have been foundwith [FeH] lt minus2 The full details of the EMBLA survey will bepresented in future works
From the first 3600 stars observed in 2012 April and July 10 wereimmediately identified as very metal-poor candidates Six of thesetargets were observed with FLAMESUVES on the VLT (Dekkeret al 2000) as part of the Gaia-ESO Survey (Randich amp Gilmore2013) in 2012 May and August The UVES observations have aresolving power of 47 000 using the 580 nm setup The data reduc-tion of the FLAMESUVES data in the survey is described in detailin Sacco et al (2014) Of these six stars the signal-to-noise ratiosof two were too poor to be able to gain any useful analysis from(SN le 8) The other four had average SN per pixel values rangingfrom 14 to 73 sufficient for the derivation of stellar parameters andchemical abundances The spectra of two of the stars are shownin Fig 3 where they are compared to the Gaia-ESO benchmark
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4243
Figure 2 Raw MDF without accounting for selection effects of the first8 611 spectra from the EMBLA survey (red) compared to the MDF of theARGOS bulge survey (blue) Both are normalized to have the same area
Figure 3 Comparison of the FLAMESUVES spectrum of Bulge-1 (black)and Bulge-3 (green) with the Gaia benchmark metal-poor halo giant (Jofreet al 2014) HD 122563 (red) over the Mg triplet wavelength region
metal-poor giant star HD 122563 ([FeH] = minus264 Jofre et al2014) which has similar stellar parameters In addition 10 haloEMP candidates similarly selected from SkyMapper photometryand intermediate resolution spectroscopy were observed throughGaia-ESO
3 ST E L L A R PA R A M E T E R S A N D C H E M I C A LA BU N DA N C E S
The FLAMESUVES spectra of the Gaia-ESO Survey are anal-ysed by 13 different nodes (Smiljanic et al 2014) each usingthe same MARCS model atmospheres (Gustafsson et al 2008) andline-lists (Heiter et al in preparation) but different analysis tech-niques Due to the metal-poor nature of the stars and with mostpipelines being optimized for solar-metallicity stars not all anal-ysis nodes were able to establish robust parameters of the bulge
and halo stars as well as the two metal-poor benchmark stars HD122563 (giant) and HD 140283 (subgiant) However three nodesprovided accurate parameters for the majority of these metal-poorstars the nodes IACAIP and Nice both used global fitting codeswhile ULB used line-by-line analysis (Smiljanic et al 2014) Inaddition we included the results of two further methods first amodified version of the Lumba node pipeline used in the Smiljanicet al analysis that uses the SME code (Valenti amp Piskunov 1996)for primarily H and Fe lines to determine stellar parameters andsecondly a similar analysis to that used in other SkyMapper EMPanalyses (eg Keller et al 2014) using the SMH code (Casey 2014)in 1D LTE but with the Gaia-ESO line lists and atmospheres andeffective temperatures (Teff) measured from H lines instead Thefinal parameters were evaluated by taking weighted averages ofthese five results these can be found in Table 1 The stellar pa-rameters abundances and radial velocities derived here have beenadopted by the Gaia-ESO survey as the recommended values Theuncertainties quoted are the calculated standard errors of the fiveparameter sets
Abundances for 12 elements were derived using the SMH code(Table 2) The uncertainties are formed from standard deviationof the line measurements taken in quadrature with the abundancedifferences due to stellar parameter uncertainties Some elements(Mg Ca Ti and Ni) were measured in all four bulge stars howeversome of the elements could not be detected in the stars with lowerSN while some elements were not detected at the lowest [FeH]Additionally barium abundances were calculated from synthesis ofthe Ba lines rather than from equivalent widths taking into accounthyperfine splitting and isotopic shifts
4 D I SCUSSI ON
41 Bulge membership
All four stars were specifically taken from fields in the outer south-ern part of the bulge (minus10 lt l lt 8 b asymp minus85) Distances havebeen estimated by calculating absolute luminosities based on our de-rived Teff log(g) and assuming Mlowast = 08 M then fitting syntheticmodel fluxes from the Teff log(g) [FeH] and E(B minus V) (Schlegelet al 1998) of each star from which the correction factor used toreconstruct the bolometric luminosities from 2MASS JHKS pho-tometry is derived using the methodology described in CasagrandePortinari amp Flynn (2006) and Casagrande et al (2012) The deriveddistances are given in Table 1 Assuming a distance to the GalacticCentre of 85 kpc and a bulge radius of 3 kpc all but one are locatedinside the bulge When considering the more complex bar structureof the bulge and given the large distance uncertainties Bulge-3 isalso consistent with residing in the bulge This is seen in Fig 4where the locations of all four stars have been projected on to anN-body model taken from Shen et al (2010)
The radial velocities of our stars split them into two groups thosewith velocities similar to that of the bulk of bulge stars and thosewith larger velocities According to Ness et al (2013b) the velocitydispersion of the bulge in the region of our stars is σ = 751 km sminus1Bulge-3 has a galactocentric velocity that would therefore be typicalof a bulge star but Bulge-1 and Bulge-4 have much larger veloci-ties (minus23768 and 21646 km sminus1 respectively) and Bulge-2 lies inbetween the two groups These velocities are more characteristic ofhalo stars and may indicate that although these stars are presentlyin the bulge they are actually halo stars passing through We intendto return to the important issue of kinematics for a much larger
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4244 L M Howes et al
Table 1 Stellar parameters for the 4 bulge stars and 10 halo stars observed as part of the Gaia-ESO Survey in 2012 May and August The SkyMapper namingconvention is SMSS J(RA2000)+(Dec2000)
Star l () b () VGC d SNlowast Teff log(g) [FeH] ξ t
(km sminus1) (kpc) (K) (cgs) (dex) (km sminus1)
SMSS J18215385-3410188 (Bulge-1) 3592 minus 93 minus 23768 70 plusmn 32 73 4947 plusmn 85 141 plusmn 049 minus 260 plusmn 031 23 plusmn 02SMSS J18361733-2700053 (Bulge-2) 71 minus 89 minus 12948 53 plusmn 19 37 4926 plusmn 137 197 plusmn 035 minus 272 plusmn 028 24 plusmn 02SMSS J17551050-4128121 (Bulge-3) 3502 minus 80 minus 4828 124 plusmn 44 14 5187 plusmn 59 223 plusmn 031 minus 257 plusmn 019 20 plusmn 02SMSS J17565243-4136128 (Bulge-4) 3502 minus 84 21646 54 plusmn 28 14 5035 plusmn 196 265 plusmn 054 minus 248 plusmn 023 15 plusmn 02SMSS J09475504-1027247 (Halo-1) 2468 318 30916 21 plusmn 11 112 5258 plusmn 195 296 plusmn 061 minus 280 plusmn 017 18 plusmn 02SMSS J10091577-4127155 (Halo-2) 2731 118 38389 36 plusmn 18 59 5266 plusmn 200 272 plusmn 055 minus 242 plusmn 018 19 plusmn 02SMSS J10142785-4052503 (Halo-3) 2736 129 28467 20 plusmn 09 142 5136 plusmn 123 264 plusmn 050 minus 199 plusmn 020 16 plusmn 02SMSS J10580638-1542390 (Halo-4) 2669 391 51906 80 plusmn 26 54 4907 plusmn 58 202 plusmn 032 minus 239 plusmn 012 23 plusmn 02SMSS J11005336-1328082 (Halo-5) 2660 413 minus 7340 94 plusmn 46 10 5194 plusmn 125 292 plusmn 056 minus 230 plusmn 040 23 plusmn 02SMSS J12555146-4507342 (Halo-6) 3038 177 49602 43 plusmn 31 54 4957 plusmn 215 200 plusmn 100 minus 262 plusmn 019 26 plusmn 02SMSS J13135851-4600123 (Halo-7) 3070 167 28404 85 plusmn 32 109 4744 plusmn 62 137 plusmn 038 minus 234 plusmn 022 25 plusmn 02SMSS J13301360-4346323 (Halo-8) 3103 185 31465 112 plusmn 36 86 4558 plusmn 54 093 plusmn 031 minus 242 plusmn 008 27 plusmn 02SMSS J14214860-4408399 (Halo-9) 3196 158 23293 125 plusmn 48 44 4761 plusmn 65 147 plusmn 040 minus 242 plusmn 018 24 plusmn 02SMSS J14410090-4007413 (Halo-10) 3247 181 minus 13896 69 plusmn 27 99 4921 plusmn 93 161 plusmn 041 minus 245 plusmn 017 22 plusmn 02
Note lowastMedian SN per pixel calculated across total wavelength range
Table 2 Chemical abundances of the four bulge stars
Element Bulge-1 Bulge-2 Bulge-3 Bulge-4
[Na IFe] 065 plusmn 023 015 plusmn 017[Mg IFe] 062 plusmn 019 023 plusmn 016 minus 003 plusmn 010 minus 007 plusmn 020[Si IFe] 050 plusmn 006[Ca IFe] 040 plusmn 007 024 plusmn 006 029 plusmn 018 032 plusmn 009[Sc IIFe] 020 plusmn 016 022 plusmn 015[Ti IIFe] 038 plusmn 016 041 plusmn 014 038 plusmn 012 084 plusmn 027[Cr IFe] minus 020 plusmn 006 minus 027 plusmn 006 minus 024 plusmn 010[Mn IFe] minus 050 plusmn 004[Ni IFe] 002 plusmn 003 019 plusmn 006 011 plusmn 025 044 plusmn 007[Zn IFe] 045 plusmn 006[Y IIFe] minus 043 plusmn 015 034 plusmn 014[Ba IIFe] minus 032 plusmn 013 041 plusmn 012 minus 007 plusmn 015 minus 006 plusmn 026
sample of stars in a future detailed analysis For the time being wecontinue to refer to all four stars as bulge stars given their locationbut recognize that they may well have different origins from thetypical bulge stars
42 Chemical composition
All four program stars are confirmed (based on high-resolutionspectroscopy) to have lower metallicities ([FeH] lt minus24) thanany previously published metal-poor bulge star We compare ourfour bulge stars in Fig 5 to metal-poor halo stars also iden-tified by SkyMapper and observed as part of the Gaia-ESOSurvey as well as other published bulge and halo stars The 10halo stars also observed with the Gaia-ESO Survey have similarmetallicities to the bulge stars and all were analysed in an identicalmanner
Our abundance analysis reveals that all four bulge stars are sig-nificantly α-enhanced (Fig 5) with average abundance ratios of[MgFe] = 019 [SiFe] = 050 (one star) [CaFe] = 034 and[TiFe] = 050 This enhancement is in line with the plateau shownin less metal-poor bulge stars of Garcıa Perez et al (2013) Alves-Brito et al (2010) and Bensby et al (2013) However one obvious
difference in this limited sample of metal-poor bulge stars is theintrinsic scatter in abundance specifically in Mg and Ti Whereasthe [CaFe] ratios are all similar matching the metal-poor halo starsand the more metal-rich bulge stars for Mg and Ti the scatter islarger Two stars have [MgFe] lt 00 and one of those is veryoverabundant in Ti ([TiFe] = 084) The scatter for these elementsis comparable to that for the halo stars analysed here although insome cases the bulge stars may have larger scatter (noticeable inMg for example) Compared to the larger sample of halo stars fromYong et al (2013) given the limited statistics and remaining abun-dance uncertainties our bulge and halo stars appear quite similarin [αFe]
Due to the wavelength region (480ndash680 nm) and SN of our spec-tra it was only possible to measure Y and Ba in two of our bulgestars In these stars both neutron-capture elements are underabun-dant and similar to our sample of halo stars Again the scatter islarger than for the more metal-rich bulge stars although this is ex-pected for neutron capture elements as seen in halo stars (Francoiset al 2007) There appear to be no obvious chemical differencesin either neutron-capture or α abundances between those stars withvery different velocities although we caution that more stars areneeded to confirm this
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4245
Figure 4 The four bulge stars plotted on to Fig 1 of Shen et al (2010) which shows the face-on and side-on views of the bulge constructed from their N-bodymodel of the BRAVA survey data The position of the sun is marked with a cross on the x-axis Underneath the position of the bulge stars as viewed from theSunrsquos location
5 C O N C L U S I O N S
Using the unique photometric capabilities of the SkyMapper tele-scope and the large field-of-view high multiplexing AAOmegaspectrograph on the AAT the EMBLA survey has alreadyfound more than 300 stars spectroscopically confirmed to have[FeH] lt minus20 We have presented an abundance analysis of fourof these observed in high resolution as part of the Gaia-ESO Sur-vey These four are all considerably more metal poor than any
previously studied bulge star (minus272 lt [FeH] lt minus248) andare chemically similar to metal-poor halo stars at similar [FeH]The four stars are the first of many which will be studied athigh resolution by the EMBLA survey using Magellan and VLTwhich will allow us to study the metal-poor tail of the bulgersquosMDF make a detailed comparison with the halo and study theoldest stars in the Universe many of which would have formedat z asymp 15
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4242 L M Howes et al
disc fragmentation could have produced smaller mass stars someof which may have survived to the present day (Clark et al 2011)
Surveys focusing on the discovery of these old and metal-poor stars have almost exclusively targeted the Galactic halo (egChristlieb et al 2008) although some more recent studies havelooked at dwarf galaxies of the Local Group (eg Frebel Kirby ampSimon 2010) The halo is known to be on average more metal-poorthan other Galactic components and some of these halo stars passthrough the solar neighbourhood making them relatively uncom-plicated to observe The number of metal-poor halo stars discoveredhas been growing and there are now chemical abundances for gt400metal-poor stars with [FeH] lt minus25 (Norris et al 2013)
It is not obvious though that the halo is the ideal place to lookfor the first stars Using cold dark matter simulations DiemandMadau amp Moore (2005) predicted that if any were to survive to thepresent day 30ndash60 per cent of them would reside within the inner3 kpc of the Galaxy a population density of first stars that would be1000 times greater than that of the solar neighbourhood Tumlinson(2010) has modelled the current spatial distribution of all stars with[FeH] lt minus35 that formed prior to z = 15 and shown that becauseof the inside-out construction of dark-matter haloes the oldest aswell as the most metal-poor stars should be more frequent in thecentral regions of the Galaxy In other words even if these stars mayhave originated elsewhere they are now most likely to be locatedwithin the central regions of the Galaxy
Few if any dedicated attempts have been made to search theGalactic bulge for extremely metal-poor stars The bulge is knownto be metal rich with a metallicity distribution function (MDF)peaking at [FeH] sim +03 (eg Gonzalez et al 2013 Ness et al2013a) Furthermore the huge number of stars in the bulge the dis-tance to the bulge and the high degree of extinction in the Galacticplane make it practically very difficult to find metal-poor stars thereThe ARGOS survey (Ness et al 2013a) spectroscopically studied28 000 stars in the bulge at R asymp 10 000 identifying only 16 starswith minus28 le [FeH] le minus20 outlining the extent of the problemof finding metal-poor stars in a metallicity unbiased survey of thebulge Similar results were found in the BRAVA survey (Kunderet al 2012) The APOGEE survey found five new metal-poor stars(Garcıa Perez et al 2013) with minus21 le [FeH] le minus16 from 2400observed bulge stars To date no bulge star with [FeH] lt minus21has been exposed to a high-resolution abundance analysis
This Letter is the first in a series of papers exploring the results ofthe EMBLA1 (Extremely Metal-poor BuLge stars with AAOmega)survey which aims to find the most metal-poor stars in the bulgeHere we present the results of our initial observations from whichwe have analysed four bulge stars with [FeH] lt minus2
2 O BSERVATIONS
Our observations are conducted in three stages We first acquireuvgriz photometry from the SkyMapper telescope identifyingmetal-poor candidates that we then confirm spectroscopically withthe AAOmega multi-object spectrograph Finally we analyse themost interesting stars using high-resolution spectra obtained with8 m class telescopes
The SkyMapper telescope is a 13 m telescope capable of imag-ing in six bandpasses with a 57-square-degree field of view (Kelleret al 2007) The filters have been designed to optimize both stellar
1 In Nordic mythology Embla was the first woman born in the middle ofthe world from the remains of giants
Figure 1 Two-colour plot using the g v and i bands of SkyMapper todemonstrate the metallicity dependence on the (v minus g) minus 2(g minus i) colourThe coloured circles are data taken from both EMBLA and the ARGOSsurvey (Ness et al 2013a) with [FeH] determined spectroscopically (fieldat (lb)=(0minus10)) The red trapezium shows our selection criteria for metal-poor candidates The arrow represents the mean reddening vector in thisfield E(B minus V) = 017 (Schlegel Finkbeiner amp Davis 1998)
and extragalactic astronomy in particular the narrow v-band fil-ter centred on the Ca II K line provides a useful stellar metallicityindicator We have obtained SkyMapper photometry taken duringcommissioning for more than 100 deg2 of the bulge with each fieldcontaining of the order of 106 stars ranging from 12th to 18th mag-nitude From (v minus g) minus 2(g minus i) (g minus i) two-colour diagrams we areable to select the most metal-poor candidates (Fig 1) Fig 1 showsour selection lsquoboxrsquo which accounts for the effects of reddening
The AAOmega spectrograph on the AAT provides spectra ofasymp350 stars (plus asymp50 sky and guide fibres) simultaneously over a2 degree diameter field-of-view (Sharp et al 2006) Approximately8500 bulge stars were observed in 2012 and 2013 All of theseobservations were taken using the 1700D grating for the red armand the 580V grating for the blue arm providing a resolving powerof about 10 000 over the 845ndash900 nm and of 1300 over the 370ndash580 nm region The data were reduced using 2DFDR and analysedwith the SICK PIPELINE (Casey 2014b) to measure radial velocitiesand to determine the stellar parameters (Teff log g [FeH] and[αFe])
Fig 2 shows the raw MDF of the EMBLA survey uncorrected forselection biases In comparison to the relatively unbiased ARGOSMDF (Ness et al 2013a) the average metallicity is approximately08 dex lower with a significant tail of stars down into the ex-tremely metal-poor regime More than 300 stars have been foundwith [FeH] lt minus2 The full details of the EMBLA survey will bepresented in future works
From the first 3600 stars observed in 2012 April and July 10 wereimmediately identified as very metal-poor candidates Six of thesetargets were observed with FLAMESUVES on the VLT (Dekkeret al 2000) as part of the Gaia-ESO Survey (Randich amp Gilmore2013) in 2012 May and August The UVES observations have aresolving power of 47 000 using the 580 nm setup The data reduc-tion of the FLAMESUVES data in the survey is described in detailin Sacco et al (2014) Of these six stars the signal-to-noise ratiosof two were too poor to be able to gain any useful analysis from(SN le 8) The other four had average SN per pixel values rangingfrom 14 to 73 sufficient for the derivation of stellar parameters andchemical abundances The spectra of two of the stars are shownin Fig 3 where they are compared to the Gaia-ESO benchmark
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4243
Figure 2 Raw MDF without accounting for selection effects of the first8 611 spectra from the EMBLA survey (red) compared to the MDF of theARGOS bulge survey (blue) Both are normalized to have the same area
Figure 3 Comparison of the FLAMESUVES spectrum of Bulge-1 (black)and Bulge-3 (green) with the Gaia benchmark metal-poor halo giant (Jofreet al 2014) HD 122563 (red) over the Mg triplet wavelength region
metal-poor giant star HD 122563 ([FeH] = minus264 Jofre et al2014) which has similar stellar parameters In addition 10 haloEMP candidates similarly selected from SkyMapper photometryand intermediate resolution spectroscopy were observed throughGaia-ESO
3 ST E L L A R PA R A M E T E R S A N D C H E M I C A LA BU N DA N C E S
The FLAMESUVES spectra of the Gaia-ESO Survey are anal-ysed by 13 different nodes (Smiljanic et al 2014) each usingthe same MARCS model atmospheres (Gustafsson et al 2008) andline-lists (Heiter et al in preparation) but different analysis tech-niques Due to the metal-poor nature of the stars and with mostpipelines being optimized for solar-metallicity stars not all anal-ysis nodes were able to establish robust parameters of the bulge
and halo stars as well as the two metal-poor benchmark stars HD122563 (giant) and HD 140283 (subgiant) However three nodesprovided accurate parameters for the majority of these metal-poorstars the nodes IACAIP and Nice both used global fitting codeswhile ULB used line-by-line analysis (Smiljanic et al 2014) Inaddition we included the results of two further methods first amodified version of the Lumba node pipeline used in the Smiljanicet al analysis that uses the SME code (Valenti amp Piskunov 1996)for primarily H and Fe lines to determine stellar parameters andsecondly a similar analysis to that used in other SkyMapper EMPanalyses (eg Keller et al 2014) using the SMH code (Casey 2014)in 1D LTE but with the Gaia-ESO line lists and atmospheres andeffective temperatures (Teff) measured from H lines instead Thefinal parameters were evaluated by taking weighted averages ofthese five results these can be found in Table 1 The stellar pa-rameters abundances and radial velocities derived here have beenadopted by the Gaia-ESO survey as the recommended values Theuncertainties quoted are the calculated standard errors of the fiveparameter sets
Abundances for 12 elements were derived using the SMH code(Table 2) The uncertainties are formed from standard deviationof the line measurements taken in quadrature with the abundancedifferences due to stellar parameter uncertainties Some elements(Mg Ca Ti and Ni) were measured in all four bulge stars howeversome of the elements could not be detected in the stars with lowerSN while some elements were not detected at the lowest [FeH]Additionally barium abundances were calculated from synthesis ofthe Ba lines rather than from equivalent widths taking into accounthyperfine splitting and isotopic shifts
4 D I SCUSSI ON
41 Bulge membership
All four stars were specifically taken from fields in the outer south-ern part of the bulge (minus10 lt l lt 8 b asymp minus85) Distances havebeen estimated by calculating absolute luminosities based on our de-rived Teff log(g) and assuming Mlowast = 08 M then fitting syntheticmodel fluxes from the Teff log(g) [FeH] and E(B minus V) (Schlegelet al 1998) of each star from which the correction factor used toreconstruct the bolometric luminosities from 2MASS JHKS pho-tometry is derived using the methodology described in CasagrandePortinari amp Flynn (2006) and Casagrande et al (2012) The deriveddistances are given in Table 1 Assuming a distance to the GalacticCentre of 85 kpc and a bulge radius of 3 kpc all but one are locatedinside the bulge When considering the more complex bar structureof the bulge and given the large distance uncertainties Bulge-3 isalso consistent with residing in the bulge This is seen in Fig 4where the locations of all four stars have been projected on to anN-body model taken from Shen et al (2010)
The radial velocities of our stars split them into two groups thosewith velocities similar to that of the bulk of bulge stars and thosewith larger velocities According to Ness et al (2013b) the velocitydispersion of the bulge in the region of our stars is σ = 751 km sminus1Bulge-3 has a galactocentric velocity that would therefore be typicalof a bulge star but Bulge-1 and Bulge-4 have much larger veloci-ties (minus23768 and 21646 km sminus1 respectively) and Bulge-2 lies inbetween the two groups These velocities are more characteristic ofhalo stars and may indicate that although these stars are presentlyin the bulge they are actually halo stars passing through We intendto return to the important issue of kinematics for a much larger
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4244 L M Howes et al
Table 1 Stellar parameters for the 4 bulge stars and 10 halo stars observed as part of the Gaia-ESO Survey in 2012 May and August The SkyMapper namingconvention is SMSS J(RA2000)+(Dec2000)
Star l () b () VGC d SNlowast Teff log(g) [FeH] ξ t
(km sminus1) (kpc) (K) (cgs) (dex) (km sminus1)
SMSS J18215385-3410188 (Bulge-1) 3592 minus 93 minus 23768 70 plusmn 32 73 4947 plusmn 85 141 plusmn 049 minus 260 plusmn 031 23 plusmn 02SMSS J18361733-2700053 (Bulge-2) 71 minus 89 minus 12948 53 plusmn 19 37 4926 plusmn 137 197 plusmn 035 minus 272 plusmn 028 24 plusmn 02SMSS J17551050-4128121 (Bulge-3) 3502 minus 80 minus 4828 124 plusmn 44 14 5187 plusmn 59 223 plusmn 031 minus 257 plusmn 019 20 plusmn 02SMSS J17565243-4136128 (Bulge-4) 3502 minus 84 21646 54 plusmn 28 14 5035 plusmn 196 265 plusmn 054 minus 248 plusmn 023 15 plusmn 02SMSS J09475504-1027247 (Halo-1) 2468 318 30916 21 plusmn 11 112 5258 plusmn 195 296 plusmn 061 minus 280 plusmn 017 18 plusmn 02SMSS J10091577-4127155 (Halo-2) 2731 118 38389 36 plusmn 18 59 5266 plusmn 200 272 plusmn 055 minus 242 plusmn 018 19 plusmn 02SMSS J10142785-4052503 (Halo-3) 2736 129 28467 20 plusmn 09 142 5136 plusmn 123 264 plusmn 050 minus 199 plusmn 020 16 plusmn 02SMSS J10580638-1542390 (Halo-4) 2669 391 51906 80 plusmn 26 54 4907 plusmn 58 202 plusmn 032 minus 239 plusmn 012 23 plusmn 02SMSS J11005336-1328082 (Halo-5) 2660 413 minus 7340 94 plusmn 46 10 5194 plusmn 125 292 plusmn 056 minus 230 plusmn 040 23 plusmn 02SMSS J12555146-4507342 (Halo-6) 3038 177 49602 43 plusmn 31 54 4957 plusmn 215 200 plusmn 100 minus 262 plusmn 019 26 plusmn 02SMSS J13135851-4600123 (Halo-7) 3070 167 28404 85 plusmn 32 109 4744 plusmn 62 137 plusmn 038 minus 234 plusmn 022 25 plusmn 02SMSS J13301360-4346323 (Halo-8) 3103 185 31465 112 plusmn 36 86 4558 plusmn 54 093 plusmn 031 minus 242 plusmn 008 27 plusmn 02SMSS J14214860-4408399 (Halo-9) 3196 158 23293 125 plusmn 48 44 4761 plusmn 65 147 plusmn 040 minus 242 plusmn 018 24 plusmn 02SMSS J14410090-4007413 (Halo-10) 3247 181 minus 13896 69 plusmn 27 99 4921 plusmn 93 161 plusmn 041 minus 245 plusmn 017 22 plusmn 02
Note lowastMedian SN per pixel calculated across total wavelength range
Table 2 Chemical abundances of the four bulge stars
Element Bulge-1 Bulge-2 Bulge-3 Bulge-4
[Na IFe] 065 plusmn 023 015 plusmn 017[Mg IFe] 062 plusmn 019 023 plusmn 016 minus 003 plusmn 010 minus 007 plusmn 020[Si IFe] 050 plusmn 006[Ca IFe] 040 plusmn 007 024 plusmn 006 029 plusmn 018 032 plusmn 009[Sc IIFe] 020 plusmn 016 022 plusmn 015[Ti IIFe] 038 plusmn 016 041 plusmn 014 038 plusmn 012 084 plusmn 027[Cr IFe] minus 020 plusmn 006 minus 027 plusmn 006 minus 024 plusmn 010[Mn IFe] minus 050 plusmn 004[Ni IFe] 002 plusmn 003 019 plusmn 006 011 plusmn 025 044 plusmn 007[Zn IFe] 045 plusmn 006[Y IIFe] minus 043 plusmn 015 034 plusmn 014[Ba IIFe] minus 032 plusmn 013 041 plusmn 012 minus 007 plusmn 015 minus 006 plusmn 026
sample of stars in a future detailed analysis For the time being wecontinue to refer to all four stars as bulge stars given their locationbut recognize that they may well have different origins from thetypical bulge stars
42 Chemical composition
All four program stars are confirmed (based on high-resolutionspectroscopy) to have lower metallicities ([FeH] lt minus24) thanany previously published metal-poor bulge star We compare ourfour bulge stars in Fig 5 to metal-poor halo stars also iden-tified by SkyMapper and observed as part of the Gaia-ESOSurvey as well as other published bulge and halo stars The 10halo stars also observed with the Gaia-ESO Survey have similarmetallicities to the bulge stars and all were analysed in an identicalmanner
Our abundance analysis reveals that all four bulge stars are sig-nificantly α-enhanced (Fig 5) with average abundance ratios of[MgFe] = 019 [SiFe] = 050 (one star) [CaFe] = 034 and[TiFe] = 050 This enhancement is in line with the plateau shownin less metal-poor bulge stars of Garcıa Perez et al (2013) Alves-Brito et al (2010) and Bensby et al (2013) However one obvious
difference in this limited sample of metal-poor bulge stars is theintrinsic scatter in abundance specifically in Mg and Ti Whereasthe [CaFe] ratios are all similar matching the metal-poor halo starsand the more metal-rich bulge stars for Mg and Ti the scatter islarger Two stars have [MgFe] lt 00 and one of those is veryoverabundant in Ti ([TiFe] = 084) The scatter for these elementsis comparable to that for the halo stars analysed here although insome cases the bulge stars may have larger scatter (noticeable inMg for example) Compared to the larger sample of halo stars fromYong et al (2013) given the limited statistics and remaining abun-dance uncertainties our bulge and halo stars appear quite similarin [αFe]
Due to the wavelength region (480ndash680 nm) and SN of our spec-tra it was only possible to measure Y and Ba in two of our bulgestars In these stars both neutron-capture elements are underabun-dant and similar to our sample of halo stars Again the scatter islarger than for the more metal-rich bulge stars although this is ex-pected for neutron capture elements as seen in halo stars (Francoiset al 2007) There appear to be no obvious chemical differencesin either neutron-capture or α abundances between those stars withvery different velocities although we caution that more stars areneeded to confirm this
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4245
Figure 4 The four bulge stars plotted on to Fig 1 of Shen et al (2010) which shows the face-on and side-on views of the bulge constructed from their N-bodymodel of the BRAVA survey data The position of the sun is marked with a cross on the x-axis Underneath the position of the bulge stars as viewed from theSunrsquos location
5 C O N C L U S I O N S
Using the unique photometric capabilities of the SkyMapper tele-scope and the large field-of-view high multiplexing AAOmegaspectrograph on the AAT the EMBLA survey has alreadyfound more than 300 stars spectroscopically confirmed to have[FeH] lt minus20 We have presented an abundance analysis of fourof these observed in high resolution as part of the Gaia-ESO Sur-vey These four are all considerably more metal poor than any
previously studied bulge star (minus272 lt [FeH] lt minus248) andare chemically similar to metal-poor halo stars at similar [FeH]The four stars are the first of many which will be studied athigh resolution by the EMBLA survey using Magellan and VLTwhich will allow us to study the metal-poor tail of the bulgersquosMDF make a detailed comparison with the halo and study theoldest stars in the Universe many of which would have formedat z asymp 15
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4243
Figure 2 Raw MDF without accounting for selection effects of the first8 611 spectra from the EMBLA survey (red) compared to the MDF of theARGOS bulge survey (blue) Both are normalized to have the same area
Figure 3 Comparison of the FLAMESUVES spectrum of Bulge-1 (black)and Bulge-3 (green) with the Gaia benchmark metal-poor halo giant (Jofreet al 2014) HD 122563 (red) over the Mg triplet wavelength region
metal-poor giant star HD 122563 ([FeH] = minus264 Jofre et al2014) which has similar stellar parameters In addition 10 haloEMP candidates similarly selected from SkyMapper photometryand intermediate resolution spectroscopy were observed throughGaia-ESO
3 ST E L L A R PA R A M E T E R S A N D C H E M I C A LA BU N DA N C E S
The FLAMESUVES spectra of the Gaia-ESO Survey are anal-ysed by 13 different nodes (Smiljanic et al 2014) each usingthe same MARCS model atmospheres (Gustafsson et al 2008) andline-lists (Heiter et al in preparation) but different analysis tech-niques Due to the metal-poor nature of the stars and with mostpipelines being optimized for solar-metallicity stars not all anal-ysis nodes were able to establish robust parameters of the bulge
and halo stars as well as the two metal-poor benchmark stars HD122563 (giant) and HD 140283 (subgiant) However three nodesprovided accurate parameters for the majority of these metal-poorstars the nodes IACAIP and Nice both used global fitting codeswhile ULB used line-by-line analysis (Smiljanic et al 2014) Inaddition we included the results of two further methods first amodified version of the Lumba node pipeline used in the Smiljanicet al analysis that uses the SME code (Valenti amp Piskunov 1996)for primarily H and Fe lines to determine stellar parameters andsecondly a similar analysis to that used in other SkyMapper EMPanalyses (eg Keller et al 2014) using the SMH code (Casey 2014)in 1D LTE but with the Gaia-ESO line lists and atmospheres andeffective temperatures (Teff) measured from H lines instead Thefinal parameters were evaluated by taking weighted averages ofthese five results these can be found in Table 1 The stellar pa-rameters abundances and radial velocities derived here have beenadopted by the Gaia-ESO survey as the recommended values Theuncertainties quoted are the calculated standard errors of the fiveparameter sets
Abundances for 12 elements were derived using the SMH code(Table 2) The uncertainties are formed from standard deviationof the line measurements taken in quadrature with the abundancedifferences due to stellar parameter uncertainties Some elements(Mg Ca Ti and Ni) were measured in all four bulge stars howeversome of the elements could not be detected in the stars with lowerSN while some elements were not detected at the lowest [FeH]Additionally barium abundances were calculated from synthesis ofthe Ba lines rather than from equivalent widths taking into accounthyperfine splitting and isotopic shifts
4 D I SCUSSI ON
41 Bulge membership
All four stars were specifically taken from fields in the outer south-ern part of the bulge (minus10 lt l lt 8 b asymp minus85) Distances havebeen estimated by calculating absolute luminosities based on our de-rived Teff log(g) and assuming Mlowast = 08 M then fitting syntheticmodel fluxes from the Teff log(g) [FeH] and E(B minus V) (Schlegelet al 1998) of each star from which the correction factor used toreconstruct the bolometric luminosities from 2MASS JHKS pho-tometry is derived using the methodology described in CasagrandePortinari amp Flynn (2006) and Casagrande et al (2012) The deriveddistances are given in Table 1 Assuming a distance to the GalacticCentre of 85 kpc and a bulge radius of 3 kpc all but one are locatedinside the bulge When considering the more complex bar structureof the bulge and given the large distance uncertainties Bulge-3 isalso consistent with residing in the bulge This is seen in Fig 4where the locations of all four stars have been projected on to anN-body model taken from Shen et al (2010)
The radial velocities of our stars split them into two groups thosewith velocities similar to that of the bulk of bulge stars and thosewith larger velocities According to Ness et al (2013b) the velocitydispersion of the bulge in the region of our stars is σ = 751 km sminus1Bulge-3 has a galactocentric velocity that would therefore be typicalof a bulge star but Bulge-1 and Bulge-4 have much larger veloci-ties (minus23768 and 21646 km sminus1 respectively) and Bulge-2 lies inbetween the two groups These velocities are more characteristic ofhalo stars and may indicate that although these stars are presentlyin the bulge they are actually halo stars passing through We intendto return to the important issue of kinematics for a much larger
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4244 L M Howes et al
Table 1 Stellar parameters for the 4 bulge stars and 10 halo stars observed as part of the Gaia-ESO Survey in 2012 May and August The SkyMapper namingconvention is SMSS J(RA2000)+(Dec2000)
Star l () b () VGC d SNlowast Teff log(g) [FeH] ξ t
(km sminus1) (kpc) (K) (cgs) (dex) (km sminus1)
SMSS J18215385-3410188 (Bulge-1) 3592 minus 93 minus 23768 70 plusmn 32 73 4947 plusmn 85 141 plusmn 049 minus 260 plusmn 031 23 plusmn 02SMSS J18361733-2700053 (Bulge-2) 71 minus 89 minus 12948 53 plusmn 19 37 4926 plusmn 137 197 plusmn 035 minus 272 plusmn 028 24 plusmn 02SMSS J17551050-4128121 (Bulge-3) 3502 minus 80 minus 4828 124 plusmn 44 14 5187 plusmn 59 223 plusmn 031 minus 257 plusmn 019 20 plusmn 02SMSS J17565243-4136128 (Bulge-4) 3502 minus 84 21646 54 plusmn 28 14 5035 plusmn 196 265 plusmn 054 minus 248 plusmn 023 15 plusmn 02SMSS J09475504-1027247 (Halo-1) 2468 318 30916 21 plusmn 11 112 5258 plusmn 195 296 plusmn 061 minus 280 plusmn 017 18 plusmn 02SMSS J10091577-4127155 (Halo-2) 2731 118 38389 36 plusmn 18 59 5266 plusmn 200 272 plusmn 055 minus 242 plusmn 018 19 plusmn 02SMSS J10142785-4052503 (Halo-3) 2736 129 28467 20 plusmn 09 142 5136 plusmn 123 264 plusmn 050 minus 199 plusmn 020 16 plusmn 02SMSS J10580638-1542390 (Halo-4) 2669 391 51906 80 plusmn 26 54 4907 plusmn 58 202 plusmn 032 minus 239 plusmn 012 23 plusmn 02SMSS J11005336-1328082 (Halo-5) 2660 413 minus 7340 94 plusmn 46 10 5194 plusmn 125 292 plusmn 056 minus 230 plusmn 040 23 plusmn 02SMSS J12555146-4507342 (Halo-6) 3038 177 49602 43 plusmn 31 54 4957 plusmn 215 200 plusmn 100 minus 262 plusmn 019 26 plusmn 02SMSS J13135851-4600123 (Halo-7) 3070 167 28404 85 plusmn 32 109 4744 plusmn 62 137 plusmn 038 minus 234 plusmn 022 25 plusmn 02SMSS J13301360-4346323 (Halo-8) 3103 185 31465 112 plusmn 36 86 4558 plusmn 54 093 plusmn 031 minus 242 plusmn 008 27 plusmn 02SMSS J14214860-4408399 (Halo-9) 3196 158 23293 125 plusmn 48 44 4761 plusmn 65 147 plusmn 040 minus 242 plusmn 018 24 plusmn 02SMSS J14410090-4007413 (Halo-10) 3247 181 minus 13896 69 plusmn 27 99 4921 plusmn 93 161 plusmn 041 minus 245 plusmn 017 22 plusmn 02
Note lowastMedian SN per pixel calculated across total wavelength range
Table 2 Chemical abundances of the four bulge stars
Element Bulge-1 Bulge-2 Bulge-3 Bulge-4
[Na IFe] 065 plusmn 023 015 plusmn 017[Mg IFe] 062 plusmn 019 023 plusmn 016 minus 003 plusmn 010 minus 007 plusmn 020[Si IFe] 050 plusmn 006[Ca IFe] 040 plusmn 007 024 plusmn 006 029 plusmn 018 032 plusmn 009[Sc IIFe] 020 plusmn 016 022 plusmn 015[Ti IIFe] 038 plusmn 016 041 plusmn 014 038 plusmn 012 084 plusmn 027[Cr IFe] minus 020 plusmn 006 minus 027 plusmn 006 minus 024 plusmn 010[Mn IFe] minus 050 plusmn 004[Ni IFe] 002 plusmn 003 019 plusmn 006 011 plusmn 025 044 plusmn 007[Zn IFe] 045 plusmn 006[Y IIFe] minus 043 plusmn 015 034 plusmn 014[Ba IIFe] minus 032 plusmn 013 041 plusmn 012 minus 007 plusmn 015 minus 006 plusmn 026
sample of stars in a future detailed analysis For the time being wecontinue to refer to all four stars as bulge stars given their locationbut recognize that they may well have different origins from thetypical bulge stars
42 Chemical composition
All four program stars are confirmed (based on high-resolutionspectroscopy) to have lower metallicities ([FeH] lt minus24) thanany previously published metal-poor bulge star We compare ourfour bulge stars in Fig 5 to metal-poor halo stars also iden-tified by SkyMapper and observed as part of the Gaia-ESOSurvey as well as other published bulge and halo stars The 10halo stars also observed with the Gaia-ESO Survey have similarmetallicities to the bulge stars and all were analysed in an identicalmanner
Our abundance analysis reveals that all four bulge stars are sig-nificantly α-enhanced (Fig 5) with average abundance ratios of[MgFe] = 019 [SiFe] = 050 (one star) [CaFe] = 034 and[TiFe] = 050 This enhancement is in line with the plateau shownin less metal-poor bulge stars of Garcıa Perez et al (2013) Alves-Brito et al (2010) and Bensby et al (2013) However one obvious
difference in this limited sample of metal-poor bulge stars is theintrinsic scatter in abundance specifically in Mg and Ti Whereasthe [CaFe] ratios are all similar matching the metal-poor halo starsand the more metal-rich bulge stars for Mg and Ti the scatter islarger Two stars have [MgFe] lt 00 and one of those is veryoverabundant in Ti ([TiFe] = 084) The scatter for these elementsis comparable to that for the halo stars analysed here although insome cases the bulge stars may have larger scatter (noticeable inMg for example) Compared to the larger sample of halo stars fromYong et al (2013) given the limited statistics and remaining abun-dance uncertainties our bulge and halo stars appear quite similarin [αFe]
Due to the wavelength region (480ndash680 nm) and SN of our spec-tra it was only possible to measure Y and Ba in two of our bulgestars In these stars both neutron-capture elements are underabun-dant and similar to our sample of halo stars Again the scatter islarger than for the more metal-rich bulge stars although this is ex-pected for neutron capture elements as seen in halo stars (Francoiset al 2007) There appear to be no obvious chemical differencesin either neutron-capture or α abundances between those stars withvery different velocities although we caution that more stars areneeded to confirm this
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4245
Figure 4 The four bulge stars plotted on to Fig 1 of Shen et al (2010) which shows the face-on and side-on views of the bulge constructed from their N-bodymodel of the BRAVA survey data The position of the sun is marked with a cross on the x-axis Underneath the position of the bulge stars as viewed from theSunrsquos location
5 C O N C L U S I O N S
Using the unique photometric capabilities of the SkyMapper tele-scope and the large field-of-view high multiplexing AAOmegaspectrograph on the AAT the EMBLA survey has alreadyfound more than 300 stars spectroscopically confirmed to have[FeH] lt minus20 We have presented an abundance analysis of fourof these observed in high resolution as part of the Gaia-ESO Sur-vey These four are all considerably more metal poor than any
previously studied bulge star (minus272 lt [FeH] lt minus248) andare chemically similar to metal-poor halo stars at similar [FeH]The four stars are the first of many which will be studied athigh resolution by the EMBLA survey using Magellan and VLTwhich will allow us to study the metal-poor tail of the bulgersquosMDF make a detailed comparison with the halo and study theoldest stars in the Universe many of which would have formedat z asymp 15
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4244 L M Howes et al
Table 1 Stellar parameters for the 4 bulge stars and 10 halo stars observed as part of the Gaia-ESO Survey in 2012 May and August The SkyMapper namingconvention is SMSS J(RA2000)+(Dec2000)
Star l () b () VGC d SNlowast Teff log(g) [FeH] ξ t
(km sminus1) (kpc) (K) (cgs) (dex) (km sminus1)
SMSS J18215385-3410188 (Bulge-1) 3592 minus 93 minus 23768 70 plusmn 32 73 4947 plusmn 85 141 plusmn 049 minus 260 plusmn 031 23 plusmn 02SMSS J18361733-2700053 (Bulge-2) 71 minus 89 minus 12948 53 plusmn 19 37 4926 plusmn 137 197 plusmn 035 minus 272 plusmn 028 24 plusmn 02SMSS J17551050-4128121 (Bulge-3) 3502 minus 80 minus 4828 124 plusmn 44 14 5187 plusmn 59 223 plusmn 031 minus 257 plusmn 019 20 plusmn 02SMSS J17565243-4136128 (Bulge-4) 3502 minus 84 21646 54 plusmn 28 14 5035 plusmn 196 265 plusmn 054 minus 248 plusmn 023 15 plusmn 02SMSS J09475504-1027247 (Halo-1) 2468 318 30916 21 plusmn 11 112 5258 plusmn 195 296 plusmn 061 minus 280 plusmn 017 18 plusmn 02SMSS J10091577-4127155 (Halo-2) 2731 118 38389 36 plusmn 18 59 5266 plusmn 200 272 plusmn 055 minus 242 plusmn 018 19 plusmn 02SMSS J10142785-4052503 (Halo-3) 2736 129 28467 20 plusmn 09 142 5136 plusmn 123 264 plusmn 050 minus 199 plusmn 020 16 plusmn 02SMSS J10580638-1542390 (Halo-4) 2669 391 51906 80 plusmn 26 54 4907 plusmn 58 202 plusmn 032 minus 239 plusmn 012 23 plusmn 02SMSS J11005336-1328082 (Halo-5) 2660 413 minus 7340 94 plusmn 46 10 5194 plusmn 125 292 plusmn 056 minus 230 plusmn 040 23 plusmn 02SMSS J12555146-4507342 (Halo-6) 3038 177 49602 43 plusmn 31 54 4957 plusmn 215 200 plusmn 100 minus 262 plusmn 019 26 plusmn 02SMSS J13135851-4600123 (Halo-7) 3070 167 28404 85 plusmn 32 109 4744 plusmn 62 137 plusmn 038 minus 234 plusmn 022 25 plusmn 02SMSS J13301360-4346323 (Halo-8) 3103 185 31465 112 plusmn 36 86 4558 plusmn 54 093 plusmn 031 minus 242 plusmn 008 27 plusmn 02SMSS J14214860-4408399 (Halo-9) 3196 158 23293 125 plusmn 48 44 4761 plusmn 65 147 plusmn 040 minus 242 plusmn 018 24 plusmn 02SMSS J14410090-4007413 (Halo-10) 3247 181 minus 13896 69 plusmn 27 99 4921 plusmn 93 161 plusmn 041 minus 245 plusmn 017 22 plusmn 02
Note lowastMedian SN per pixel calculated across total wavelength range
Table 2 Chemical abundances of the four bulge stars
Element Bulge-1 Bulge-2 Bulge-3 Bulge-4
[Na IFe] 065 plusmn 023 015 plusmn 017[Mg IFe] 062 plusmn 019 023 plusmn 016 minus 003 plusmn 010 minus 007 plusmn 020[Si IFe] 050 plusmn 006[Ca IFe] 040 plusmn 007 024 plusmn 006 029 plusmn 018 032 plusmn 009[Sc IIFe] 020 plusmn 016 022 plusmn 015[Ti IIFe] 038 plusmn 016 041 plusmn 014 038 plusmn 012 084 plusmn 027[Cr IFe] minus 020 plusmn 006 minus 027 plusmn 006 minus 024 plusmn 010[Mn IFe] minus 050 plusmn 004[Ni IFe] 002 plusmn 003 019 plusmn 006 011 plusmn 025 044 plusmn 007[Zn IFe] 045 plusmn 006[Y IIFe] minus 043 plusmn 015 034 plusmn 014[Ba IIFe] minus 032 plusmn 013 041 plusmn 012 minus 007 plusmn 015 minus 006 plusmn 026
sample of stars in a future detailed analysis For the time being wecontinue to refer to all four stars as bulge stars given their locationbut recognize that they may well have different origins from thetypical bulge stars
42 Chemical composition
All four program stars are confirmed (based on high-resolutionspectroscopy) to have lower metallicities ([FeH] lt minus24) thanany previously published metal-poor bulge star We compare ourfour bulge stars in Fig 5 to metal-poor halo stars also iden-tified by SkyMapper and observed as part of the Gaia-ESOSurvey as well as other published bulge and halo stars The 10halo stars also observed with the Gaia-ESO Survey have similarmetallicities to the bulge stars and all were analysed in an identicalmanner
Our abundance analysis reveals that all four bulge stars are sig-nificantly α-enhanced (Fig 5) with average abundance ratios of[MgFe] = 019 [SiFe] = 050 (one star) [CaFe] = 034 and[TiFe] = 050 This enhancement is in line with the plateau shownin less metal-poor bulge stars of Garcıa Perez et al (2013) Alves-Brito et al (2010) and Bensby et al (2013) However one obvious
difference in this limited sample of metal-poor bulge stars is theintrinsic scatter in abundance specifically in Mg and Ti Whereasthe [CaFe] ratios are all similar matching the metal-poor halo starsand the more metal-rich bulge stars for Mg and Ti the scatter islarger Two stars have [MgFe] lt 00 and one of those is veryoverabundant in Ti ([TiFe] = 084) The scatter for these elementsis comparable to that for the halo stars analysed here although insome cases the bulge stars may have larger scatter (noticeable inMg for example) Compared to the larger sample of halo stars fromYong et al (2013) given the limited statistics and remaining abun-dance uncertainties our bulge and halo stars appear quite similarin [αFe]
Due to the wavelength region (480ndash680 nm) and SN of our spec-tra it was only possible to measure Y and Ba in two of our bulgestars In these stars both neutron-capture elements are underabun-dant and similar to our sample of halo stars Again the scatter islarger than for the more metal-rich bulge stars although this is ex-pected for neutron capture elements as seen in halo stars (Francoiset al 2007) There appear to be no obvious chemical differencesin either neutron-capture or α abundances between those stars withvery different velocities although we caution that more stars areneeded to confirm this
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4245
Figure 4 The four bulge stars plotted on to Fig 1 of Shen et al (2010) which shows the face-on and side-on views of the bulge constructed from their N-bodymodel of the BRAVA survey data The position of the sun is marked with a cross on the x-axis Underneath the position of the bulge stars as viewed from theSunrsquos location
5 C O N C L U S I O N S
Using the unique photometric capabilities of the SkyMapper tele-scope and the large field-of-view high multiplexing AAOmegaspectrograph on the AAT the EMBLA survey has alreadyfound more than 300 stars spectroscopically confirmed to have[FeH] lt minus20 We have presented an abundance analysis of fourof these observed in high resolution as part of the Gaia-ESO Sur-vey These four are all considerably more metal poor than any
previously studied bulge star (minus272 lt [FeH] lt minus248) andare chemically similar to metal-poor halo stars at similar [FeH]The four stars are the first of many which will be studied athigh resolution by the EMBLA survey using Magellan and VLTwhich will allow us to study the metal-poor tail of the bulgersquosMDF make a detailed comparison with the halo and study theoldest stars in the Universe many of which would have formedat z asymp 15
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
Most metal-poor stars in the Galactic bulge 4245
Figure 4 The four bulge stars plotted on to Fig 1 of Shen et al (2010) which shows the face-on and side-on views of the bulge constructed from their N-bodymodel of the BRAVA survey data The position of the sun is marked with a cross on the x-axis Underneath the position of the bulge stars as viewed from theSunrsquos location
5 C O N C L U S I O N S
Using the unique photometric capabilities of the SkyMapper tele-scope and the large field-of-view high multiplexing AAOmegaspectrograph on the AAT the EMBLA survey has alreadyfound more than 300 stars spectroscopically confirmed to have[FeH] lt minus20 We have presented an abundance analysis of fourof these observed in high resolution as part of the Gaia-ESO Sur-vey These four are all considerably more metal poor than any
previously studied bulge star (minus272 lt [FeH] lt minus248) andare chemically similar to metal-poor halo stars at similar [FeH]The four stars are the first of many which will be studied athigh resolution by the EMBLA survey using Magellan and VLTwhich will allow us to study the metal-poor tail of the bulgersquosMDF make a detailed comparison with the halo and study theoldest stars in the Universe many of which would have formedat z asymp 15
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
4246 L M Howes et al
Figure 5 Comparison of the abundances in our bulge stars (blue circles)and the metal-poor halo stars from Gaia-ESO (red circles) Also shownare bulge stars from Bensby et al (2013 light green triangles) Alves-Britoet al (2010 dark green triangles) APOGEE (Garcıa Perez et al 2013 purplesquares) as well as halo stars from Yong et al (2013 black dots)
AC K N OW L E D G E M E N T S
We greatly acknowledge generous funding from the AustralianResearch Council (grants FL110100012 and DP120101237) TBwas funded by grant no 621-2009- 3911 from The Swedish Re-search Council Support from the Swedish National Space Boardis acknowledged Based on data products from observations madewith ESO Telescopes at the La Silla Paranal Observatory underprogramme ID 188B-3002 This work was partly supported by theEuropean Union FP7 programme through ERC grant number320360 and by the Leverhulme Trust through grant RPG-2012-541 We acknowledge the support from INAF and MinisterodellrsquoIstruzione dellrsquoUniversita e della Ricerca (MIUR) in the formof the grant lsquoPremiale VLT 2012rsquo The results presented here ben-efit from discussions held during the Gaia-ESO workshops andconferences supported by the ESF (European Science Foundation)through the GREAT Research Network Programme MZ acknowl-edges support from Fondecyt Regular 1110393 the BASAL CATA
PFB-06 and by the Chilean Ministry for the Economy Develop-ment and Tourismrsquos Programa Iniciativa Cientıfica Milenio throughgrant IC12009 awarded to the Millennium Institute of Astro-physics
R E F E R E N C E S
Abel T Bryan G Norman M 2002 Science 295 93Alves-Brito A Melendez J Asplund M Ramırez I Yong D 2010 AampA
513 35Bensby T et al 2013 AampA 549 147Bromm V Loeb A 2006 ApJ 642 382Casagrande L Portinari L Flynn C 2006 MNRAS 373 13Casagrande L Ramırez I Melendez J Asplund M 2012 ApJ 761 16Casey A R 2014 preprint (arXiv14055968)Casey A R 2014b ApJS submittedChristlieb N Schorck T Frebel A Beers T C Wisotzki L Reimers D
2008 AampA 484 721Clark P C Glover S C O Smith R J Greif T H Klessen R S Bromm
V 2011 Science 331 1040Cooke R Pettini M Murphy M T 2012 MNRAS 425 347Dekker H DrsquoOdorico S Kaufer A Delabre B Kotzlowski H 2000 Proc
SPIE 4008 534Diemand J Madau P Moore B 2005 MNRAS 364 367Francois P et al 2007 AampA 476 935Frebel A Norris J E 2013 Planets Stars and Stellar Systems Vol 5
Springer-Verlag Berlin p 55Frebel A Kirby E N Simon J D 2010 Nature 464 72Garcıa Perez A E et al 2013 ApJ 767 L9Gonzalez O A Rejkuba M Zoccali M Valent E Minniti D Tobar R
2013 AampA 552 110Gustafsson B Edvardsson B Eriksson K Joslashrgensen U G Nordlund Aring
Plez B 2008 AampA 486 951Jofre P et al 2014 AampA 564 133Keller S C et al 2007 PASA 24 1Keller S C et al 2014 Nature 506 463Kunder A et al 2012 AJ 143 57Nakamura F Umemura M 2001 ApJ 548 19Ness M et al 2013a MNRAS 430 836Ness M et al 2013b MNRAS 432 2092Norris J E et al 2013 ApJ 762 25Randich S Gilmore G 2013 The Messenger 154 47Sacco G G et al 2014 AampA 565 113Schlegel D Finkbeiner D Davis M 1998 ApJ 500 525Sharp R et al 2006 Proc SPIE 6269 14Shen J Rich R M Kormendy J Howard C D de Propris R Kunder A
2010 ApJ 720 L72Smiljanic R et al 2014 preprint (arXiv14090568)Tumlinson J 2010 ApJ 708 1398Valenti J A Piskunov N 1996 AampAS 118 595Yong D et al 2013 ApJ 762 26
This paper has been typeset from a TEXLATEX file prepared by the author
MNRAS 445 4241ndash4246 (2014)
at The A
ustralian National U
niversity on March 24 2015
httpmnrasoxfordjournalsorg
Dow
nloaded from
![Signatures of Peculiar Supernova Nucleosynthesis in ...Galactic metal-poor stars with [Fe/H] < -1 Old stars that were formed 10 Gyrs ago Mostly enriched by core-collase SNe (CCSNe)](https://img.pdfslide.net/doc/110x75/60b73154628b076b26556f41/signatures-of-peculiar-supernova-nucleosynthesis-in-galactic-metal-poor-stars.jpg)
![Bechtol and Xiyan Peng arXiv:1711.07469v1 [astro-ph.GA] · PDF fileRR Lyrae stars are pulsating variable stars found in old, metal-poor populations that can be used as stan-dard candles](https://img.pdfslide.net/doc/110x75/5abb65177f8b9af27d8cb3b7/bechtol-and-xiyan-peng-arxiv171107469v1-astro-phga-lyrae-stars-are-pulsating.jpg)
![Astronomy c ESO 2013 Astrophysics - Open Research: Home · 2013. 5. 24. · metal-poor stars. Still, at [Fe=H] < 2, the predicted level of 6 Li=7](https://img.pdfslide.net/doc/110x75/60f690bce7d2b321a6695a8e/astronomy-c-eso-2013-astrophysics-open-research-home-2013-5-24-metal-poor.jpg)
