Documentos de Académico
Documentos de Profesional
Documentos de Cultura
CHEMICAL ABUNDANCES IN
THE UNIVERSE:
CONNECTING FIRST STARS
TO PLANETS
Edited by: KATIA CUNHA,
MONIQUE SPITE and BEATRIZ BARBUY
Advisers
IAN F. CORBETT, IAU General Secretary,
European Southern Observatory, Germany
U. GROTHKOPF, European Southern Observatory, Germany
CHRISTIANN STERKEN, University of Brussels, Pleinlaan 2, 1050 Brussels, Belgium
Members
IAUS260
DAVID VALLS-GABAUD, GEPI Observatoire de Paris, 5 Place Jules Janssen, 92195 Meudon, France
IAUS261
S. A. KLIONER, Lohrmann Observatory, Dresden Technical University, Mommsenstr. 13, 01062 Dresden, Germany
IAUS262
G. BRUZUAL A., CIDA, Apartado Postal 264, 5101-A Merida, Venezuela
IAUS263
J. A. FERNANDEZ, Departamento de Astronoma, Facultad de Ciencias, Igua 4225, 11400 Montevideo, Uruguay
IAUS264
A. KOSOVICHEV, Stanford University, 452 Lomita Mall, Stanford, CA 94305-4085, USA
IAUS265
K. CUNHA, NOAO, 950 N. Cherry Avenue, Tucson, AZ 85719, USA
IAUS266
R. DE GRIJS, Kavli Institute for Astronomy and Astrophysics, Peking University,
Yi He Yuan Lu 5, Hai Dian District, Beijing 100871, China
and Department of Physics & Astronomy, The University of Sheffield, Sheffield S3 7RH, UK
IAUS267
B. PETERSON, Department of Astronomy, 140 West 18th Ave, Ohio State University, Columbus, OH 43219, USA
IAUS268
C. CHARBONNEL, Geneva Observatory, 51, chemin des Maillettes, 1290 Versoix, Switzerland
IAU
CHEMICAL ABUNDANCES IN
THE UNIVERSE:
CONNECTING FIRST STARS
TO PLANETS
PROCEEDINGS OF THE 265th SYMPOSIUM OF THE
INTERNATIONAL ASTRONOMICAL UNION
HELD IN RIO DE JANEIRO, BRAZIL
AUGUST 1014, 2009
Edited by
KATIA CUNHA
NOAO, USA & ON/MCT, BRAZIL
MONIQUE SPITE
OBSERVATOIRE DE PARIS-MEUDON, FRANCE
and
BEATRIZ BARBUY
PAULO, SAO
PAULO, BRAZIL
UNIVERSIDADE DE SAO
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xiv
Organizing committee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xvi
Conference participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xvii
Plenary Session
Chair: Beatriz Barbuy
Nucleosynthesis now and then. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Stanford E. Woosley, A. Heger, L. Roberts, and R. D. Hoffman
15
23
27
34
42
46
54
57
Poster Papers
Enrichment of thorium (Th) and lead (Pb) in the early galaxy. . . . . . . . . . . . . . .
Wako Aoki and Satoshi Honda
61
63
The importance of initial conditions and metallicity for the fragmentation of protogalactic gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Anne-Katharina Jappsen, Simon C. O. Glover, Mordecai-Mark Mac Low,
and Ralf S. Klessen
65
vi
Contents
Silver stars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Camilla Juul Hansen and Francesca Primas
67
Mass and angular momentum loss of first stars via decretion disks. . . . . . . . . . . .
Jir Krticka; Stanley P. Owocki, and Georges Meynet
69
71
73
75
77
81
90
94
98
106
111
117
Poster Papers
s/r ratios in carbon-enhanced metal-poor stars . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dinah M. Allen, Sean G. Ryan, Silvia Rossi, and Stelios A. Tsangarides
118
120
Contents
vii
122
124
126
EMP stars with high mass IMF and hierarchical galaxy formation. . . . . . . . . . . .
Yutaka Komiya, Takuma Suda, Asao Habe, and Masayuki Y. Fujimoto
128
130
132
134
139
147
155
159
163
Possibility of measuring the amount of intergalactic metals with 14-N VII HFS line
Dmitrijs Docenko and Rashid A. Sunyaev
167
171
viii
Contents
Poster Papers
Metallicity of the high-redshift Universe traced by radio galaxies . . . . . . . . . . . . .
K. Matsuoka, T. Nagao, R. Maiolino, A. Marconi, and Y. Taniguchi
179
181
183
187
197
201
205
Poster Papers
Can we trust elemental abundances derived in late-type giants with the classical
1D stellar atmosphere models?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A. Kucinskas, V. Dobrovolskas, A. Ivanauskas, H.-G. Ludwig, E. Caffau, K.
Blazevicius, J. Klevas, and D. Prakapavicius
Problems in abundance determination from UV spectra of hot supergiants . . . . .
M. Sarta Dekovic, D. Kotnik-Karuza, T. Jurkic, and D. Dominis Prester
The determination of the abundances of the Fe group elements in early B stars
from high resolution FUV Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geraldine J. Peters, Saul J. Adelman, Ivan Hubeny, and Thierry Lanz
Accurate fundamental stellar parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hans Bruntt
209
211
213
215
219
227
Contents
ix
233
237
Poster Papers
Feh-Duf: very high-velocity low-metallicity star with peculiar chemical abundance
Natalia A. Drake and Claudio B. Pereira
241
Ver
onica Firpo, Guillermo Bosch, Guillermo H
agele, Angeles
I. Daz, and
Nidia Morrell
243
Chemical evolution models for local group dwarf spheroidal galaxies: the evolution
of Fe-peak elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gustavo A. Lanfranchi, Francesca Matteucci and Gabriele Cescutti
245
247
Spitzer finds cosmic neons and sulfurs sweet spot: part III, NGC 6822. . . . . . . .
R. H. Rubin, I. A. McNabb, J. P. Simpson, R. J. Dufour, A. W. A.
Pauldrach, S. W. J. Colgan, T. W. Craven, E. D. Gitterman, and C. C. Lo
249
The effect of the corotation on the radial gradient of metallicity of spiral galaxies
Sergio Scarano Jr. and Jacques R. D. Lepine
251
255
263
Structure and kinematics of the stellar halos and thick disks of the Milky Way
based on calibration stars from SDSS DR7 . . . . . . . . . . . . . . . . . . . . . . . . . . .
D. Carollo, T. C. Beers, M. Chiba, J. E. Norris, K. C. Freeman, and Y. S.
Lee
267
271
279
285
289
The galactic thin and thick disks in the context of galaxy formation . . . . . . . . . .
Thomas Bensby and Sofia Feltzing
300
Contents
304
Planetary nebulae and star formation history in the galactic disk and bulge . . . .
Yulia Milanova and Alexander Kholtygin
313
317
325
336
Poster Papers
Chemical similarities between the galactic bulge and local thick disk red giant
stars: analysis from optical data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Alan Alves-Brito, Jorge Melendez, and Martin Asplund
342
344
346
348
350
352
354
356
358
On the physical existence of the Zeta Reticuli moving group: a chemical composition analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Letcia D. Ferreira, Gustavo F. Porto de Mello, and Licio da Silva
Chemical analysis of B stars within 9 - 11 kpc from the galactic center . . . . . . . .
Maria Isela Zevallos Herencia and Simone Daflon
360
362
Contents
Chemical fingerprinting and chemical analysis of galactic halo substructure . . . .
Steven R. Majewski, Mei-Yin Chou, Katia Cunha, Verne V. Smith, Richard
J. Patterson, and David Martnez-Delgado
Uncovering the evolutionary sequences for the C-J stars based on their chemical
abundances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ana Beatriz de Mello and Silvia Lorenz-Martins
xi
364
366
368
370
Nucleosynthesis in the Hyades open cluster: evidence for the enhanced depletion
of 12 C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simon C. Schuler, Jeremy R. King, and Lih-Sin The
Lithium abundances in southern associations containing young stars . . . . . . . . . .
Licio da Silva, Carlos Alberto Torres, Ramiro de la Reza, Germano Quast,
Claudio de Melo, and Michael Sterzik
372
374
376
Investigation of the chemical structure of our galaxy using radial pulsating stars
as tracers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Marian Doru Suran
378
Evolution of [O/Mg], [Na/Mg], [Al/Mg], and [K/Mg] in the Galaxy, from a NLTE
analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
M. Spite, F. Spite, P. Bonifacio, V. Hill, S. Andrievsky, R. Cayrel, P.
Francois, and S. Korotin
FEROS abundance analysis of 21 bulgelike SMR stars. . . . . . . . . . . . . . . . . . . . . .
Marina Trevisan, Beatriz Barbuy, M. Grenon, B. Gustafsson and L.
Pompeia
Metal-rich infall onto the inner disk through the interaction between bulge winds
and gaseous halos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Takuji Tsujimoto and Kenji Bekki
Superbubble H ii regions: how self-enriched should they be? . . . . . . . . . . . . . . . . .
Aida Wofford
380
382
384
386
391
399
xii
Contents
403
408
412
416
Poster Papers
Planetary populations according to the orbital angular momentum . . . . . . . . . . .
Jo
ao A. S. Amarante and Helio J. Rocha-Pinto
Lithium abundance as a boundary condition for age and mass determination of
solar twin stars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
M. Castro, J.-D. do Nascimento Jr., J. S. da Costa, J. Melendez, M. Bazot,
S. Theado, G. F. Porto de Mello, and J. R. De Medeiros
Distribution of refractory and volatile elements in COROT planet host stars . . .
C. Chavero, R. de la Reza, R. C. Domingos, N. A. Drake, C. B. Pereira,
and O. C. Winter
Photospheric parameters and C abundances in solar-like stars with and without
planets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ronaldo Da Silva and Andre Milone
Irradiation effects in CO and CO2 ices induced by swift heavy Ni ions at 46 MeV
and 537 MeV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A. Domaracka, E. Seperuelo Duarte, P. Boduch, H. Rothard, E. Balanzat,
E. Dartois, S. Pilling, L.S. Farenzena, and E. F. da Silveira
420
422
424
426
428
430
432
434
Photospheric and coronal abundances of solar-type stars with planets: the case of
Bootis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Antonio Maggio, Jorge Sanz-Forcada, and Luigi Scelsi
Evolution of the abundance of biomolecules in the interstellar medium at the gas
phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
436
438
Contents
xiii
440
442
447
453
How galaxies form: mass assembly from chemical abundances in the era of large
surveys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rosemary Wyse (Invited Review)
461
470
476
Poster Papers
The Apache Point Observatory Galactic Evolution Experiment (APOGEE) in
Sloan Digital Sky Survey III (SDSS-III) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Steven R. Majewski, John C. Wilson, Fred Hearty, Ricardo R. Schiavon,
and Michael F. Skrutskie
Author index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
480
482
xiv
Preface
Chemical Abundances in the Universe, theme of IAU Symposium 265, is a broad, diverse and rapidly evolving field, due to the ever-expanding capabilities delivered by new
arrays of instruments on large telescopes, together with significant advances in modelling
and increasing access to large and accurate nuclear, atomic, and molecular databases. In
recent years, connections between seemingly different areas in astrophysics have become
clearer. For example, the first generation of extremely massive stars that drove reionization left their nucleosynthetic yields chemically imprinted in the oldest low-mass stars,
which still exist today. IAU Symposium 265 aimed to provide a unified picture of the
production of chemical elements over cosmic time, and how this chemical evolution links
together the early Universe of the first stars, through the formation of galaxies and their
diverse populations of stars, to a Universe of heavy-element rich stars and planets.
The sessions held during the symposium followed this evolution and provided a framework within which a modern picture of the chemical structure of the Universe could be
built. The program started with the big bang nucleosynthesis, sailed through the first
stars and the reionization era; moved to the discussion of chemical abundances in the
high redshift Universe and then closer to home focused on the chemical abundance constraints on mass assembly and star formation in Local Group Galaxies and in particular
the Milky Way. The chemical abundance connection was then further explored via the
link with the formation of planetary systems.
Even as the observations continue to flow in, and improving analysis tools are used
to extract accurate chemical abundances, new classes of extremely large telescopes and
instruments are currently being planned. The closing session of Symposium 265 framed
the next sets of questions to be posed and highlighted some of the current large programs
and future surveys needed to tackle these questions using the next generation of very
large telescopes and their instruments, which are already being designed.
IAU Symposium 265 took place during the second week of the IAU General Assembly in
the beautiful city of Rio de Janeiro from 10-14 August 2009. A General Assembly, which
naturally attracts astronomers from all fields in the astronomical community, was the
perfect venue for a symposium that included a variety of topics with chemical abundances
being the common link among them. The symposium was very well attended. According
to the database collected by the IAU, Symposium 265 was the one with the largest number
of participants, as measured from the intention of participation at the time of registration
for the General Assembly. The list of participants in these proceedings includes all those
who registered for the symposium during the IAU General Assembly. However, it should
be noted that since there were many parallel events happening at one time there was a
floating audience who attended only parts of the different meetings but the core audience
of IAUS265 was quite large.
Due to its broad subject matter the symposium was supported by a number of IAU
commissions and two working groups. The SOC membership had a perfect gender balance, and involved most continents, highlighting the importance of international collaborations and interconnections that will be increasingly important to present and future
projects in Astronomy.
xv
Last, but not least, we would like to thank the organizers of the IAU in Rio for the
local support; the IAU for supporting this symposium and for providing grants; all the
participants and in particular the speakers for delivering great talks. Also thanks to
all the contributors to these proceedings and to the sponsors of the meeting: the agencies Conselho Nacional de Pesquisas CNPq, under Ministry of Science and Technology,
CAPES under Ministry of Education, FAPESP and FAPERJ, and the institutions Observat
orio Nacional, Rio de Janeiro; National Optical Astronomy Observatory, Observatoire
de Paris and Instituto de Astronomia, Geofsica e Ciencias Atmosfericas da Universidade
de S
ao Paulo.
Katia Cunha, Monique Spite and Beatriz Barbuy, co-chairs SOC
Tucson, Paris, and S
ao Paulo, December 2009
xvi
Scientific
Katia Cunha (USA & Brazil, chair)
Beatriz Barbuy (Brazil, co-chair)
Timothy Beers (USA)
Cristina Chiappini (Switzerland)
Bengt Gustafsson (Sweden)
Dante Minniti (Chile)
Max Pettini (UK)
Stan Woosley (USA)
Acknowledgements
The symposium was sponsored and supported by the IAU Divisions IV (Stars),
VII (Galactic System), VIII (Galaxies and the Universe); and by the IAU Commissions No. 28 (Galaxies), No. 29 (Stellar Spectra), No. 35 (Stellar Constitution),
No. 36 (Theory of Stellar Atmospheres), No. 37 (Star Clusters and Associations),
No. 45 (Stellar Classification), No. 53 (Extrasolar Planets).
The Local Organizing Committee acknowledges
funding by the
International Astronomical Union,
Ministerio de Ciencia e Tecnologia,
Conselho Nacional de Pesquisas (CNPq),
Fundacao de Amparo a` Pesquisa do Estado de Sao Paulo (FAPESP),
Fundacao de Amparo a` Pesquisa do Estado do Rio de Janeiro (FAPERJ),
Coordenacao de Aperfeicoamento de Pessoal de Nvel Superior (CAPES)
xvii
Participants
Ummi Abbas, INAF-Osservatorio Astronomico de Torino, Italy
abbas@oato.inaf.it
Roberto Abraham, University of Toronto, Canada
abraham@astro.utoronto.ca
France Allard, Centre de Recherche Astrophysique de Lyon, France
fallard@ens-lyon.fr
Dinah Allen, IAG-USP, Brazil
dimallen@astro.iag.usp.br
Carlos Allende Prieto, Moscow State University, UK
callende@astro.as.utexas.edu
Athem Alsabti, University College London, Iraq
a.alsabti@ucl.ac.uk
Alan Alves-Brito, Swineburne University, Australia
abrito@astro.swin.edu.au
Joao Antonio Amarante, UFRJ, Brazil
amarante@astro.ufrj.br
Peter Anders, University Utrecht, Netherlands
p.anders@uu.nl
Michael Andersen, Dark Cosmology Centre, Denmark
mia@dark-cosmology.dk
Johannes Andersen, The Niels Bohr Institute, Denmark
ja@astro.ku.dk
Diana Andrade-Pilling, PUC Rio de Janeiro, Brazil
dianaufrj@gmail.com
Wako Aoki, National Astronomical Observatory of Japan, Japan
aoki.wako@nao.ac.jp
Immo Appenzeller, University of Heidelberg, Germany
iappenze@lsw.uni-heidelberg.de
Ronald Armando, UFRJ, Brazil
ronald09@astro.ufrj.br
David Arnett, University of Arizona, USA
darnett@as.arizona.edu
Martin Asplund, Max Planck Institute for Astrophysics, Germany
asplund@mpa-garching.mpg.de
Marcelo Assafin, Observat
orio de Valongo - UFRJ, Brazil
massaf@astro.ufrj.br
Bruno L. Astorina, UFRJ, Brazil
b.lastorina@gmail.com
Victor Avila, Observat
orio Nacional, Brazil
victor@on.br
Aycin Aykutalp, Kapteyn Astronomical Institute, Netherlands
aykutalp@astro.rug.nl
Willem Baan, ASTRON, Netherlands
baan@astron.nl
Octavian Badescu, Astronomical Institute of Romanian Academy, Romania
octavian@aira.astro.ro
Alexandre Bagdonas, Universidade de S
ao Paulo, Brazil
alebagdonas@gmail.com
Pascal Ballester, European Southern Observatory, Germany
pballest@eso.org
Andrzej Baran, Cracow Pedagogical University, Poland
sfbaran@cyf-kr.edu.pl
Carmen Adriana Martinez Barbosa, Universidad Nacional, Colombia
anamabo3@gmail.com
Beatriz Barbuy, Universidade de S
ao Paulo, Brazil
barbuy@astro.iag.usp.br
Guillermo Barro Calvo, Universidad Complutense de Madrid, Spain
gbc@astrax.fis.ucm.es
Stanislava Bartasiute, Vilnius University, Lithuania
stanislava.bartasiute@ff.vu.lt
Sarbani Basu, Yale University, USA
sarbani.basu@yale.edu
Manuel Bautista, Virginia Tech, USA
bautista@vt.edu
Pedro Paulo Bonetti Beaklini, IAG-USP, Brazil
beaklini@astro.iag.usp.br
Tim Bedding, University of Sydney, Australia
bedding@physics.usyd.edu.au
Timothy Beers, Michigan State University, USA
beers@pa.msu.edu
Natalie Behara, Observatoire de Paris-Meudon, France
Natalie.Behara@obspm.fr
Peter Beiersdorfer, Lawrence Livermore National Laboratory, USA
beiersdorfer@llnl.gov
Mohamed Hedi Ben Ismail, Tunis Science City, Tunisia
b.ismail@cst.rnu.tn
Thomas Bensby, European Southern Observatory, Chile
tbensby@eso.org
Maria Bergemann, Max Planck Institute for Astrophysics, Germany
mbergema@mpa-garching.mpg.de
Jacqueline Bergeron, Institut dAstrophysique de Paris, France
bergeron@iap.fr
Elise Servajean Bergoeing, Universidad de Chile, Chile
eliseservajean@gmail.com
Peter Bernath, University of York, UK
pfb500@york.ac.uk
Melina Bersten, Universidad de Chile, Chile
melina@das.uchile.cl
Isadora Chaves Bicalho Domingos, UNB, Brazil
isadora.bicalho@gmail.com
Emile Biemont, University of Li
ege, Belgium
e.biemont@ulg.ac.be
Michael Blanton, New York University, USA
michael.blanton@gmail.com
Heloisa Maria Boechat-Roberty, Universidade Federal do Rio de Janeiro, Brazil
heloisa@astro.ufrj.br
Wilfried Boland, NOVA, Netherlands
boland@strw.leidenuniv.nl
Jade Bond, University of Arizona, USA
jbond@lpl.arizona.edu
Piercarlo Bonifacio, Observatoire de Paris-Meudon, France
piercarlo.bonifacio@obspm.fr
Vinicius Bordalo, Observat
orio Nacional/MCT, Brazil
vschmidt@on.br
Alan Boss, Department of Terrestrial Magnetism, Carnegie Institution of Washington, USA
boss@dtm.ciw.edu
Rychard Bouwens, University of California - Santa Cruz, USA
bouwens@ucolick.org
Richard P. Boyle, Vatican Observatory, USA
rpboyle@mac.com
Gustavo Bragan
ca, Observat
orio Nacional, Brazil
braganca@on.br
Andre Brahic, Universit
e de Paris, France
brahic@cea.fr
Fabio Bresolin, University of Hawaii, USA
bresolin@ifa.hawaii.edu
Jarle Brinchmann, Leiden Observatory, Netherlands
jarle@strw.leidenuniv.nl
Volker Bromm, University of Texas, USA
vbromm@astro.as.utexas.edu
Gustavo Bruzual, CIDA, Venezuela
bruzual@cida.ve
Jan Budaj, Astronomical Institute Slovak Academy of Sciences, Slovakia
budaj@ta3.sk
Martin Bureau, University of Oxford, UK
bureau@astro.ox.ca.uk
Giorgia Busso, Leiden Observatory, Netherlands
busso@strw.leidenuniv.nl
Claudio Caceres, European Southern Observatory, Chile
ccaceres@eso.org
Elisabetta Caffau, Observatoire de Paris-Meudon, France
elisabetta.caffau@obspm.fr
Nelson Callegari Jr, UNESP, Brazil
calleg@rc.unesp.br
Luis E. Campusano, Universidad de Chile, Chile
luis@das.uchile.cl
Matteo Cantiello, Institute for Astronomy Utrecht, Netherlands
m.cantiello@uu.nl
Leticia Carigi, Universidad Nacional Autonoma, Mexico
carigi@astroscu.unam.mx
Joleen Carlberg, University of Virginia, USA
jkm9n@virginia.edu
Daniela Carollo,RSAA-Mount Stromlo, Australia
carollo@mso.anu.edu.au
Luis Carrasco, INAOE, Mexico
carrasco@inaoep.mx
Jorge Marcio Ferreira Carvano, Observat
orio Nacional, Brazil
carvano@on.br
Luca Casagrande, Max Planck Institute for Astrophysics, Germany
luca@mpa-garching.mpg.de
Gabriela Castelleti, Institute for Astronomy and Space Physics (IAFE), Argentina
gcastell@iafe.uba.ar
Bruno Vaz Castilho, Laborat
orio Nacional de Astrofsica / MCT, Brazil
bruno@lna.br
Matthieu Castro, Universidade Federal do Rio Grande do Norte, Brazil
mcastro@dfte.ufrn.br
Denise Castro, Observat
orio Nacional, Brazil
denise@on.br
Oscar Cavichia, Universidade de S
ao Paulo, Brazil
cavichia@astro.iag.usp.br
Diego Cesarsky, CEA, France
dac.viroflay@free.fr
Catherine Cesarsky, CEA, France
catherine.cesarsky@cea.fr
Thyagarajan Chandrasekhar, Physical Research Laboratory (PRL), India
chandra@prl.res.in
William Chaplin, University of Birmingham, UK
w.j.chaplin@bham.ac.uk
Corinne Charbonnel, Geneva Observatory & CNRS, Switzerland
corinne.charbonnel@unige.ch
Carolina Chavero, Observat
orio Nacional, Brazil
carolina@on.br
Fuzhen Cheng, Centre for Astrophysics, China Nanjing
fzhen@ustc.edu.cn
Alexandre Cherman, Rio de Janeiro Planetarium, Brazil
acherman@rio.rj.gov.br
Ana Leonor Chies-Santos, Utrecht University, Netherlands
a.l.chies@uu.nl
Paul Clark, Zentrum fuer Theo. Astrophysik der Universitaet Heidelberg, Germany
pcc@ita.uni-heidelberg.de
Alain Coc, CSNSM, France
alain.coc@csnsm.in2p3.fr
xviii
Participants
Participants
xix
xx
Participants
Participants
xxi
xxii
Participants
Plenary Session
S. E. Woosley et al.
-19
-18
-17
-16
-15
2000
4000
6000
8000
Photospheric Velocity (at day 50)
10000
-19
-18
-17
-16
-15
2000
Observed SNe
E = 0.3 B
E = 0.6 B
E = 1.2 B
E = 2.4 B
E = 4.8 B
4000
6000
8000
Photospheric Velocity (at day 50)
10000
Figure 1. Constraining the kinetic energy of Type IIp supernovae. This figure, taken from
Kasen & Woosley (2009), shows the correlation between the V-band plateau luminosity of a
set of supernova models measured at 50 days and the expansion speed of the photosphere (left
frame). In the right frame, the model results are compared with the observational data of Hamuy
(2003). The models include five supernovae with masses from 12 to 25 solar masses (color coded
in the web version) and two different metallicities for the 15 solar mass model (the square is
for 15 M with 0,1 solar metallicity). The higher mass models have larger radii and are the
brightest in each cluster of points. The model presupernovae were exploded with five different
choices of kinetic energy at infinity, 0.3, 0.6, 1.2, 2.4 and 4.8 1051 erg. Model data points tend
to group by energy with the most energetic points on the right and least on the left. Data is
most consistent with an explosion energy of 0.6 1.2 1051 erg, though there are examples of
explosions with higher and lower energies.
well fit by a range of explosion energies from 0.3 to 5 1051 erg, with most points falling
in the range 0.6 to 1.2 1051 erg. The scatter due to main sequence mass and metallicity
does not affect the result much.
The other major parameter in the explosion model is the mass cut. After the shock
goes through, what stays behind and what is ejected? This parameter is also constrained
by observations, namely the mass of the gravitationally bound remnants left by the
explosions. These are usually neutron stars for solar metallicity progenitors. Observations
show that typical neutron star masses are around 1.35 M (Thorsett & Chakrabarty,
1999). When corrected for neutrino losses, that mass limits the average baryonic mass
inside the mass cut. Similarly, the constraint that very neutron-rich elements in the iron
core not be frequently ejected limits the minimum mass inside the mass cut (Weaver
et al. 1978). The mass cut must be outside the iron core. Finally, all groups average
over some distribution of stellar masses, typically using a Salpeter initial mass function
(IMF), and that tends to smooth out some of the variability at a given mass. Of course
there is some residual uncertainty in the synthesis of many individual isotopes because
of unknown reaction rates, sensitivity to conditions at the inner boundary, contributions
from sources other than massive stars, etc., but, by and large, elemental abundances
below krypton are well understood.
An important exception, however, is the r-process above A = 90 and the p-process
between mass A = 90 and 130 (including the intriguing species 92 Mo). Fig 8 of Woosley
& Heger, in particular, which shows the isotopic yields from a generation of massive stars
with solar metallicity and a Salpeter initial mass function, illustrates this inadequacy. For
the r-process, the general mechanism is generally agreed upon - the explosive expansion
and cooling of nucleonic matter with a neutron excess (Hoyle & Fowler 1960). The favored
site for years has been the neutrino-powered wind of a young (1 - 10 s) proto-neutron
star (Woosley et al. 1994). Neutrinos drive mass loss composed of neutrons and protons
with a ratio set by the fluxes and spectra of electron neutrinos and antineutrinos. Early
on, the wind may actually be proton-rich and could contribute to the p-process (Pruet
et al. 2006; Fr
ohlich et al. 2006). Later, the flux of antineutrinos is larger and the wind
becomes neutron-rich, possibly producing the r-process.
Unfortunately, all realistic calculations since 1994 have failed to give a high enough
entropy in the wind to make more than the lightest r-process isotopes, and it is now
thought that the entropy in the study by Jim Wilson that was used then may have been
unphysically high (Sam Dalhed, private communication). The entropy matters because
decreasing density when the temperature is between 2 and 5 109 K, i.e., raising the
entropy, results in fewer -particles reassembling into heavy elements. If only a few
reassemble, there are a lot of neutrons (or protons for the p-process) for each heavy seed,
and a robust flow to heavy elements ensues. For the same reason, a short expansion
time scale, i.e., a rapid freeze out from nuclear statistical equilibrium, also helps. If, on
the other hand, most of the -particles reassemble, the neutrons are incorporated into
neutron-rich nuclei in the mass range 60 to 90 and the r-process does not even make it
Figure 2. Nucleosynthesis in a 15 solar mass supernova including the contribution of the neutrino wind. The upper panel shows production factors, Xi /Xi, , for just the wind. The production factors are large but the amount of mass involved is small. Synthesis below A = 60 and
above A = 105 is negligible. This figure, taken from Roberts et al. (2009), integrates over the
complete ten second history of the wind as the neutrino luminosities and spectra evolve. The
bottom panel shows this nucleosynthesis combined with that from the rest of the supernova,
properly normalized. Standard values for the neutrino physics give a yield that is not grossly
out of line with nucleosynthesis in the rest of the supernova. However, the model completely
fails to produce the r-process except for the closed shell isotopes of 88 Sr, 89 Y, and 90 Zr. It also
produces the light p-process elements 78 Kr and 84 Sr.
S. E. Woosley et al.
to the first peak. There is a potentially more serious problem in that the wind might not
only not make the r-process, but could grossly overproduce a few tightly bound nuclei
with a closed neutron shell at N = 50, 88 Sr, 89 Y, and 90 Zr (Hoffman et al. 1996).
The current situation is illustrated in Fig. 2. This calculation takes the neutrino fluxes
and spectrum from Woosley et al (1994), but uses a modified version of the Kepler
code (Weaver et al 1978) to compute the hydrodynamics. Post-Newtonian corrections to
gravity are included and the wind typically has a final entropy S/NA k 120 as opposed
to about 400 in the Wilson calculation. Modern neutrino cross sections are included and
the composition of the wind is calculated using a large reaction network. Of particular
import are weak magnetism corrections to the neutrino-nucleon interactions (Horowitz
2002). The wind is included in the explosion of a 15 M star (1.2 1051 erg) and the
total nucleosynthesis determined, i.e., the wind is integrated over its entire duration and
its contribution is plotted with the nucleosynthesis from the rest of the star.
Fig 2 shows the results when weak magnetism is included. The effect of this correction
is to make the wind more proton-rich at earlier times when there is larger mass loss and
to diminish the large overproduction of 88 Sr, 89 Y, and 90 Zr, which is otherwise seen. The
good news is that the wind doesnt overproduce anything. It successfully makes the light
p-process isotopes 78 Kr and 84 Sr along with 88 Sr, 89 Y, and 90 Zr. The bad news is that
it makes nothing heavier. A caveat is necessary here, however. The fluxes and spectra of
neutrinos were calculated in a model that ignored weak magnetism and hence the results
are internally inconsistent. In a revised model the neutron excess may be a bit larger,
but it is still doubtful that any appreciable r-process will result.
So there is a problem making the observed r-process. There are three possible solutions.
First is that the neutrino spectra may be greatly revised in a fortuitous way so as to
produce a much larger neutron excess. Flavor mixing might help do this, but the difference
in the average energies of -, -, and e-neutrinos is not as great as it used to be. Of course
if the electron-neutrinos magically disappeared and the anti-neutrinos did not (so that
e + p n + e+ greatly dominated over e + n p + e ), the problem would be
solved. Or, there may be other ways of increasing the entropy in the wind to high values
or decreasing the dynamical timescale. Magnetic fields and rotation are one possibility
(Qian & Woosley 1996; Metzger et al 2007) yet to be demonstrated. Vibrations from
the newly formed neutron star are another (Qian & Woosley 1996; Burrows et al 2007;
Otsuki et al 2008).
The third possibility is that one gives up on supernovae as the dominant production
site of the heaviest r-process isotopes. The leading alternate contender is merging neutron
stars, where the large neutron excesses and rapid expansion time scales occur naturally.
The historical problem with this site is not that it cant make the r-process, but that
it makes to much of it, and too infrequently (Argast et al. 2004). It may be worth
revisiting this issue though, with a more modern view of binary neutron star formation
and evolution, including the effects of kicks, multiple pulsar velocity distributions, etc.
The early history of the Galaxy may also have been more complex than in the Argast et
al. mode, with mergers, a time-variable IMF, etc.
However it is resolved, the fact is that the production of the r-process has been a
problem for a long time. It would be good to see some real progress.
2. Metal-Poor Stars
The stars that were born with far less than their solar complement of heavy elements
may have had a different history from their modern counterparts. Even if the initial mass
function was the same, their mass loss history was probably different. Line-driven and
Figure 3. Model abundances vs. the metal poor data set of Cayrel et al. (2004). The model grid
consists of 120 zero metallicity stars in the mass range 12 to 100 M . Each star was exploded
with an energy of 1.21051 erg and moderately mixed. Model data was averaged over a Salpeter
IMF. Since these stars included no s-process or neutrino wind contribution, Sc and Zn may be
underproduced. This figure is taken from Heger & Woosley (2009).
grain-driven mass loss, in particular, would have been less, and consequently stars born
with the same mass as today would have died with higher masses (Eldridge & Vink 2006,
Meynet & Maeder 2005). These larger mass stars would have been more tightly bound
and harder to explode (Woosley et al. 2002). They may also have been more compact
and prone to fall back and diminished mixing (Church et al 2009). Together, these facts
suggest a more massive population of presupernova stars, but at the same time, less
efficient ejection of their heavy elements.
The situation is further complicated by other forms of mass loss and by primary nitrogen production. Low metallicity stars may still lose their mass by rotational instabilities
(Meynet, Ekstr
om, & Maeder 2006). Rotation also enhances the possibility that the outer
part of the convective core will mix with the hydrogen shell during core helium burning
(Ekstr
om et al. 2008). When this happens large amounts of nitrogen are produced at
the base of the envelope and a super-charged hydrogen burning shell causes a previously
zero-metallicity blue supergiant to expand and become a red supergiant. Carbon dredge
up increases the grain opacity and the possibility of mass loss. Also, mass loss as a luminous blue variable may lead to envelope removal in very massive stars, just as it is doing
today in Eta Carina.
Fortunately, the nucleosynthesis, except of nitrogen, is much more sensitive to what
happens in the helium core than the hydrogen envelope (so long as the helium core
explodes - see next section). It is thus worth exploring whether the same sorts of models
that give good agreement with the major solar abundances might also work well at low
metallicity.
Fig. 3 shows the model yields from Heger & Woosley (2009) compared with the observations of metal-deficient stars (Fe/H] -3) by Cayrel et al. (2004). Considering that only
S. E. Woosley et al.
model stars with zero metallicity were included in the fit, hence ignoring the s-process
and the neutrino wind, the agreement is quite good. The overproduction of Cr probably
reflects a difficulty with atomic physics and ionization stage (Sobeck et al. 2007, Lai et
al. 2008). The observed abundance of Na is prone to large non-LTE corrections. In any
case, it would be difficult to make less Na than in zero metal stars with no s-process and
little CNO cycling. Potentially more problematic are the underproductions of Co and Zn.
In the present calculation, both are due to the -rich freeze out in the innermost zones
that are ejected. Variations in the location of the mass cut and the way the explosion
is simulated (thermal energy vs. piston) could change their production appreciably. Zinc
might also be partly produced in the neutrino wind (Hoffman et al. 1996, 1998; Wanajo
et al. 2009). All in all, the fit is as good as the previous one was to solar abundances and
additional efforts to remedy a two-sigma difference do not appear warranted.
Indeed, better agreement is obtained with the observational data set of Lai et al. (2008)
who studied 28 metal poor stars with [Fe/H] between -2 and -4 (13 were < -2.6). In that
paper the same stellar model set from Heger & Woosley (2009) was employed, but mixing
and explosion energy were allowed to float so as to obtain the best fit. The best fit was
obtained with an explosion energy of only 6 1050 (M/20)0.5 erg and a low value of
mixing. This low energy its decline with mass tends to emphasize the contribution of
low mass supernovae below 30 solar masses. The rest made black holes that swallowed
much of their nucleosynthesis. The good fit included Sc, Cr, Co, and Zn, but mysteriously
overproduced Cu. This is still not understood, but it would be very difficult to produce
the Co, Ni, and Zn abundances seen in the observations without making a comparable
production of Cu. All four are made in the -rich freeze out.
An additional product of stellar evolution at low metallicity is the compact remnants
they leave. Even if low metal stars of high mass lose their envelopes because of rotation,
LBV outbursts, and red giant winds, the Wolf-Rayet stars they leave behind are likely
to have both a larger typical mass and a lower mass loss rate than for solar metallicity
stars. Wolf-Rayet mass loss depends sensitively upon the iron abundance down to low
values of [Fe/H] (Vink 2005). It is thus likely that the first stars produced more black
holes and black holes of a higher average mass than the stars dying today. Zhang et
al. (2008) predict an average black hole mass of 8.5 7.6 M produced about 50% of
the time. This is to be compared with 3.8 1.0 M produced 9% of the time for solar
metallicity stars. In both cases, the numbers assume a maximum neutron star mass of
2.0 M . The maximum black hole mass depends on the explosion energy but is unlikely
to be larger than the biggest helium core that does not experience the pulsational pair
instability, about 40 M .
Since these simple considerations suggest an increasing frequency of black hole formation with decreasing metallicity, the comparative paucity of black hole candidates in the
SMC (Liu et al. 2005) is a mystery.
from a binary companion (e.g., Komiya et al., 2007). However, so far, these two stars
show no evidence for binary membership or for the s-process.
Several explanations have been offered (Tumlinson 2007). The earliest was that these
stars represent the nucleosynthetic products of one or a few supernovae of large mass
that experienced little mixing and a large amount of fallback (Umeda & Nomoto, 2003;
Iwamoto et al., 2005). More recent calculations by Heger & Woosley (2009) support this
conclusion, but multi-dimensional calculations of the mixing by Joggerst et al. (2009) suggest that quantitative agreement may be difficult. The large ratio of O/Mg, in particular,
is hard to achieve.
It has also been suggested that the ultra-iron-poor stars, like the CEMP stars, may
have acquired their light element enhancements only at their surfaces from a companion
that passed through an AGB phase (Suda et al., 2004), but the lack of evidence for binary
membership or s-process enhancement is troubling. On the other hand, Tumlinson (2007)
has emphasized that the depletion of lithium seen in HE1327-2326, favors the binary
mass-transfer hypothesis.
Meynet et al.(2006) have proposed that the large CNO abundances reflect the mass
loss from a massive population of stars (around 60 M ) that lost their envelopes through
rotational mass shedding and then collapsed to black holes to avoid a large production of
heavier elements. Frebel et al.(2008) finds, however, that the elements other than nitrogen
are produced in inadequate quantities in this model. In short, there doesnt seem to be
any single explanation that satisfies everyone.
We are struck by the fact that no stars with extremely low CNO (commensurate with
Fe in the ultra-iron-poor stars) have been discovered. The CEMP stars are a minor
fraction of all metal-poor stars. It is thus surprising that the first two UMP (ultra-metalpoor) stars discovered are carbon-rich. One possibility is that the UMP stars actually
10
S. E. Woosley et al.
reflect the composition of the galaxy (or its precursors) following the first generation
of stellar nucleosynthesis. That is, they are not explained by one or a few events or by
binary companions, but by the yields of all Pop III stars. The lack of iron constrains
the IMF on both the top and the bottom. In particular there could have been few stars
lighter than about 30 M and few heavier than about 150 - 175 solar masses. Lighter
stars would have made supernovae like today with abundances like today. Heavier stars
would have made copious heavy elements, especially iron, in pair-instability supernovae
(Heger & Woosley 2002). Such an IMF has been suggested for other reasons by Tan &
McKee(2004).
Between about 30 and 90 M one might have the sort of evolution described by
Meynet et al. Neutron stars and iron would occasionally be produced, but mostly the
helium cores of the stars would collapse to black holes after losing their nitrogen-rich
envelopes to winds. Above about 80 - 90 M , depending on rotation, one encounters
the pulsational pair-instability supernovae (Heger & Woosley 2002; Woosley, Blinnikov
& Heger 2007). The end point of such a star, 110 M on the main sequence, is illustrated
in Fig. 4. Prior to the time shown, all mass exterior to 45 M has already been ejected by
violent, supernova-like outbursts episodes. That mass was rich in CNO and Ne (Na was
not followed in the calculation). Specifically, 0.84 M of carbon, 0.22 M of nitrogen,
1.9 M of oxygen and 0.42 M of neon was ejected in the pulses.
The remaining 45 M core, at the time the star finally formed an iron core and
collapsed, was very tightly bound, BE = 4.8 1051 erg outside the base of the oxygen
shell. It will be very difficult to explode with neutrinos and the iron core itself, 2.1
M , is not far from the maximum neutron star mass. Neutron star rotation with a period
longer than 2 ms would also provide inadequate energy to eject most of the matter outside
the iron core. The most likely outcome is that the whole core becomes a black hole.
Calculations in progress suggest that this sort of behavior might characterize all zero
metallicity stars from about 90 to 150 solar masses, with the exact limits sensitive to the
rotation rate and the uncertain characterization of convective overshoot and rotational
mixing. If so, the first generation of stars may have left us chiefly CNO and a lot of black
holes with masses around 40 M . The prediction then is that no UMP stars will ever
be found that are not CNO rich and, conversely, no stars with much less CNO than the
the UMP stars will be found. The transition that happened between [Fe/H] = -5 and -3,
namely the filling in of the IMF below 30 M , may have happened because of cooling due
to light elements, not Fe. The little bit of iron that existed then came from a few stars
near 30 M , or, more exciting, it may have come from a few pair-instability supernovae
above 150 M . In the latter case, the odd-Z to even-Z abundances above Mg might show
a distinctive non-solar pattern.
We acknowledge helpful conversations with Candace Joggerst and Dan Kasen. SEW is
most grateful to the Organizing Committee, especially Beatriz Barbuy, for all their help.
This work has been supported by the National Science Foundation (AST-0909129) and by
the US Department of Energy (DE-FC02-06ER41438, DE-FC02-09ER41618, DE-FG0287ER40328). Luke Roberts acknowledges support from the DOE NNSA Stewardship
Science Graduate Fellowship Program.
References
Aoki, W. et al. 2006, ApJ, 639, 897
Argast, D., Samland, M., Thielemann, F.-K., & Qian, Y.-Z. 2004, Astron. & Ap., 416, 997
11
Burbidge, E. M., Burbidge, G. R., Fowler, W. A., & Hoyle, F. 1957, Reviews of Modern Physics,
29, 547
Burrows, A., Livne, E., Dessart, L., Ott, C. D., & Murphy, J. 2007, ApJ, 655, 416
Cameron, A. G. W. 1957, Chalk River Report, CRL 41
Cayrel, R., et al. 2004, Astro. & Ap., 416, 1117
Christlieb, N., et al. 2002, Nature, 419, 904
Christlieb, N., et al. 2004, ApJ, 603, 708
Ekstr
om, S., Meynet, G., Chiappini, C., Hirschi, R., & Maeder, A. 2008, Astro. & Ap., 489, 685
Eldridge, J. J., & Vink, J. S. 2006, ApJ, 452, 295
Frebel, A., et al. 2005, Nature, 434, 871
Frebel, A., Collet, R., Eriksson, K., Christlieb, N., & Aoki, W. 2008, ApJ, 684, 588
Fr
ohlich, C., Martnez-Pinedo, G., Liebend
orfer, M., Thielemann, F.-K., Bravo, E., Hix, W. R.,
Langanke, K., & Zinner, N. T. 2006, Physical Review Letters, 96, 142502
Hamuy, M. 2006, ApJ, 582, 905
Heger, A., & Woosley, S. E. 2002, ApJ, 567, 532
Heger, A., & Woosley, S. E. 2009, ApJ, submitted, astroph 0803.3161
Hoffman, R. D., Woosley, S. E., Fuller, G. M., & Meyer, B. S. 1996, ApJ, 460, 478
Hoffman, R. D., M
uller, B., & Janka, H.-T. 2008, ApJ Lettr., 676, L127
Horowitz, C. J. 2002, Phys. Rev. D, 65, 043001
Hoyle, F., & Fowler, W. A. 1960, ApJ, 132, 565
Iwamoto, N., Umeda, H., Tominaga, N., Nomoto, K., & Maeda, K. 2005, Science, 309, 451
Joggerst, C. C., Almgren, A., Bell, J., Heger, A., Whalen, D., & Woosley, S. E. 2009, ApJ, in
press, astroph 0907.3885
Kasen, D., & Woosley, S. E. 2009, ApJ, 703, 2205
Kobayashi, C., Umeda, H., Nomoto, K., Tominaga, N., & Ohkubo, T. 2006, ApJ, 653, 1145
Komiya, Y., Suda, T., Minaguchi, H., Shigeyama, T., Aoki, W., & Fujimoto, M. Y. 2007, ApJ,
658, 367
Lai, D. K., Bolte, M., Johnson, J. A., Lucatello, S., Heger, A., & Woosley, S. E. 2008, ApJ, 681,
1524
Limongi, M., & Chieffi, A. 2009, Memorie della Societa Astronomica Italiana, 80, 151
Liu, Q. Z., van Paradijs, J., & van den Heuvel, E. P. J. 2005, Astron. & Ap., 442, 1135
Metzger, B. D., Thompson, T. A., & Quataert, E. 2007, ApJ, 659, 561
Meynet, G., & Maeder, A. 2005, Astron. & Ap., 429, 581
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, Astron. & Ap., 447, 623
Otsuki, K., Burrows, A., & Matos, M. 2008, Proceedings of the 10th Symposium on Nuclei in
the Cosmos. Available on line at http://pos.sissa.it/cgi-bin/reader/conf.cgi?confid=53,
Pruet, J., Hoffman, R. D., Woosley, S. E., Janka, H.-T., & Buras, R. 2006, ApJ, 644, 1028
Qian, Y.-Z., & Woosley, S. E. 1996, ApJ, 471, 331
Roberts, L., Woosley, S. E., & Hoffman, R. D. 2009, in preparation for ApJ
Sobeck, J. S., Lawler, J. E., & Sneden, C. 2007, ApJ, 667, 1267
Suda, T., Aikawa, M., Machida, M. N., Fujimoto, M. Y., & Iben, I. J. 2004, ApJ, 611, 476
Tan, J. C., & McKee, C. F. 2004, ApJ, 603, 383
Thorsett, S. E., & Chakrabarty, D. 1999, ApJ, 512, 288
Tumlinson, J. 2007, ApJ, 665, 1361
Umeda, H., & Nomoto, K. 2003, Nature, 422, 871
Vink, J. S., & de Koter, A. 2005, Astron. & Ap., 442, 587
Wanajo, S., Nomoto, K., Janka, H.-T., Kitaura, F. S., Mller, B. 2009, ApJ, 695, 208
Weaver, T. A., Zimmerman, G. B., & Woosley, S. E. 1978, ApJ, 225, 1021
Woosley, S. E., Wilson, J. R., Mathews, G. J., Hoffman, R. D., & Meyer, B. S. 1994, ApJ, 433,
229
Woosley, S. E., Heger, A., & Weaver, T. A. 2002, Reviews of Modern Physics, 74, 1015
Woosley, S. E., & Heger, A. 2007, Physics Reports, 442, 269
Woosley, S. E., Blinnikov, S., & Heger, A. 2007, Nature, 450, 390
Session I
Primordial Nucleosynthesis
and the First Stars in the Universe
1. Introduction
Primordial, or Big Bang, Nucleosynthesis (BBN) provides a key probe of the physics
and early evolution of the Universe, as do observations of the cosmic microwave background (CMB) radiation. These probes offer windows on the early Universe at two widely
separated epochs: BBN when the Universe was only 20 minutes old and the CMB some
400 thousand years later. BBN and the CMB provide complementary tests of the consistency of the standard, hot big bang cosmology and offer observational tests of its
quantitative predictions. For a recent review, see Steigman (2007); for a comparison
between the predictions of BBN and the CMB, see Simha & Steigman (2008).
The standard models of cosmology, and of standard, big bang nucleosynthesis (SBBN),
employ the general theory of relativity (GR) to describe an expanding Universe, filled
with radiation (including three flavors of light neutrinos) and matter (including nonbaryonic dark matter). For our purposes here, the presence or not of dark energy is
irrelevant since dark energy (or a cosmological constant) plays no role in the physics of
the early Universe which concerns us here. For BBN, the relic abundances of D, 3 He, 4 He,
and 7 Li are predicted as a function of two cosmological parameters: the baryon density
parameter B and the expansion rate parameter S H 0 /H, where H is the standard
model value of the Hubble parameter and S 6= 1 allows for a large class of non-standard
models of particle physics and/or cosmology. For SBBN it is assumed that the Hubble
parameter assumes its standard model value (S = 1), so that the relic abundances depend
on only one cosmological parameter, B .
Here, to assess the current status of the standard models of particle physics and cosmology, well ask if the BBN-predicted and observationally inferred relic abundances of the
light nuclides agree and, if the BBN-determined values of B and H are consistent with
the values inferred from independent (non-BBN) cosmological observations, including
those of the CMB and of the Large Scale Structure (LSS) observed in the Universe.
15
16
Gary Steigman
(2.1)
N (0 )/ .
(2.2)
where
N parameterizes any difference from the standard model, early Universe predicted
energy density, normalized to the contribution from one additional light neutrino. For
SBBN, N = 3 and S = 1 (N = 0). However, since any departures from the standard
models may arise from new particle physics and/or new cosmology, N does not necessarily count additional flavors of neutrinos and, indeed, N may be < 3 or > 3 (i.e., S < 1
or S > 1). N (or N ) is simply a convenient way to parameterize a non-standard, early
Universe expansion rate. For example, if the particle content of the standard model of
particle physics (0 = ) is adopted,
G0 /G = 1 + 7N /43,
(2.3)
the BBN or CMB determined values of N can constrain any deviations between the early
Universe magnitude of Newtons constant and the value measured terrestrially today.
3. Primordial Nucleosynthesis
Standard BBN
For standard BBN (SBBN) the relic abundances of the light nuclides produced during
Primordial Nucleosynthesis
17
primordial nucleosynthesis depend only on the baryon density parameter B . Over a limited range in B 1010 10 61, the SBBN-predicted abundances (Kneller & Steigman
1.6
(2004), Steigman (2007)) depend on the baryon density parameter as, (D/H)P 10
,
0.6
3
7
2.0
4
( He/H)P 10 , and ( Li/H)P 10 . A good fit to the SBBN-predicted He mass
fraction is YP = 0.2485 0.0005 + 0.0016(10 6). The SBBN-predicted relic abundances
of D and 7 Li are most sensitive to the baryon density parameter, while those of 3 He and
4
He are less (for the latter, much less) sensitive to B .
Non-Standard BBN
For non-standard BBN (S 6= 1, N 6= 3), it is the primoridal abundance of 4 He which
provides the most sensitive probe. According to Kneller & Steigman (2004), for 0.85 <
1.6
<
<
<
<
S <
1.15
(1.3
N
5.0)
and
5
7,
(D/H)
,
where
10
P
D
10
D
2.0
Figure 1. The deuterium abundances, yD 105 (D/H), inferred from observations of high-redshift, low-metallicity QSOALS, as a function of the log of the corresponding H I column densities.
The solid line is at the value of the weighted mean of the seven yD values, hyD i = 2.7, and the
dotted lines show the estimated uncertainty, (yD ) = 0.2.
18
Gary Steigman
(3.1)
Note that the Pettini et al. (2008) value of log(yDP ) = 0.45 0.03 corresponds to yDP =
2.8 0.2, consistent, within the errors, with the weighted mean of the individual yD
values. For quantitative comparisons, the value of yDP from eq. (3.1) is adopted here.
For SBBN, this value of the primordial D abundance corresponds to 10 = 6.0 0.3
or, B h2 = 0.022 0.001 (Steigman (2007)), a 5% determination of the baryon density
parameter. For comparison, if the Pettini et al. (2008) value were adopted, a slightly
lower (but consistent) value, 10 = 5.8 0.3 or, B h2 = 0.021 0.001, would be found.
Non-BBN Determinations Of The Baryon Density Parameter: CMB And LSS
The baryon density parameter determined by SBBN and the deuterium observations
reflects the value of this parameter when the Universe is some 20 minutes old. According to standard model physics and cosmology, the value of this parameter should be
unchanged some 400 thousand years later, at recombination (and, at present, some
14 Gyr later). The comparison between the BBN-determined baryon density parameter and that inferred from observations of the CMB (see, e.g., Spergel et al. (2007) and
further references therein) and of LSS provides a test of the standard models of particle
physics and cosmology (Steigman (2007)). According to Simha & Steigman (2008), the
combination of the CMB plus LSS data results in 10 = 6.10.2 or, B h2 = 0.0220.001.
More recent results from the WMAP team (Dunkley et al. (2009), Komatsu et al. (2009))
and others (Sanchez et al. (2009)) suggest a slightly higher value of 10 = 6.22 0.16 or,
B h2 = 0.0227 0.0006. Within the uncertainties, B (BBN) B (CMB/LSS); the number of baryons (nucleons) in a comoving volume of the Universe is unchanged between
BBN and recombination (as it should be for the standard model).
4. Testing SBBN
Having found agreement between B (BBN) and B (CMB/LSS), we still need to test
the consistency of SBBN. That is, do the abundances of the other light nuclides (3 He,
4
He, 7 Li) predicted by SBBN, using the D-determined value of B , agree with their
observationally inferred primordial abundances?
Helium-3:
In contrast to deuterium, the post-BBN evolution of 3 He is complex. When gas is
cycled through stars, D is burned to 3 He, any prestellar 3 He (prestellar D + 3 He) is
burned away in the hot interiors (of all but the least massive stars), but preserved in
their cooler, outer layers and, new 3 He is synthesized via stellar nucleosynthesis in the
interiors of lower mass stars (e.g., Iben (1967), Rood (1972), Rood, Steigman, & Tinsley
(1999)). Competition between destruction, preservation, and synthesis, complicates the
Primordial Nucleosynthesis
19
process of using the observations to infer the primordial 3 He abundance. The data (see
Bania, Rood, & Balser (2002)) dont help. Observations of 3 He are limited to the solar
system Geiss & Gloeckler (1998) and to H II regions in the Galaxy (see Steigman (2006a)
and Steigman (2007) for further discussion and references). If, in the course of Galactic
chemical evolution there is a net increase in the abundance of 3 He from its primordial
value, the 3 He abundance and metallicity should be correlated (and, the 3 He abundance
in the interstellar medium (ISM) of the Galaxy at present should exceed the presolar
nebula abundance). As may be seen in Figure 2, no clear trend with metallicity is revealed and, while many of the observed 3 He abundances do exceed the solar abundance,
some dont. All that may be inferred from this data (in this authors opinion) is that
the lowest 3 Heabundances observed are consistent with the SBBN-predicted primordial
abundance and, the remaining 3 He abundances suggest net production of 3 He in the
course of Galactic chemical evolution. D, 3 He, and the CMB/LSS are in agreement with
SBBN.
Lithium-7:
7
Li is a relatively fragile, weakly bound nuclide, easily destroyed at the high temperatures inside most stars. However, as with 3 He, some 7 Li may survive in the cooler,
outer layers of stars and, stellar production and cosmic ray nucleosynthesis likely increase the post-BBN abundance of 7 Li. The data reveal that while lithium is depleted
in many stars, the overall trend is a lithium abundance which increases with metallicity. As the metallicity approaches zero (primordial), the 7 Li abundances are expected
to plateau (the Spite plateau) at the primordial abundance. In Figure 3 the lithium
abundances, [Li] 12+log(7 Li/H), are shown as a function of the iron abundances (on
a log scale, normalized to the solar iron abundance) for the most metal-poor halo stars
(and for the globular cluster NGC 6397). Where is the Spite plateau? The data in Figure 3 (see, also, the contributions to these proceedings by Sbordone et al. and Melendez
et al.) fail to reveal clear evidence for a plateau as [Fe/H] 0. Even more disturbing
is the fact that none of the lithium abundances inferred from these observations of the
oldest, most metal-poor, most nearly primordial stars in the Galaxy, come even close
20
Gary Steigman
to the SBBN-predicted abundance. The observed abundances are too low by factors of
3 5. This gap seems too wide to be closed by observational uncertainties. Either
this conflict between the predictions of SBBN and the observations is pointing to new
physics and/or cosmology or, our understanding of the structure and evolution of the
oldest, most metal-poor stars in the Galaxy is seriously incomplete.
Helium-4:
After hydrogen, helium (4 He) is the most abundance element in the Universe. In the
post-BBN Universe, as gas cycles through stars, the helium abundance (the mass fraction of 4 He, Y) increases from its primordial value YP . While the correction for stellar
produced 4 He is uncertain, its contribution can be minimized by restricting attention to
the most metal-poor sites, the low-metallicity, extragalactic H II regions (Blue Compact
Galaxies). The current status of the search for YP is an object lesson in the difference
between quantity and quality. With more than 100 helium abundance determinations,
statistical uncertainties are very small: (YP )stat <
0.001 (Izotov et al. (2007)). However, most analyses fail to deal adequately with the many identified (but often ignored)
sources of systematic errors, whose values are estimated to be larger, (YP )syst >
0.006
(see Steigman (2007) for a discussion of these and other related issues and, for further
references). Following Steigman (2007) and Simha & Steigman (2008), here we adopt
YP = 0.240 0.006 as an estimate of the primordial 4 He mass fraction. For SBBN, this
corresponds to a very small value of the baryon density parameter, 10 (4 He) <
3 (note
that the simple fitting formula for YP in 3 is not valid for such a low helium abundance
and the result here is from the full BBN code). Such a low estimate of the baryon density
parameter is clearly in conflict with the value determined by SBBN and D, which is
otherwise consistent with the CMB/LSS determined value. How serious is this conflict?
That is, given the estimate of the error in the observationally determined value of YP ,
how bad is the disagreement? For 10 (D) = 6.0, the SBBN-predicted primordial helium
abundance is YP = 0.249. The observationally-inferred abundance differs from the predicted abundance by only 1.5, a not very serious disagreement. However, this tension
Figure 3. The log of the 7 Li abundances, [Li] 12+log(7 Li/H), from observations of old, very
metal-poor halo stars, as a function of the stellar iron abundances. Blue filled circles (Asplund
et al. (2006)), red, filled triangles (Boesgaard et al. (2005)), green filled squares (Aoki et al.
(2009)). The black filled circle (Lind et al. (2009)) is for the globular cluster NGC6397. The solid
lines shows 1 range around the SBBN-predicted primordial abundance [Li]SBBN = 2.700.06.
Primordial Nucleosynthesis
21
between the predicted and observed helium abundances could be a hint of non-standard
physics and/or cosmology.
6. Conclusions
The very good agreement between the values of the baryon density and universal
expansion rate parameters determined by BBN, when the Universe was 20 minutes
old, and by the CMB/LSS, some 400 thousand to 14 billion years later, leads to
constraints on some extensions of the standard models of particle physics and cosmology.
3.5
3
4
N
2.5
3
2
1.5
1
0
4
6
10
1
5
5.5
6
10
6.5
Figure 4. Left panel: The 68% and 95% contours in the N vs. 10 plane (from Simha &
Steigman (2008)) for BBN (D & 4 He) (solid) and for the CMB/LSS (dashed). The crosses
indicate the best fit values (see the text). Right panel: The joint BBN/CMB/LSS contours; note
the expanded scales for N and 10 .
22
Gary Steigman
Entropy Conservation?
For example, the numbers of baryons and CMB photons in a comoving volume are
related by the baryon density parameter, NB = B N . In the standard model of particle
physics, NB is unchanged from the BBN to recombination and the present. Comparing B
at BBN and at recombination leads to the constraint: N (CMB)/N (BBN) = 0.920.07,
limiting any post-BBN entropy production.
Extra, Post-BBN Radiation Density?
Since 0R /R = 1 + 7N /43, the BBN and CMB constraints on N limit the radiation
energy densities at these widely separated epochs. In the absence of the creation of
new radiation (e.g., by the late decay of a massive particle), N (BBN) = N (CMB).
Comparing N at BBN and at recombination constrains any possible difference between
these values. This comparison reveals that 0.94 6 N 6 1.23, consistent with no extra,
post-BBN radiation density.
Variation Of the Gravitational Constant?
The BBN and CMB constraints on N also limit any difference between the magnitude
of the gravitational constant at BBN or at recombination and that observed today terrestrially since G0 /G = 1 + 7N /43. From N at BBN, GBBN /G0 = 0.91 0.07, while
the CMB/LSS bound on N leads to an even tighter constraint, GCMB /G0 = 0.99 0.12.
7. Summary
For N 3, BBN agrees with the observations of the CMB (and LSS and H0 ), confirming the consistency of the standard models of particle physics and cosmology. But,
lithium remains a problem whose origin may lie with stellar depletion/dilution or with
new particle physics and/or cosmology. When BBN is combined with the CMB and LSS,
interesting constraints on some non-standard models of particle physics and cosmology
can be obtained.
References
Aoki, W. et al. (2009), ApJ, In Press, (arXiv:0904.1448).
Asplund, M., et al. 2006, ApJ, 644, 229.
Bania, T.M., Rood, R.T., & Balser, D.S. 2002, Nature, 415, 54.
Boesgaard, A.M., Stephens, A., & Deliyannis, C.P. 2005, ApJ, 633, 398.
Dunkley, J., et al. 2009, ApJS, 180, 306.
Geiss, J. & Gloeckler, J.G. 1998, Space Sci. Rev. 84, 239.
Iben, I., Jr. 1967, ApJ, 147, 624.
Izotov, Y., Thuan, T.X., & Stasinska, G. 2007, ApJ, 662, 15.
Komatsu, E., et al. 2009, ApJS, 180, 330.
Kneller, J.P. & Steigman, G. 2004, New J. Phys., 6, 117.
Lind, K., et al. 2009, A&A, In Press, (arXiv:0906.2876).
Pettini, M., Zych, B.J., Murphy, M.T., Lewis, A., & Steidel, C.C. 2008, MNRAS, 391, 1499.
Rood, R.T. 1972, ApJ, 177, 681.
Rood, R.T., Steigman, G., & Tinsley, B.M. 1999, ApJL, 207, L57.
Sanchez, A.G., et al. 2009, MNRAS, In Press (arXiv:0901.2570).
Simha, V. & Steigman, G. 2008, JCAP, 06, 016.
Spergel, D.N., et al. 2007, ApJS, 170, 377.
Steigman, G. 2006, Int. J. Mod. Phys. E, 15, 1.
Steigman, G. 2006, JCAP, 10, 016.
Steigman, G. 2007, Ann. Rev. Nucl. Part. Sci., 57, 463.
1. Introduction
The spectroscopic signature of the presence of 6 Li in the atmospheres of metal-poor
halo stars is a subtle extra depression in the red wing of the 7 Li doublet, which can
only be detected in spectra of the highest quality. Based on high-resolution, high signalto-noise VLT/UVES spectra of 24 bright metal-poor stars, Asplund et al. (2006) report
the detection of 6 Li in nine of these objects. The average 6 Li/7 Li isotopic ratio in the
nine stars in which 6 Li has been detected is q(Li) 0.04 and is very similar in each
of these stars, defining a 6 Li plateau at approximately log n(6 Li) = 0.85 (on the scale
log n(H) = 12). A convincing theoretical explanation of this new 6 Li plateau turned
out to be problematic. Even when the depletion of the 6 Li isotope during the pre-mainsequence phase would be ignored, the high abundances of 6 Li at the lowest metallicities
cannot be explained by current models of galactic cosmic-ray production (for a concise
review see e.g. Christlieb 2008, and references therein).
A possible solution of the so-called second Lithium problem was suggested by Cayrel
et al. (2007), who point out that the intrinsic line asymmetry caused by convection in the
photospheres of cool stars is almost indistinguishable from the asymmetry produced by
a weak 6 Li blend on a presumed symmetric 7 Li profile. As a consequence, the derived 6 Li
abundance should be significantly reduced when the intrinsic line asymmetry in properly
taken into account. Using 3D non-LTE line formation calculations based on 3D hydrodynamical model atmospheres computed with the CO5 BOLD code (Freytag et al. 2002,
Wedemeyer et al. 2004, see also http://www.astro.uu.se/bf/co5bold main.html),
we quantify the theoretical effect of the convection-induced line asymmetry on the resulting 6 Li abundance as a function of effective temperature, gravity, and metallicity, for
a parameter range that covers the stars of the Asplund et al. (2006) sample.
23
24
M. Steffen et al.
1.00
1.00
0.95
0.95
25
Flux
1.05
Flux
1.05
0.90
0.85
0.90
FWHM= 8.497 [km/s]
FWHM= 7.681 [km/s]
0.80
20
10
0
v [km/s]
10
0.85
V= 0.188 [km/s]
V= 0.308 [km/s]
0.80
0.6
20
0.4
0.2
0.0
V [km/s]
0.2
0.4
0.6
Figure 1. Comparison of 3D LTE (dashed) and 3D non-LTE (solid) line profile (left) and line
bisector (right) of a single 7 Li component, computed for a typical metal-poor turn-off halo star
(Teff = 6300 K, log g = 4.0, [Fe/H] = 2). Non-LTE effects strongly reduce the equivalent
width of the line (W: 35.5 15.6 m
A) while they increase the half width of the line profile
(F W HM : 7.7 8.5 km/s); the line asymmetry is diminished in non-LTE (velocity span of
bisector v : 0.31 0.19 km/s).
3.6
3.6
3.8
3.8
log g [cgs]
log g [cgs]
by F W HM of the Gaussian kernel, and a global line shift, v. The value q (Li) of the
best fit is then identified with the correction q(Li) that has to be subtracted from
the 6 Li/7 Li isotopic ratio determined from the 1D LTE analysis in order to correct for
the bias introduced by the intrinsic line asymmetry: q(Li) = q (Li), and q (3D) (Li) =
q (1D) (Li) - q(Li). The procedure takes saturation effects properly into account.
We have determined q(Li) in the relevant range of stellar parameters according to
the method outlined above. The results are displayed in Fig. 2 for [Fe/H]=1 and 2. At
given metallicity, the corrections are largest for low gravity and high effective temperature, increasing towards higher metallicity. We note that q(Li) is essentially insensitive
to the choice of v sin i. The downward correction of the 6 Li/7 Li isotopic ratio is typically
in the range 0.01 < q(Li) < 0.02 for the stars of the Asplund et al. (2006) sample (see
Fig. 2). After subtracting for each of these stars the individual q(Li), according to Teff ,
log g, and [Fe/H], the mean 6 Li/7 Li isotopic ratio of the sample is reduced from 0.0212 to
0.0059, as illustrated in Fig. 3. If we keep the error bars given by Asplund et al. (2006), the
number of stars with a 6 Li detection above the 2 level decreases from 9 to 4. One of them,
HD 106038, survives only because of its particularly small error bar of 0.006, another
one, CD-30 18140, just barely fulfills the 2 criterion. The remaining two stars are G020024, which shows the clearest evidence for the presence of 6 Li (q(Li) = 0.052 0.017),
and HD 102200 with a somewhat weaker 6 Li signal (q(Li) = 0.033 0.013). The spectra
of these stars should be reanalyzed with 3D non-LTE line profiles.
4.0
4.2
4.4
6400
4.2
4.4
[Fe/H]=2
6500
4.0
6300
6200 6100
Teff [K]
6000
5900
[Fe/H]=1
6500
6400
6300
6200 6100
Teff [K]
6000
5900
Figure 2. Contours of q(Li) in the Teff log g plane for metallicities [Fe/H]=2 (left) and
[Fe/H]=1 (right). White symbols mark the positions of the stars from the Asplund et al. (2006)
sample with 2.5 < [Fe/H] < 1.5 (left), and 1.5 < [Fe/H] < 0.5 (right).
M. Steffen et al.
: 0.0212
0.10 Mean
Sigma: 0.0208
26
0.05
0.00
0.05
5800
6000
6200
: 0.0059
0.10 Mean
Sigma: 0.0212
0.05
0.00
0.05
6400
5800
6000
Teff [K]
6200
6400
Teff [K]
Figure 3. 6 Li/7 Li isotopic ratio, and 1 error bars, as a function of effective temperature as
derived by Asplund et al. (2006) before (left) and after (right) subtraction of q(Li) to correct
for the bias due to the intrinsic line asymmetry. Filled circles denote 6 Li detections above the
2 level, open circles denote non-detections.
Table 1. Fitting three observed Li I 6707
A spectra with 1D LTE and 3D non-LTE synthetic
line profiles, respectively. Columns (7)-(10) show the results for v sin i = 0.0 / 2.0 km/s.
Star
Teff
[K]
HD 74000
6203 4.03
-2.05
600
-2.30
550
HD 84937
-2.40
630
Notes:
1)
6310 4.10
1)
A(Li)
2)
q(Li)
[%]
2)
F W HM 3)
[km/s]
3.1
5.9
3.6
6.2
3.7
6.1
/
/
/
/
/
/
2.1
5.4
2.9
5.7
2.7
5.6
3)
v
[km/s]
0.64
0.42
0.04
-0.17
0.08
-0.17
/
/
/
/
/
/
0.64
0.42
0.05
-0.17
0.07
-0.14
Gaussian kernel
spectra with
As a consistency check, we have also fitted a few observed Li I 6707 A
1D LTE and 3D non-LTE synthetic line profiles, respectively. The fitting parameters are
again A(Li), q(Li), F W HM , and v. As expected, the 3D analysis yields lower q(Li) by
roughly 0.02. Details are compiled in Table 1. HD 74000 and G271162 are considered
non-detections, while HD 84937 remains a clear 6 Li detection with q (3D) (Li) 0.04.
4. Conclusions
The present study indicates that only 2 or at most 4 out of the 24 stars of the Asplund
et al. (2006) sample remain significant 6 Li detections when subjected to a 3D non-LTE
analysis, suggesting that the presence of 6 Li in the atmospheres of galactic halo stars is
rather the exception than the rule. This would imply that it is no longer necessary to
look for a global mechanism accounting for a 6 Li enrichment of the galactic halo, but
that it is sufficient to explain only a few exceptional cases, which is probably much easier.
References
Asplund, M., Lambert, D. L., Nissen, P. E., Primas, F., & Smith, V. V. 2006, ApJ, 644, 229
Barklem, P.S., Belyaev, A.K., Asplund, M. 2003, A&A, 409, L1
Castelli, F., & Kurucz, R. L. 2004, arXiv:astro-ph/0405087
Cayrel, R., Steffen, M., Chand, H., Bonifacio, P., Spite, M., Spite, F., Petitjean, P., Ludwig,
H.-G., & Caffau, E. 2007, A&A, 473, L37
Christlieb, N. 2008, Journal of Physics G: Nuclear Physics, 35, 014001
Freytag, B., Steffen, M., & Dorch, B. 2002, AN, 323, 213
Ludwig, H.-G., Caffau, E., Steffen, M., Freytag, B., Bonifacio, P., & Kucinskas, A. 2009,
MemSAI, 80, 708
Sbordone, L., Bonifacio, P., Caffau, E., et al. 2009, A&A (submitted)
Wedemeyer, S., Freytag, B., Steffen, M., Ludwig, H.-G., & Holweger, H. 2004, A&A, 414, 1121
Abstract. One of the key challenges for the next 10 years is to understand the first sources of
light, the first stars and possibly accreting black holes. Their formation ended the cosmic dark
ages at redshifts z 20 30, and signaled the transition from the simple initial state of the
universe to one of ever increasing complexity. We here review recent progress in understanding
the formation process of the first stars with numerical simulations, starting with cosmological
initial conditions and modelling the detailed physics of accretion. Once formed, the first stars
exerted crucial feedback on the primordial intergalactic medium, due to their input of radiation
and of heavy chemical elements in the wake of supernova explosions. The current theoretical
model posits that the first stars were predominantly very massive, typically 100M . Our
predictions will be tested with upcoming near-infrared observatories, such as the James Webb
Space Telecope, in the decade ahead.
Keywords. cosmology: theory, early universe galaxies: formation stars: formation, supernovae
1. Introduction
An important open question in cosmology is to understand how and when the cosmic
dark ages ended (Bromm et al. 2009). Within the current CDM paradigm, the first
stars, the so-called Population III (Pop III), are predicted to have formed at redshifts
z 20 30. Their emergence signals the rapid transformation of the universe into an
increasingly complex, hierarchical system, due to the energy and heavy element input
from Pop III stars and accreting black holes (Barkana & Loeb 2001; Bromm & Larson
2004; Ciardi & Ferrara 2005; Miralda-Escude 2003). Currently, we can directly probe
the state of the universe roughly a million years after the Big Bang by detecting the
temperature anisotropies in the cosmic microwave background (CMB), thus providing
us with the initial conditions for subsequent structure formation. Complementary to the
CMB observations, we can probe cosmic history all the way from the present-day universe
to roughly a billion years after the Big Bang, using the best available ground- and spacebased telescopes. In between lies the remaining frontier, and the first stars and galaxies
are the sign-posts of this early, formative epoch.
To simulate the build-up of the first stellar systems, we have to address the feedback
from the very first stars on the surrounding intergalactic medium (IGM), and the formation of the second generation of stars out of material that was influenced by this
feedback. There are a number of reasons why addressing the feedback from the first stars
and understanding second-generation star formation is crucial:
(i) The first steps in the hierarchical build-up of structure provide us with a simplified
laboratory for studying galaxy formation, which is one of the main outstanding problems
in cosmology.
(ii) The initial burst of Pop III star formation may have been rather brief due to the
strong negative feedback effects that likely acted to self-limit this formation mode (Greif
27
28
Volker Bromm
& Bromm 2006; Yoshida et al. 2004). Second-generation star formation, therefore, might
well have been cosmologically dominant compared to Pop III stars.
(iii) A subset of second-generation stars, those with masses below 1 M , would have
survived to the present day. Surveys of extremely metal-poor Galactic halo stars therefore
provide an indirect window into the Pop III era by scrutinizing their chemical abundance
patterns, which reflect the enrichment from a single, or at most a small multiple of,
Pop III SNe (Beers & Christlieb 2005; Frebel et al. 2007; Karlsson et al. 2008). Stellar
archaeology thus provides unique empirical constraints for numerical simulations, from
which one can derive theoretical abundance patterns to be compared with the data.
Existing and planned observatories, such as HST, Keck, VLT, and the James Webb
Space Telescope (JWST), planned for launch around 2014, yield data on stars and quasars
less than a billion years after the Big Bang. The ongoing Swift gamma-ray burst (GRB)
mission provides us with a possible window into massive star formation at the highest
redshifts (Lamb & Reichart 2000; Bromm & Loeb 2002, 2006). Measurements of the nearIR cosmic background radiation, both in terms of the spectral energy distribution and
the angular fluctuations provide additional constraints on the overall energy production
due to the first stars (Santos et al. 2002; Magliocchetti et al. 2003; Dwek et al. 2005;
Fernandez & Komatsu 2006; Kashlinsky et al. 2005). Understanding the formation of the
first sources of light is thus of great interest to observational studies conducted both at
high redshifts and in our local Galactic neighborhood.
First Stars
29
Figure 1. Cosmological environment for Pop III star formation (from Stacy et al. 2009). Shown
is the gas density distribution on progressively smaller scales, as labeled in each panel. In the
bottom two panels, the asterisk denotes the location of the first sink formed, which acts as a
proxy for the growing protostar. The top two panels show the emergence of a minihalo, inside
of which the dissipating gas assumes a quasi-hydrostatic morphology, just prior to the onset of
runaway collapse.
et al. 2009). Again, the Pop III case might be less different from present-day massive
star formation as previously thought. However, the task still remains to couple the 3dimensional Pop III assembly with the radiation-hydrodynamics of protostellar feedback.
It is to be expected that such radiative effects will hinder fragmentation into secondary
clumps, once the first few protostars have emerged.
While the formation of the very first, Pop III.1, stars in minihalos relied on H2 cooling,
the HD molecule can play an important role in the cooling of primordial gas in several
situations, allowing the temperature to drop well below 200 K (Abel et al. 2002; Bromm
et al. 2002). In turn, this efficient cooling may lead to the formation of primordial stars
with masses of the order of 10 M , the so-called Pop III.2 (Johnson & Bromm 2006).
In general, the formation of HD, and the concomitant cooling that it provides, is found
to occur efficiently in primordial gas which is strongly ionized, owing ultimately to the
high abundance of electrons which serve as catalyst for molecule formation in the early
30
Volker Bromm
Figure 2. Disk formation and protostellar accretion (from Stacy et al. 2009). Left panel: Velocity
field of the gas in the center of a minihalo. Sinks are denoted as follows: The asterisk marks the
location of the most massive sink, crosses that of the second most massive one, and diamonds
those of the other, smaller sinks. Shown is the situation 5000 yr after initial sink formation. At
this point, an ordered, nearly Keplerian velocity structure has been established within the disk.
Right panel: Sink mass vs. time. The solid line shows the mass of the first sink particle, fitted
by a power law according to M t0.55 (red line). The dash-dotted line depicts the accretion
history found in Bromm & Loeb (2004). The dotted line traces the mass growth of the second
largest sink. It is evident that the sinks grow to masses & 10M within a few 1,000 yr.
universe (Shapiro & Kang 1987). Efficient cooling by HD can be triggered within the
relic H ii regions that surround Pop III.1 stars at the end of their brief lifetimes (Alvarez
et al. 2006), owing to the high electron fraction that persists in the gas as it cools and
recombines (Johnson et al. 2007; Yoshida et al. 2007b). The efficient formation of HD
can also take place when the primordial gas is collisionally ionized, such as behind the
shocks driven by the first SNe or in the virialization of massive DM halos (Machida et
al. 2005; Johnson & Bromm 2006; Shchekinov & Vasiliev 2006).
There might thus be a progression of characteristic masses of the various stellar populations that form in the early universe. In the wake of Pop III.1 stars (typically with
M 100M ) formed in DM minihalos, Pop III.2 star formation (with M 10M )
ensues in regions which have been previously ionized, typically associated with relic H ii
regions left over from massive Pop III.1 stars collapsing to black holes, while even later,
when the primordial gas is locally enriched with metals, Pop II (with M 1M ) stars
begin to form (Bromm & Loeb 2003; Greif & Bromm 2006). Recent simulations confirm
this picture, as Pop III.2 star formation ensues in relic H ii regions in well under a Hubble time, while the formation of Pop II stars after the first SN explosions is delayed by
more than a Hubble time (Greif et al. 2007; Yoshida et al. 2007a,b; but see Whalen et
al. 2008).
First Stars
31
& Haiman 2002). These so-called atomic-cooling halos did not rely on the presence of
molecular hydrogen to enable cooling of the primordial gas. In addition, their potential
wells were sufficiently deep to retain photoheated gas, in contrast to the shallow potential
wells of minihalos (Dijkstra et al. 2004). These are arguably minimum requirements to
set up a self-regulated process of star formation that comprises more than one generation
of stars, and is embedded in a multi-phase interstellar medium.
One of the important consequences of atomic cooling is the softening of the equation of
state below the virial radius, allowing a fraction of the potential energy to be converted
into kinetic energy (Wise & Abel 2007; Greif et al. 2008). This implies that perturbations
in the gravitational potential can generate turbulent motions on galactic scales, which are
then transported to the centre of the galaxy. In this context the distinction between two
fundamentally different modes of accretion becomes important (Greif et al. 2008). Gas
accreted directly from the IGM is heated to the virial temperature and comprises the sole
channel of inflow until cooling in filaments becomes important. This mode is termed hot
accretion, and dominates in low-mass haloes at high redshift. The formation of the virial
shock and the concomitant heating in an atomic cooling halo are visible in Fig. 3. The
second mode, termed cold accretion, becomes important as soon as filaments are massive
enough to enable molecule reformation, which allows the gas to cool and flow into the
central regions of the nascent galaxy with high velocities. Although the assembly of the
first galaxies provides us with an idealized laboratory for galaxy formation in general,
the degree of complexity that is exhibited in the corresponding merger tree is already
considerable. Still, current supercomputer simulations have just reached the capacity to
address the first galaxy formation process in an a-priori fashion, one star at a time.
Cold accretion is a viable agent for driving turbulence, due to the large amount of
kinetic energy it brings to the center of the galaxy. Two physically distinct mechanisms
are responsible for creating shocks (Greif et al. 2008). The virial shock forms where the
ratio of infall velocity to local sound speed approaches unity, while a multitude of unorganized shocks forms near the center of the galaxy and is mostly caused by accretion of
cold, high-velocity gas from filaments. These are more pronounced than the virial shock
and have a significantly higher angular component. They create transitory density perturbations that could in principle become Jeans-unstable and trigger the gravitational
collapse of individual clumps. In concert with metal enrichment by previous star formation in minihaloes, metal mixing in the first galaxies will likely be highly efficient and
could lead to the formation of the first low-mass star clusters (Clark et al. 2008), in extreme cases possibly even to metal-poor globular clusters (Bromm & Clarke 2002). Some
of the extremely iron-deficient, but carbon and oxygen-enhanced stars observed in the
halo of the Milky Way may thus have formed as early as redshift z 10 (Karlsson et al.
2008).
4. Outlook
Understanding the formation of the first sources of light marks the frontier of highredshift structure formation. It is crucial to predict their properties in order to develop
the optimal search and survey strategies for the JWST. Whereas ab-initio simulations
of the very first stars can be carried out from first principles, and with virtually no free
parameters, one faces a much more daunting challenge with the first galaxies. Now, the
previous history of star formation has to be considered, leading to enhanced complexity
in the assembly of the first galaxies. One by one, all the complex astrophysical processes
that play a role in more recent galaxy formation appear back on the scene. Among them
are external radiation fields, comprising UV and X-ray photons, and possibly cosmic
32
Volker Bromm
Figure 3. Assembly of the first galaxy (from Greif et al. 2008). Gravity is assembling the dark
matter (top row) and pristine gas (middle row) into a primordial galaxy. Shown is the evolution
at three different stages (from left to right). White crosses indicate individual Pop III stars, which
form prior to the first galaxy. The assembly process is accompanied by powerful virialization
shocks that heat up the infalling gas to > 104 K (bottom row).
rays produced in the wake of the first SNe. There will be metal-enriched pockets of gas
which could be pervaded by dynamically non-negligible magnetic fields, together with
turbulent velocity fields built up during the virialization process. However, the goal of
making useful predictions for the first stars and galaxies is now clearly drawing within
reach, and the pace of progress is likely to be rapid. A few years from now we should
have a better idea of whether our theoretical ideas are on track, or need to be refined.
Acknowledgements
I would like to thank the organizers for putting together a stimulating program of talks
and discussions, which gave me numerous ideas for future exploration. Support from NSF
grant AST-0708795 and NASA ATFP grant NNX08AL43G is gratefully acknowledged.
The simulations presented here were carried out at the Texas Advanced Computing
Center (TACC).
First Stars
References
Abel, T., Bryan, G. L., & Norman, M. L. 2002, Science, 295, 93
Alvarez, M. A., Bromm, V., & Shapiro, P. R. 2006, ApJ, 639, 621
Barkana, R., & Loeb, A. 2001, Phys. Rep., 349, 125
Beers, T. C., & Christlieb, N. 2005, ARA&A, 43, 531
Bromm, V., & Clarke, C. J. 2002, ApJ, 566, L1
Bromm, V., Coppi, P. S., & Larson, R. B. 2002, ApJ, 564, 23
Bromm, V., & Larson, R. B. 2004, ARA&A, 42, 79
Bromm, V., & Loeb, A. 2002, ApJ, 575, 111
. 2003, Nature, 425, 812
. 2004, New Astronomy, 9, 353
. 2006, ApJ, 642, 382
Bromm, V., Yoshida, N., Hernquist, L., & McKee, C. F. 2009, Nature, 459, 49
Ciardi, B., & Ferrara, A. 2005, Space Science Reviews, 116, 625
Clark, P. C., Glover, S. C. O., & Klessen, R. S. 2008, ApJ, 672, 757
Dijkstra, M., Haiman, Z., Rees, M. J., & Weinberg, D. H. 2004, ApJ, 601, 666
Dwek, E., Arendt, R. G., & Krennrich, F. 2005, ApJ, 635, 784
Fernandez, E. R., & Komatsu, E. 2006, ApJ, 646, 703
Frebel, A., Johnson, J. L., & Bromm, V. 2007, MNRAS, 380, L40
Greif, T. H., & Bromm, V. 2006, MNRAS, 373, 128
Greif, T. H., Johnson, J. L., Bromm, V., & Klessen, R. S. 2007, ApJ, 670, 1
Greif, T. H., Johnson, J. L., Klessen, R. S., & Bromm, V., 2008, MNRAS, 387, 1021
Johnson, J. L., & Bromm, V. 2006, MNRAS, 366, 247
Johnson, J. L., Greif, T. H., & Bromm, V. 2007, ApJ, 665, 85
Johnson, J. L., Greif, T. H., & Bromm, V. 2008, MNRAS, 388, 26
Karlsson, T., Johnson, J. L., & Bromm, V. 2008, ApJ, 679, 6
Kashlinsky, A., Arendt, R. G., Mather, J., & Moseley, S. H. 2005, Nature, 438, 45
Lamb, D. Q., & Reichart, D. E. 2000, ApJ, 536, 1
Machida, M. N., Tomisaka, K., Nakamura, F., & Fujimoto, M. Y. 2005, ApJ, 622, 39
Magliocchetti, M., Salvaterra, R., & Ferrara, A. 2003, MNRAS, 342, L25
McKee C. F., & Tan, J. C. 2008, ApJ, 681, 771
Miralda-Escude, J. 2003, Science, 300, 1904
Oh, S. P., & Haiman, Z. 2002, ApJ, 569, 558
Santos, M. R., Bromm, V., & Kamionkowski, M. 2002, MNRAS, 336, 1082
Shapiro, P. R., & Kang, H. 1987, ApJ, 318, 32
Shchekinov, Y. A., & Vasiliev, E. O. 2006, MNRAS, 368, 454
Spergel, D. N. et al. 2007, ApJS, 170, 377
Stacy, A. R., Greif, T. H., & Bromm, V. 2009, MNRAS, submitted (arXiv:0908.0712)
Tan, J. C., & McKee C. F. 2004, ApJ, 603, 383
Turk, M. J., Abel, T., & OShea, B. 2009, Science, 325, 601
Whalen, D., van Veelen, B., OShea, B. W., & Norman, M. L. 2008, ApJ, 682, 49
Wise, J. H., & Abel, T. 2007, ApJ, 665, 899
Yoshida, N., Bromm, V., & Hernquist, L. 2004, ApJ, 605, 579
Yoshida, N., Oh, S. P., Kitayama, T., & Hernquist, L. 2007a, ApJ, 663, 687
Yoshida, N., Omukai, K., & Hernquist, L. 2007b, ApJ, 667, L117
Yoshida, N., Omukai, K., Hernquist, L., & Abel, T. 2006, ApJ, 652, 6
Yoshida, N., Omukai, K., & Hernquist, L. 2008, Science, 321, 669
33
Institute for the Physics and Mathematics of the Universe (IPMU), University of Tokyo
Kashiwanoha 5-1-5, Kashiwa, Chiba 277-8568, Japan
email: nomoto@astron.s.u-tokyo.ac.jp
2
Department of Astronomy, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
3
Department of Physics, Konan University, Okamoto, Kobe 658-8501, Japan
Abstract. We review the properties of supernova (SN) as a function of the progenitors mass
M . (1) Mup - 10 M stars are super-AGB stars and resultant electron capture SNe may be
Faint supernovae like Type IIn SN 2008S. (2) 10 - 12 M stars undergo Fe-core collapse to form
neutron stars (NS) and Faint supernovae. (3) 12 M - MBN stars undergo Fe-core collapse to
form NSs and normal core-collapse supernovae. (4) MBN - 90 M stars undergo Fe-core collapse
to form Black Holes. Resultant supernovae are bifurcate into Hypernovae and Faint supernovae.
The observed properties of SN 2008ha can be explained as this type of Faint supernovae. (5)
90 - 140 M stars produce Luminous SNe, like SNe 2007bi and 2006gy. (6) 140 - 300 M
stars become pair-instability supernovae which could be Luminous supernovae (SNe 2007bi and
2006gy). (7) Very massive stars with M >
300M undergo core-collapse to form intermediate
mass black holes. Some SNe could be more Luminous supernovae (like SN 2006gy).
Keywords. Galaxy: halo gamma rays: bursts nuclear reactions, nucleosynthesis, abundances stars: abundances stars: AGB supernovae: general
35
(4) MBN - 90 M stars: Hypernovae and FSNe: These stars undergo Fe-core
collapse to form a black hole (BH). SNe seem to be bifurcate into two branches, Hypernovae and Faint SNe. If the BH has little angular momentum, little mass ejection would
take place and be observed as FSNe. On the other hand, a rotating BH could eject a
matter in a form of jets to make a Hypernova. The latter explosions produce a large
amount of heavy elements from -elements and Fe-peak elements. Nucleosynthesis in
these jet-induced explosions is in good agreement with the abundance patterns observed
in extremely metal-poor stars.
(5) 90 - 140 M stars: Luminous SNe (LSNe): These massive stars undergo
nuclear instabilities and associated pulsations (-mechanism) at various nuclear burning
stages depending on the mass loss and thus metallicity. Eventually, these stars undergo
Fe-core collapse. Depending on the angular momentum, Hypernova-like energetic SNe
could occur to produce large amount 56 Ni. (Because of the large ejecta mass, the expansion velocities may not be high enough to form a broad line features.) Thanks to the
large E and 56 Ni mass, the SNe could be a LSNe. The possible presence of circumstellar
matter (CM) leads to energetic SN IIn. Pulsation could also cause luminous event.
(6) 140 - 300 M stars: LSNe: If these very massive stars (VMS) do not lose much
mass, they become pair-instability supernovae (PISN). The star is completely disrupted
without forming a BH and thus ejects a large amount of heavy elements, especially 56 Ni.
Radioactive decays could produce LSNe.
(7) Stars with M >
300M: LSNe: These VMSs are too massive to be disrupted
by PISN but undergo core collapse (CVMS), forming an intermediate-mass black hole
(IMBH). Some mass ejection could be possible, associated with the possible jet-induced
explosion, which becomes a very luminous SNe (LSNe).
In the following sections, we summarize the properties of the above supernovae in some
detail. For their chemical yields, see Nomoto et al. (2009) for a recent review.
36
1993J (IIb)
-16
2.00E-016
[O I]
-15
1.60E-016
-14
-16
2007Y (Ib)
-14
-13
20
40
60
2005cz (Ib)
-12
[Ca II]
SN Ib 2004dk, +392*
1994I (Ic)
2008S (IIn)
-10
-18
SN Ic 1994I, +147
8.00E-017
4.00E-017
2008ha (I?)
SN Ib 2005cz, +179
0.00E+000
-8
0
60
120
180
240
300
360
4000
5000
6000
7000
8000
9000
Rest Wavelength ()
Figure 1. (Left): The absolute R-band light curve of faint supernovae: SN IIn 2008S (black
open circles), SN Ib 2005cz (red circles), SN I 2008ha (orange open squares), and SN Ib 2007Y
(green squares) as compared with those of SN IIb 1993J (cyan triangles) and SN Ic 1994I (blue
stars). Also shown is the light curve of SN 1994I, but dimmed by 1.5 magnitudes (magenta open
stars).
Figure 2. (Right): The late-time spectrum of SN Ib 2005cz (t = +179 days). Also shown are
SN Ib 2004dk at t 390 days, SN IIb 1993J at t = +203 days, SN Ic 1994I at t = +147 days,
peculiar SN Ia 2005hk at t = +232 days, and peculiar SN I 2008ha at t = +65 days. It is very
unique that SN 2005cz shows only weak [O I] 6300, 6364 and much stronger [Ca II] 7291,
7323 than [O I].
37
-14
Bolometric LC
25HeG
-13.5
-13
-12.5
-12
-11.5
-11
0
10
20
30
40
50
Days since the explosion
60
70
Figure 3. (Left): The ejected mass of 56 Ni as a function of the main sequence mass M of the
progenitors for several supernovae/hypernovae.
Figure 4. (Right:) The bolometric light curves of SN 2008ha and the model with M = 25M ,
E = 1.0 1048 erg, Mej = 0.12M , and 0.003 M 56 Ni.
As for the host galaxy problem, the 10M star model is found to be consistent with
the properties recently-inferred for the host galaxy of SN 2005cz. It is still a genuine E2
galaxy, but has a relatively young stellar population with life times of 107 108 years
and SN Ib 2005cz is likely the end product of one of these young stars (see Kawabata
et al. 2009 and references therein).
38
Ne
Mg
[X/Fe]
Si
Ar
Ca
Ti
Cr
Fe
Ni
Zn
-1
B
5
Na
10
Al
Cl
15
Sc
20
Mn Co
25
Cu
Ga
30
Figure 5. (Left): Averaged elemental abundances of stars with [Fe/H] = 3.7 (Cayrel et al.
2004) compared with the hypernova yield (20 M , E51 = 10).
Figure 6. (Right): The peculiar abundance pattern of the EMP stars in the Hercules dwarf
spheroidal galaxy (Koch et al. 2008) is compared with the Hypernova yield (Tominaga et al. in
prep.).
Foley et al. 2009). This SN is of type I and the peak V magnitude is only 14.2 mag.
The rise and decline of the LC is quite fast. Line velocities are such low as 2, 000
km s1 . Moriya et al. (2009) have shown that these features can be reproduced by the
core-collapse supernova model. Figure 4 shows the bolometric LC of the model with
M = 25M , E = 1.0 1048 erg, and 0.003 M 56 Ni. The ejecta of this explosion model
undergoes large fallback because of low E, so that the ejecta mass is only 0.12 M . The
LC of this model well-reproduces SN 2008ha. (SN 2008ha is not included in Figure 4
because the model with M = 13M and E = 1.4 1048 erg also well-reproduces SN
2008ha (Moriya et al. 2009)).
The fallback SN (e.g., Iwamoto et al. 2005, Fryer et al. 2009) should also undergo
mixing of 56 Ni before the occurrence of fallback in order to reproduce the observed light
curve. Tominaga (2009) has shown that such mixing and fallback in spherical explosion
is equivalent to the jet-induced nucleosynthesis.
In the jet-induced nucleosynthesis and mass ejection, the important parameter is the
energy deposition rate E dep (Tominaga et al. 2007). The variation of E dep in the range
of E dep,51 E dep /1051 ergs s1 = 0.3 1500 leads to the following variation of the
properties of GRBs and associated SNe. For low energy deposition rates (E dep,51 < 3), the
ejected 56 Ni masses (M (56 Ni) < 103 M ) are smaller than the upper limits for non-SN
GRBs 060505 and 060614 (Iwamoto et al. 2005). For intermediate energy deposition rates
56
3
<
(3 <
Edep,51 < 60), the explosions eject 10 M M ( Ni) < 0.1M , and the final BH
56
masses are 10.8M <
MBH < 15.1M. The resulting SN is faint (M ( Ni) < 0.01M )
56
or sub-luminous (0.01M <
M ( Ni) < 0.1M).
Faint SN as a result of large fallback has been suggested to be responsible to produce
the peculiar abundance patterns of extremely metal-poor (EMP) stars (Umeda & Nomoto
2002, Iwamoto et al. 2005). In the jet-induced explosion model, the abundance patterns
of EMP stars (esp. [C/Fe]) are related to E dep as follows. Lower E dep yields larger MBH
and thus larger [C/Fe], because the infall reduces the amount of inner core material (Fe)
relative to that of outer material (C).
The observed abundance patterns of extremely metal-poor (EMP) stars are classified
into three groups according to [C/Fe]: (1) [C/Fe] 0, normal EMP stars (4 < [Fe/H]
< 3, e.g., Cayrel et al. 2004); (2) [C/Fe] >
+1, Carbon-enhanced EMP (CEMP) stars
(4 < [Fe/H] < 3, e.g., CS 2294937, Depagne et al. 2002); (3) [C/Fe] +4, hyper
39
Ne Mg
Si
Ar
Ca
Ti
Cr
Fe
Ni
Zn
EMP stars
CS22949-37
2
1
10
Fe-decomposition region
GR instability region
9.5
Na
Al
Cl
Sc
Mn Co Cu Ga
e+e- pair
instability region
5
C
Ne Mg
Si
Ar
Ca
Ti
Cr
Fe
Ni
Zn
HE1327-2326
HE0107-5240
log Tc (K)
[X/Fe]
-1
GR instability
region
8.5
00
10
YII (40M )
M-2 (137M )
-1
B
Na
Al
Cl
Sc
Mn Co Cu Ga
7.5
10
15
20
25
30
log c (g cm-3)
Figure 7. (Left): A comparison of the abundance patterns between metal-poor stars and models
(Tominaga et al. 2007). Upper: typical EMP stars (red dots, Cayrel et al. 2004) and CEMP (blue
triangles, CS 2294937, Depagne et al. 2002) and models with E dep,51 = 120 (solid line) and
= 3.0 (dashed line). Lower: HMP stars: HE 13272326, (red dots, e.g., Frebel et al. 2005), and
HE 01075240, (blue triangles, Christlieb et al. 2002, Bessell & Christlieb 2005) and models with
E dep,51 = 1.5 (solid line) and = 0.5 (dashed line).
Figure 8. (Right): Evolutionary tracks of the central temperature and central density of very
massive stars (Ohkubo et al. 2009). The numbers in brackets are the final masses for models
YII and M-2. The 1000M stars (Ohkubo et al. 2006) are also shown.
metal-poor (HMP) stars ([Fe/H] < 5, e.g., HE 01075240, Christlieb et al. 2002; Bessell
& Christlieb 2005; HE 13272326, Frebel et al. 2005).
Figure 7 shows that the abundance patterns of the averaged normal EMP stars, the
CEMP star CS 2294937, and the two HMP stars (HE 01075240 and HE 13272326)
are well reproduced by the models with E dep,51 = 120, 3.0, 1.5, and 0.5, respectively.
The model for the normal EMP stars ejects M (56 Ni) 0.2M , i.e., a factor of 2 less
than SN 1998bw. On the other hand, the models for the CEMP and the HMP stars eject
M (56 Ni) 8 104 M and 4 106 M , respectively.
To summarize, (1) the explosions with large energy deposition rate, E dep , are observed
as GRB-HNe, and their yields can explain the abundances of normal EMP stars, and
(2) the explosions with small E dep are observed as GRBs without bright SNe and can be
responsible for the formation of the CEMP and the HMP stars. We thus propose that
GRB-HNe and GRBs without bright SNe belong to a continuous series of BH-forming
massive stellar deaths with relativistic jets of different E dep .
40
-21
lcv.du 1:3
Young et al. (2009) bolometric
Gal-Yam et al. (2009) R
absolute bolometric magnitude
absolute magnitude
-20
no mixing
PISN 270Msun
SN 2007bi
SN 2006gy (Y09)
SN 2006gy (Agnoletto+ 2009)
SN 2006gy (Smith+ 2007) R
-20
-19
-18
-17
-16
-19
-18
-17
-16
-15
-15
0
100
200
300
400
500
-14
-100
100
200
300
400
Figure 9. (Left): The bolometric light curve of the C+O core SN models
(Mej = 39 M , Ekin = 3.3 1052 erg, and M (56 Ni) = 6.1 M ) compared with the
bolometric LC of SN 2007bi (Moriya et al. 2009).
Figure 10. (Right): The PISN model (M = 270M ) for the LC of SNe 2006gy and 2007bi
(Moriya et al. 2009).
Nomoto 2008; Ohkubo et al. 2009). In the extreme case, the pulsation could induce
dynamical mass ejection and optical brightening as might be observed in the brightest
SN 2006gy (Woosley et al. 2007).
If the explosion energy is large enough, the mass of 56 Ni can be as large as 6M
(Umeda & Nomoto 2008). The resultant light curve can be consistent with Luminous
Supernovae such as SNe 2006gy and 2007bi (Figs. 9 and 10: Moriya et al. 2009).
8. Very Massive Stars (> 300M ) and Intermediate Mass Black Holes
It is possible that the First Stars were even more massive than 300M, if rapid mass
accretion continues during the whole main-sequence phase of Pop III stars (Ohkubo et al.
2006, 2008).
Such massive stars undergo core-collapse (CVMS: core-collapse VMS) as seen from
the 1000 M star track in Figure 8. If such stars formed rapidly rotating black holes,
jet-like mass ejection could produce processed material (Ohkubo et al. 2006). In fact, for
moderately aspherical explosions, the patterns of nucleosynthesis match the observational
data of both intracluster medium and M82 (Ohkubo et al. 2006).
It is also possible that LSNe 2006gy and/or 2007bi can be the explosion of the above
CVMS.
41
This result suggests that explosions of CVMS contribute significantly to the chemical
evolution of gases in clusters of galaxies. This result may support the view that Pop III
CVMS could be responsible for the origin of intermediate mass black holes (IMBH).
References
Arnett, W. D. 1996, Supernovae and Nucleosynthesis (Princeton: Princeton Univ. Press)
Barkat, Z., Rakavy, G., & Sack, N. 1967, PRL 18, 379
Bessell, M. S., & Christlieb, N. 2005, in V. Hill et al. (eds.), From Lithium to Uranium, Proc.
IAU Symposium No. 228 (Cambridge: Cambridge Univ. Press), 237
Cayrel, R., et al. 2004, A&A 416, 1117
Christlieb, N., et al. 2002, Nature 419, 904
Depagne, E., et al. 2002, A&A 390, 187
Foley, R. J., et al. 2009, AJ 138, 376
Frebel, A., et al. 2005, Nature 434, 871
Fryer, C., et al. 2009, arXiv:0908.0701
Gal-Yam, A., et al. 2009, Nature 462, 624
Heger, A., & Woosley, S.E. 2002, ApJ 567, 532
Heger, A., & Woosley, S.E. 2008, ApJ submitted (arXiv:0803.3161)
Iwamoto, K., Mazzali, P.A., Nomoto, K., et al. 1998, Nature 395, 672
Iwamoto, N., Umeda, H., Tominaga, N., Nomoto, K., & Maeda, K. 2005, Science 309, 451
Kawabata, K., Maeda, K., Nomoto, K., et al. 2009, arXiv:0906.4573
Kitaura, F.S., Janka, H.-Th., & Hillebrandt, W. 2006, A&A 450, 345
Koch, A., et al. 2008, ApJ 688, L13
Kobayashi, C., Umeda, H., Nomoto, K., Tominaga, N., & Ohkubo, T. 2006, ApJ 653, 1145
Limongi, M., Straniero, & Chieffi, A. 2000, ApJS 129, 625
Maeda, K., & Nomoto, K. 2003, ApJ 598, 1163
Moriya, T., et al. 2009, ApJ submitted
Nomoto, K. 1984, ApJ 277, 791
Nomoto, K. & Hashimoto, M. 1988, Phys. Rep. 163, 13
Nomoto, K., et al. 2005, in The Fate of Most Massive Stars, ed. R. Humphreys & K. Stanek
(ASP Ser. 332), 374 (astro-ph/0506597)
Nomoto, K., et al. 2006, Nuclear Phys A 777, 424 (astro-ph/0605725)
Nomoto, K., et al. 2009, in IAU Symp. 254, The Galaxy Disk in Cosmological Context, ed. J.
Andersen, et al. (Cambridge: Cambridge Univ. Press), 355 (arXiv: 0901.4536)
Ohkubo, T., Umeda, H., Maeda, K., Nomoto, K., Suzuki, T., Tsuruta, S., & Rees, M. J. 2006,
ApJ 645, 1352
Ohkubo, T., Nomoto, K., Umeda, H., Yoshida, N., & Tsuruta, S. 2009, ApJ 706, 1184
Prieto, J. L., et al. 2008, ApJ 681, L9
Pumo, M. L., et al. 2009, ApJ 705, L138
Smartt, S. J. 2009, ARA&A 47, 63
Thompson, T. A., et al. 2008, ApJ submitted (arXiv:0809.0510)
Tominaga, N., Maeda, K., Umeda, H., Nomoto, K., Tanaka, et al. 2007, ApJ 657, L77
Tominaga, N. 2009, ApJ 690, 526
Umeda, H., & Nomoto, K. 2002, ApJ 565, 385
Umeda, H., & Nomoto, K. 2008, ApJ 673, 1014
Valenti, S., et al. 2009, Nature 459, 674
Wanajo, S., Nomoto, K., Janka, H.-T., Kitaura, F. S., & M
uller, B. 2009, ApJ 695, 208
Woosley, S. E., Blinnikov, S., & Heger, A. 2007, Nature 450, 390
Young, D. R., et al. 2009, arXiv:0910.2248
Yoshida, N., Omukai, K., & Hernquist, L. 2008, Science 321, 669
1. Introduction
Primordial supernovae (SNe) polluted the early universe with the first metals by z
20 - 30. Although numerical models suggest that the first stars were 25 - 500 M (Bromm
et al. 1999, Bromm et al. 2002, Nakamura & Umemura 2001, Abel et al. 2000, Abel et
al. 2002, OShea & Norman 2007), the Pop III initial mass function (IMF) is yet to be
observationally constrained. The chemical imprint of Pop III SNe on low-mass secondgeneration stars may yield indirect measures of the primordial IMF. Remnants of the
second generation are being sought in surveys of ancient extremely metal-poor (EMP)
and hyper-metal poor (HMP) stars in the Galactic halo (Beers & Christlieb 2005, Frebel
et al. 2005). HMP stars ([Fe/H] < 4) are thought to be enriched by only one or a
few SNe. EMP stars (4 < [Fe/H]< 3) are more common and exhibit less scatter in
their abundance ratios, implying that they form from gas enriched by a sample of SNe
progenitors with a well-defined IMF.
Stellar evolution models suggest that 15 - 40 M primordial stars die in core-collapse
SNe and that 140 - 260 M stars explode in pair-instability supernovae (PISN) with
up to a hundred times more energy (Heger & Woosley 2002). Stars from 40 - 60 M
may end their lives as hypernovae (HNe), with energies in between those of conventional
SNe and PISN (Nomoto et al. 2006). To date, numerical attempts to reconcile elemental
42
43
abundances from these explosions with those measured in metal-poor stars have met
with only limited success (Umeda & Nomoto 2002, Umeda & Nomoto 2003, Umeda &
Nomoto 2005, Tominaga et al. 2007, Tominaga 2009, Joggerst 2009). In particular, the
odd-even nucleosynthetic yields predicted for PISN have yet to be detected in HMP or
EMP stars. Furthermore, HMP stars exhibit skewed [C/Fe] ratios, implying that carbon
abundances in some stars were enhanced either by a binary companion or by formation
in chemically stratified environments. These observations highlight the difficulty with
directly equating elemental yields from one generation with the chemical abundances of
the next: intervening hydrodynamical processes such as mixing within the SN explosion
itself or mass transfer between two stars may complicate the uptake of elements between
generations.
44
Figure 2. IMF averages of yields compared to observations of EMP stars from Cayrel et al.
(2004) and Lai et al. (2008). Higher-explosion energy rotating Z = 0 stars reproduce EMP
abundances well if 15 M stars are included.
45
allows RT instabilities more time to develop and promotes mixing between adjacent
shells. The delay also enhances fallback of Fe and Ni onto the compact remnant, in
some cases preventing any of it from escaping into the IGM. Because the structure
is relatively insensitive to the magnitude of the rotation once it is present, so is the
degree of mixing and fallback. Less mixing occurs in our Z = 104 Z models because
the progenitors expand less with rotation. As expected, fallback decreases and mixing
increases as explosion energy rises, but at a given energy fallback increases and mixing
decreases with increasing progenitor mass. We show the former two trends for 15 M
Z = 0 and 104 Z stars in Figure 1.
3. Conclusion
One of our models, a zero-metallicity 15 M 2.4 B SN, reproduces the abundances of
the HMP star HE0557-4840 well, but none of the others match HE0107-5240 or HE13272326 exactly, which have higher C, N, O, and Na abundances and can be explained by
models with more exotic explosion mechanisms. Simulations of jet-driven explosions in
40 M Pop III stars have successfully reproduced the abundances of HE0107-5240 and
HE1327-2326 (Tominaga et al. 2007). However, as we show in Figure 2, a Salpeter IMF
average of the elemental yields of our models are a good match to the abundances found
in the much larger sample of EMP stars in the Cayrel et al. (2004) and Lai et al. (2008)
surveys. These abundances likely originated from a population of early SN progenitors
with a well-defined IMF that was built up over a relatively narrow range of redshifts, and
are therefore more representative of primordial SNe than the one or two singular events
that imprinted HMP stars. Our models suggest that rotation and mixing in 15 - 40 M
primordial SNe can account for the majority of metal production in the early universe.
References
Abel, T., Bryan, G. L., & Norman, M. L. 2000, ApJ, 540, 39
Abel, T., Bryan, G. L., & Norman, M. L. 2002, Science, 295, 93
Beers, T. C., & Christlieb, N. 2005, ARAA, 43, 531
Bromm, V., Coppi, P. S., & Larson, R. B. 1999, ApJL, 527, L5
Bromm, V., Coppi, P. S., & Larson, R. B. 2002, ApJ, 564, 23
Cayrel, R. et al 2004, A&A, 416, 1117
Frebel, A., et al. 2005, Nature, 434, 871
Heger, A., & Woosley, S. E. 2002, ApJ, 567, 532
Heger, A. & Woosley, S. E. 2009, ArXiv e-prints
Joggerst, C. C., Woosley, S. E., & Heger, A. 2009, ApJ, 693, 1780
Joggerst, C. C., Almgren, A., Bell, J., Heger, A., Whalen, D., & Woosley, S. E. 2009,
arXiv:0907.3885
Lai, D. K. et al 2008, ApJ, 681, 1524
Nakamura, F., & Umemura, M. 2001, ApJ, 548, 19
Nomoto, K. et al. 2006, Nuclear Physics A, 777, 424
OShea, B. W., & Norman, M. L. 2007, ApJ, 654, 66
Tominaga, N., Umeda, H., & Nomoto, K. 2007, ApJ, 660, 516
Tominaga, N. 2009, ApJ, 690, 526
Umeda, H. & Nomoto, K. 2002, ApJ, 565, 385
Umeda, H. & Nomoto, K. 2003, Nature, 422, 871
Umeda, H. & Nomoto, K. 2005, ApJ, 619, 427
1. Introduction
Stars create most of the lighter elements in charged-particle fusion reactions. The major
isotopes of heavier elements, here defined as those with atomic numbers Z > 30, must
be created in reactions of the type A Z + n A+1 Z, followed by nuclear transformations
of the type A Z A Z+1 + . Collectively, the content of these so-called neutroncapture (n-capture) elements is negligible. In the solar system for example, the most
abundant n-capture element germanium (Z = 32) is about 10,000 times less abundant
than iron, and most other n-capture elements are orders of magnitude less abundant than
germanium. In spite of their small abundances, the n-capture elements are important in
the quest to understand Galactic chemical evolution. Nearly 20 presentations at this
Symposium prominently feature observations and/or interpretations of the n-capture
elements. Here we comment of some of the unique aspects of these elements at low
metallicity. A comprehensive overview is not possible here; we highlight some issues that
are discussed at length in the recent review by Sneden, Cowan, & Gallino (2008).
The creation of n-capture elements is easily (and probably too neatly) divided into
two extreme cases of neutron ingestion and -decay timescales. If n >> then unstable nuclei will have time to execute all -decays between successive neutron captures by
target nuclei, and element production proceeds along the valley of stability. This relatively slow nucleosynthesis mechanism is labeled the s-process. It normally is a byproduct
of quiescent He-burning, most easily accomplished during the asymptotic giant branch
(AGB) phase of low-intermediate mass stars. At the opposite extreme, if n << , target
46
47
nuclei are (for a few seconds at most) overwhelmed with neutrons, ingesting them out to
the neutron drip line. When the neutron flood stops, decays occur until stability is
reached. This rapid-burst mechanism is labeled the r-process. The site of the r-process
is still uncertain, but must be associated with the explosive deaths of high-mass stars.
The r- and s-processes yield different, easily distinguished abundance distributions in
low metallicity stars.
2. r-Process Stars
Solar-system n-capture abundances result from many generations of stellar element
donors of a large mass range. Not surprisingly, they are a complex mix of r- and sprocess abundance contributions, as first quantitatively estimated by Cameron (1982),
and further developed with data improvements by, e.g., Arlandini et al. (1999) and Simmerer et al. (2004). Only a few prior nucleosynthesis events could have contributed to
the n-capture abundances of very metal-poor stars. This is manifest first in their large
star-to-star bulk variations, i.e., the n-capture/Fe values can range over a factor of 1000
at the same metallicity. The detailed abundance distributions also vary, from r-processdominant, to s-process-dominant, and to many intermediate mixes.
1.00
BD+173248
CS 22892052
CS 31082001
HD 115444
HD 221170
SS rprocess (Simmerer)
SS rprocess (Arlandini)
Ba
Relative log
0.60
0.20
Nd
Ce
Dy
Gd
Er
Sm
Yb
0.20
Hf
0.60
La
Eu
1.00
Ho
Pr
Tb
1.40
55
60
65
Tm
70
75
Atomic Number
Figure 1. Figure 8 of Sneden et al. (2009): rare-earth abundances in five r-rich metal-poor
stars (points), and two estimates of the solar-system r-only abundance distribution (lines). All
abundance sets have been normalized to agree at Eu. See Sneden et al. (2009) for further details.
Enrichment of metal-poor stars by the r-process has been known since the pioneering
work of Spite & Spite (1978) and Truran (1981). Abundance studies of halo giant stars
HD 115444 (Griffin et al. 1982) and HD 122563 (Sneden & Parthasarathy 1983) strengthened the r-process interpretation. Recent studies have demonstrated in excruciating detail
the near perfect match between the n-capture abundances of many metal-poor stars and
the solar-system r-process-only distribution, especially at the heavy end (Z > 56). We
illustrate this agreement for the rare-earth elements in Figure 1; star-to-star scatter in
the relative abundances in r-rich stars is now at the 0.05 dex level.
48
Figure 2. A Periodic Table highlighting the useful n-capture elements, defined here as ones that
are accessible in metal-poor stellar spectra and that generally have reliable laboratory transition
data.
A prime factor in the increased n-capture abundance precision has been the resurgence
of laboratory atomic physics effort on neutral and first ionized species of these elements.
In Figure 2 we show a Periodic Table that calls attention to those n-capture elements
that have detectable lines in high resolution U V and visible spectra of metal-poor stars.
Nearly all of these elements have enjoyed lab analyses since the early 1980s that have
resulted in transition probabilities of large numbers of lines to better than 10% accuracy,
and in hyperfine and isotopic structure information that largely meet the needs of stellar
spectroscopists. Many investigators have contributed to this work, and in particular credit
is due to groups in Li`ege and Wisconsin (especially for the lighter and rare-earth ncapture elements), and Lund (especially for the heaviest elements). Uncertainties in basic
transition data have now become fairly minor contributors to the overall error budgets
of n-capture abundances in metal-poor stars.
Given the early and repeated attention to HD 122563 and HD 115444, here we consider
this pair instead of the more exotic extremely rrich stars such as CS 31082-001 (Hill et
al. 2002) or CS 22892-052 (Sneden et al. 2003). HD 122563 and HD 115444 seem to have
typical r-process patterns, characterized by, e.g, log (Ba/Eu) +1.0 compared with
the solar-system r + s total ratio log (Ba/Eu) +1.7. But while their lighter n-capture
abundances are similar (e.g., log (Sr) +0.2 in HD 115444 and 0.1 in HD 122563),
their rare-earth contents differ by more than an order of magnitude (e.g. log (Eu) 1.6
and 2.8, respectively). In Figure 3 we illustrate this contrast by plotting the abundances
normalized to Sr. The discrepancy grows with increasing atomic number, as clearly shown
by Honda et al. (2006, 2007) in a comparison between HD 122563 (and the similar
HD 88609) and CS 22892-052.
Not all n-capture elements will cooperate by having accessible transitions in the visible
spectral region. Further insights must rely on gathering high-resolution spectra in the
U V , and the Hubble Space Telescope STIS is the instrument of choice. In Figure 4 we
continue the HD 115444 and HD 122563 comparison with the spectra of two prominent
r-process lines. The Eu II line contrast in the left-hand panel reinforces the large abundance difference for this element discussed previously. The Pt I line strength contrast
in the right-hand panel is just as large; the line in HD 122563 cannot be reliably de-
49
Figure 3. n-capture abundances in HD 115444 (Westin et al. 2000, Sneden et al. 2009) and
HD 122563 (Honda et al. 2007), normalized to the Sr abundances for each star.
tected. But the abundance analyses of this and most other n-capture lines that occur
at short wavelengths ( < 4000
A) is difficult due to line contamination that increases
with decreasing wavelength. Certainly caution is warranted in interpretation of n-capture
abundances determined from a single transition in the U V .
Figure 4. n-capture lines in HD 115444 and HD 122563. The visible-wavelength spectra (left
hand panel) were obtained with the Keck I HIRES, and the U V spectra were obtained with the
HST STIS. Cowan et al. (2005) has details about these spectra and their analyses.
Concerns about the reliability of individual n-capture abundances cannot obscure the
general trends seen in the HST spectra. In Figure 5 we summarize the abundances for
11 r-rich giants from Cowan et al. (2005). The correlations are with Europium, the traditional r-process indicator. Relative [X/H] values are used instead of log in order
to inter-compare the abundance enhancements of different elements. In the left-hand
panel we plot the abundances of Ge, Zr, and La against Eu. While [La/H] tracks [Eu/H]
extremely well, the trends with Zr and especially Ge are poor. This is in line with the differences between lighter and heavier n-capture abundances of HD 115444 and HD 122563
(Figure 3). Much cleaner correlations are shown between the 3rd r-process-peak elements
50
Os, Ir, and Pt with Eu (Figure 5, right-hand panel. This panel also shows the La abundance with an addition of 0.45 dex. This offset corrects it for the solar-system s-process
fraction, so that [La/H] + 0.45 = [La/H]ronly . The [Laronly ,Os,Ir,Pt/H] abundance
trends with [Eu/H] are all consistent with a single, nearly one-to-one relationship.
Abundance comparisons like those shown in Figure 5 have lent support to suggestions
for multiple astrophysical nucleosynthesis processes. It seems clear that the heaviest ncapture elements , i.e., Ba and above, are correlated, and are consistent with a (scaled)
Solar System r-process abundance pattern. This r-process pattern exists in a very large
number of r-process-rich stars (see Sneden et al. 2008, and references therein). On the
other hand the abundances of the lighter n-capture elements in these same stars are
not correlated thus, the large scatter of the abundances of Ge and Zr with respect to
Eu in the left-hand-panel of Figure 5. Much recent work has explored various scenarios
to explain these lighter n-capture elemental abundances. There have been a number of
suggestions that it will require at least two r-processes a main one for the heavier
n-capture elements and a second, or weak r-process for the lighter n-capture elements
from Z=40 to 6 50 (Kratz et al. 2007).
The abundances of the elements Sr-Y-Zr also seem to require multiple synthesis sources,
suggesting the possibility of a Light Element Primary Process (LEPP), which could
result from either neutrons or charged particles (Travaglio et al. 2004). A combination
of the main r-process and the LEPP could then explain the abundances of the lighter
n-capture elements (Montes et al. 2007). The Ge abundances of Cowan et al. (2005)
see Figure 5 do not correlate with r-process abundances, and instead scale with
the metallicities of stars, at least for [Fe/H] 6 2. From this, Fr
ohlich et al. (2006) have
suggested a new synthesis origin for Ge early in the history of the Galaxy: the -p process,
involving neutrino scattering in supernovae that could produce protons to be captured
on Fe nuclei. Finally, very recent work suggests that various conditions (i.e., entropies) in
the high entropy wind of supernovae might be able to reproduce both the heavy and light
n-capture abundances (Farouqi et al. 2009). Clearly more observational and theoretical
studies will be required to discern the origin of these various n-capture elemental regimes.
51
3. s-Process Stars
A large fraction (20%) of halo stars with metallicity [Fe/H] < 2 are C-rich, [C/Fe] > 1
(e.g., Beers & Christlieb 2005; Lucatello et al. 2006). Such stars are called Carbon Enriched Metal Poor (CEMP) stars. High resolution spectroscopic studies further divide
CEMP stars into four subclasses: CEMP-s, enhanced in s-process elements; CEMP-sr,
enhanced in both s-process and r-process elements; CEMP-r, enhanced in r-process elements, and CEMP-no, with no n-capture element overabundances. The majority of
CEMP stars are enhanced in s-process abundances (both CEMP-s or CEMP-sr).
Heavy s-process nuclei (YPb) are synthesized by a long-lived low mass (1.35 M )
star during its late Asymptotic Giant Branch (AGB) phases, where the star undergoes
a series of recurrent Thermal Pulse (TP) instabilities in the He shell. The TP-AGB
phase lasts 106 years, while the envelope, enriched in carbon and s-process elements, is
progressively lost by intense stellar winds (see Sneden et al. 2008 and references therein).
The contribution of s-process elements to the Galactic interstellar medium thus can
occur only later in the halo or in the disk, not early in the halo, at [Fe/H] < 2. Nearly
all CEMP-s stars belong to close binary systems, where the observed star, a dwarf or a
giant far from the TP-AGB phase, of mass 0.9 M , became enriched in C and s-process
elements by mass transfer from the winds of the more massive companion while on the
TP-AGB phase (which now is presumably an unseen white dwarf).
Elements mostly fed by the s-process are grouped in three peaks around neutron magic
numbers: the light peak (ls: Sr, Y, Zr) at N = 50; the heavy peak (hs: Ba. La, Pr, Ce,
Nd) at N = 82; and Pb at N = 126, at the termination of the s-process path. The
relative elemental abundances of the three s-process peaks mainly depends on the initial
Fe content (the seed for the s-process) and the strength of the major source of neutrons,
driven by the 13 C(,n)16 O reaction. A small number of protons are assumed to penetrate
from the bottom of the convective envelope into the radiative He-rich and C-rich inner
zone during a third dredge up (TDU) episode. Production of 13 C proceeds from proton
capture on 12 C at H re-ignition (13 C pocket; Bisterzo et al. 2009). A spread in 13 Cpocket efficiencies, defined as a multiple/fraction of the ST case that reproduces the
solar-system main s-process component, is needed in order to account for different sprocess distributions observed at a given metallicity (Bisterzo et al. 2009b, Husti et al.
2009). The 13 C neutron source in principle does not depend on metallicity, being the
result of proton capture on newly synthesized 12 C. Consequently, at low metallicities the
number of neutrons captured by Fe seeds favors the production of very heavy Pb at the
termination of the s-process path with respect to ls and hs abundances.
Here we give three examples of how the detailed observed abundance patterns can be
used to constrain the parameters of C-rich, s-rich binary stars Figure 6 compares an sprocess distribution generated from an AGB model of initial mass 1.5 M generated with
the FRANEC code (Chieffi & Straniero 1989), to observed abundances of the CEMP-s
star HE 2158-0348 Cohen et al. (2006). For this star the observed [La/Eu] =0.8 is in
agreement with s-process predictions. La is mostly contributed by the s-process (70%
in the solar-system), whereas Eu is mostly contributed by the r-process (95% in the
solar system). An initial [Eu/Fe] = 0.5 has been assumed, consistent with the average enhancement of Eu in unpolluted low metallicity stars. The agreement between theoretical
predictions and observations are obtained after diluting one part of the AGB envelope
over 20 parts of the envelope of the envelope of the present star (dil = 1.35 dex).
Figure 7 compares s-process predictions with abundances in the CEMP-sr star CS
22898-027. The initial AGB mass has been assumed to be 1.3. The exceptional [Eu/Fe] 2,
and similarly large enhancements of other hs elements (e.g., [La/Eu] 0) can only be
52
MAGBini
ini
~ 1.5 Mo (n20)
Pb
[El/Fe]
Sm
Ce
La
Ba
Zr
Nd
N
Mg
Ca
Eu
Ti
Sr
Co
Mn
Al
Cr
case ST/6
-1
0
10
20
30
40
50
60
70
80
90
Atomic Number
Figure 6. Observed HE 2158-0348 abundances compared with the best-fit model abundance
distribution (computed with mass 1.5 M , 13 C pocket efficiency ST/6, and dilution factor
dil = 1.35 dex.
CS 22898-027
[El/Fe]
Pb
Ba
La Nd
Ce Sm
Eu
[Fe/H] = -2.26
Er
2
C
Sr Zr
Y
Dy
1
CaTi
Ni
Cr
Mg
Na
Al
Mn
-1
0
10
20
30
40
50
60
70
80
90
Atomic Number
Figure 7. Observed and best-fit model abundance distributions for CS 22898-027.
reconciled by assuming that the initial cloud from which the binary system formed was
already locally polluted by ejecta from a very r-process-rich Type II SN.
Figure 8 shows the theoretical prediction of Pb for the CEMP-s star HE 1135+0139. All
predicted n-capture element abundances are in excellent agreement with the observation
by Barklem et al. (2005). We emphasize that the observed abundances in these CEMP
stars, such as those illustrated here, can be reproduced by models with relatively few
parameters and under reasonable physical conditions.
Acknowledgements
We thank the friends and colleagues who have participated in the studies that inform
our review, and S. Bisterzo for providing several s-process figures. This work has been
supported in part by the National Science Foundation through grants AST-0607708 to
CS and AST-0707447 to JJC, and from the Italian MIUR-PRIN06 Project Late phases
of Stellar Evolution: Nucleosynthesis in Supernovae, AGB stars, Planetary Nebulae.
53
MAGBini
ini
~ 1.4 - 1.5 Mo
[r/Fe] = 0.0
[El/Fe]
2
Ba
Ce
Nd
La
Sr
Mg
Zr
Y
Ca Ti
Sc
Eu
Co
0
Al
V
Cr
Ni
Mn
-1
0
10
20
30
50
60
70
80
90
Atomic Number
Figure 8. Observed Observed and best-fit model abundance distributions for HE 1135+0139.
References
Aoki, W. et al. 2007, ApJ 655, 492
Aoki, W., Norris, J. E., Ryan, S. G., Beers, T. C., & Ando, H. 2002a, PASJ 54, 933
Aoki, W. et al. 2002, ApJ 580, 1149
Arlandini, C. et al. 1999, ApJ 525, 886
Barklem, P. S., et al. 2005, A&A 439, 129
Beers, T. C., & Christlieb, N. 2005, ARAA 43, 531
Bisterzo, S., Gallino, R., Straniero, O., & Cristallo, S. 2009, MNRAS in press
Bisterzo, S., Gallino, R., Straniero, O., & Aoki, W. 2009, PASA 26, 314
Cameron, A. G. W. 1973, Sp. Sci. Rev. 15, 121
Chieffi, A., & Straniero, O. 1989, ApJS 71, 47
Cowan, J. J., et al. 2005, ApJ 627, 238
Farouqi, K. et al. 2009, ApJ 694, L49
Fr
ohlich, C. et al. 2006, Phys. Rev. Lett. 96 142502
Griffin, R., Gustafsson, B., Vieira, T., & Griffin, R. 1982, MNRAS 198, 637
Hill, V., et al. 2002, A&A 387, 560
Honda, S., Aoki, W., Ishimaru, Y., & Wanajo, S. 2007, ApJ 666, 1189
Honda, S., Aoki, W., Ishimaru, Y., Wanajo, S., & Ryan, S. G. 2006, ApJ 643, 1180
Husti, L., Gallino, R., Bisterzo, S., Straniero, O., & Cristallo, S. 2009, PASA 26, 176
Kratz, K.-L. et al. 2007, ApJ 662, 39
Lucatello, S., et al. 2006, ApJ 652, L37
Montes, F., et al. 2007, ApJ 671, 1685
Simmerer, J. et al. 2004, ApJ 617, 1091
Sneden, C., Cowan, J. J., & Gallino, R. 2008, ARAA 46, 241
Sneden, C., et al. 2003, ApJ 591, 936
Sneden, C., Lawler, J. E., Cowan, J. J., Ivans, I. I., & Den Hartog, E. A. 2009, ApJ 182, 80
Sneden, C., & Parthasarathy, M. 1983, ApJ 267, 757
Spite, M., & Spite, F. 1978, A&A 67, 23
Travaglio, C. et al. 2004, ApJ 601, 864
Truran, J. W. 1981, A&A 97, 391
Westin, J., Sneden, C., Gustafsson, B., & Cowan, J. J. 2000, ApJ 530, 783
Globular clusters continue to provide ideal laboratories for testing the predictions
of stellar evolution (e.g., Gratton et al. 2004). With the exception of Centauri and
M15 (e.g., Norris & Da Costa 1995, Sneden et al. 2000), in general only a handful of
neutron-capture elements have been measured in globular clusters. For example, the sprocess is usually represented by measurements of Y, Zr, Ba and/or La whereas Eu
measurements are often the sole indicator of the r-process. Possible explanations for the
lack of abundance measurements for additional neutron-capture elements include limited
wavelength coverage and signal-to-noise ratios.
M4 and M5 are particularly well suited for refining our understanding of stellar nucleosynthesis of the s-process elements. This pair of clusters was extensively studied by
Ivans et al. (1999, 2001) who showed that they have essentially identical metallicities,
[Fe/H] = 1.2, and very similar abundance ratios [X/Fe] for many - and Fe-peak elements. However, the s-process elements Ba and La are overabundant in M4 relative to
M5. We have obtained very high quality spectra (R=55,000 and S/N=800 per resolution
element at 8000
A) using the MIKE spectrograph (Bernstein et al. 2003) on the Magellan
Telescope of 12 bright giants in M4 and 2 bright giants in M5 to further probe the relative
abundances of neutron-capture elements in these two clusters.
Using standard analysis techniques, we measured the relative abundances of some 16
neutron-capture elements from Rb to Th in both clusters. This comprehensive analysis
is only possible due to the large wavelength coverage, high spectral resolution, and high
S/N ratios. We confirm that all s-process elements are overabundant in M4 relative to
M5 and find that the r-process elements may be slightly overabundant in M5 relative to
M4 (Yong et al. 2008a, 2008b).
Since there is no detectable star-to-star abundance dispersion for any neutron-capture
element, we suppose that the abundance differences between these two clusters were
present between the clusters natal clouds. Relative to M5, the M4 stars formed from gas
enriched in s-process products but deficient in r-process material. This conjecture may
54
55
Figure 1. Abundance ratios [X/Fe] for the elements from Rb to Th in M5 (upper), M4 (middle),
and M13 (lower). We show the best-fit predictions to the elements from Ba to Hf (filled circles)
using scaling factors s and r which were multiplied by the solar s-process and r-process
abundances, respectively.
56
D. Yong et al.
elements, both clusters can be well explained by simply taking linear combinations of
scaled versions of the solar s- and r-process material.
For M13, existing measurements (Yong et al. 2006a, 2006b) were supplemented with
additional neutron-capture elements. Based on these preliminary measurements (made
with the assistance of two summer research scholars, Benjamin Sparkes and Andrew
Cameron), we again find that we are able to fit the measurements using scaled versions
of the solar s- and r-process abundances. However, for all three clusters, the lighter
elements (Rb-Mo) are not well matched by the predictions.
We have additional observations that will allow us to perform a similar dissection of the
neutron-capture abundances for other clusters and field stars. To complement these new
measurements, it is important to extend the stellar yields (e.g., Chieffi & Limongi 2003,
Karakas & Lattanzio 2007) and chemical evolution models (e.g., Travaglio et al. 2004,
Karakas et al. 2006, Marcolini et al. 2009) to include a larger suite of neutron-capture
elements if possible.
References
Bernstein, R., Shectman, S. A., Gunnels, S. M., Mochnacki, S., & Athey, A. E., 2003, in Proc.
SPIE, 4841, 1694
Chieffi, A., & Limongi, M., 2004, ApJ, 608, 405
Gratton, R., Sneden, C., & Carretta, E., 2004, ARA&A, 42, 385
Ivans, I. I., Sneden, C., Kraft, R. P., Suntzeff, N. B., Smith, V. V., Langer, G. E., & Fulbright,
J. P., 1999, AJ, 118, 1273
Ivans, I. I., Kraft, R. P., Sneden, C., Smith, G. H., Rich, R. M., & Shetrone, M., 2001, AJ, 122,
1438
Karakas, A. I., Fenner, Y., Sills, A., Campbell, S. W., & Lattanzio, J. C., 2006, ApJ, 652, 1240
Karakas, A., & Lattanzio, J. C., 2007, PASA, 24, 103
Marcolini, A., Gibson, B. K., Karakas, A. I., & S
anchez-Bl
azquez, P., 2009, MNRAS, 395, 719
Norris, J. E., & Da Costa, G. S., 1995, ApJ, 447, 680
Simmerer, J., Sneden, C., Cowan, J. J., Collier, J., Woolf, V. M., & Lawler, J. E., 2004, ApJ,
617, 1091
Sneden, C., Johnson, J., Kraft, R. P., Smith, G. H., Cowan, J. J., & Bolte, M. S., 2000, ApJ,
536, L85
Travaglio, C., Gallino, R., Arnone, E., Cowan, J., Jordan, F., & Sneden, C., 2004, ApJ, 601, 864
Yong, D., Aoki, W., & Lambert, D. L., 2006, ApJ, 638, 1018
Yong, D., Aoki, W., Lambert, D. L., & Paulson, D. B., 2006, ApJ, 639, 918
Yong, D., Lambert, D. L., Paulson, D. B., & Carney, B. W., 2008, ApJ, 673, 854
Yong, D., Karakas, A. I., Lambert, D. L., Chieffi, A., & Limongi, M., 2008, ApJ, 689, 1031
Mount Stromlo Observatory, Australian National University, Weston Creek ACT 2611,
Australia
email: akarakas@mso.anu.edu.au
2
Centre for Stellar and Planetary Astrophysics, Monash University, Clayton 3800, Victoria,
Australia
email: Maria.Lugaro@sci.monash.edu.au
3
Astronomy and Astrophysics Group, Department Fisica i Enginyeria Nuclear, Universitat
Politecnica de Catalunya, Barcelona, Spain
email: simwcampbell@gmail.com
Abstract. Elements heavier than iron are produced in asymptotic giant branch (AGB) stars via
the slow neutron capture process (s process). Recent observations of s-process-enriched Carbon
Enhanced Metal-Poor (CEMP) stars have provided an unprecedented wealth of observational
constraints on the operation of the s-process in low-metallicity AGB stars. We present new
preliminary full network calculations of low-metallicity AGB stars, including a comparison to
the composition of a few s-process rich CEMP stars. We also discuss the possibility of using
halo planetary nebulae as further probes of low-metallicity AGB nucleosynthesis.
Keywords. stars: abundances stars: AGB and post-AGB
1. Introduction
Asymptotic giant branch (AGB) stars play a crucial role in the chemical evolution
of elements heavier than iron. The operation of the s-process in AGB stars synthesized
up to half of all elements heavier than Fe in the solar system, with most of the Pb on
Earth coming from low-metallicity AGB stars. Heavy elements, such as Zr, Ba, and Pb,
are mixed from the deep interior to the stellar surface by recurrent mixing episodes.
Strong stellar winds then expel this enriched matter into the interstellar medium, thus
contributing to the chemical evolution of galaxies and stellar systems.
Detailed neutron-capture abundance information is now known for many extremely
metal-poor field and globular cluster stars in the Galactic halo. Many of the metalpoor field halo stars are also greatly enhanced in carbon, the Carbon Enhanced Metal
Poor (CEMP) stars, along with elements that can be produced by the slow neutron
capture process (the s process) and/or rapid neutron capture process (the r process; see
reviews by Beers & Christlieb 2005 and Sneden, Cowan, & Gallino 2008). That many
of the CEMP stars with s process enhancements (hereafter CEMP-s stars) are also in
long-period binary systems (Lucatello et al. 2005) has led to the hypothesis that these
stars obtained their carbon and heavy element abundances via mass transfer from a
now extinct asymptotic giant branch (AGB) companion. Comparison between CEMP-s
abundance data with theoretical stellar models of AGB stars has further supported this
hypothesis (e.g., Ivans et al. 2005).
In these proceedings, we report on new preliminary theoretical predictions of s process
nucleosynthesis from low-mass, low-metallicity AGB stars. These models are compared
to the composition of a couple of CEMP stars with enhancements of neutron-capture
57
58
elements. We also discuss the potential of using halo planetary nebulae as probes of
heavy element nucleosynthesis in low-mass AGB stars.
3. Comparison to CEMPS
In Figs. 1 and 2 we show the predicted [X/Fe] abundances versus atomic mass for the
1.5M and 2M , Z = 0.0001 models, with initial Ba and Eu abundances as indicated in
the figure caption. In Fig. 1 the mean abundances for the CEMP star HD 196944 from
the studies of Van eck et al. (2001) and Aoki et al. (2002) are shown. Overlaid on Fig. 2
are the derived abundance of the CEMP star CS 29497-030, with data from Ivans et
al. (2005). Observational errors are not included in either figure but are on the order of
0.1-0.3 dex depending on the element, see the references for further details.
In these figures we are not trying to provide a perfect match to the observational data,
but to show preliminary examples of the model predictions. In Fig. 1 the general pattern
of the observed distribution is matched by the model but all of the observed data points
lie at the lower end of the predicted values. This indicates that there was substantial
dilution of the AGB wind with the envelope of the accretor star. This is plausible as HD
196944 is a giant star, thus it has an extended convective envelope. As discussed in Ivans
et al. (2005), CS 29497-030 requires an AGB model pre-enriched with r-process material
to provide the best match to the observed data.
In a future study we plan to compare our model predictions to larger numbers of
http://groups.nscl.msu.edu/jina/reaclib/db/
59
[X/Fe]
2.5
2
1.5
1
0.5
0
40
50
60
Atomic Number
70
80
Figure 1. Predicted temporal evolution of [X/Fe] versus atomic number during the thermally-pulsing AGB phase. The [X/Fe] values evolve from lower to higher values during the TP-AGB
phase. Shown are results for the 1.5M , Z = 0.0001 model with initial [Ba/Fe] = 0.0 and [Eu/Fe]
= 0.0. Overlaid are abundances derived for the CEMP star HD 196944, taking the mean of abundances from Aoki et al. (2002) and Van Eck et al. (2001).
4
3.5
[X/Fe]
2.5
2
1.5
1
0.5
0
-0.5
40
50
60
70
80
Atomic Number
Figure 2. Predicted temporal evolution of [X/Fe] versus atomic number during the thermally-pulsing AGB phase. Shown are results for the 2M , Z = 0.0001 model with initial [Ba/Fe] =
1.5 and [Eu/Fe] = 2.0. Overlaid are abundances derived for the CEMP star CS 29497-030 from
Ivans et al. (2005).
CEMP-s and CEMP-r + s stars, covering a larger range in initial progenitor AGB mass
and metallicity.
60
spectra provide a complimentary data set to the abundances derived from the spectra
of cool evolved stars. Observations have revealed that some PNe are enriched in the
heavy elements that can be produced by the s-process including Ge, Se, Kr, Xe, and
Ba (e.g., Sterling & Dinerstein 2008). Sterling & Dinerstein (2008) obtained Se and Kr
abundances for 120 Galactic PNe, and investigated trends between s-process enrichments
and PN morphology and other nebular and stellar characteristics.
All of the PNe considered by Sterling & Dinerstein (2008) had metallicities appropriate
for the Galactic disk. There are a few PNe that have been identified in the Galactic halo.
Neutron capture-element information from these objects could be used as a complimentary data set to the CEMP stars, allowing these low-metallicity PNe to be used as tracers
of the formation of the Galaxy at the earliest times, and to constrain the evolution and
nucleosynthesis of low-metallicity AGB stars.
In Karakas & Lugaro (2009) we compared s-process predictions from low-mass (1.25
and 2M ) AGB stars with [Fe/H] = 2.3 to Se and Kr abundances found in Galactic
PNe from Sterling & Dinerstein (2008). Predictions for elements between Fe and Rb
were also given. Oxygen was used as a proxy for metallicity, i.e., [X/O] instead of Fe,
to be consistent with observations. Taking Zn as an example, the predicted abundance
of [Zn/O] is sub-solar, with [Zn/O] 1 (starting at [Zn/O] = 0.4, scaled solar for
Zn). At these low metallicities, the dredge-up of oxygen during the AGB substantially
increases the surface composition, from [O/Fe] = 0.4 (typical of the halo) to [O/Fe] 1.0.
In all cases, Zn is actually produced by the s-process but in smaller quantities to oxygen,
hence the [X/O] ratios do not reflect the degree of production. For example, the final
[Zn/Fe] = 0.45. Note that the iron abundance remains essentially unchanged.
Our results indicate that at these low metallicities, oxygen is no longer a suitable
proxy for the initial metallicity of the progenitor star. Argon, if available, would be a
more suitable proxy, as its abundance remains unaltered by AGB nucleosynthesis. For
PNe neutron-capture elements to be useful, the issue of a suitable metallicity proxy would
need to be resolved.
5. Summary
We have computed new full network predictions from low-mass AGB stars of low
metallicity. These predictions are to be compared to the derived composition of carbonenhanced metal-poor stars of similar metallicity as the models. We also discuss the possibility of using halo planetary nebulae as further probes of low-metallicity AGB nucleosynthesis. Such comparisons allow the physics and nucleosynthesis of AGB models to be
constrained, and provide insight into the formation of the Galaxy at the earliest times.
References
Aoki, W., et al. 2002, ApJ, 580, p1149
Beers, T. C. and Christlieb, N. 2005, ARA&A, 43, p531
Heil, M. et al. 2008, Physical Review C, 78, 025803
Ivans, I. I., Sneden, C., Gallino, R., Cowan, J. J. & Preston, G. W. 2005, ApJ, 627, L145
Karakas, A. I. & Lattanzio, J. C. 2007, PASA, 24, p103
Karakas, A. I. & Lugaro M. 2009, PASA, accepted.
Karakas, A. I., et al. 2009, ApJ, 690, p1130
Lucatello, S. et al. 2005, ApJ, 625, p824
Sneden, C., Cowan, J. J., & Gallino, R. 2008, ARA&A, 46, p241
Sterling, N. C. & Dinerstein, H. L. 2008, ApJS, 174, 158
Van Eck, S., Goriely, S., Jorissen, A., & Plez, B. 2001, Nature, 412, p793
Abstract. We have been determining abundances of Th, Pb and other neutron-capture elements in metal-deficient cool giant stars to constrain the enrichment of heavy elements by the
r- and s-processes. Our current sample covers the metallicity range between [Fe/H]= 2.5 and
1.0. (1) The abundance ratios of Pb/Fe and Pb/Eu of most of our stars are approximately
constant, and no increase of these ratios with increasing metallicity is found. This result suggests that the Pb abundances of our sample are determined by the r-process with no or little
contribution of the s-process. (2) The Th/Eu abundance ratios of our sample show no significant scatter, and the average is lower by 0.2 dex in the logarithmic scale than the solar-system
value. This result indicates that the actinides production by the r-process does not show large
dispersion, even though r-process models suggest high sensitivity of the actinides production to
the nucleosynthesis environment.
Keywords. Galaxy:halo, stars:abundances, stars:Population II
1. Introduction
Th and Pb are key elements to understand the production of heavy elements by
neutron-capture reactions. The difficulty in observational studies for these elements are
the weakness of spectral lines in optical range, and also contaminations of spectral features of other elements and molecules. For this reason, measurements of these elements
have been limited to very metal-poor stars with relatively large enhancements in neutroncapture elements (see Roederer et al. (2009) and references therein).
In order to investigate the enrichment of Th and Pb, we have been extending the
measurements for higher metallicity range than that previously studied. The spectral
lines of these elements are relatively strong in cool giants (the effective temperature
lower than 4500 K), and the Th line at 5989
A is little blended with other features. This
paper reports the preliminary results of our measurements of Pb and Th abundances
obtained by high resolution spectroscopy with the Subaru Telescope High Dispersion
Spectrograph (HDS) for bright metal-poor giants.
2. Measurements
Pb abundances are determined from the Pb I 4057
A line, which is the almost unique
line found in optical spectra of metal-poor stars. Although CH molecular lines exist in
this spectral range, the contamination is not severe because of low carbon abundances
in evolved red giants in our sample.
Th abundances have been measured from the Th II 4019
A line for most of metal-poor
stars in previous studies. This line is, however, severely blended with other absorption
61
62
Figure 1. left: [Pb/Fe] and [Pb/Eu] as a function of [Fe/H] for halo stars (circles) and globular
clusters (triangles). This sample includes that of Aoki & Honda (2008). The dotted line in the
lower panel indicates the r-process component of the solar-system material estimated by Arlandini et al. (1999). right: Th/Eu ratio as a function of [Fe/H]. Filled circles are the preliminary
results obtained by the present work. The solid line indicates the solar Th/Eu value.
lines in mildly metal-poor stars. In such stars (in particular cool red giants), the Th II
5989
A line is sufficiently strong to measure the Th line. This line is almost free from
contaminations of other species (Aoki et al. 2007).
Abstract. In this study we consider the pre-enrichment of minihalos, and study the impact
of metallicity on pre-collapsing minihalos by using the cosmological, N-body simulation code
Enzo. The metallicities that we consider are assumed to be the result of pre-enrichment by
earlier star formation. In the simulations of 103 and 101 Z we see a big difference for the
collapse of the minihalo. In the high metallicity case the minihalo is more compact compared to
the low metallicity case and we reach higher densities due to the efficient cooling. Also in the
high metallicity case the gas cools down to lower temperatures and we see cold, dense gas which
indicates a multi-phase ISM. This leads us to think that there is a transition region between
metallicities of 103 and 101 Z which lowers the mass scale of the next generation of stars.
Furthermore, because the gas cools more efficiently in the high metallicity case there is less
pressure support against gravity and therefore we see higher velocities.
1. Introduction
In the last two decades, cosmological simulations have become an important tool for
theoreticians to simulate the formation of structure from primordial fluctuations. According to the CDM model of hierarchical structure formation, the first (PopIII) stars have
formed at redshifts of z =20-30, in dark matter halos of 106 M (Tegmark et al.1997).
The formation of a star depends on the ability of interstellar gas to cool and form dense
molecular gas. Therefore chemical composition and metallicity of the interstellar gas are
the key parameters to study. To investigate the effect of metallicity on the transition from
a top-heavy IMF to the current IMF we have run simulations of pre-collapsing minihalos
with different metallicities by using the adaptive mesh refinement code, Enzo (Bryan &
Norman 1997).
In our simulation we use 1283 grid cells on the top grid with three nested subgrids,
each refining by a factor of two. The box size of the simulation is 0.25 h 1 Mpc. The
resolution of our simulations is 0.1 pc at redshift 15. We run simulations from redshift 99
to 5 and use three different cooling prescriptions for metallicities of 103 , 102 , 101 Z .
In these simulations, we ignore the UV background radiation and feedback effects from
the PopIII stars. The metallicity dependent cooling includes finestructure emissions from
carbon and oxygen as well as molecular lines from species like CO. All level populations
are computed under statistical equilibrium. The chemistry includes gas phase and grain
surface formation of H2 and HD.
In the near future, in order to understand the nature of the first stars and their feedback effects we are going to implement the PDR and XDR codes of Spaans & Meijerink
(2008) with updated grain surface formation of H2 and HD from Cazaux & Spaans (2009)
into Enzo.
63
64
2. Results
Figure 1 shows the density-temperature profile for a minihalo for metallicities of 103
and 101 Z at z=15. In the high metallicity case we see cold dense gas which indicates a
multiphase ISM while we lack this gas in the low metallicity case. This is because in the
low metallicity case the gas cannot cool efficiently enough to lower temperatures. Also,
due to the shorter cooling time scale we see that in the high metallicity case the minihalo
evolves faster dynamically. This can be seen in figure 2 which shows the magnitudes of
the total velocities in a slice of the simulation. Because the gas cools more efficiently there
is less pressure support against gravity and therefore the high metallicity case reaches
higher velocities.
Figure 2. Magnitudes of the total velocities of a minihalo for the central 2 kpc for
metallicities of 101 (left) and 103 (right) Z at redshift 15.
References
Bryan, G. & Norman, M. 1997, ASPC, 123, 363B
Cazaux, S. & Spaans, M. 2009, A&A, 496, 365C
Spaans, M., & Meijerink, R. 2008, ApJ, 678L,5
Tegmark, M., Silk, J., Rees, M. J., Blanchard, A., Abel, T., Palla, F. 1997, ApJ, 474, 1
In a seminal paper, Bromm et al. (2001) performed simulations of the collapse of cold
gas in a top-hat potential that included the metallicity-dependent effects of atomic finestructure cooling. In the absence of molecular cooling, they found that fragmentation
suggestive of a present-day initial mass function only set in at metallicities above a
threshold value of Z 103.5 Z . However, they noted that the neglect of molecular
cooling could be significant. Omukai et al. (2005) argued, based on the results of their
detailed one-zone models, that molecular cooling would indeed dominate the cooling over
many orders of magnitude in density.
The effects of molecular cooling at densities up to n 500 cm3 were studied by
Jappsen et al. (2007) in three-dimensional collapse simulations of warm ionized gas in
minihalos for a wide range of environmental conditions. This study used a time-dependent
chemical network running alongside the hydrodynamic evolution as described in Glover
& Jappsen (2007). The physical motivation was to investigate whether minihalos that
formed within the relic Hii regions left by neighboring Pop III stars could form subsequent
generations of stars themselves, or whether the elevated temperatures and fractional
ionizations found in these regions suppressed star formation until larger halos formed. In
this study, it was found that molecular hydrogen dominated the cooling of the gas for
abundances up to at least 102 Z . In addition, there was no evidence for fragmentation
at densities below 500 cm3 .
Jappsen et al. (2009b) investigated the behaviour of hot, initially ionized gas in a fixed
DM potential, with varying values of rotation, turbulence and metallicity. These models
showed no fragmentation during collapse to number densities as high as 105 cm3 , for
metallicities reaching as high as 101 Z , far above the threshold suggested by previous
work. Jappsen et al. (2009a) showed that gas in simulations with low initial temperature, moderate initial rotation, and a top-hat DM overdensity, will readily fragment into
65
66
A. Jappsen et al.
4
0.5
log10 T [K]
0.4
3
0.3
0.2
0.1
0.5
log10 T [K]
0.4
3
0.3
0.2
0.1
0.5
log10 T [K]
0.4
3
0.3
0.2
0.1
2
-2
0.0
0
2
4
log10 n [cm-3]
2
3
4
log10 M [MSUN]
Figure 1. Gas temperature versus H density (left-hand side) and mass distribution of clumps
(right-hand side) for gas collapsing in a typical minihalo, shown for the primordial case and
pre-enriched to Z = 103 and Z = 101 Z , from top to bottom (see also Jappsen et al. 2009a).
The temperature evolution of primordial gas is very similar to the Z = 103 Z case, showing
that metal-line cooling becomes important only for very high metallicities. The fraction of low
mass fragments increases with higher metallicity, since more gas can cool to the temperature
of the CMB before becoming Jeans-unstable. However, the fragments are still very massive,
suggesting that metal-line cooling might not be responsible for the transition to low-mass stars.
Silver Stars
Camilla Juul Hansen1 and Francesca Primas1
1
Abstract. The rapid neutron-capture process (r-process), which produces some of the heaviest
elements, is not well understood. Obtaining accurate abundances of these heavy elements (Z >
38) is important, both in the context of the chemical evolution of the Galaxy and for understanding the site(s) and process(es) of formation of those elements. We have determined elemental
abundances for several r-process elements, notably silver, from high resolution VLT/UVES spectra. Silver was chosen because it is predominantly a light r-process element (38 < Z < 50), and
little is known about its formation and evolution in the Galaxy. Here, we present our preliminary
results.
Keywords. Silver, abundances, r-process
1. Introduction
Recent observational studies (Francois et al. 2007, Montes et al. 2007) of neutroncapture elements in metal-poor stars have provided indications that the r-process has
a second component (as is the case for the slow neutron-capture process (s-process)
Travaglio et al. 2004). Of the heavier elements (Z > 38), the lighter of these (Z = 38 50) are seen to be over-abundant compared to the scaled solar abundances of elements
with Z > 50 (such as Ba, La,... Eu etc.) at low metallicity, hence a second process (the
so-called Lighter Element Primary Process - LEPP) has been proposed to explain this
relative excess. In order to test the LEPP hypothesis, we chose to target silver (previously
studied in only seven stars by Crawford et al. 1998), since the r-process dominates its
production (79%, Arlandini et al. 1999), making it a good tracer in the interval 38 < Z
< 50.
2. The Sample
The sample consists of dwarf stars spanning a broad range in temperature (T: 5400
- 6500K), gravity (log g: 3.73 - 4.74 dex) and metallicity ([Fe/H]: -0.54 to -3.37). All
67
68
the stellar spectra have been reduced using the UVES pipeline and shifted, added, and
normalized in IRAF. MARCS (Gustafsson et al. 2008) stellar model atmospheres and
MOOG spectrum synthesis code were used in order to derive the abundances. In Figure
1, a fit to the silver line at 3382
A in HD189558 (a cool dwarf) is shown. Silver has two
main resonance lines, both in the UV.
Figure 2. [Ag/Ba] vs [Ba/H] (left) and [Ag/Eu] vs [Eu/H] (right) abundance ratios and
associated uncertainties. Arrows indicate abundance upper limits.
At metallicities of our sample ([Fe/H] > -1.6), it is not surprising that Ag shows a
tighter correlation with Eu than with Ba (see Figure 2). Eu is a r-process element (94%
main r-process, Arlandini et al. 1999) while Ba is mainly produced by the s-process like
other elements we have looked at (e.g. Zr).
At our stellar metallicities, it is not possible to characterize the LEPP and silver does
not seem to be able to prove the existence of this process, the impact from the LEPP at
[Fe/H] > -1.6 have been erased and contributions from other processes, such as the main
r- and s-process, are seen instead.
Future work will explore metal-poor giants, thereby increasing the sample and hopefully obtaining silver at lower metallicities allowing us to trace and characterize the LEPP
through determinations of key r- and s-process elements.
References
Arlandini et al. 1999, AJ,525, 866
Crawford et al. 1998, AJ, 116, 2489
Francois et al. 2007, A&A,476,935
Gustafsson et al. 2008, A&A, 486, 951
Montes et al. 2007, AJ, 671, 1685
Travaglio et al. 2004, AJ, 601, 864
Abstract. Although the first stars were likely very hot and luminous, their low or zero metallicity implies that any mass loss through winds driven by line-scattering of radiation in metal ions
was likely small or non-existent. Here we examine the potential role of another possible mechanism for mass loss in these first stars, namely via decretion disks associated with near-critical
rotation induced from evolution of the stellar interior. In this case the mass loss is set by the
angular momentum needed to keep the stellar rotation at or below the critical rate. In present
evolutionary models, that mass loss is estimated by assuming effective release from a spherical
shell at the surface. Here we examine the potentially important role of viscous coupling of the
decretion disk in outward angular momentum transport, emphasizing that the specific angular
momentum at the outer edge of the disk can be much larger than at the stellar surface. The net
result is that, for a given stellar interior angular momentum excess, the mass loss required from
a decretion disk can be significantly less than invoked in previous models assuming a direct,
near-surface release.
Keywords. Stars: mass-loss stars: evolution stars: rotation
70
. .
J / J0
v / vK
0.1
rcrit
10-2
10-3
vr / a
10-4
10-5
=1
= 0.1
= 0.01
1
10
100
1000
r / Req
Figure 1. The dependence of the radial (in units of the sound speed a) and longitudinal (in
units of the Keplerian velocity vK ) velocities, and the angular momentum loss in units of the
equatorial angular momentum loss J0 on the radius in a viscous disk. The results are plotted
for different viscosity paremeters of Shakura & Sunyaev (1973).
Its extension is fixed at the point where matter decouples, keeping its angular momentum
constant. Fig. 1 shows that for a thermal expansion of an isothermal disk the typical disk
extension is a few hundred times the stellar equatorial radius.
The mass loss rates needed to maintain the star just below the critical limit is thus
much lower when a decretion disk is formed than in previous models that assume direct
loss from the stellar surface. In the example of Fig. 1, one infers that the disk reduces
the mass loss rate by a factor ten or more.
We provide prescriptions for computing the instantaneous mass loss rates in the presence of a decretion disk and for different physical circumstances:
case A: very short radiative ablation timescale of the disk
case B: disk extension is by viscosity and thermal expansion alone
case C: both viscosity and radiative effects determine the extension of the disk.
In the first stars, case B is the most likely one.
Acknowledgement
205/07/0031. SPO acknowledges support
This work was supported by grant GA CR
of NSF grant AST-0507581 and NASA grant NNG05GC36.
References
Krticka, J., & Kub
at, J. 2006, A&A, 446, 1039
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, A&A, 447, 623
Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, 337
Centro de Astrofsica da Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal
2
Max-Planck-Institut f
ur Astrophysik, Germany
Abstract. We present Li abundances for 73 stars in the metallicity range -3.5 < [Fe/H] <
-1.0 using improved IRFM temperatures (Casagrande et al. 2009) with precise E(B-V) values
obtained mostly from interstellar NaI D lines, and high-quality equivalent widths (EW 3%).
At all metallicities we uncover a fine-structure in the Li abundances of Spite plateau stars,
which we trace to Li depletion that depends on both metallicity and mass. Models including
atomic diffusion and turbulent mixing seem to reproduce the observed Li depletion assuming
a primordial Li abundance ALi = 2.64 dex (MARCS models) or 2.72 (Kurucz overshooting
models), in good agreement with current predictions (ALi = 2.72) from standard BBN.
Keywords. nucleosynthesis cosmology: observations stars: abundances
72
J. Melendez et al.
1. Introduction
There are three subgroups of the gamma-ray bursts (GRBs) separable with respect to
the duration and hardness (analogy of the color in the gamma band). The three subgroups
of GRBs (see Figure 1) were confirmed in the databases of different satellites: Compton
(BATSE) - Horv
ath (1998), Horv
ath (2002), Horv
ath et al. (2006); Swift - Horv
ath et
pa et al. (2009); BeppoSAX - Horv
al. (2008), Huja et al. (2009); RHESSI - R
ath (2009);
for more aspects about the subgroups of GRBs see also Gehrels et al. (2006).
Figure 1. On the figure the RHESSIs GRBs are separated into three subgroups with respect
to the duration T90 and hardness H (T90 is in seconds, CL means confidence level, H is dimensionless).
73
74
2. Overview
Short GRBs give the fraction of (10 20)% of the detected GRBs, and on the sky
they are distributed anisotropically (Vavrek et al. 2008). Intermediate GRBs form a fraction of 10% of the detected GRBs. On the sky they are also distributed anisotropically
(Mesz
aros A. et al. 2000). The redshifts are not clarified yet for the intermediate subgroup. The short GRBs should be till redshift 1 challenging the cosmological principle
due to their anisotropic distribution (Mesz
aros A. et al. 2009).
The long GRBs give the majority ( (60 80)%) of the observed GRBs. Contrary
to the remaining two subgroups, they seem to be distributed isotropically on the sky
(Vavrek et al. 2008). They are strongly related to Type Ic supernovae and to the WolfRayet stars (Mesz
aros P. 2006). They can be at very high redshifts (up to redshift 8.2)
(Bagoly et al. 2006, Krimm et al. 2009).
3. Long GRBs, star-formation rate and the impact on the first stars
The redshift distribution of the long GRBs follows the star-formation rate (Mesz
aros A.
et al. 2006, Le & Dermer 2007). This result has an interesting impact on the cosmology:
Because the long bursts are coupled to supernovae (hence, to massive stars), and the long
bursts are at very high redshifts, the long GRBs support observationally the existence of
the stars at very high redshifts (up to 8). Hence, they may well serve as milestones in
the early Universe.
4. Acknowledgements
Thanks are due to the valuable discussions with Z. Bagoly, L.G. Balazs, W. Hajdas, I.
Horv
ath, R. Hudec, S. Klose, S. Larsson, P. Mesz
aros, F. Ryde, G. Tusnady, R. Vavrek,
P. Veres and C. Wigger. The study was supported by GAUK grant No.46307, by OTKA
grant No. K077795, by the Grant Agency of the Czech Republic grant No. 205/08/H005,
and by the Research Program MSM0021620860 of the Ministry of Education of the Czech
Republic.
References
Bagoly, Z., Mesz
aros, A., Bal
azs, L. G., Horv
ath, I., Klose, S., Larsson, S., Mesz
aros, P., Ryde,
F., & Tusn
ady, G., 2006 A&A, 453, 797
Gehrels, N., & et al. 2006, Nature, 444, 1044
Horv
ath, I. 1998, ApJ, 508, 757
Horv
ath, I. 2002, A&A, 392, 791
Horv
ath, I. 2009, Ap&SS, 323, 83
Horv
ath, I., Bal
azs, L.G., Bagoly, Z., Ryde, F., & Mesz
aros, A. 2006, A&A, 447, 23
Horv
ath, I., Bal
azs, L.G., Bagoly, Z., & Veres, P. 2008, A&A, 489, L1
pa, J. 2009, A&A, 504, 67
Huja, D., Mesz
aros, A., & R
Krimm, H.A., & et al., 2009 GCN, 9198
Le, T., & Dermer, C.D., 2007, ApJ, 661, 394
Mesz
aros, A., Bagoly, Z., Horv
ath, I., Bal
azs, L.G., & Vavrek, R. 2000, ApJ, 539, 98
Mesz
aros, A., Bagoly, Z., Bal
azs, L.G., & Horv
ath, I., 2006, A&A, 455, 785
Mesz
aros, A., Bal
azs, L.G., Bagoly, Z., & Veres, P., 2009, Gamma-Ray Burst: Sixth Huntsville
Symposium, AIP Conf. Proc., 1133, 483
Mesz
aros, P. 2006, Rep.Progr.Phys. 69, 2259
pa, J., Mesz
R
aros, A., Wigger, C., Huja, D., Hudec, R., & Hajdas, W. 2009, A&A, 498, 399
Vavrek, R., Bal
azs, L.G., Mesz
aros, A., Horv
ath, I., & Bagoly, Z. 2008, MNRAS, 391, 1741
Abstract. We present the largest sample available to date of lithium abundances in extremely
metal poor (EMP) Halo dwarfs. Four Teff estimators are used, including IRFM and H wings
fitting against 3D hydrodynamical synthetic profiles. Lithium abundances are computed by
means of 1D and 3D-hydrodynamical NLTE computations. Below [Fe/H]-3, a strong positive
correlation of A(Li) with [Fe/H] appears, not influenced by the choice of the Teff estimator. A
linear fit finds a slope of about 0.30 dex in A(Li) per dex in [Fe/H], significant to 2-3 , and
consistent within 1 among all the Teff estimators. The scatter in A(Li) increases significantly
below [Fe/H]-3. Above, the plateau lies at hA(Li)3D,NLTE i=2.1990.086. If the primordial
A(Li) is the one derived from standard Big Bang Nucleosynthesis (BBN), it appears difficult
to envision a single depletion phenomenon producing a thin, metallicity independent plateau
above [Fe/H]=-2.8, and a highly scattered, metallicity dependent distribution below.
Keywords. nuclear reactions, nucleosynthesis, abundances, Galaxy: halo, Galaxy: abundances,
stars: abundances, stars: Population II
1. Introduction
In this work we present lithium abundances for a sample of 28 stars in the -3.6<[Fe/H]<2.4 range, 10 of which have [Fe/H]6-3.0. All the stars were observed with UVES@VLT
with S/N80-100 in the Li doublet range. We employed four temperature scales: H
wings fitting against a grid of synthetic profiles from 1D models using Barklem et
al.(2000) or Ali & Griem(1966) self-broadening theories (BA and ALI scales); H fitting
against a 3D-hydrodynamical grid based on CO5 BOLD models (Freytag et al.(2002),
Wedemeyer et al.(2004)) with Barklem et al.(2000) self-broadening (3D scale); InfraRed
Flux method (IRFM, Gonz
alez Hernandez & Bonifacio(2009)). Fe I and Fe II lines were
used to establish metallicity and set gravity and microturbulence. Hydrodynamical 3D
spectral synthesis was employed together with an 8-levels Li model atom to derive 3D
NLTE Li I 670.8nm doublet line profiles (see Sbordone et al.(2009) for details).
75
76
Figure 1. A(Li) vs. [Fe/H] from some recent studies. Blue circles, Asplund et al.(2006) data, red
triangles, Aoki et al.(2009) data, magenta squares, CS 22876032 from Gonz
alez Hern
andez et
al.(2008), filled symbol primary star, open symbol secondary star. Black diamonds, this work, BA
temperature scale. Dot-dashed gray line, best linear fit to Asplund et al.(2006) data, continuous
dark gray line, best fit to our data. Typical error bars for our data are displayed.
2. Results
Although up to 400 K difference in Teff exist between the coolest scale (BA) and the
hottest (ALI and IRFM), the results are remarkably similar. In all cases, a linear fit
in the A(Li) vs [Fe/H] plane bears a strong positive slope of about 0.3 dex in A(Li)
per dex in[Fe/H], significant at 2-3 and consistent within 1 among the four scales.
Lithium abundances scatter increases significantly at [Fe/H]<-3.0, while a thin plateau
exists above this metallicity (see Fig. 1). The low-metallicity distribution appears roughly
triangular, with some stars staying at (or close to) the plateau value, and others showing
Li depletions which are in general larger at lower metallicities. No star is found above
the plateau level: this asymmetric increase in the scatter leads to the slope observed
in the linear fit. It appears thus that the Spite plateau is progressively disrupted below
[Fe/H]-3, where an increasing number of stars show significantly sub-plateau A(Li). We
suggest a two process scenario to explain this finding: i) the stars in our sample have all
formed with a Li content corresponding to the plateau (be it the primordial value, or the
result of some previous depletion) and ii) some further (atmospheric?) depletion process
acts preferentially on more metal poor objects, but becomes negligible above [Fe/H]-3.
References
Ali, A. W., & Griem, H. R. 1966, Physical Review , 144, 366
Aoki, W., et al., 2009, ApJ, 698, 1803
Asplund, M., et al., 2006, ApJ, 644, 229
Barklem, P. S., Piskunov, N., & OMara, B. J. 2000, A&A, 355, L5
Bonifacio, P., et al. 2007, A&A, 462, 851
Freytag, B., Steffen, M., & Dorch, B. 2002, Astronomische Nachrichten, 323, 213
Gonz
alez Hern
andez, J. I., & Bonifacio, P. 2009, A&A, 497, 497
Gonz
alez Hern
andez, J. I., et al. 2008, A&A, 480, 233
Sbordone, L., et al., 2009, A&A submitted
Wedemeyer, S., et al., 2004, A&A, 414, 1121
1. Introduction
Currently, over 1,000 objects are regarded as planetary nebulae (PNe) in our Galaxy,
while 14 of them have been identified as halo members from their location and kinematics.
Of them, 5 objects are known as C- and N-rich ([C,N/O]>0) halo PNe with [Ar/H]<1.7.
K648 (in M15), H4-1, and BoBn1 are C- and N-rich halo PNe with [Ar/H]2.1. The
progenitors of halo PNe are thought to be 0.8 M stars. These C- and N-rich halo PNe,
however, show signatures that they have evolved from massive progenitors.
For example, these halo PNe can become N-rich, but not C-rich, if they have evolved
from single 0.8 M stars. To become C-rich PNe, the third dredge-up must take place
in the late AGB phase. However, the third dredge-up takes place in stars with initial
masses >1.2-1.5 M . Also, current stellar evolutionary models predict that the post-AGB
evolution of a star with an initial mass 0.8 M proceeds too slowly for a visible PN to
be formed. Unless these issues are resolved, we will not be able to take full advantage of
halo PNe as a proof of low-mass star evolution and the Galactic chemical evolution.
How these progenitor stars became C and N-rich halo PNe? To answer this key question, we have been observing these halo PNe using the 8.2-m Subaru telescope/HighDispersion Spectrograph (HDS) and collecting archival data and carefully analyzing
spectra of these objects. Our recent research reveals that [C/Fe] and [N/Fe] abundances
of K648, H4-1, and BoBn1 are compatible with those of carbon-enhanced metal poor
(CEMP) stars (Otsuka et al. 2008a). Interestingly, BoBn1 is the most fluorine rich among
77
78
M. Otsuka et al.
Figure 1. (left panel) [Kr/Ar]-[C/Ar] diagram. Galactic PNe: the data of 37 objects taken from
Sterling & Dinerstein (2008). (right panel) [Xe or Ba/Ar]-[C/Ar] diagram. Galactic PNe: taken
from Sharpee et al. (2007). CEMP-s and C-rich AGB: taken from the SAGA data base (Suda
et al. 2008).
F-detected PNe, and its F-abundance is comparable with the s-process enhanced CEMP
star (CEMP-s) HE1305+0132 (Otsuka et al. 2008b). These C-, N-rich halo PNe might
share their origins with CEMP-s stars. If these halo PNe had evolved from binary stars
such as CEMP-s stars, s-process element might be enhanced as well as C and N. To examine enhancement of s-process elements and verify our hypothesis, we have performed
deep HDS spectroscopy for H4-1 and BoBn1. In this paper, we present our recent results
of chemical abundance analysis of these objects.
Session II
Abstract. The cosmic microwave background and the cosmic expansion can be interpreted as
evidence that the Universe underwent an extremely hot and dense phase about 14 Gyr ago. The
nucleosynthesis computations tell us that the Universe emerged from this state with a very simple
chemical composition: H, 2 H, 3 He, 4 He, and traces of 7 Li. All other nuclei where synthesised at
later times. Our stellar evolution models tell us that, if a low-mass star with this composition
had been created (a zero-metal star) at that time, it would still be shining on the Main
Sequence today. Over the last 40 years there have been many efforts to detect such primordial
stars but none has so-far been found. The lowest metallicity stars known have a metal content,
Z, which is of the order of 104 Z . These are also the lowest metallicity objects known in the
Universe. This seems to support the theories of star formation which predict that only high mass
stars could form with a primordial composition and require a minimum metallicity to allow the
formation of low-mass stars. Yet, since absence of evidence is not evidence of absence, we cannot
exclude the existence of such low-mass zero-metal stars, at present. If we have not found the
first Galactic stars, as a by product of our searches we have found their direct descendants, stars
of extremely low metallicity (Z 6 103 Z ). The chemical composition of such stars contains
indirect information on the nature of the stars responsible for the nucleosynthesis of the metals.
Such a fossil record allows us a glimpse of the Galaxy at a look-back time equivalent to redshift
z = 10, or larger. The last ten years have been full of exciting discoveries in this field, which I
will try to review in this contribution.
Keywords. hydrodynamics, line: formation, nucleosynthesis, stars: abundances, stars: Population II, Galaxy: abundances, Galaxy: evolution, Galaxy: halo
1. Introduction
The quest for the First Stars and their immediate descendants has been a field of
very active research, both in the high redshift and in the local Universe. In this review
I will only deal with advances in the local Universe, mainly focusing on literature which
appeared in the last four years. I refer the reader to the reviews of Beers & Christlieb
(2005) and Bonifacio (2007) for the earlier literature. I also largely omit the results on
neutron-capture elements, since this is covered by another review in this Symposium
(Sneden et al. 2009) and I refer to the review Sneden et al.(2008) for older literature.
I will touch only briefly on the topic of Carbon Enhanced Metal Poor stars, which is
covered by the review of Aoki (2009) in this Symposium. Finally I will largely ignore
the abundant literature on lithium, which should be discussed in this volume by the
contributions of Melendez et al. (2009), Sbordone et al. (2009) and Steffen et al. (2009).
I will try to concentrate on the observations, without trying to review their theoretical
interpretation.
81
82
Piercarlo Bonifacio
83
low silicon abundance (1/10 of the iron), but a normal Mg abundance. Should oxygen
be under-abundant like Si, this would be the most metal-poor object known.
84
Piercarlo Bonifacio
(Masseron et al. 2006), whose extremely high luminosity (log g 6 0.3) qualifies it as
a TP-AGB star. The abundance pattern of this star suggests an intermediate mass of
2M or larger. However, its distance (about 50 kpc) implies it belongs to the outer Halo,
where no recent star formation has occurred.
85
Coma Berenices, Norris et al.(2008) discovered eight stars with [Fe/H] 3 in Bootes I
and one with [Fe/H] 3.5. On the other hand Sextans does not show any stars below
[Fe/H]= 3, although Aoki et al.(2009) found six below 2.5. The conclusion is that EMP
stars are to be found everywhere and their detailed abundances will tell us something on
the first stars in their host galaxies.
86
Piercarlo Bonifacio
Figure 1. 3D corrections for the [OI] 630 nm line, for giant stars computed from CO5 BOLD
models, compared with those computed by Collet et al. (2007), from models computed with the
Stein& Nordlund code (Stein & Nordlund 1998). The difference is non-negligible, and it is likely
rooted in the different microphysics of the two codes, but also in the different models used as
1D reference to compute the 3D corrections.
discrepancy, however the NLTE computations of Takeda et al.(2005) imply small NLTE
corrections for Zn and in any case not significantly different for giants and dwarfs. Also
granulation effects are unlikely to be the cause of this discrepancy. In the case of zinc we
are also facing the problem of small number statistics, since for very few dwarf stars zinc
has been measured. The issue should be further investigated both from the theoretical
and observational point of view.
The NLTE effects may be relevant also for neutron capture elements. Up to now results
for Ba have been published (Mashonkina et al. 2008, Andrievsky et al. 2009) and in this
case the large star-to-star scatter at any metallicity is confirmed by the NLTE analysis,
pointing to a poor mixing of these elements in the early Galaxy.
6. Granulation effects
Besides LTE, the most important simplifying assumption made in the analysis of stellar spectra is that of a static atmosphere. Thus the majority of analysis rely on 1D
hydrostatic model atmospheres. In the last 10 years a considerable advance has come
through the use of three dimensional hydrodynamical simulations of stellar atmospheres
(3D models for short). All the analysis published so far rely on simulations computed
either with the code of Stein & Nordlund(1998) or with the CO5 BOLD code (Freytag
et al. 2002, Freytag et al. 2003, Wedemeyer et al. 2004). Such models are more physically
motivated than 1D models, although there is still considerable work to validate them
and bring them at the level of maturity of current 1D models. The treatment of opacity
in such models is based on an opacity binning scheme (Nordlund 1982, Ludwig 1992,
Ludwig, Jordan, & Steffen 1994), however the optimal number of bins to employ and
their definition is still a matter of investigation. Behara et al.(2009a) found significant
differences in the temperature structure of the outer layers for models computed using
six or twelve opacity bins.
87
88
Piercarlo Bonifacio
89
Ito, H., Aoki, W., Honda, S., & Beers, T. C. 2009a, ApJ, 698, L37
Ito, H., Aoki, W., Honda, S., Beers, T. C., & Tominaga, N. 2009b, IAU Symp. 265, p. 124
Kucinskas, A., Ludwig, H. -G., Caffau, E., & Steffen, M. 2009, MemSAI, 80, 720
Lai, D. K., Bolte, M., Johnson, J. A., & Lucatello, S. 2004, AJ, 128, 2402
Lai, D. K., Bolte, M., Johnson, J. A., Lucatello, S., Heger, A., & Woosley, S. E. 2008, ApJ, 681,
1524
Lai, D. K., Rockosi, C. M., Bolte, M., Johnson, J. A., Beers, T. C., Lee, Y. S., Allende Prieto,
C., & Yanny, B. 2009, ApJ, 697, L63
Ludwig, H.-G. 1992, PhDT, University of Kiel
Ludwig, H.-G., Jordan, S., & Steffen M. 1994, A&A, 284, 105
Ludwig, H. -., Caffau, E., Steffen, M., Freytag, B., Bonifacio, P., & Kucinskas, A. 2009, MemSAI,
80, 708
Ludwig, H.-G., Behara, N. T., Steffen, M., & Bonifacio, P. 2009, A&A, 502, L1
Mashonkina, L., et al. 2008, A&A, 478, 529
Masseron, T., et al. 2006, A&A, 455, 1059
McWilliam, A., Preston, G. W., Sneden, C., & Searle, L. 1995, AJ, 109, 2757
Melendez, J., & Cohen, J. G. 2009, ApJ, 699, 2017
Melendez, J., Casagrande, L., Ramirez, I. 2009 IAU Symp. 265, p. 71
A. 1982, A&A, 107, 1
Nordlund,
Norris, J. E., Christlieb, N., Korn, A. J., Eriksson, K., Bessell, M. S., Beers, T. C., Wisotzki,
L., & Reimers, D. 2007, ApJ, 670, 774
Norris, J. E., Gilmore, G., Wyse, R. F. G., Wilkinson, M. I., Belokurov, V., Evans, N. W., &
Zucker, D. B. 2008, ApJ, 689, L113
Perryman, M. A. C., et al. 2001, A&A, 369, 339
Prantzos, N. 2008, EAS Publications Series, 32, 311
Rastegaev, D. A. 2009, Astronomy Letters, 35, 466
Rich, J. A., & Boesgaard, A. M. 2009, ApJ, 701, 1519
Roederer, I. U., et al. 2008, ApJ, 679, 1549
Roederer, I. U. 2009 IAU Symp. 265, p. 368
Sbordone, L., Bonifacio, P., Caffau, E., et al. 2009 IAU Symp. 265, p. 75
Shetrone, M. D., C
ote, P., & Sargent, W. L. W. 2001, ApJ, 548, 592
Shi, J. R., Gehren, T., Mashonkina, L., & Zhao, G. 2009, A&A, 503, 533
Smiljanic, R., Pasquini, L., Bonifacio, P., Galli, D., Gratton, R. G., Randich, S., & Wolff, B.
2009a, A&A, 499, 103
Smiljanic, R., Pasquini, L., Bonifacio, P., Galli, D., Barbuy, B., Gratton, R., & Randich, S.
2009b, IAU Symp. 265, p. 134
Sneden, C., Cowan, J.J. & Gallino, R. 2009 IAU Symp. 265, p. 46
Sneden, C., Cowan, J. J., & Gallino, R. 2008, ARAA, 46, 241
Spite, M., Spite, F., Bonifacio, P., Andrievsky, S., Cayrel R., Francois, P., Korotin, S. 2009 IAU
Symp. 265, p. 380
Steffen, M., Cayrel, R., Bonifacio, P., Ludwig, H.-G., Caffau, E. 2009 IAU Symp. 265, p. 23
Stein, R. F., & Nordlund, A. 1998, ApJ, 499, 914
Takeda, Y., Hashimoto, O., Taguchi, H., Yoshioka, K., Takada-Hidai, M., Saito, Y., & Honda,
S. 2005, PASJ, 57, 751
Tan, K. F., Shi, J. R., & Zhao, G. 2009, MNRAS, 392, 205
Thompson, I. B., et al. 2008, ApJ, 677, 556
Venn, K. A., & Lambert, D. L. 2008, ApJ, 677, 572
Wedemeyer, S., Freytag, B., Steffen, M., Ludwig, H.-G., & Holweger, H. 2004, A&A, 414, 1121
Yong, D., Lambert, D. L., & Ivans, I. I. 2003, ApJ, 599, 1357
Yong, D., Lambert, D. L., Allende Prieto, C., & Paulson, D. B. 2004, ApJ, 603, 697
Yong, D., Aoki, W., & Lambert, D. L. 2006, ApJ, 638, 1018
York, D. G., et al. 2000, AJ, 120, 1579
1. Introduction
Extremely metal-poor (EMP) stars are the useful probes for the star formation history
in the very early Universe. These stars were born in iron-poor or perhaps the primordial
gas cloud produced soon after the Big Bang and considered to retain the information on
the early epoch of the universe. Thanks to the detailed abundance analyses of large-scale
surveys of metal-poor stars (e.g., Aoki et al. 2007), we have thousands of data of EMP
candidates with element abundances of many species known.
One of the important characteristics in observed EMP stars is an abundance anomalies and star to star variations among stars at [Fe/H] . 2. In order to understand
comprehensively the origins of elements and star formation history of our Galaxy, we
developed the database of observed EMP stars by compiling observed abundances and
related information from the literature (The SAGA database which is available online
at http://saga.sci.hokudai.ac.jp, Suda et al. 2008). The SAGA database contains more
than 20,000 records of abundance data for more than 1,300 unique stars.
In this paper, we compare the stellar evolution models with the observed EMP abundances using the SAGA database. The present paper explores the possible extra mixing
in EMP stars by considering the dredge-up by surface convection of low-mass and lowmetallicity stars. We focus on the lithium depletion in EMP stars in the next section and
discuss the possible extra mixing in red giants in 3.
90
91
3.0
2.5
2.0
log (Li)
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
7000
Carbon normal
C-rich, [C/Fe]>0.5
[Fe/H]=-2.3
[Fe/H]=-3
[Fe/H]=-4
[Fe/H]=-5
6500
6000
5500
5000
4500
4000
3500
Teff
Figure 1. Lithium abundance as a function of effective temperature. Lines represent the expected lithium abundance using stellar models by assuming that the lithium are completely
burnt at T > 2.5 106 K. Sample stars are divided into carbon-rich and carbon-normal subgroups depending on whether a star has [C/Fe] > 0.5 or not.
92
2.0
[C/N]
1.0
0.0
-1.0
MP
EMP,RGB
EMP,MS
Crich,RGB
Crich,MS
CEMP,RGB
CEMP,MS
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
[C/Fe]
Figure 2. Abundance relation between [C/N] and [C/Fe] for 172 stars taken from the SAGA
database. Sample stars are divided into seven subgroups by metallicity (boundary is [Fe/H]
= 2.5), carbon abundance as in Fig. 1, and evolutionary status (red giants or dwarf). The
notations of the symbols are the same as in Suda et al.(2008). Solid lines denote constant values
of [N/Fe] = 0, 1, 2 and 3 from left to right.
93
corresponding mass shell (in Msun) below the bottom of surface convection
0.010.02 0.03
0.04
0.05
0.06
0.0
[Fe/H]=-4
[C/N] in shell
-0.5
[Fe/H]=-3
-1.0
[Fe/H]=-2.3
-1.5
-2.0
-2.5
-3.0
0
6
8
ln (Pshell/Pconv)
10
12
Figure 3. Abundance profile of 0.8M model for various metallicities as a function of the
pressure normalized by the pressure at the bottom of the surface convection. The corresponding
amount of mixed mass for the case of [Fe/H] = 3 is labeled on the top of panel. Models are
taken at the maximum depth of surface convection during the first ascent on the giant branch.
Each line denotes the [C/N] in the shell (metallicity is labeled next to the line).
dredge-up the layers with active CN cycles, but not deep enough to dredge-up the layers
where ON cycles and NeNa cycles are active. This is also the case for AGB models.
At present, it is difficult to reproduce both the lithium depletion and C/N ratio from
the standard stellar evolution models. It seems that at least two different types of mixing
have been operated in metal-poor halo stars. The stellar models that satisfy all the
requirements from observed abundance pattern are desired, which will be explored in the
future works.
Acknowledgement
This work has been supported by Grant-in-Aid for Scientific Research (18104003), from
Japan Society of the Promotion of Science. T. S. has been supported by a Marie Curie
Incoming International Fellowship of the European Community FP7 Program under
contract number PIIF-GA-2008-221145.
References
Aoki W., Beers T. C., Christlieb N., Norris J. E., Ryan S. G., Tsangarides S., 2007, ApJ, 655,
492
Deliyannis, C. P., Demarque, P., & Kawaler, S. D. 1990, ApJS, 73, 21
Fujimoto, M. Y., Ikeda, Y., & Iben Jr., I. 2000, ApJ, L25
Gratton, R. G., Sneden, C., Carretta, E., & Bragaglia, A. 2000, A&A, 354, 169
Komiya, Y., Suda, T., Minaguchi, H., Shigeyama, T., Aoki, W., & Fujimoto, M. Y. 2007, ApJ,
367
Nishimura, T., Aikawa, M., Suda, T., & Fujimoto, M. Y. 2009, PASJ, in press
Spite, M., Cayrel, R., Plez, B., Hill, V., Spite, F., Depagne, E., Francois, P., Bonifacio, P.,
Barbuy, B., Beers, T., Andersen, J., Molaro, P., Nordstr
om, B., & Primas, F. 2005, A&A,
430, 655
Suda, T., Aikawa, M., Machida, M. N., Fujimoto, M. Y., & Iben Jr., I. 2004, ApJ, 476
Suda, T. & Fujimoto, M. Y. 2009, MNRAS, submitted
Suda, T., Katsuta, Y., Yamada, S., Suwa, T., Ishizuka, C., Komiya, Y., Sorai, K., Aikawa, M.,
& Fujimoto, M. Y. 2008, ApJ, 60, 1159
Abstract. TS 01 is an exceptional planetary nebula (PN) in the Galactic halo: it is the most
oxygen-poor and has a double-degenerate core with mass close to 1.4M , possibly a Supernova Ia
progenitor. With data from the far UV to the IR we can pin down the abundances of half a dozen
of elements. The oxygen abundance is by 1.9(0.3) dex lower than in the Sun. Standard AGB
models with appropriate mass and metalicity cannot explain the observed chemical composition.
We find that additional mixing, induced by stellar rotation and/or by the presence of the close
companion can explain most of the features of the abundance pattern in TS 01.
Keywords. ISM: planetary nebulae: individual ISM: abundances Stars: AGB and postAGB Stars: binaries
1. Introduction
PN G 135.9+55.9/SBS 1150+599A, now referred to as TS 01, has an oxygen abundance
much lower than any other known PN. Its stellar core is a double degenerate close binary
system with period 3.9 h and total mass close to the Chandrasekhar limit. The story of its
discovery is presented in Tovmassian et al. (2001, 2004). Here we summarize our present
knowledge on its chemical composition and its interpretation in the light of current AGB
theory. A full presentation of this and other aspects of our work regarding the past and
future of this object can be found in Stasi
nska et al. (2009) and Tovmassian et al. (2009).
Standard methods for abundance determinations in PNe cannot be used for TS 01,
since the electron temperature cannot be obtained from the observations. Instead, photoionization models have to be constructed to make use of the largest possible set of observational constraints. Table 1 presents the status of the abundances derived for TS 01
since 2001. These abundances were based on an increasing set of observations, improving
both in quality and wavelength coverage. The study of Tovmassian et al. (2001) relied on
optical spectra where the only line from a heavy element was [O iii] 5007. Richer et al.
(2002) and Jacoby et al. (2002) independently acquired deeper spectra extending more
to the ultraviolet and showing the [Ne v] 3426 line. As a matter of fact, both sets of line
intensities were later shown to be incorrect. Pequignot & Tsamis (2005) made the best
use of the merged optical data sets, but did not include in their analysis the ultraviolet
FUSE and HST data that were already available, and interpreted very loosely some upper
limits from the published optical spectra. They thus determined a much higher oxygen
abundance for TS 01 (however still holding the record of lowest O/H found in PNe). Using still deeper optical spectra together with the UV data, Stasi
nska et al. (2005) found,
again, an oxygen abundance of about 1/100 that of the Sun, and provided estimates of
the C and N abundances.
94
95
T01
R02
J02
PT05
S05
S09
T09
Sun
10.82
10.91
10.95 0.04 10.95 10.93
<8.30
<8.0
7.50.3
8.0 0.3 7.2 0.3 8.4
<8.65
6.9.2
7.2 0.3
< 6.9
7.8
6.30.5 6.20.4 6.930.4 7.50.3 6.850.25 6.8 0.3
<6.8
8.7
5.9
7.470.4 6.70.18
6.8 0.3
6.5.3
7.8
< 5.1
<7.00
<5.6
<5.7
7.1
< 5.5
<5.00
<3.9
<4.7
6.2
references: T01: Tovmassian et al. (2001); R02: Richer et al. (2002); J02: Jacoby et al. (2002);
PT05: Pequignot & Tsamis (2005); S05: Stasi
nska et al. (2005); S09: Stasi
nska et al. (2009);
T09 : Tovmassian et al. (2009) (chemical composition derived from stellar atmosphere analysis).
96
G. Stasi
nska et al.
Figure 2. Comparison of TS 01 (filled red circle) with other Galactic PNe (black symbols) in
abundance ratios diagrams. The Sun is represented by the Solar symbol.
97
Figure 3. Comparison of the chemical composition in TS 01 with the results of stellar evolution
models for a star with initial mass 1M , and initial composition as indicated by the black dot.
Continuous line: without rotation; Dashed line: with rotation.
4. Conclusion
The planetary nebula TS 01 has been confirmed as the most oxygen-poor among known
PNe, with an oxygen abundance of 1/70 of the Solar value. From upper limits on its S
and Ar abundances, its metallicity is estimated to be at most 1/30 of Solar.
Standard AGB models with initial masses and metallicities corresponding to the progenitor of TS 01 are not able to reproduce the observed abundance pattern. Taking into
account stellar rotation with an initial velocity of 100 km s1 , as we have done in a
model built for this purpose, greatly improves the situation. In TS 01, one may expect
even faster rotation, due to mass transfer in the binary, therefore bringing model results
and observations in even better agreement.
Acknowledgements
Many thanks to M. Richer, M. Pe
na, R. Szczerba, C. Charbonnel, L. Yungelson, R.
Napiwotzki, S. Simon-Diaz, L. Jamet, Y. Izotov, A. Fullerton, V. Suleimanov, J. Tomsick
who provided important input at various stages in our study of TS 01.
References
Cristallo S., Straniero O., Gallino R. et al., 2009, ApJ, 696, 797
Decressin T., Charbonnel C., Siess L. et al., 2009, arXiv:0907.5200
Jacoby, G. H., Feldmeier, J. J., Claver, C. F., et al. 2002, AJ, 124, 3340
Karakas A. I., Lattanzio J. C., 2003, PASA, 20, 393
Morisset C., 2006, IAUS, 234, 467
Pequignot, D., Tsamis, Y. 2005, A&A, 430, 187
Richer M. G., Tovmassian G., Stasi
nska G. et al., 2002, A&A, 395, 929
Stasi
nska G., Tovmassian G. H., Richer, et al., 2005, IAUS, 228, 323
Stasi
nska et al. 2009, A&A in press
Tovmassian G. H., Stasi
nska G., Chavushyan V. H., et al., 2001, A&A, 370, 456
Tovmassian G. H., Napiwotzki R., Richer, M. G., et al., 2004, ApJ, 616, 485
Tovmassian et al. 2009, to be submitted
1. Introduction
The study of the most iron poor stars in the halo offers a unique window on the
nature and the evolution of the first generations of stars which provided the matter
from which these halo stars were formed. Constraints on the first stellar generations may
provide interesting views on topical questions as the nature of the stars which reionize
the Universe, the initial mass function of the first stellar generations, the frequency of
long soft Gamma Ray Bursts at very low metallicity, the timescale for the mixing of the
newly synthesized products with interstellar medium material, or the conditions of star
formation in the very early life of our Galaxy.
In this context, many observed features of the very iron poor stars are quite surprising
and were not at all expected:
One expected that halo stars, having a very low [Fe/H], should show an important
scatter in their abundances. Indeed, at these very early times, stars are expected to form
from not well mixed clumps and thus should bear the nucleosynthetic signatures of a
few peculiar events. But Cayrel et al. (2004) find that the scatter in the abundances
The critical velocity is the surface equatorial velocity such that the centrifugal acceleration
compensates for the local gravity.
98
99
Figure 1. Left panel : variation of the abundances of various elements (in mass fraction) as
a function of the Lagrangian mass in the outer layers of a 60 M model at Z = 105 with
ini = 800 km s1 at the end of the core He-burning phase. The grey area covers the CO core.
Right panel : same as left part for a non rotating 60 M model at Z = 105 . Models computed
by Meynet & Maeder (2002) and Meynet et al. (2006).
of several elements ratios (e.g. [Cr/Fe]) is very small. It can be as low as 0.05 dex. In
contrast, r-process elements and nitrogen for instance reveal large star-to-star scatters
(Ryan et al. 1996; Honda et al. 2004; Andrievsky et al. 2009). Explanation are proposed
in Ishimaru et al. (2004, 2006); Cescutti (2008).
One expected that the very first generations of Pop. III stars were very massive and
contributed to the enrichment of the interstellar medium through Pair Instability Supernovae (PISN, Barkat et al. 1967; Bond et al. 1984). The nucleosynthetic pattern produced
by such events is well understood and presents specific features (Heger & Woosley 2005),
which were expected to be observed at least in some iron-poor halo stars. At the moment
no trace of PISN has been found (Cayrel et al. 2004). Some authors have suggested an
explanation why this signature has, up to now, escaped detection (Karlsson et al. 2008;
Ekstrom et al. 2008b).
Spectroscopic observations of halo stars (e.g. Spite et al. 2005) indicate a primary
production of nitrogen over a large metallicity range at low metallicity. According to the
chemical evolution models of Chiappini et al. (2006, 2008), primary nitrogen production
by non-rotating or slow-rotating Pop III stars is not sufficient to explain the observations.
For instance, primary nitrogen needs to be produced on larger ranges of masses and
metallicities than expected in non rotating standard models. A similar plateau is observed
in Damped Lyman Alpha (DLA) systems (Pettini et al. 2008).
Halo stars with log(O/H)+12 inferior to about 6.5 present higher C/O ratios than
halo stars with log(O/H)+12 between 6.5 and 8.2 (Akerman et al. 2004; Spite et al. 2005).
Again a similar trend is observed in DLAs (Pettini et al. 2008). This is not predicted by
slow rotating models.
The features listed above concern the bulk of the halo field stars. Now in addition
to this normal population, there exists a group of stars showing a very different composition. These stars are collectively called Carbon-Enhanced Metal Poor stars (CEMP)
and present high [C/Fe] ratios (see the review by Beers & Christlieb 2005). The scatter
in [C/Fe] is quite important indicating that these stars were not formed from a well
mixed reservoir but acquired their peculiar high carbon abundance from locally enriched
100
G. Meynet et al.
Table 1. Comparison of the outputs of various stellar models.
ini
MMS MpostMS Max. of Xs (14 N) Reference
km s1
M
M
Mass
M
60
60
60
60
0
108
105
5 104
0
0
0
0
0.00
0.18
0.21
0.78
0.00
0.09
0.22
13.29
0.00
2.34 1010
2.34 107
2.21 104
Ekstr
om et al. (2008a)
Meynet et al. (2006)
Meynet et al. (2006)
Decressin et al. (2007b)
60
60
60
60
0
108
105
5 104
800
800
800
800
2.32
2.38
6.15
20.96
2.41
33.64
16.57
21.79
3.54
1.02
2.07
3.86
104
102
103
102
Ekstr
om et al. (2008a)
Meynet et al. (2006)
Meynet et al. (2006)
Decressin et al. (2007b)
material. Some of these stars show also strong overabundances (with respect to iron) of
nitrogen, oxygen and other heavy elements. At first order, these stars do appear to be
formed from material having been processed mainly by H- and He-burning processes.
Very interestingly, halo stars in clusters also present some surprising features. While
part of the cluster population follows the same trend as the normal population of the
field, another part presents very different chemical patterns, with for instance, strong
depletions of oxygen and strong enhancements in sodium (see the review by Gratton
et al. 2004). These stars do appear to be formed from material having been processed
only by H burning.
Valid models should provide explanations for all these features. Ideally they should also
provide some predictions for not yet observed characteristics. In the following we discuss
the possible role of stellar axial rotation and we show how it can deeply affect the nucleosynthetic outputs of stars both in the massive (> 8 M ) and intermediate mass range
(2 < M/M < 8) and how it can explain some of the above observed features.
101
22
them are C, Ne. This is shown in Fig. 1, where the compositions of a non-rotating
and a rotating 60 M model at Z=105 are plotted.
A second consequence is that, during the core He-burning phase, the surface of the
star can be strongly enriched in CNO elements, increasing its surface metallicity by
many orders of magnitudes (Meynet & Maeder 2002; Hirschi 2007). This strong surface
metallicity enhancement can trigger important mass losses driven by radiatively line
driven winds. The material ejected in that way will bear the nucleosynthetic signature
of both H- and He-burning processes.
Another effect which occurs preferentially at low Z is the loss of mass through mechanical mass loss. What we call here mechanical mass loss is the loss of mass when the
surface velocity is so high that, at the equator, the centrifugal acceleration is equal to
the gravity. When such circumstances are realized, a tiny kick suffices to launch matter
into a keplerian orbit around the star and thus to form an equatorial disk. What does
happen to that material remains quite speculative. A reasonable hypothesis is that this
material is lost. This process does occur in near solar metallicity for sufficiently rapidly
rotating B-type stars. Not for O-type stars, because, at solar metallicity, they lose too
much mass and therefore angular momentum by stellar winds (Ekstr
om et al. 2008c).
At very low metallicity, one expects that the reaching of the critical limit will also occur
for more massive stars as O-type stars, because, in contrast to what happens at high
metallicities, radiatively driven winds are weaker during the MS phase and thus do not
remove the angular momentum brought to the surface by the meridional currents.
What is the importance of these various effects? How do they vary as a function of the
initial mass, metallicity and initial rotational velocities? Some elements of response can
be obtained from Table 1. We can note the following trends
The mass lost during the MS phase (MMS ), in rotating models, is due, at these
very low metallicities, mainly to the reaching of the critical limit (see column 4 of Table
1). The effect of mechanical mass loss can be estimated by comparing MMS for the
rotating and the non-rotating models. We see that the mechanical mass loss has a kind
of metallicity dependence. The model at higher metallicity loses much more mass by this
process than models at lower metallicities. The reason for this is due to three facts: first,
a given value of the velocity on the ZAMS corresponds, at high Z, to a higher value of the
ratio /crit than at low Z (see Fig. 12 in Ekstrom et al. 2008c). Second, the meridional
currents slow down the internal regions and accelerate the surface and are thus the main
agents which bring the velocity of the surface near the critical one. Third, these currents
are slower in more compact stars. Therefore, in low metallicity stars, it takes more time,
starting from a given initial velocity, to reach the critical limit . When the critical limit
is reached, the surface velocity remains near the critical value until the end of the MS
phase. This is due to the following behaviour: the star encounters the critical limit for the
first time, mass is lost at the equator, angular momentum is removed, the star evolves
away from the critical limit. Then meridional currents bring again angular momentum
to the surface making the star to evolve back against the critical limit. The cycle begins
again. The timescale for the evolution back to the critical limit will be shorter at higher
than at lower metallicities for the same reason as above, and thus this will favor higher
mass losses by this process. The material released by this mechanical mass loss is only
Enhancement in the abundance of 22 Ne in the He-core of the rotating model has a strong
impact on the s-process according to the work by Pignatari et al. (2008).
Let us note that, due to these three effects, without mass loss, stars would reach the critical
limit more rapidly at high metallicity. However, when the metallicity is near the solar metallicity,
the strong line driven stellar winds remove rapidly the angular momentum brought to the surface
by the meridional currents and prevent the high mass losing stars to reach the critical limit.
102
G. Meynet et al.
enriched in H-burning products, since it is ejected during the MS phase. Decressin et al.
(2007b,a) have studied the consequences of such models for explaining the chemically
peculiar stars observed in globular clusters.
In the models at Z equal to 108 and 105 , a phase exists where the star is in
the red part of the HR diagram, and presents strong enhancements of its surface CNO
content. The last column of Table 1 indicates the maximum value of the abundance of
14
N reached at the surface. In non-rotating models, no change of the surface abundance
occurs for Z 6 105 . In the non-rotating Z = 5 104 model, nitrogen enhancement
occurs as a result of mass loss. Only secondary nitrogen is produced in non metal-free,
non-rotating models. In the rotating models, primary nitrogen is produced. The most
efficient producers are models with Z = 108 and 105 . In these models, the apparition
at the surface of primary CNO elements triggers high mass loss rates (see column 5 in
Table 1). The material ejected in that way has been processed by both H- and He-burning
processes (Meynet et al. 2006; Hirschi 2007). For the velocities and mass considered
here, this only occurs for non-zero very low metallicity, but not for Pop III stars. At
near solar metallicities, this does not occur for models with the same initial angular
momentum content. Why? For Z=0, we shall see below that the mixing during the
core He-burning phase is less efficient than in non-zero metallicity stellar models. At
higher metallicities, rotational mixing is less efficient. Furthermore the line driven winds
become more important, and in very massive stars, they may remove the H-burning
regions early during the core He-burning phase preventing primary nitrogen production.
In the following, we argue that the material ejected under the form of this wind triggered
by the self enrichment of the surface, presents strong similarities with the abundance
pattern observed at the surface of the CEMP stars.
The rotating models produce primary nitrogen (and also primary 13 C and 22 Ne).
The most efficient producers are the models with Z equal to 108 and 105 . The pop III
stars produce less primary nitrogen for the following reason: due to the absence of CNO
elements, these models begin to convert H into He through the pp chains. The energy
output from these chains is not sufficient to compensate for the high luminosity of the
star. Therefore the rest of the energy has to be extracted from the gravitational reservoir
and the star must contract. Contraction occurs until the central temperatures reach
sufficient high values to activate the triple alpha reaction. Some carbon is then produced
and the CNO cycle can be activated. From this stage on, the H-burning is pursued as
in more metal rich stars, through the CNO cycle. A consequence of this is that the core
H-burning occurs at temperatures very similar to that of the core He-burning. Thus, in
Pop III stars, at the end of the core H-burning phase, the core does not need to contract
a lot in order to re-activate the He-burning reactions. This absence of strong contraction
maintains the star in the blue part of the HR diagram during most of the core He-burning
phase. It also prevents a strong gradient of angular velocity to form at the border of the
core and therefore efficient shear mixing. Less primary nitrogen is formed. In more metal
rich models, less primary nitrogen is produced both because mixing is less efficient and
because the H-burning shell is more distant from the He-burning core.
103
Figure 2. Left panel : Composition of the mixture composed of wind material of 60 M models at
Z = 105 and of layers ejected by the supernova diluted with 10 times more interstellar material.
The different curves correspond to the model WIND+SUPERNOVA in Table 2, various values
of the mass-cut are considered. The continuous and short dashed lines curves correspond to
the rotating model with ini = 800 km s1 . The dotted lines and long dashed curves curves
correspond to the non-rotating models (models from Meynet & Maeder 2002). As the mass cut
increases (downwards), the amount of carbon and oxygen ejected by the supernova decreases.
The stars show the observed values for the most iron poor star known today (Frebel et al.
2008). The vertical hatched zones show the range of observed values for CEMP-no stars with
log g superior to 3.8 (left grey zone, 5 stars in the sample) and with log g inferior to 3.8 (right
red zone, 20 stars in the sample). The values were taken from Tables 1 and 2 of Masseron
et al. (2009). Only upper limits are given for [O/H], this is why the columns extend down to
the bottom of the figure. Right panel : The continuous line shows the composition obtained by
mixing the outer envelope (mass above the CO core) at the E-AGB phase of our rotating 7 M
stellar model with 100 times more of interstellar medium. The dotted curve with squares shows
the composition obtained in the same way (mass above the CO core and interstellar medium)
using our analog non-rotating 7 M stellar model.
In order to compute the evolution with time or with [Fe/H] of such stars, detailed
chemical evolution models have to be used. The models by Chiappini et al. (2006) show
that the N/O plateau observed at low Z can be well reproduced when yields of rotating
models (Vini = 800 km s1 ) are used. The slower rotating models can produce a plateau
for the nitrogen to oxygen ratio at low metallicities but the level is two to three orders of
magnitude below the observed one. Non-rotating models would still be much lower and
would not produce a plateau at low Z. (see Fig. 4 Chiappini et al. 2006). Not only the N/O
ratio is very well reproduced by these models but also the C/O upturn mentionned in
the introduction. It would be extremely interesting to obtain information of the 12 C/13 C
ratio in very metal poor stars, whose surface still reflects the abundances of the cloud
from which these stars were formed. Chiappini et al. (2008) show that the slow and fast
rotating models predict very different values for this ratio.
Stars can also form from interstellar material enriched by one or a very small number
of nucleosynthetic events. In that case one expects some important scatter from one star
to another and also some strong differences with respect to stars formed from the well
mixed reservoir. The CEMP stars are likely formed in that way.
We show in Table 2 the chemical composition of various components (winds, wind plus
supernova ejecta, envelope of an early-AGB star) of material ejected by various stellar
models. The composition indicated in that table would be the one of stars formed from
pure ejecta. Looking at the results of Table 2, the following trends can be deduced :
104
G. Meynet et al.
Table 2. Chemical composition of the ejecta of various models in mass fraction. The number
ratio 12 C/13 C is also indicated.
Mini
Zini
ini Meje
XH
XHe
XC12
12
C/
13
XN14
XO16
WIND
85
60
40
108 800
108 800
108 700
11.4
5.4
6.8
3.0e-02 5.0e-02
5.5e-03 1.7e-04
8.6e-04 5.8e-04
60
105 800
4.6
4.2e-04 1.2e-05
60
60
60
60
60
60
60
10
105
105
105
105
105
105
0
0
800
800
800
800
800
37
41.8
27
29
30
32
34.5
0.48
0.42
0.42
0.39
0.38
0.37
0.33
0.52
0.57
0.58
0.61
0.61
0.60
0.55
1.7e-07
6.4e-03
2.0e-05
3.4e-05
8.6e-04
2.4e-03
5.4e-03
3.65e+00
1.55e+05
4.70e+00
7.32e+00
1.90e+02
5.39e+02
1.37e+03
4.7e-06
4.3e-06
7.3e-04
8.5e-04
8.4e-04
8.1e-04
7.3e-04
2.1e-06
5.1e-03
1.7e-05
8.5e-05
6.9e-03
2.3e-02
1.0e-01
7
7
105
0
105 800
6.1
5.7
E-AGB ENVELOPE
Stars made of pure ejecta (or diluted with small amount of interstellar material)
would be He-rich. This is true whatever the source of the ejecta, rotating or non-rotating,
wind material, wind plus supernova material, or envelope of an early AGB star. Such
stars would be depleted in lithium.
Low values for the 12 C/13 C ratio indicates that the ejecta are rich in CNO-processed
material. This conclusion does not depend on the degree of dilution with interstellar
material in case we consider ejecta of spinstar models. Indeed. in rotating models, both
12
C and 13 C are produced through primary channels in so large quantities that the
dilution should be enormous for changing it.
We see that, when the contribution of the supernova increases, the 12 C/13 C ratio
becomes larger. Also the abundances of carbon and oxygen increase. This is because some
layers, rich in He-burning products, are ejected. We conclude that stars presenting high
[N/C] and [N/O] ratios cannot be mainly made from material processed by He burning
and thus from supernova ejecta (except if the supernova only eject the outermost layers).
This is also independent of the degree of dilution with interstellar material. It is also
independent of the models considered (rotating or non rotating).
We see that rotating models are the only ones, among those presented in Table 2,
showing strong enhancements of the three CNO elements simultaneously. The reason for
this is that, in these models, mixing of He-burning products with H-burning products
occurred in the star which has ejected the material.
Comparisons with observed values are made in Fig. 2. Composition of pure ejecta mixed
with some amount of interstellar material are plotted. An important point to mention in
order to correctly interpret this figure is the following: for the rotating models, the positions of the different curves in this diagram do not much depend on the initial metallicity.
Here the quantities are plotted for models at Z = 105 . Would we have plotted data
obtained from the rotating model at Z = 108 , similar results would have been obtained.
This is because all the three isotopes (and also 13 C) are produced by (quasi) metallicity
independent channels. For the non-rotating models, we have a different situation. While
12
C and 16 O have a strong primary component, 14 N and 13 C are only secondary. This
105
means that, would we have plotted the results for a model with Z = 10 for instance,
the point corresponding to 14 N would have shifted downwards by about 3 dex! Keeping
this in mind, we see that that non-rotating massive stars cannot fit the observed values
of the Frebel star for instance (Z 106 ). In contrast, rotating models produce situations where the three elements are enhanced. This is true for massive stars but also for
intermediate mass stars. We also see that with a dilution factor of 10, CEMP stars made
up of winds and of a small amounts of supernova ejecta would be He-rich. The 12 C/13 C
ratios would be as indicated in Table 2 (the mass ejected is equal to 60 minus Mcut ).
Greater dilutions factors (about 10 times higher) seem to be required in case CEMP
stars result from the mixture of an E-AGB envelope with interstellar material. In that
case, these stars would not be He-rich. They would also show high lithium abundances
at least equal to the Spite Li plateau in case the depletion process in those very metal
poor stars occurred as it occurred in the stars of the Spite plateau. The 12 C/13 C ratios
would be equal to 105 and 2560 for the rotating and non-rotating case respectively.
References
Akerman, C. J., Carigi, L., Nissen, P. E., Pettini, M., & Asplund, M. 2004, A&A, 414, 931
Andrievsky, S. M., Spite, M., Korotin, S. A., et al. 2009, A&A, 494, 1083
Barkat, Z., Rakavy, G., & Sack, N. 1967, Physical Review Letters, 18, 379
Beers, T. C. & Christlieb, N. 2005, ARA&A, 43, 531
Bond, J. R., Arnett, W. D., & Carr, B. J. 1984, ApJ, 280, 825
Cayrel, R., Depagne, E., Spite, M., et al. 2004, A&A, 416, 1117
Cescutti, G. 2008, A&A, 481, 691
Chiappini, C., Ekstr
om, S., Meynet, G., et al. 2008, A&A, 479, L9
Chiappini, C., Hirschi, R., Meynet, G., et al. 2006, A&A, 449, L27
Decressin, T., Charbonnel, C., & Meynet, G. 2007a, A&A, 475, 859
Decressin, T., Meynet, G., Charbonnel, C., Prantzos, N., & Ekstr
om, S. 2007b, A&A, 464, 1029
Ekstr
om, S., Meynet, G., Chiappini, C., Hirschi, R., & Maeder, A. 2008a, A&A, 489, 685
Ekstr
om, S., Meynet, G., & Maeder, A. 2008b, in IAU Symposium, Vol. 250, IAU Symposium,
ed. F. Bresolin, P. A. Crowther, & J. Puls, 209216
Ekstr
om, S., Meynet, G., Maeder, A., & Barblan, F. 2008c, A&A, 478, 467
Gratton, R., Sneden, C., & Carretta, E. 2004, ARA&A, 42, 385
Heger, A. & Woosley, S. 2005, in IAU Symposium, Vol. 228, From Lithium to Uranium: Elemental Tracers of Early Cosmic Evolution, ed. V. Hill, P. Francois, & F. Primas, 297302
Hirschi, R. 2007, A&A, 461, 571
Honda, S., Aoki, W., Kajino, T., et al. 2004, ApJ, 607, 474
Ishimaru, Y., Wanajo, S., Aoki, W., & Ryan, S. G. 2004, ApJ, 600, L47
Ishimaru, Y., Wanajo, S., & Prantzos, N. 2006, in International Symposium on Nuclear Astrophysics - Nuclei in the Cosmos
Karlsson, T., Johnson, J. L., & Bromm, V. 2008, ApJ, 679, 6
Maeder, A. 2009, Physics, Formation and Evolution of Rotating Stars, ed. A. Maeder
Maeder, A. & Meynet, G. 2001, A&A, 373, 555
Maeder, A., Meynet, G., Ekstr
om, S., & Georgy, C. 2009, Communications in Asteroseismology,
158, 72
Masseron, T., Johnson, J. A., Plez, B., et al. 2009, ArXiv e-prints
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, A&A, 447, 623
Meynet, G. & Maeder, A. 2002, A&A, 390, 561
Meynet, G. & Maeder, A. 2005, A&A, 429, 581, paperXI
Pettini, M., Zych, B. J., Steidel, C. C., & Chaffee, F. H. 2008, MNRAS, 385, 2011
Pignatari, M., Gallino, R., Meynet, G., et al. 2008, ApJ, 687, L95
Ryan, S. G., Norris, J. E., & Beers, T. C. 1996, ApJ, 471, 254
Spite, M., Cayrel, R., Plez, B., et al. 2005, A&A, 430, 655
1. Introduction
John von Neumann (von Neumann 1948) proposed a way to deal with the intractable
problem of hydrodynamic turbulence, by (1) using numerical simulation on computers
to construct turbulent solutions of the hydrodynamic equations, (2) building intuition
from study of these solutions, and (3) constructing analytic theory to describe them.
He proposed that iterating this procedure could lead to a practical understanding of
turbulent flow. The computer power available at that time was totally inadequate to
compute hydrodynamics on sufficiently refined grids to produce turbulent flow; numerical
viscosity restricts the effective Reynolds number. Today, computing power is adequate for
the simulations of truly turbulent, three-dimensional (3D), time dependent, compressible
flows, so we have begun a program based upon von Neumans proposal.
Turbulent flow in its many guises (e.g., convection, overshooting, shear mixing, semiconvection, etc.) is probably the weakest aspect of our theoretical description of stars
(and accretion disks). The full problem that faces us includes rotation, magnetic fields,
and multi-fluids (to account for compositional heterogeneity, diffusion, radiative levitation, and nuclear burning). In this paper we describe the progress made toward von
Neumanns goal. We plan to replace the venerable mixing-length theory (MLT) with a
physics-based mathematical theory which can be tested by refined simulations and terrestrial experiment (e.g., laboratory fluid experiments, meteorological and oceanographic
observations). Particularly relevant are high-energy density plasma (HEDP) experiments,
which now can access regions of temperature and density that overlap stellar conditions
up to helium burning (Remington et al. 2000, 2006, Drake 2006), and deal with plasma
and magnetic fields, just like star matter, not with an unionized fluid like air or water.
107
(2.1)
where the term on the left-hand side is the Lagrangian time derivative of the composition
in a comoving spherical shell with velocity v, the first term on the right-hand side is the
mixing due to rotation and turbulent velocities u across the Lagrangian shell boundary, and the last term is the composition change due to nuclear reactions which change
species i. Thus,
Ri = Yi i + Yj j
Yi Yk NA hvi + Yl Ym NA hvi +
(2.2)
where the terms on the right-hand side represent all the ways in which species i can be
made or destroyed. The advection operator
u Yi
(2.3)
involves a velocity field u which is determined non-locally and a first order spatial gradient
Yi . This is replaced by
Yi
[(4r2 )2 D(
)],
(2.4)
m
m
which has a second order derivative in space and a phenomenological local diffusion
coefficient D. Except for contrived cases, these are the same (zero) only in the limit that
composition is homogeneous. We need a major community effort to base stellar mixing
algorithms on physics, comparable to the efforts led by Willy Fowler for nuclear reaction
rates, so that both the advection and reaction terms are reliable.
108
include multiple burning shells show interactions between the shells; 3D simulations of
multiple shells are planned. Neither have the pulsations, nor the interaction of burning
shells, been included in any 1D progenitor models to date.
Another novel feature found by Meakin & Arnett (2007b) is entrainment at convection
boundaries. The physics of the process is interesting; it involves the erosion of a stablystratified layer by a turbulent velocity field, mediated by nonlinear g-mode waves (Meakin
& Arnett 2007a).
While these simulations do not contain an entire star, and thus limit the accuracy of the
description of low-order modes, whole star simulations are developing enough resolution
to exhibit turbulent flows (Brun 2009). Since we find that even modest resolution will
give reliable average quantities (see below), we expect the simulation step to be soon
generalized and extended to include rotation and magnetic fields.
5. Future Prospects
5.1. Comparison to Stars
These insights are being implemented into an algorithm for stellar evolution. The idea is
to use fully 3D phenomena, found in simulations and captured by analytic theory, by projecting them onto a 1D geometry, as used in stellar models. Unlike MLT, the algorithm is
nonlocal and time-dependent, not static. It should be applicable to deep, nearly adiabatic
convection without modification. Because it have some time dependence it should be useful for models of pulsating stars. It will replace overshooting and semi-convection
because it uses the bulk Richardson criterion for the extent of convection (Meakin &
Arnett (2007b)). Because the turnover flow in the convection zone is averaged over, this
algorithm is not limited by the corresponding Courant condition, and is appropriate for
John Lattanzio has dubbed this the 321D algorithm.
109
stellar evolution over long time scales. We emphasis that failure is possible now that free
parameters are being eliminated, so that inadequacies of the theory will be evident.
5.2. 3D Hydrodynamic Atmospheres
The 321D approach merges naturally with work on 3D atmospheres (Arnett, Meakin,
& Young 2009 and above) and work on accretion disks (Balbus & Hawley 1998, Balbus
2009, Blackman 2010, and Blackman & Pessah 2010). These approaches all use mean
field equations, starting from the same general equations of mass, momentum and energy
conservation for fluids, and use averaging to derive general properties. Because of this,
physical processes are not introduced in patchwork fashion, but a logical necessities of the
conservation laws. Insights into MHD in disks can spark insight into angular momentum
transport in stars, and insights into stellar turbulence should do the same for accretion
disk theory. As Bohdan Paczynski was fond of saying, accretion disks are just flat stars.
5.3. Rotation and Magnetic Fields
Perhaps the greatest challenge for stellar evolution is the treatment of angular momentum
transport. The rigid rotation of the Suns radiative core, and the differential rotation of
the convective envelope, inferred from helioseismology, seem to have been a surprise. If
we wish to understand GRBs and hypernovae, most workers seem to assume that a key
role is played by rotation in the gravitational collapse and explosion (an idea dating back
to Fred Hoyle, at least). We expect to have little success if we extrapolate from the sun,
using algorithms that give the wrong qualitative behavior. The von Neumann proposal,
generalized to include rotation and magnetic fields, offers hope.
Figure 1 shows the results of a first step toward understanding that problem. Our
convection in a box simulation is continued, but with the box being rotated around the
polar axis. The initial rotation is rigid body, so that the specific angular momentum is
Figure 1. Specific angular momentum versus radius. The convection zone readjusts to
constant specific angular momentum, not rigid body rotation (Meakin & Arnett 2010).
110
quadratic in the radius. After a few turnover times, the results in Figure 1 is obtained,
in which the specific angular momentum tends toward a constant in the convection zone,
while remaining rigid body outside. Further, magnetic instabilities (MRI, etc.) seem to
cause radiative regions to tend toward rigid body rotation, even if they initially have
some other rotation law. A perusal of the literature suggests that in stellar evolution, the
opposite is often assumed.
6. Conclusion
The von Neumann proposal of using computation and theory together seems to work
well for stellar turbulence, and promises to be of value for the more complex problem
which includes rotation and magnetic fields. Perhaps the best aspect of this approach
is that it certainly will make new predictions of phenomena which hitherto have been
essentially in the realm of observation only.
References
Arnett, D., 1996, Supernovae and Nucleosynthesis, Princeton University Press, Princeton NJ
Arnett,D., Meakin, C., Young, P. A., 2009a, ApJ, 690,1715
Arnett,D., Meakin, C., Young, P. A., 2009b, ApJ, submitted, arXiv0910.0821
Asplund, M., 2005, A&A Rev., 43, 481
Balbus, S. A., 2009, MNRAS, 395, 2056
Balbus, S. A., & Hawley, J. F., 1998, Rev. Mod. Phys., 70, 1
Blackman, E., 2010, Astron. Nachr., in press
Blackman, E., & Pessah, M., 2009, ApJL, 704, L113
Boris, J., 2007, in Implicit Large Eddy Simulations, ed. F. F. Grinstein, L. G. Margolin, & W.
J. Rider, Cambridge University Press, p. 9
Brun, A. S., 2009, ApJ, 702, 1078
Drake, R. P., 2006, High-Energy-Density Physics, Springer, Berlin, 2006, p. 6
Fontenla, J. M., Avrett, E., Thuillier, G., & Harder, J., 2006, ApJ, 639, 441
Kolmogorov, A. N., 1941, Dokl. Akad. Nauk SSSR, 30, 299
Kolmogorov, A. N.,1962, J. Fluid Mech., 13, 82
Landau, L. D. & Lifshitz, E. M., 1959, Fluid Mechanics, Pergamon Press, London
Meakin, C., & Arnett, D., 2006, ApJL, 637, L53
Meakin, C., & Arnett, D., 2007a, ApJ, 665, 690.
Meakin, C., & Arnett, D., 2007b, ApJ, 667, 448
Meakin, C., & Arnett, D., 2010, in preparation.
Nordlund, A., & Stein, R., 2000, The Impact of Large-Scale Surveys on Pulsating Star Research,
ASP Conf. Series, 203, 362
Remington, B. A., Drake, R. P., Takabe, H., & Arnett, D., 2000, Phys. Plasmas, 7, 1641
Remington, B. A., Drake, R. P., & Ryutov, D. D., 2006, Rev. Mod. Phys., 78, 755
Stein, R. F., and Nordlund,
A., 1998, ApJ, 499, 914
von Neumann, J., 1948, in Collected Works, Volume VI, 1963, Pergamon Press, Oxford, p. 467-9
Woodward, P. R., 2007, in Implicit Large Eddy Simulantions, ed. F. F. Grinstein, L. G. Margolin,
& W. J. Rider, Cambridge University Press, p. 130
Abstract.
A significant fraction of metal-poor stars have large over-abundances of carbon, and are called
Carbon-Enhanced Metal-Poor (CEMP) stars. Most of CEMP stars also show excesses of heavy
neutron-capture elements like Ba, indicating that their origin is the nucleosynthesis in AGB
stars. Remaining CEMP stars that have Ba abundances as low as non-carbon-rich stars appear
in the lowest metallicity range ([Fe/H]. 2.5), and connections with the two most iron-deficient
stars (so-called Hyper Metal-Poor stars) are suggested. Although the origins of the carbonexcesses in these objects have not been well identified, some objects suggest contributions of
faint supernovae. Remaining problems on CEMP stars, such as the binary fraction, excess of
r-process elements, are discussed.
Keywords. stars:abundances, stars:AGB and post-AGB, stars:carbon, supernovae
1. Introduction
Carbon-Enhanced Metal-Poor stars (CEMP stars) in the Galactic halo have been
known as CH stars (Keenan 1942) or subgiant CH stars (Bond 1974) that show strong
CH absorption bands compared to other stars having similar temperatures. A number
of stars showing strong CH bands have been identified by the HK survey (Beers et
al. 1992), suggesting that the fraction of carbon-enhanced stars is much higher in the
halo (the metal-poor range) than that in the disk (the metal-rich range). The recent
estimates of the fraction of carbon-rich stars have confirmed this, although the derived
values distribute rather wide range (1025%; e.g. Cohen et al. 2006; Marsteller et al.
2006; Frebel et al. 2006).
The importance of CEMP stars was widely recognized by the discoveries of Hyper
Metal-Poor (HMP) stars HE 01075240 and HE 13272326 by Christlieb et al. (2002)
and Frebel et al. (2005), respectively. These stars are called most metal-poor because of
their low iron abundances ([Fe/H]< 5), but their carbon abundance ratios are extremely
high ([C/Fe] +4). Although the estimate of carbon abundance ratios are sensitive to
the 3D effects in stellar atmospheres, the large overabundance of carbon in these objects
is clear, and the origin of this peculiar abundance pattern is considered as a key to
understanding the nucleosynthesis of first generations of stars (Beers & Christlieb 2005).
In more general, CEMP stars are believed to contain useful information to understand
the nucleosynthesis in the early Galaxy. The criterion of the carbon abundance ratio
in the definition of carbon-enhanced objects is dependent on the authors of previous
studies: some assume [C/Fe]= 1.0 but others do 0.5 or 0.7, and some study takes the
evolutionary stage of the object into consideration. However, as seen in Figure 1, the
distributions of carbon abundance ratios clearly split into two groups, and the estimate
of the fraction of carbon-enhanced stars is not significantly changed by the criterion.
Here we simply define CEMP stars as those having [C/Fe]> +1.0.
111
112
Wako Aoki
2. Origins of carbon-excesses
For understanding the origins of the carbon over-abundances, the neutron-capture
element Ba is a key element. This element is efficiently produced by the s-process in
thermally pulsing AGB stars, which are also a major source of carbon. Singly ionized Ba
has strong resonance doublet lines as well as other weaker lines in the optical range, that
makes it possible to detect this element for a wide abundance range.
Figure 2 (left) shows the Ba abundance ratio ([Ba/Fe]) as a function of [Fe/H]. A
majority ( 80%) of CEMP stars show large over-abundances of carbon. A correlation
between [C/Fe] and [Ba/Fe] are seen in Figure 2 (right) for these Ba-enhanced stars. This
means that the carbon-excesses of most of CEMP stars are attributed to AGB stars in
which Ba are also synthesized by the s-process. These stars are called as CEMP-s stars
(Beers & Christlieb 2005). It should be noted that the Ba-excess in some CEMP stars is
possibly due to the r-process rather than the s-process. Measurements of other neutroncapture elements (e.g. Eu) are necessary to separate the two possibilities. However, among
more than 10 CEMP stars for which detailed abundance measurements have been made,
only one object (CS 22892052) is known to have a large Ba-excess that is attributed to
the r-process.
On the other hand, there are a small fraction of CEMP stars that have Ba abundances
as low as other non-carbon-enhanced stars. These stars are called CEMP-no stars
(Beers & Christlieb 2005). Although the Ba abundances have not been determined for
the two HMP stars, the low upper limits suggest that these two stars are likely belong
to the class of CEMP-no stars.
It is clear in Figure 2 (left) that CEMP-no stars appear in the lowest metallicity range
([Fe/H]< 2.5), while CEMP-s stars are seen in [Fe/H]& 3. The difference of the
metallicity distribution between the two classes is evident in Figure 3 (left) that shows
the histogram of [Fe/H] for CEMP-s and CEMP-no stars. We note that CEMP stars for
which Ba is not detected are not included in the histogram for CEMP-no stars. These
objects have [Fe/H]< 3, including the two HMP stars. The difference of the metallicity
distribution is more remarkable if these stars are included in the comparison.
113
Figure 2. left: [Ba/Fe] as a function of [Fe/H] for CEMP stars (filled circles and squares) and
carbon-normal stars (crosses). For some extremely metal-poor stars, including the two HMP
stars, only upper limits of Ba abundance are determined. right: Same as the left panel, but for
[Ba/Fe] as a function of [C/Fe].
Another difference between CEMP-s and CEMP-no stars is found in the carbon abundance distribution. Figure 3 (right) shows the histograms of [C/H]. The [C/H] distribution
of CEMP-s stars has a peak at [C/H]= 0.5 0.0 and a cut-off at [C/H] 0. This cut-off
probably indicates the carbon abundance produced by AGB stars, which is not significantly dependent on metallicity (Ventura et al. 2002). The tail found in the lower [C/H]
side can be interpreted as the result of dilution of the carbon-rich material produced by
an AGB star in the stellar envelope after the mass transfer from the companion AGB
star.
By contrast, the [C/H] values of CEMP-no stars distribute a wide range, suggesting the
origins of the carbon-excess of these stars are different from that of CEMP-s stars. This is
a problem probably related to the origins of HMP stars. To explain the carbon-excesses of
HMP stars, several models have been proposed, including faint supernovae (see below),
rotating massive stars from which significant amount of carbon is ejected (Maynet et al.
2006), and AGB stars that produce carbon but do not yield heavy elements. These models
might also be applied to CEMP-no stars.
Among CEMP-no stars, the two extremely metal-poor stars CS 22949037 and CS 29498
043 have large over-abundances of elements as well as C, N and O (McWilliam et al.
1995; Aoki et al. 2002). elements are usually considered to be products of core-collapse
supernovae. Hence, the progenitors of these two stars are most likely massive stars that
exploded as faint supernovae, explosions producing only small amount of Fe. Such supernovae are proposed to explain the abundance patterns of HMP stars, and successfully
explain also these -enhanced stars (e.g., Umeda & Nomoto 2003).
Another example that is well explained by the faint supernova models is recently
discovered. That is the bright CEMP-no subgiant BD+44 493 that has extremely low
[Fe/H] (= 3.7) and moderately high [C/H](= +1.3). The high O/C ratio as well as low
Ba abundance of this object are not explained by AGB nucleosynthesis models, while
the low N abundance does not support the rotating massive star scenario. The details of
this object is reported by Ito et al. (in this volume) as well as by Ito et al. (2009).
Our conclusion here is that we have at least some evidence for the contribution of faint
supernovae to some carbon-enhanced, extremely metal-poor stars having low Ba abundances. However, other possibilities are not excluded for other CEMP-no stars. Indeed,
114
Wako Aoki
Figure 3. Histograms of [Fe/H] (left) and [C/H] (right) for CEMP-no stars (hatched) and
CEMP-s stars (blank with solid lines).
the wide [C/H] distribution and the variety of -elements abundance ratios suggest that
the origins of carbon-excesses in these objects are not unique. Further investigations of
these objects are clearly desired to understand the origins of CEMP stars and HMP stars.
115
different from that at higher metallicity. These objects are recently discussed by Masseron
et al. (2009) as CEMP-low-s stars, and further observational studies are desired.
3.3. CEMP-rs stars
A significant fraction of Ba-enhanced CEMP stars also show large over-abundance of
Eu compared to the value expected from the Ba and La abundances assuming s-process
nucleosynthesis. Such stars are called CEMP-rs stars, as both s- and r-process elements
seem to be enhanced. A simple interpretation is that these are CEMP-s stars, but are
also affected by the r-process before or after the star formation. However, the fraction of
CEMP-rs stars among Ba-enhanced CEMP stars is high, while r-process-enhanced stars
with normal carbon abundance are quite rare in metal-poor stars. This class of objects
has also been discussed by Masseron et al. (2009) as well as other studies referred in the
paper.
3.4. Li abundances in some CEMP-s stars
Li abundances of CEMP-s stars are usually low even in turn-off and subgiant stars. This
is interpreted as the result of mass accretion from an AGB star in which Li is depleted.
However, at least two CEMP-s stars (LP 706-7 and CS 22898-027) have Li abundance
as high as the Spite plateau value. One possibility is that Li is enriched in the AGB
star that provides carbon and heavy elements. The model calculation of Iwamoto et al.
(2004) suggests production of Li in metal-poor AGB stars. However, the reason for the
fact that the Li abundance becomes similar to the Spite plateau value is not clear (Ryan
et al. 2005). More detailed studies of Li abundances in CEMP stars, e.g., more accurate
determination of the abundance, are required to understand the origin of Li in these
stars.
4. Summary
CEMP stars are related to many topics discussed in the symposium. A majority of
CEMP stars show large over-abundances of Ba, and the origin of carbon-excesses is
explained by AGB nucleosynthesis. However, some problems (e.g. large enhancement
of r-process elements or high Li abundance in some objects) remain to be solved, and
further studies of these stars will shed new light on AGB models. Although the formation
mechanism of CEMP-no stars is still unclear, we found some evidence for the scenario
that faint supernovae are a source of carbon-enhanced material in the very early Galaxy.
Since the sample of CEMP-no stars is quite small, further searches for these objects,
most of which have very low metallicity, and detailed follow-up studies are indispensable.
References
Aoki, W., Beers, T. C., Christlieb, N., Norris, J. E., Ryan, S. G., & Tsangarides, S. 2007, ApJ,
655, 492
Aoki, W., Norris, J. E., Ryan, S. G., Beers, T. C., & Ando, H. 2002, ApJ, 576, L141
Beers, T. C., & Christlieb, N. 2005, ARAA, 43, 531
Beers, T. C., Preston, G. W., & Shectman, S. A. 1992, AJ, 103, 1987
Bond, H. E. 1974, ApJ, 194, 95
Christlieb, N., et al. 2002, Nature, 419, 904
Cohen, J. G., et al. 2006, AJ, 132, 137
Keenan, P. C. 1942, ApJ, 96, 101
Frebel, A., et al. 2005, Nature, 434, 871
Frebel, A., et al. 2006, ApJ, 652, 1585
116
Wako Aoki
Ito, H., Aoki, W., Honda, S., & Beers, T. C. 2009, ApJ, 698, L37
Iwamoto, N., Kajino, T., Mathews, G. J., Fujimoto, M. Y., & Aoki, W. 2004, ApJ, 602, 377
Lucatello, S., Beers, T.C., Christlieb, N., Barklem, P.S., Rossi, S., Marsteller, B., Sivarani, T.,
& Lee, Y., ApJ, 625, 825
Marsteller, B. E., Beers, T. C., Sivarani, T., Rossi, S., Knapp, J., Plez, B., Johnson, J., &
Masseron, T. 2006, Bulletin of the American Astronomical Society, 38, 1229
Masseron, T., Johnson, J. A., Plez, B., Van Eck, S., Primas, F., Goriely, S., & Jorissen, A. 2009,
arXiv:0901.4737
McClure, R. D. 1984, ApJ, 280, L31
McWilliam, A., Preston, G. W., Sneden, C., & Searle, L. 1995, AJ, 109, 2757
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, A&A, 447, 623
Preston, G. W., & Sneden, C. 2001, AJ, 122, 1545
Ryan S. G., Aoki W., Norris J.E., Beers T.C. 2005, ApJ, 635, 349
Umeda, H., & Nomoto, K. 2003, ApJ, 422, 871
Ventura, P., DAntona, F., & Mazzitelli, I. 2002, A&A, 393, 215
Sterrekundig Instituut Utrecht, P.O. Box 80000, NL-3584 TA Utrecht, The Netherlands,
email: O.R.Pols@uu.nl
2
Institut dAstronomie et dAstrophysique, Universite Libre de Bruxelles, CP226, Boulevard
du triomphe, B-1050 Bruxelles, Belgium
3
Dept. of Physics and Astronomy, McMaster University, Hamilton, Ontario, L8S 4M1, Canda
4
School of Mathematical Sciences, Monash University, P.O. Box 28M, Victoria 3800, Australia
A large fraction, between 10 and 25%, of very metal-poor stars in the Galactic halo are
carbon-rich objects, with enhancements of carbon relative to iron exceeding a factor 10.
The majority of these carbon-enhanced metal-poor (CEMP) stars show enhancements of
heavy s-process elements and have been found to be spectroscopic binary systems. Many
of their properties are well explained by the binary mass transfer scenario, in which a
former asymptotic giant branch (AGB) companion star has polluted the low-mass star
with its nucleosynthesis products. The same scenario predicts the existence of nitrogenrich metal-poor (NEMP) stars, with [N/C] > 0.5, from AGB companions more massive
than about 3 solar masses. In contrast to CEMP stars, however, such NEMP stars are very
rare. Recent studies suggest that the high frequency of CEMP stars requires a modified
initial mass function (IMF) in the early Galaxy, weighted towards intermediate-mass
stars. Such models also implicitly predict a large number of NEMP stars which is not
seen.
Here we investigate whether the observed incidence of CEMP and NEMP stars among
metal-poor stars and their abundance patterns can be understood without invoking a
change in the IMF. We study the formation and evolution of CEMP stars by means of
a binary population synthesis technique, in which we simulate the evolution of 2 106
binaries at metallicity [Fe/H] = 2.3, following the evolution and surface abundances
of both binary components as well as their mutual interactions. This approach allows
us to explore uncertainties in the CEMP-star formation scenario by parameterization
of uncertain input physics. In particular, we consider the uncertainty in the physics of
third dredge up in the AGB primary, binary mass transfer and mixing in the secondary
star. We confirm earlier findings that with current detailed AGB models, in which third
dredge up is limited to stars more massive than about 1.25 M, the large observed CEMP
fraction cannot be accounted for. We find that efficient third dredge up in low-mass (less
than 1.25 M) metal-poor AGB stars may offer at least a partial explanation to the large
observed CEMP frequency, while remaining consistent with the small observed NEMP
frequency. Our models show that most CEMP stars are also expected to be enriched in
fluorine, of which the recent detection of a large F abundance in the CEMP star HE
1305+0132 is a clear example. For full details we refer to the paper by Izzard et al.
(2010).
References
Izzard, R. G., Glebbeek, E., Stancliffe, R. J., Pols, O. R. 2010, A&A in press (arXiv:0910.2158)
117
1. Introduction
Most of CEMP stars have Ba abundance available, whereas only a few of them have
Eu abundance published and, some of them, only an upper limit. The lack of accurate
abundances for a more pure r-element such as Eu has made the abundance pattern of
many stars unclear, since the ratio of the two elements are necessary to correctly classify
them. As an example, some stars classified as CEMP-s through Ba abundance were only
found to be CEMP-r/s after the determination of Eu abundance by Tsangarides (2005).
CEMP stars with [Ba/Fe] < 0 could be CEMP-no or CEMP-r and those with [Ba/Fe]
> 1 could be CEMP-s or CEMP-r/s, depending on the Eu abundance.
119
Figure 1. Abundance ratios for CEMP stars from the literature and 10 stars from this work.
In panel a), dotted lines represent the region of CEMP-r/s and the dashed line is the limit
for r-process only. The dotted line in panel b) represents [Ba/Fe] = [Eu/Fe]. Symbols: starred
circles: this work; filled squares: 0 < [Ba/Eu] < 0.5 (CEMP-r/s); open squares: [Ba/Eu] > 0.5
(CEMP-s); crosses: [Ba/Fe] < 0 (CEMP-no); filled circles: [Ba/Eu] < 0 (r); dashes: upper limit
for both, Ba and Eu.
3. Results
The panel a) of Fig. 1 shows that one star of our sample might be CEMP-s. Although
its Ba abundance is not so high ([Ba/Fe] = 0.75), the low Eu abundance ([Eu/Fe] =
-0.01) increases the ratio [Ba/Eu]. This star is close to the line [Eu/Fe] 0 in panel b).
Four stars are in the lower limit of the range 0 < [Ba/Eu] < 0.5, the CEMP-r/s zone
represented by the dotted lines in panel a). These stars are in the dotted line in panel
b), since their [Ba/Fe] [Eu/Fe]. The other five stars have [Ba/Eu] < 0, the region
of CEMP-no and CEMP-r, where three of them are below the dashed line in panel a),
which represents the production of Ba only through the r-process, [Ba/Eu] -0.7. Two
stars with [Ba/Eu] < -0.7 have [Eu/Fe] higher than the ones from the literature called
CEMP-r (filled circles), whereas the third could be called CEMP-no, since its [Ba/Fe]
< 0. Two stars with [Ba/Fe] 0.3 might be CEMP-r since [Ba/Eu] < 0 for them. The
very high [Eu/Fe] of four stars deserves further investigation in order to confirm the
classification inferred here.
References
Alonso A., Arribas S., & Martinez-Roger C. 1998, A&AS, 131, 209
Green, E.M., Demarque, P., & King, C.R. 1987, in The revised Yale isochrones and luminosity
functions, New Haven: Yale Observatory
Lejeune, T., Cuisinier, F., & Buser, R. 1998, A&AS, 130, 65
Hakkila, J., Myers, J.M., Stidham, B.J., & Hartmann, D.H. 1997, AJ, 114, 2043
Tsangarides, S.A. 2005, Ph.D. Thesis, The Open University, UK
Universidade de S
ao Paulo, Brazil
2
Observatoire de Paris, France
3
Observatoire de la C
ote dAzur, France
4
European Southern Observatory, Germany
5
Universite de Montpellier, France
6
University of Texas, USA
7
Michigan State University, USA
5
Niels Bohr Institute, Denmark
6
Osservatorio Astronomico di Trieste, Italy
Abstract. The abundance derivation of heavy r-elements may provide a better understanding
of the r-process, and the determination of several reference r-elements should allow a better
determination of the stars age. The spatial ultraviolet (UV) region presents a large number of
lines of heavy elements, and in some cases such as Bi, Pt, Au, detectable lines are only available
in the UV. The extreme r-process star CS 31082-001 ([Fe/H]=-2.9) was observed in the spatial
UV in order to determine abundances of the heavy elements, using STIS on board HST.
Keywords. metal-poor stars, r-elements, ultraviolet
1. Introduction
The ESO large programme FIRST STARS was devoted to the accurate determination
of abundances in very metal-poor stars. A very interesting star was revealed (Hill et al.
2002; Cayrel et al. 2001), to be extremely enhanced in r-process elements, in particular
two radioactive elements of the third peak of the r-process: thorium and uranium. The
uranium line at 3859.6
A was measurable, so that the ratio of the abundances of Th and
U was well determined, and could thus provide a particularly reliable determination of
the age of this star. Indeed, possible variations of the r-process will affect much less the
ratio of these two elements, which have very similar atomic and nuclear structures. It
is however desirable to check the decay of the radioactive elements by comparison with
other stable elements of the third r-process peak, together with Pb, a key element (Plez
et al. 2004), and elements of the second peak. Besides CS 31082-001 and CS 22892-052
(Sneden et al. 1994), Frebel et al. (2007) reported the identification of another such star,
HE 1523-0901. This latter object has V=11.1, [Fe/H]=-2.95, [r/Fe]1.8, and an age of
13.2 Gyr was deduced from the U/Th, U/Ir, Th/Eu, and Th/Os ratios.
121
Os I
2838.6
76
1.45
0.10
Os I
3058.6
76
1.45
0.30
Pt I
2929.8
78
1.80
0.65
A(Sun) A(Star)
Pt I
3064.7
78
1.80
0.40
Au I
2675.9
79
1.01
-0.90
Pb I
2833.0
82
1.95
-0.65
Bi I
3067.7
83
0.68
-0.50
(U/Th)init = -0.16, but should improve with the knowledge of other r-process elements
involved in the third peak, where Lead and Bismuth are the obvious targets. Plez et al.
(2004) were able to measure the Pb abundance in CS 31082-001 from near-UV spectra
obtained with the VLT-UVES at ESO. An abundance of log (Pb/H)+12=-0.55 was
obtained, very close to the value of -0.73, expected from the decay of 238 U and 232 Th
alone. This low value of Pb leaves little room for an s-process contribution in the solar
system.
Figure 1. Pb 2833.053
A line: crosses: observed spectrum, solid red and blue lines: synthetic
spectra computed with [Pb/Fe]=none, -0.65, -0.15.
The code turbospectrum (Alvarez & Plez 1998) was employed to compute synthetic
spectra in the UV region. In this spectral region it is crucial to properly account for
scattering in the continuum. So far, we have derived abundances for the elements shown
in Table 1. We confirm that the Pb abundance in CS 31082-001 is low, whereas only
upper limits are available for CS 22892-052 and HE 1523-0901. Os and Pt are enhanced,
as U and Th, whereas Bi and Au are less enhanced. Further abundances are being derived
for other elements (GdII, DyII, ErII, LuII among others).
Fig. 1 shows the fit of synthetic spectra to the observed Pb 2833.052
A line.
References
Alvarez, R., & Plez, B. 1998, A&A, 330, 1109
Cayrel, R., Hill, V., Beers, T.C., Barbuy, B, Spite, M. et al. 2001, Nature, 409, 691
Hill, V., Plez, B., Cayrel, R., Beers, T.C., Nordstr
om, B. et al. 2002, A&A, 387, 560
Plez, B., Hill, V., Cayrel, R., Spite, M., Barbuy, B. et al. 2004, A&A, 428, L9
Sneden, C., Preston, G.W., McWilliam, A., Searle, L. 1994, ApJ, 431, L27
Toenjes, R., Schatz, H., Kratz, K.-L., Pfeiffer, B., Beers, T.C. et al. 2002, ASPC, 245, 376
1
CIFIST Marie Curie Excellence Team
GEPI, Observatoire de Paris CNRS, Universite Paris Diderot 3 Istituto Nazionale di
Astrofisica - Osservatorio Astronomico di Trieste
email: natalie.behara@obspm.fr
Abstract. Approximately 20% of very metal-poor stars ([Fe/H] < 2.0) are strongly enhanced
in carbon ([C/Fe] > +1.0). Such stars are referred to as carbon-enhanced metal-poor (CEMP)
stars. We present a chemical abundance analysis based on high resolution spectra acquired with
UVES at the VLT of three dwarf CEMP stars: SDSS J1349-0229, SDSS J0912+0216 and SDSS
J1036+1212. These very metal-poor stars, with [Fe/H] < -2.5, were selected from our ongoing
survey of extremely metal-poor dwarf candidates from the SDSS.
Among these CEMPs, SDSS J1349-0229 has been identified as a carbon star ([C/O] > +1.0).
First and second peak s-process elements, as well as second peak r-process elements have been
detected in all stars. In addition, elements from the third r-process peak were detected in one
of the stars, SDSS J1036+1212. We present the abundance results of these stars in the context
of neutron-capture nucleosynthesis theories.
Keywords. stars: abundances, stars: fundamental parameters, stars: AGB and post-AGB.
Teff
J13490229 6200
J0912+0216 6500
J1036+1212 6000
log g [Fe/H]
4.00
4.50
4.00
3.0
2.5
3.2
[C/Fe]
[N/Fe]
[O/Fe]
122
2.26
1.58
1.26
1.67
1.25
1.31
123
Figure 1. Left figure: Observed spectra of the CH G band. Right figure: Equivalent width
contribution function plotted as a function of optical depth for the 3D model (grey) and the
1DLHD model (black) for two different C/O ratios. Scaled solar C/O is plotted as a solid line,
while a C/O ratio typical for a CEMP star is plotted as a dot-dashed line. Overplotted are the
average temperature profile of the 3D model (solid line) and of the 1DLHD model (dashed line).
Figure 2. Left figure: We compared [C/Fe] of our stars (star symbols) to a sample of CEMP
stars from Sivarani et al. (2006). Excluding the two most metal-poor stars, a clear correlation is
seen between [C/Fe] and [Fe/H]. The two exceptions are SDSS J1036+1212 from this work and a
CEMP-no/s star CS 29528-041. Right figure: We attempt to classify our stars by comparing their
[Ba/Sr] abundance against the different families of CEMP stars. We classify SDSS J1349-0229
and SDSS J0912+0216 as CEMP-r+s stars due to their high Ba and Eu (both > 1.0). SDSS
J1036+1212 becomes the third member of the CEMP-no/s class, due to its low Sr abundance.
Abstract. We performed a chemical abundance analysis of the very bright (V = 9.1) carbonenhanced metal-poor (CEMP) star BD+44 493, which is the first star found with metallicity
[Fe/H] < 3.5 and an apparent magnitude V < 12. The star is classified as a CEMP-no
subgiant, and its abundance pattern implies that a first-generation faint supernova is the most
likely origin of its carbon excess. We set an very low upper limit on this stars beryllium abundance, which demonstrates that high C and O abundances do not necessarily imply high Be
abundances.
Keywords. stars: abundances, stars: individual(BD+44 493), stars: Population II
125
A(Be)
1
0
-1
-2
-4
-3
-2
-1
[Fe/H]
Figure 1. A(Be) vs. [Fe/H]. Our result for BD+44 493 is shown by the filled circle.
Our analysis is the first attempt to measure a Be abundance for a CEMP star. Since
Be is produced via the spallation of CNO nuclei, their abundances, especially O abundances, have been expected to correlate with Be abundances. However, our low Be upper
limit shows that the high C and O abundances in BD+44 493 are irrelevant to its Be
abundance, which offers a new insight into the origin of CEMP-no stars.
References
Boesgaard, A. M., Deliyannis, C. P., King, J. R., Ryan, S. G., et al. 1999 AJ, 117, 1549
Carney, B. W., Latham, D. W., Stefanik, R. P., Laird, J. B., & Morse, J. A. 2003, AJ, 125, 293
Ito, H., Aoki, W., Honda, S., & Beers, T. C. 2009, ApJL, 698, L37
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, A&A, 447, 623
Nishimura, T., Aikawa, M., Suda, T., & Fujimoto, M. Y. 2009, accepted for PASJ
Tominaga, N., Maeda, K., Umeda, H., Nomoto, K., et al. 2007, ApJL, 657, L77
Department of Physics & Astronomy and JINA: Joint Institute for Nuclear Astrophysics,
Michigan State University, East Lansing, MI 48824
email: kenne257@msu.edu
2
Indian Institute of Astrophysics, Koramangala, Bangalore 560034, India
3
IAG, University of S
ao Paulo, Brazil
4
Department of Astronomy, Ohio State University, Columbus, OH
Abstract. We report on medium-resolution near-IR spectroscopy of a sample of over 60 CarbonEnhanced Metal-Poor (CEMP) stars observed with SOAR/OSIRIS, selected from the HK survey of Beers and colleagues and the Hamburg/ESO Survey of Christlieb and colleagues. Oxygen
abundances from the molecular CO lines as well as rough estimates of 12 C/13 C ratios are estimated from the near-IR spectra of these stars. Near-IR model spectra with varying oxygen
abundances, in combination with previously determined parameters from optical spectra are
used for the estimation of abundances for this sample. As both oxygen abundances and 12 C/13 C
ratios are tracers of nucleosynthesis, we hope to gain information about Galactic nucleosynthesis
through the analysis of this sample.
Keywords. stars: abundances, stars: carbon, Galaxy: halo, surveys
1. Introduction
Carbon, nitrogen, and oxygen abundances, in addition to 12 C/13 C ratios, are important in order to constrain properties of different types of carbon-enhanced, metal-poor
stars in the Galactic halo. There are two categories of CEMP stars: those with neutroncapture enhancement (CEMP-s, CEMP-r, CEMP-r/s), and those without (CEMP-no).
Abundance patterns of those stars with neutron-capture enhancement are posited to be
the result of mass transfer from AGB companion stars. The origin of the abundance
patterns in CEMP-no stars is less certain. Proposed models include low-metallicity AGB
mass-transfer (in which the s-process is suppressed), mass loss by rapidly rotating mega
metal-poor ([Fe/H]<-6.0) stars (Hirshi et al. 2006; Meynet et al. 2006), or pollution by
early supernovae. The 12 C/13 C ratios in CEMP-no stars tend to be quite low (Aoki et
al. 2007), which suggests that substantial mixing has occurred in the progenitor object.
The 12 C/13 C and [O/Fe] abundances are crucial to distinguish the origin of the different
types of CEMP stars, and they are not easily available through optical observations due
to the weakness of the [OI] 6300
A lines. The near-IR region or the spectrum is ideal
for such abundance measurements.
127
Figure 1. New [O/Fe] estimates with error bars for the entire sample as compared to [C/Fe]
estimated from optical spectra.
The 12 C/13 C ratios are estimated using a similar technique, although our results include only rough estimates for 20 of the stars in the sample. A refined technique will be
employed in the near future that will allow for more accurate estimates of this value for
a larger sample of stars. In addition to revised 12 C/13 C estimates, future analysis of this
sample will include comparison of our estimates to values predicted by AGB models as
well as models of rapidly-rotating, mega metal-poor stars.
References
Aoki, W., Beers, T. C., Christlieb, N., Norris, J. E., Ryan, S. G., & Tsangarides, S. 2007, ApJ,
655, 492
Beers, T. C. & Christlieb, N. 2005, ARA&A, 43, 531
Beers, T. C., Sivarani, T., Marsteller, B., Lee, Y. S., Rossi, S., Plez, B. 2007, AJ, 133, 1193
Hirschi, R., Fr
ohlich, C. Liebend
orfer, M., and Thielemann, F.-K. 2006, RvMA, 19, 101
Meynet, G., Ekstr
om, S., & Maeder, A. 2006, A&A, 447, 623
National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo, 181-8588, Japan
2
Astrophysics Group, EPSAM, Keele University, Keele, Staffordshire ST5 5BG, UK
3
Department of Cosmoscience, Hokkaido University, Sapporo, Hokkaido 060-0810, Japan
Abstract. Extremely metal-poor (EMP) stars in the Galactic halo are stars formed in the
very early stage of the chemical evolution of the Galaxy. In previous study, we proposed that
typical mass of EMP stars are massive, based on observations of carbon-enhanced EMP stars.
In this study, we build a merger tree of the Galaxy semi-analytically and follow the chemical
evolution along the merger tree. We also consider the effect of binary and high-mass initial mass
function(IMF). Resultant theoretical metallicity distribution function (MDF) and abundance
distribution are compared with observed metal-poor halo stars.
Keywords. stars: abundances, stars: Population II, Galaxy: evolution, Galaxy: halo
1. Introduction
In the Galactic halo, many EMP stars with metallicity [Fe/H] < 2.5 are detected
thanks to large scaled surveys Beers et al.(1992), Christlieb et al.(2001). These stars are
expected to be probes to the first stars and galaxy formation in the early universe.
129
EMP stars in the early universe can be estimated from number of EMP stars observed by
large scale surveys when we assume the IMF. Thus, from observation, we can estimate
the averaged iron yield, < YF e >, per massive EMP stars as a function of Mmd .
Onthe other hand, it is said that iron yield of type II supernova is 0.07 M from
observations. Theoretical studies about supernova nucleosynthesis derive the dependence
of the iron yield on the mass of progenitor stars (Woosley & Weaver(1995), Umeda &
Nomoto(2002)). We calculate IMF weighted average of iron yield as a function of Mmd .
These two estimation of averaged iron yield < YF e > should be same. As a result,
< YF e > from observation of EMP stars are consistent with theoretical value at Mmd
10 M. From two constraints, from CEMP stars and iron yield, we conclude that Mmd
10 M and 0.4 for EMP stars.
Abstract. We determined the atmospheric parameters and abundance pattern of two chemically
peculiar metal-poor stars: HD 10613 and BD+04 2466 and discuss the nature of these two
objects.
Keywords. stars: fundamental parameters; stars: Population II; stars: chemically peculiar
1. Introduction
Barium and CH stars belong to a class of chemically peculiar stars where binarity
is an essential requirement to explain their overabundance of carbon and the elements
heavier than iron. Regarding the stellar population type, CH stars are clearly members
of the halo population, they have high radial velocities and are metal-poor objects. CH
stars have been regarded as population II counterparts of the barium stars. Population
studies done for barium stars show that they differ from the CH stars with respect to
their distribution in the Galaxy. Barium stars are found in the disk and in the halo of
the Galaxy (G
omez et al. 1987). This study shows that barium stars can also be divided
into groups according to their luminosities, kinematic and spatial parameters. It was also
shown that barium stars in the halo are very rare, only 6% of the total sample. Here
we continue our investigation on metal-deficient barium stars candidates searching for
possible candidates in the literature. In a previous search we identified and analyzed
HD 206983 (Junqueira & Pereira 2001, Drake & Pereira 2008). We now analyze one star
that was suspected to be a metal-poor barium star by Catchpole et al. (1977) and also
classified as a member of the halo population by G
omez et al. (1987), HD 10613. In
addition we analyze BD+04 2466 which belongs to this small sample of metal-deficient
barium stars of Luck & Bond (1991).
131
log
[X/H] [X/Fe]
log
[X/H] [X/Fe]
Ci
Ni
Oi
Na i
Mg i
Si i
Ca i
Sc ii
Ti i
Cr i
Mn i
Ni i
Cu i
Zn i
2
1
1
2
2
6
7
6
8
6
2
10
1
1
8.250.06
7.430.22
8.530.04
5.35
7.03
6.930.11
5.750.12
2.450.26
4.420.12
4.970.12
4.29
5.430.08
3.51
3.72
-0.27
-0.49
-0.30
-0.98
-0.55
-0.62
-0.61
-0.72
-0.60
-0.70
-1.10
-0.82
-0.70
-0.88
+0.20 2
6.24
-1.31
+0.21 11 4.970.20 -1.39
+0.10 5 1.250.12 -1.92
+0.22 2
3.27
-1.75
+0.12 4 3.630.06 -2.04
-0.28 1
3.09
-2.30
0.00 3 4.130.18 -2.12
+0.12
-0.06 2
2.92
-1.68
+1.17
+1.10
+0.30
+0.02
+0.61
+0.53
0.00
+0.17
-0.12
-0.38
-0.20
+0.24
Y ii
Zr i
Zr ii
Ba ii
La ii
Ce ii
Nd ii
Eu ii
Pb i
7
3
2
1
6
6
16
1
1
2.200.20
2.710.05
2.610.43
2.75
1.670.23
1.910.20
2.230.21
0.41
2.48
-0.04
+0.11
0.00
+0.62
+0.50
+0.33
+0.73
-0.10
+0.48
+0.78 6
+0.93
+0.82 4
+1.44 1
+1.32 4
+1.15 5
+1.55 12
+0.72
+1.30 1
+0.47
+0.79
+1.70
+1.20
+1.07
+1.35
+1.92
0.790.13
1.470.19
1.91
0.450.05
0.730.15
0.930.12
2.00
-1.45
-1.13
-0.22
-0.72
-0.85
-0.57
0.00
3. Conclusions
Our analysis of the chemical abudances of these stars showed that:
(a) HD 10613 is another metal-poor barium star, not already shown as a binary system.
However, its luminosity and abundance ratios give support to the interpretation that the
observed overabundances of carbon and s-process elements in the photosphere of this
star are due to mass transfer from a companion, formerly a TP-AGB star.
(b) BD+04 2466 is a CH star since its carbon-to-oxygen ratio is larger than unity
(C/O = 3.63). In fact, BD+04 2466 displays the main characteristics of the mass-transfer
paradigm, i.e. presents overabundances of the elements created by the slow neutron
capture reactions and carbon as well and has already been proved to be a member of a
binary system (Jorissen et al. 2005). We also shown that BD+04 2466 is another lead
star, since its [Pb/Ce] ratio closely follows the theoretical predictions for a star at such
metallicity.
References
Drake, N. A. & Pereira, C. B., 2008, AJ, 135, 1070.
Catchpole, R. M., Robertson, B. S. C. & Warren, P. R. 1977, MNRAS, 181, 391
G
omez, A.E., Luri, X. Grenier, S. et al 1997, A&A, 319, 881
Jorissen, A., Zacs, L., Udry, S. et al. 2005, A&A, 441, 1135
Junqueira, S. & Pereira, C.B. 2001, AJ, 122, 360
Kurucz, R. L. 1993, CD-ROM 13, Atlas9 Stellar Atmosphere Programs and 2 km/s Grid (Cambridge: Smithsonian Astrophys. Obs).
Sneden, C. 1973, Ph.D. Thesis, Univ. of Texas
1. Introduction
Although CEMP stars have been found previously among the candidate metal-poor
stars selected from the HES, the selection on metallicity undersamples the population of
intermediate-metallicity CEMP stars (2.5 6[Fe/H]6 1.0); such stars are of importance
for constraining the onset of the s-process in metal-deficient Asymptotic Giant-Branch
stars (thought to be associated with the origin of carbon for roughly 80% of CEMP stars).
The new candidates also include substantial numbers of warmer carbon-enhanced stars,
which were missed in previous HES searches for carbon stars due to selection criteria
that emphasized stars with cooler temperatures.
The primary goal of the present work is to demonstrate the efficacy of searching for
these intermediate-metallicity CEMP stars. The inclusion of warmer carbon-enhanced
candidates (which do not exhibit CN and C2 bands) also enables investigations between
the observed levels of carbon enhancement and evolutionary stage. It should also be kept
in mind that the inventory of ultra ( [Fe/H] < 4.0) and hyper ( [Fe/H] < 5.0) metalpoor stars is almost certainly incomplete. Such extreme stars may have been overlooked
in previous searches due to noisy spectra in the region of CaII K on objective-prism
plates, but they could reveal themselves by the presence of strong CH G-bands that are
commonly associated with the most metal-deficient stars.
132
133
,
We have developed a new line index for the region of the carbon G-band at 4304 A
GPE (4200-4400
A), which has the advantage of capturing more information concerning
the abundance of carbon, since its width takes into account the wings of the band, which
includes other nearby carbon features, and is not subject to confounding of previously
employed narrower indices, due to sidebands that fall in regions of the spectrum for which
carbon features are present.
Aoki 2007
2.5
[C/Fe]
1.5
0.5
0.5
3.5
2.5
1.5
0.5
[Fe/H]
Figure 1. Behavior of the metallicity with the carbon abundance [C/Fe] for the observed
candidates and the stars from Aoki et al. (2007). The arrows represent upper limits.
Since the goal of this work is to target CEMP stars, we have confirmed that the high
fraction found, based on our selection method, is more than two times higher than the
one claimed in Lucatello et al. (2006) for all stars with [Fe/H]<2.0 observed to date
(20%). It is also important to note that the majority of metal-poor stars in our candidate
pool with [Fe/H]<1.0 (70%) present carbon enhancements ([C/Fe] > +0.5).
Acknowledgements
V.M.P. and S.R. acknowledge CAPES (PROEX), CNPq, JINA and FAPESP funding
(2007/04356-3). C.R.K. and T.C.B. acknowledge JINA.
References
Aoki, W., Beers, T. C., Christlieb, N. et al. 2007, ApJ, 655, 492
Christlieb, N., Sch
orck, T., Frebel, A. et al. 2008, A&A, 484, 721
Beers, T. C. & Christlieb, N. 2005, ARA&A, 43, 531
Lee, Y. S., et al. 2008, AJ, 136, 205
Lucatello, S., Beers, T. C., Christlieb, N. et al. 2006, ApJ, 652, L37
IAG, Universidade de S
ao Paulo, S
ao Paulo, Brazil
email: rsmiljan@eso.org
2
ESO, Garching bei M
unchen, Germany, 3 GEPI Observatoire de Paris - Meudon, France,
4
INAF, Osservatorio di Trieste, Trieste, Italy, 5 CIFIST Marie Curie Excellence Team, 6 INAFOsservatorio di Arcetri, Firenze, Italy 7 INAF-Osservatorio di Padova, Padova, Italy,
Abstract. Beryllium stellar abundances were suggested to be a good tracer of time in the early
Galaxy. In an investigation of its use as a cosmochronometer, using a large sample of local halo
and thick-disk dwarfs, evidence was found that in a log(Be/H) vs. [/Fe] diagram the halo
stars separate into two components. One is consistent with predictions of evolutionary models
while the other is chemically indistinguishable from the thick-disk stars. This is interpreted as a
difference in the star formation history of the two components and suggests that the local halo
is not a single uniform population where a clear age-metallicity relation can be defined.
Keywords. stars: abundances stars: late-type Galaxy: halo
1. Introduction
Be is a pure product of cosmic-ray spallation in the ISM involving mostly CNO nuclei.
Abundances of Be in metal-poor stars show a linear relation with Fe (and O) with a
slope close to one, implying that Be behaves as a primary element (Smiljanic et al. 2009,
and references therein). If cosmic rays are globally transported across the Galaxy, the
production of Be is a widespread process and Be abundances should have a smaller scatter
than the products of stellar nucleosynthesis at a given time in the early Galaxy (Suzuki
& Yoshii 2001). In other words Be should be a good tracer of time.
Pasquini et al. (2004, 2007) calculated Be abundances in turn-off stars of two globular
clusters, NGC 6397 and NGC 6752. The Be ages derived from a model of the evolution of
Be with time (Valle et al. 2002) were shown to agree with those derived from theoretical
isochrones, supporting the use of Be as a cosmochronometer.
Pasquini et al. (2005) analyzed a sample of 20 halo and thick-disk stars and found
a possible separation between stars of the two components in a log(Be/H) vs. [/Fe]
diagram. This was interpreted as a difference in the time scales of star formation.
Smiljanic et al. (2009) analyzed the largest sample of halo and thick-disk stars to date,
extending the investigation of Be as cosmochronometer and its role as a discriminator of
different stellar populations in the Galaxy. The thick disk was found to be a homogeneous
population. The halo stars were found to divide into two different components (Fig. 1).
134
135
Figure 1. (a) log(Be/H) vs. [/Fe] for the halo stars of Smiljanic et al. (2009). The solid line
is the model prediction of Valle et al. (2002). (b) log(Be/H) vs. [O/Fe] with new preliminary
oxygen abundances for a subsample of the halo stars., corrected for NLTE (Fabbian et al. 2009)
were not available, so mean abundances of -elements were used instead. Here we present
new preliminary oxygen abundances determined from the infrared triplet at 777nm.
As shown in Fig. 1, using either or oxygen abundances, the halo stars define two
clear distinct sequences. One sequence is chemically similar to the thick disk, the other
agrees with the models of Valle et al. (2002). The latter, however, with [/Fe] 6 0.25
and log(Be/H) 6 11.4, have similar kinematics. The stars have mostly velocity in the
direction of the rotation of the Galaxy V 0 and the perigalactic distance Rmin 6 1 kpc
(see Smiljanic et al. 2009 for details), as expected for accreted stars.
The splitting into two components may be related to the accretion of external systems
or to variations of star formation in different and initially independent regions of the
early halo. The interpretetion is still open, it is however clear that the halo is not a single
uniform population with a single age-metallicity relation. A similar division was found
by Nissen & Schuster (1997, 2009) but using Fe as a tracer of time. The division is clearer
when Be is used as a time scale. In the same line, recent simulations of the formation of
disk galaxies in a CDM universe by Zolotov et al. (2009) show that the inner halo has
a dual nature, it is composed both by in situ stars formed in the potential well of the
galaxy and by accreted stars, formed in subhalos later accreted by the galaxy.
Acknowledgements
R.S. acknowledges financial support from FAPESP (04/13667-4 and 08/55923-8).
References
Fabbian, D., Asplund, M., Barklem, P. S., et al. 2009, A&A, 500, 1221
Nissen, P. E. & Schuster, W. J. 1997, A&A, 326, 751
Nissen, P. E. & Schuster, W. J. 2009, in: Proc. IAU Symposium No. 254, p. 103
Pasquini, L., Bonifacio, P., Randich, S., Galli, D., & Gratton, R. G. 2004, A&A, 426, 651
Pasquini, L., Bonifacio, P., Randich, S., et al. 2007, A&A, 464, 601
Pasquini, L., Galli, D., Gratton, R. G., et al. 2005, A&A, 436, L57
Smiljanic, R., Pasquini, L., Bonifacio, P., et al. 2009, A&A, 499, 103
Suzuki, T. K. & Yoshii, Y. 2001, ApJ, 549, 303
Valle, G., Ferrini, F., Galli, D., & Shore, S. N. 2002, ApJ, 566, 252
Zolotov, A., Willman, B., Brooks, A. M., et al. 2009, ApJ, 702, 1058
Session III
Abstract. Gamma-ray bursts (GRBs) are the brightest events in the universe. They have been
used in the last ve years to study the cosmic chemical evolution, from the local universe to
the rst stars. The sample size is still relatively small when compared to eld galaxy surveys.
However, GRBs show a universe that is surprising. At z > 2, the cold interstellar medium in
galaxies is chemically evolved, with a mean metallicity of about 1/10 solar. At lower redshift
(z < 1), metallicities of the ionized gas are relatively low, on average 1/6 solar. Not only is there
no evidence of redshift evolution in the interval 0 < z < 6.3, but also the dispersion in the 30
objects is large. This suggests that the metallicity of host galaxies is not the physical quantity
triggering GRB events. From the investigation of other galaxy parameters, it emerges that active
star-formation might be a stronger requirement to produce a GRB. Several recent striking results
strongly support the idea that GRB studies open a new view on our understanding of galaxy
formation and evolution, back to the very primordial universe at z 8.
Keywords. Gamma rays: bursts, observations, ISM: abundances, cosmology: observations.
1. Introduction
During the last decade, the chemical evolution of the universe has been investigated
using a new class of objects: gamma-ray bursts (GRBs). GRBs are the brightest sources
in the universe, but were rst detected only in 1967 by a US military satellite (Klebesadel
et al. 1973), because their emission does not last long. For this reason, their cosmological
origin was demonstrated only in 1997, when the rst redshift was measured (Metzger et
al. 1998). Today, after more than twelve years, the number of events with spectroscopic
redshift is still relatively low, about 200. Nevertheless, on April 23 2009 the highest
spectroscopic redshift ever was measured, and this happened to be a GRB, GRB 090423,
at z = 8.2 (Salvaterra et al. 2009; Tanvir et al. 2009). This is not only a very exciting
success for GRB science, but it also demonstrates that the eld is still potentially and
eectively crucial for the understanding of our universe.
GRBs are very luminous, but do not shine for very long (it cannot be any dierent,
otherwise we would not be here to tell). Their -ray emission lasts at most a few minutes,
during which they radiate the same energy emitted by the Sun over its entire life, 10 Gyr.
Long-duration GRBs (more than a few seconds, the majority of those detected) originate
from the nal core collapse of a massive star, a supernova (Woosley, 1993). Short-duration
GRBs (shorter than a few seconds) have likely a dierent progenitor (Katz & Canel 1996):
the coalescence of two compact objects (neutron stars or black holes). In both classes,
rotation is the key ingredient producing the collimated emission. The GRB rate is of the
order of one event every 105 years in a galaxy. This means that, integrating over the
entire universe and considering the collimated emission, few events are detectable from
-ray satellites.
During the last two years, a number of particularly interesting discoveries have shown
139
140
Sandra Savaglio
that the universe probed by GRBs is surprisingly exciting. Apart from the already mentioned GRB 090423, in March 2008 the brightest source ever was recorded. This was
GRB 080319 at z = 1.9 (7.5 Gyrs after the Big Bang), nicknamed the naked eye
GRB because it had an optical magnitude m = 5.6 at its maximum (Bloom et al.
2009). In September 2008, the at-the-time second most distant object ever was detected,
GRB 080913B at z = 6.7 (Greiner et al. 2009).
Thanks to this rich phenomenology and the large redshift range spanned, there is no
doubt that GRBs are very eectively probing, among other things, the chemical evolution
of the universe, all the way from the local universe to the epoch of rst stars, more than
13 Gyr ago. In this paper, we will summarize the results obtained in the last ve years.
QSO-DLAs
GRB-DLAs
Figure 1. Fraction of GRB-DLAs (lled histograms) and QSO-DLAs (empty histograms) per
HI and ZnII column-density bin (left- and right-hand side panels, respectively). The QSO-DLA
histograms are complete for log NHI > 20.2 and log NZnII > 12.4. The completeness level for
GRB-DLAs is not well determined. It is apparent that column densities in GRB-DLAs are
generally higher than in QSO-DLAs. This can either indicate that GRB-DLAs originate in
bigger galaxies, or that the volume density of the gas is higher (e.g., the GRB sightline is
crossing a region closer to the galaxy center) than QSO-DLAs, or both.
141
cating that the neutral-gas regions crossed by GRBs are larger, or denser, or both, than
those crossed by QSOs. In Fig. 2 we show the metal abundances measured in the DLAs
detected in GRB hosts and in QSO sight lines. It was claimed, from a smaller sample, that GRB-DLAs have generally higher metallicity than QSO-DLAs (Berger et al.
2006; Savaglio 2006; Prochaska et al. 2007). The most up-to-date sample of GRB-DLAs,
shown in Fig. 2, contains 17 measurements and two lower limits in the redshift interval
2 < z < 6.3. The average value (and statistical dispersion) for the 15 GRB-DLAs in
2.0 < z < 4.5 is <[Z/H]> = 1.0 0.7, whereas for the 156 QSO-DLAs in the same redshift interval this is <[Z/H]> = 1.4 0.6. The new large sample of GRB-DLAs tends
to show still a higher metal content than QSO-DLAs, but the gap is getting smaller.
This indicates that the observational bias that prevents us from measuring abundances
when metal lines are too weak might aect our results. The dierence with QSO-DLAs
is that GRB afterglows, when spectroscopically observed, are on average several magnitudes fainter than the typical QSO, and they cannot be observed for too long because
they disappear quickly.
For lower redshift, z < 1, metallicities are measured with emission lines from HII
regions in the host galaxy. Emission line metallicities rely on dierent calibrators (Kewley
& Ellison 2008) used depending on the set of lines available, according to the GRB redshift
Figure 2. Redshift evolution of the metallicity relative to solar values, for 17 GRB-DLAs at
z > 2, 16 GRB hosts at z < 1 and 250 QSO-DLAs in the interval 0 < z < 4.4. Error
bars are not available for all GRB-DLAs. Errors for GRB hosts are not estimated. Errors for
QSO-DLAs are generally smaller than 0.2 dex. The dashed line is the best-t linear correlation
for QSO-DLAs. The solid line is the mean metallicity predicted by semi-analytic models for
galaxy formation (Somerville et al. 2001). The GRB-DLAs metallicity in 2 < z < 4.5 is on
average 2.5 times higher than the average value in QSO-DLAs in the same redshift interval.
142
Sandra Savaglio
and instrument setting. Results on less than 20 hosts indicate metallicities between solar
and 1/14 times solar values (Savaglio, Glazebrook & Le Borgne 2009; Levesque et al.
2009). The average value and dispersion in 16 hosts (median redshift z = 0.44) is <[Z/H]>
= 0.75 0.29 (Fig. 2; Savaglio et al. 2009). This is somehow surprising, as we do not see
evidence of redshift evolution from GRB-DLA metallicities at z > 2. On the other hand,
evolution is observed in QSO-DLAs, where metallicity at z < 1 is <[Z/H]> = 0.3 0.5,
a factor of at least 10 times higher than at z > 2.
It was recently proposed that the dierent metallicities in GRB-DLAs and QSO-DLAs
could be due to the dierent regions probed by the two populations. GRBs tend to occur
in regions with high star-formation, therefore in regions closer to the galaxy center,
where metallicity is on average larger than in a random galaxy sightline. QSOs are
background sources not associated with the galaxy hosting the DLA, therefore their
sightline is crossing the galaxy in a random location, not necessarily close to a region
of star formation (Fynbo et al. 2008). This is conrmed by the larger dust content and
extinction measured in GRB-DLAs with respect to QSO-DLAs (Kann et al. 2006; Kr
uhler
et al. 2008; Prochaska et al. 2009). However, such a sensible conclusion collides with the
relatively low metallicities found in low-z GRB hosts. The large dispersion of the metal
content in GRB hosts in a large redshift interval is indicative that perhaps metallicity is
not driving the GRB phenomenon. For this reason, we consider in the following sections
the other two fundamental physical quantities characterizing galaxies: the star-formation
rate (SFR) and the stellar mass M .
143
galaxies have failed (e.g. Berger et al. 2007; Chen et al. 2009; Levesque et al. 2009;
Savaglio et al. 2009). The main problem seems to be the small number statistics.
We can compare median values of stellar mass and metallicity for z < 1 GRB hosts
with the MZ relations for eld galaxies by Tremonti et al. (2004) at z 0.07 and Savaglio
et al. (2005) at z 0.7. At z 6 0.45 (9 GRB hosts, median redshift z = 0.17) the median
metallicity and stellar mass are log Z/Z = 0.56 and M = 109.21 M , respectively.
In the interval 0.55 6 z 6 0.97 (7 GRB hosts, median redshift z = 0.69) the median
metallicity and stellar mass are log Z/Z = 1.06 and M = 109.73 M , respectively.
To compare these values with eld-galaxy relations, we convert the MZ relations using
the newly published converters of metallicity calibrators (Kewley & Ellison 2008). The
expected metallicities in the two redshift and stellar mass bins (z = 0.17, 0.69 and M =
109.21 , 109.73 M ) are higher than in GRB hosts, in both cases log Z/Z = 0.16. The
dierence is signicant, especially for the high-redshift bin. This issue needs further
investigations with more objects, especially at higher redshifts, where the MZ relation
shows a strong redshift evolution (Erb et al. 206; Maiolino et al. 2007). X-Shooter is the
best instrument available at this time to measure metallicity in high-z GRB hosts.
From the mass and the SFR, it is possible to derive a meaningful galaxy physical
parameter: the specic star-formation rate SSFR = SFR/M , that is the SFR per unit
stellar mass. Its inverse, the growth time-scale = M /SFR, gives the time interval
required to a galaxy to reach the observed stellar mass, assuming that the measured
GRB hosts
Figure 3. Growth timescale = M /SFR (left y-axis) or its inverse, the specic star-formation
rate SSFR = SFR/M (right y-axis) as a function of redshift (Savaglio et al. 2009). Filled circles
and triangles are GRB hosts with SFRs measured from emission lines and UV luminosities,
respectively. Small, medium, and large symbols are hosts with M 6 109.0 M , 109.0 M
< M 6 109.7 M , and M > 109.7 M , respectively. The curve shows the age of the universe as
a function of redshift, and indicates the transition from bursty to quiescent mode for galaxies.
Dots are eld galaxies at 0.5 < z < 1.7 (Juneau et al. 2005). Crosses are Lyman break galaxies
at 1.3 < z < 3 (Reddy et al. 2006). The big and small stars at zero redshift represent the growth
timescale for the Milky Way and the Large Magellanic Cloud, respectively.
144
Sandra Savaglio
SFR is constant over its past history. Fig. 3 shows and SSFR as a function of redshift
for GRB hosts, and the comparison with eld galaxies. GRB hosts are almost all starforming galaxies, half of them are in the bursty regime.
Figure 4. The cosmic star formation density of the universe (from Kistler et al. 2009). Light
circles are the data from Hopkins & Beacom (2006). Crosses are contributions from Lyman-
emitters (LAEs; Ota et al. 2008). Down and up triangles are Lyman-break galaxies (LBGs) for
two UV luminosity functions: integration down to 0.2L at z = 3 (Bouwens et al. 2008) and
complete (up triangles), respectively. The latter shows a better match with values inferred from
GRBs (red diamonds; Kistler et al. 2009), indicating the strong contribution from small galaxies
generally not accounted for in the observed LBG luminosity function.
145
established that the SFRD in massive galaxies was much higher in the past than it is
today (at z 2 a factor of 6 higher than at z 0), whereas the redshift evolution has
been milder for low-mass galaxies (Juneau et al. 2005). At z > 5, we might expect SFR
to be mainly in massive galaxies. Finally, the SFRD from GRBs does not show a clear
decline for z > 5, and it is much higher than in Lyman- emitters (LAEs).
5. Conclusions
It is well known that GRBs shine through a universe that is hard to see in other ways.
They are incredibly bright and last only a short time. These two properties make them
very dierent from QSOs, which are not as bright, and do not fade away, making the
investigation of nearby galaxies much more complicated. From luminous GRB afterglows,
it is possible to measure the redshift and localize faint galaxies.
Most GRBs are associated with the death of a massive star, thus with a star-forming
region. It is well known that the SFR of the universe was much higher in the past than it
is today (Hopkins & Beacom 2006), therefore GRBs might be the most ecient way of
identifying the evolution of the SFR density. We also know that the SFRD is dominated
by small star-forming galaxies, which are probably the most common galaxies in the
distant universe (Pozzetti et al. 2009).
Identication of distant galaxies with GRBs is aected by a dierent bias than traditional galaxy surveys, because GRBs are not detected through optical instruments,
but with -ray and X-ray satellites. Some of these galaxies can be faint, because dust
extinguished, or because too far, or because intrinsically faint. With GRBs it is possible
to explore extreme regimes of galaxy parameters, thus they are important to understand
galaxy formation and evolution. GRB hosts identied in the optical and NIR at z < 2
are generally small (on average the stellar mass of the Large Magellanic Cloud) and
star-forming galaxies, although SFRs span a large interval.
GRBs are probes of the state of the chemical enrichment of the universe, from the
local universe, back to the time of the formation of rst stars. Metallicities of the cold
ISM in host galaxies at z > 2 is not low. The measured average value is 1/10 solar, and
the dispersion is large, about a factor of ve. Relatively high metallicity is conrmed
also for the highest redshift detections (Savaglio 2006; Totani et al. 2006; Price et al.
2007), which means that there is no indication of redshift evolution. Low metallicities of
the GRB progenitor are theoretically predicted (no mass loss) in order to keep a high
angular momentum, and have a highly collimated jet.
Relatively high metallicities, in this case from ionized gas of the host galaxies, are
conrmed also at z < 1. The average value is 1/6 solar, with a dispersion of a factor
of two, indicating that the metal content in host galaxies is not evolving so fast. The
sample is not very large and systematic uncertainties are still not totally under control,
therefore more observations and detections are very important. Nevertheless, the lack of
evidence of redshift evolution and the observed large dispersion suggest that GRBs do
not happen necessarily in metal poor galaxies. Star formation, on the other hand, might
be a more important physical trigger.
GRBs are extremely important for our understanding of the primordial universe and
the formation and evolution of heavy elements. The enlightening discoveries of the last few
years are a clear indication that the investigation is aected by our technical capabilities
which have dramatically improved recently. Dedicated instruments and observational
programs have opened a new window in the hidden universe and show that this is more
surprising and fascinating than expected.
146
Sandra Savaglio
References
Berger, E., Penprase, B. E., Cenko, S. B., Kulkarni, S. R., Fox, D. B., Steidel, C. C., & Reddy,
N. A. 2006, ApJ, 642, 979
Berger, E., et al. 2007, ApJ, 665, 102
Bloom, J. S., et al. 2009, ApJ, 691, 723
Bouwens, R. J., Illingworth, G. D., Franx, M., & Ford, H. 2008, ApJ, 686, 230
Chary, R., Berger, E., & Cowie, L. 2007, ApJ, 671, 272
Chen, H.-W., et al. 2009, ApJ, 691, 152
de Ugarte Postigo, A., et al. 2009, A&A, submitted
Erb, D. K., Shapley, A. E., Pettini, M., Steidel, C. C., Reddy, N. A., & Adelberger, K. L. 2006,
ApJ, 644, 813
Fruchter, A. S., et al. 2006, Nature, 441, 463
Fynbo, J. P. U., Prochaska, J. X., Sommer-Larsen, J., Dessauges-Zavadsky, M., & Mller, P.
2008, ApJ, 683, 321
Fynbo, J. P. U., et al. 2009, ApJS, in press, arXiv:0907.3449
Greiner, J., et al. 2009, ApJ, 693, 1610
Gorosabel, J., et al. 2005, A&A, 444, 711
Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651, 142
Juneau, S., et al. 2005, ApJ, 619, L135
Kann, D. A., Klose, S., & Zeh, A. 2006, ApJ, 641, 993
Katz, J. I. & Canel, L. M. 1996, ApJ, 471, 915
Kelly, P. L., Kirshner, R. P., & Pahre, M. 2008, ApJ, 687, 1201
Kewley, L. J., & Ellison, S. L. 2008, ApJ, 681, 1183
Kistler, M. D., Y
uksel, H., Beacom, J. F., Hopkins, A. M., & Wyithe, J. S. B. 2009, ApJ, 705,
L104
Klebesadel, R. W., Strong, I. B., Olson, R. A. 1973, ApJ, 182, L85
Kr
uhler, T., et al. 2008, ApJ, 685, 376
Levesque, E. M., Berger, E., Kewley, L. J., & Bagley, M. M. 2009, AJ, submitted,
arXiv:0907.4988
Maiolino, R., et al. 2008, A&A, 488, 463
Metzger, M. R., Djorgovski, S. G., Kulkarni, S. R., Steidel, C. C., Adelberger, K. L., Frail, D. A.,
Costa, E., & Frontera, F. 1997, Nature, 387, 878
Michalowski, M. J., Hjorth, J., Castro Cer
on, J. M., & Watson, D. 2008, ApJ, 672, 817
Ota, K., et al. 2008, ApJ, 677, 12
Pozzetti, L., et al. 2009, A&A, submitted, arXiv:0907.5416
Price, P. A., et al. 2007, ApJ, 663, L57
Prochaska, J. X., Chen, H.-W., Dessauges-Zavadsky, M., & Bloom, J. S. 2007, ApJ, 666, 267
Prochaska, J. X., et al. 2009, ApJL, 691, L27
Reddy, N. A., Steidel, C. C., Erb, D. K., Shapley, A. E., & Pettini, M. 2006, ApJ, 653, 1004
Salvaterra, R., et al. 2009, Nature, 461, 1258
Savaglio, S. 2006, New Journal of Physics, 8, 195
Savaglio, S., et al. 2005, ApJ, 635, 260
Savaglio, S., Fall, S. M., & Fiore, F. 2003, ApJ, 585, 638
Savaglio, S., Glazebrook, K., & LeBorgne, D. 2009, ApJ, 691, 182
Somerville, R. S., Primack, J. R., & Faber, S. M. 2001, MNRAS, 320, 504
Tanvir, N. R., et al. 2009, Nature, 461, 1254
Th
one, C. C., et al. 2008, ApJ, 676, 1151
Totani, T., Kawai, N., Kosugi, G., Aoki, K., Yamada, T., Iye, M., Ohta, K., & Hattori, T. 2006,
PASPJ, 58, 485
Tremonti, C. A., et al. 2004, ApJ, 613, 898
Woosley S. E. 1993, ApJ, 405, 277
1. Introduction
A galaxys metal content is intimately tied to its star formation history and its interactions with the surrounding intergalactic medium (IGM), and thus the measurement
of chemical abundances in a galaxy provides, in principle, a detailed fossil record of
galaxy evolution. As succeeding generations of stars are born and die, they enrich the
surrounding interstellar medium. These stars may also drive gas out of the galaxy via
powerful supernova-driven winds; such winds carry metals out into the IGM along with
the gas. Galaxies may also accrete new gas from the surrounding IGM; if this gas is
low in metallicity, it will decrease the average metallicity of the gas in the galaxy. Thus
an understanding of chemical abundances is a key component of the study of galactic
evolution.
Unfortunately, metallicity is also one of the more difficult properties of galaxies to
determine at high redshift, since relatively high signal-to-noise spectra are required and
the familiar rest-frame optical emission lines are redshifted into the infrared. However,
enormous progress has been made in this area in recent years, as the advent of sensitive
optical and near-IR spectrographs on 810 m class telescopes has allowed researchers
to obtain rest-frame UV and optical spectra of both emission and absorption lines in
galaxies at 1 . z . 3. In this article I review the most widely-used methods of abundance
determination in high redshift galaxies and their results, and discuss the mass-metallicity
relation at high redshift, its evolution, and its uses in constraining inflows and outflows
of gas.
148
Dawn K. Erb
the electron temperature and hence the metallicity. Most commonly, the auroral line
[O iii] 4363 is used; however, this line is weak under the best conditions and becomes
undetectable at metallicities above 0.5 solar, making its use extremely difficult at high
redshifts. We must therefore turn to other methods.
2.1. Strong line methods from rest-frame optical emission lines
Most common among these are the strong line indicators, which are based on the ratios
of collisionally excited forbidden lines to hydrogen recombination lines. Such methods are
calibrated either with reference to the Te method or with photoionization modeling. There
are many such methods, using a variety of different lines (see, e.g., Kewley & Ellison 2008
for a review and comparison of many of the diagnostics), but perhaps the most widely
used for high redshift work to date are the R23 and N 2 methods.
The R23 diagnostic (the ratio of the sum of the oxygen lines [O ii] 3727 and [O iii]
4959, 5007 to H; e.g. McGaugh 1991, Zaritsky et al. 1994) is probably the most
commonly employed metallicity indicator at low redshifts. It therefore benefits from being
well-studied and used for a wide variety of large samples. However, the R23 method comes
with well-known hazards: it is sensitive to extinction; different calibrations of the ratio can
differ in metallicity by as much as 0.5 dex (Kennicutt et al. 2003, Kewley & Ellison 2008);
and it is famously double-valued, rising at low metallicities, peaking somewhat below solar
metallicity, and falling as the metallicity increases further. It is therefore not very useful
for precise determinations of metallicity near this turnover region; unfortunately, this
region is exactly where the metallicities of typical star-forming galaxies at high redshift
have been shown to lie. The wide wavelength separation of the [O ii] and [O iii] lines
is a further pitfall for high redshift work, since they fall in different near-IR bands (J
and H respectively at z 2, and H and K at z 3). Most near-IR spectrographs are
therefore unable to observe them simultaneously, meaning that additional observing time
is required to observe in two bands and significant relative calibration uncertainties are
introduced.
The N 2 index, the ratio of [N ii] 6584 to H (Denicolo et al. 2002, Pettini & Pagel
2004), is free from many of the difficulties associated with R23 . Because it depends only
on two closely spaced lines, it is unaffected by extinction or flux calibration uncertainties, and observations in only one band are required. On the other hand, its saturation
at approximately solar metallicity makes it ineffective for very metal-rich objects, the
[N ii]/H ratio is sensitive to both the ionization parameter and AGN contamination,
and it may be affected by delayed nitrogen production in young, rapidly star-forming
galaxies (Perez-Montero & Contini 2009). Further difficulties arise because the weak
[N ii] line is difficult to detect in the typical low S/N spectra of high redshift objects;
because of this, most N 2 measurements to date have been made either for relatively
luminous and metal-rich galaxies or for stacked galaxy spectra.
Clearly both methods suffer from significant uncertainties. The comparison of the two
methods introduces an additional complication. Thus far, most metallicity measurements
at z 2 have used N 2, while most of those at z 3 are based on the oxygen lines. There
is likely to be a significant systematic offset between the two calibrations, and although
the best efforts have been made to understand and correct for this offset, it may depend
on the physical conditions in the galaxies. These conditions may be different at high
redshift, meaning that a low redshift comparison sample may not be appropriate. This
issue will not be resolved until large samples of high redshift galaxies with metallicities
determined with both methods exist; and even then, significant absolute uncertainty will
remain, until these indicators can be tied to metallicities determined with the Te method
for the same galaxy sample.
149
150
Dawn K. Erb
First and still the best-studied among these is the z = 2.7 galaxy MS 1512-cB58, a typical
Lyman break galaxy magnified by a factor of 30 by a foreground cluster. Initial optical
(Teplitz et al. 2000) and UV (Pettini et al. 2000) measurements indicated a metallicity of
1/41/3 Z ; more interestingly, however, the brightness of the galaxy allows for high
resolution spectra from which relative abundances of different elements can be measured.
These measurements refine the abundance of oxygen and other elements to 2/5 solar,
while nitrogen and the iron peak elements are underabundant by a factor of 3. These
values are consistent with the rapid star formation and young age of cB58, as such a
pattern would be expected if most of the metal enrichment had occurred within the
last 300 Myr, on the timescale for the release of nitrogen from intermediate mass
stars (Pettini et al. 2002). High resolution UV spectra have now been obtained for other
lensed galaxies as well, but this kind of elemental abundance analysis has proven difficult
because the Ly profile does not always allow for a measurement of the hydrogen column
density, and intervening absorption systems may make other lines difficult to measure
(Quider et al. 2009, 2010).
Gravitational lensing also allows the detection of galaxies that would otherwise be too
faint to observe at all. Yuan & Kewley (2009) have recently reported the first detection of
the [O iii] 4363 line at high redshift, in a lensed low mass (M = 4.4 108 M ) galaxy
at z = 1.7. The electron temperature metallicity derived from this measurement is quite
low, 12 + log (O/H) = 7.5; this is 0.6 dex lower than the metallicity derived from the
R23 method for the same galaxy, a discrepancy which is perhaps worrying; on the other
hand, offsets of this magnitude between metallicity diagnostics, or even between different
calibrations of the same diagnostic, are not uncommon, and R23 -derived metallicities may
be high compared to many other indicators (Kewley & Ellison 2008).
151
At z 0.7, Savaglio et al. (2005) find that galaxies are 0.10.2 dex lower in metallicity
at a given mass than in the local universe, with considerable scatter and a larger offset
at lower masses. At z 1, Perez-Montero et al. (2009) find an offset of 0.3 dex,
comparable to that found at z 2 by Erb et al. (2006a), while at z 3.5, Maiolino et al.
(2008) and Mannucci et al. (2009) find that galaxies are 6 times lower in metallicity
than in the local universe.
There are a variety of reasons to expect that galaxies at higher redshifts may be less
enriched: they are on average younger than galaxies in the local universe and probably
have higher gas fractions (Erb et al. 2006b); they may be accreting significant amounts
of metal-poor gas (e.g. Keres et al. 2005); and they drive powerful outflows, expelling
gas and metals into the IGM (Pettini et al. 2000, 2001). Nevertheless, the comparison of
metallicities across a wide redshift range is difficult, and there are several issues which
must be addressed before the redshift evolution of the mass-metallicity relation can be
reliably quantified. First, it is usually not possible to observe the same set of emission lines
for all of the galaxy samples, meaning that a variety of different metallicity indicators
have been used for the above studies. While the best efforts have been made to adjust
these calibrations to be consistent with each other, the conversions between metallicity
indicators are necessarily based on local galaxy samples (e.g. Kewley & Ellison 2008),
and physical conditions affecting the ratios of the strong emission lines may evolve with
redshift.
Second, it should also be realized that this shift in metallicity with redshift does not
represent an evolutionary sequence: the higher redshift samples are probably not the
progenitors of the galaxies at lower redshifts (Erb et al. 2006a, Maiolino et al. 2008,
Conroy et al. 2008). Rather, it represents the average metallicity of luminous star-forming
galaxies at a given epoch. A further issue related to the above two points is the apparently
large evolution between z 2.3 and z 3.5. Over the 11 Gyr between z 0 and
z 2, the average metallicities decrease by a factor of 22.5; the same decrease is seen
over the 1 Gyr between z 2.3 and z 3.5. Models and simulations currently fail
to explain this rapid evolution (Maiolino et al. 2008). Finally, several studies of the high
redshift mass-metallicity relation have noted that the offset with respect to local galaxies
appears to be larger at low masses; in other words, massive galaxies have metallicities
closer to those of massive galaxies in the local universe. Such a pattern would not be
surprising if the typical mass of galaxies hosting the bulk of star formation shifts to lower
masses at lower redshifts (e.g. Bundy et al. 2006).
(5.1)
where y is the yield, the ratio of the mass of metals produced and ejected by stars to
the mass locked in long-lived stars and remnants (Edmunds 1990). This relation can be
inverted to define the effective yield yeff :
yeff =
Z
.
ln(1/)
(5.2)
152
Dawn K. Erb
If the galaxy indeed behaves as a closed box, the effective yield will be equal to the true
yield, and either the expulsion of gas and metals from the galaxy or the accretion of metalpoor gas lower the effective yield (Edmunds 1990). Thus, measurements of metallicity and
gas fraction can reveal the presence of inflows or outflows of gas. In the local universe,
yeff is observed to be lower in lower mass galaxies; this is interpreted as a signature
of preferential loss of gas and metals from the shallower potential wells of lower mass
galaxies (Tremonti et al. 2004; see also Dalcanton 2007).
At high redshifts, current technology does not yet allow direct measurement of the
gas fractions of typical star-forming galaxies. Gas masses and gas fractions are therefore usually estimated from the star formation rate surface density, assuming the local
Kennicutt-Schmidt law (Kennicutt 1998). Using gas fractions determined in this way, effective yields have been calculated for galaxies at z 2 (Erb et al. 2006a) and at z 3.5
(Maiolino et al. 2008, Mannucci et al. 2009). In both cases, the effective yield is seen to
increase at lower masses, the opposite of the trend seen in the local universe.
Further estimates of the relative inflow and outflow rates can be obtained from the
Kennicutt-Schmidt law: it can be shown that if the local K-S law holds at high redshift,
the extended star formation histories observed in galaxies require the accretion of new gas
at approximately the rate it is lost to outflows and star formation (Erb 2008). Equation
5.1 can then be modified to include both inflows and outflows of gas (parameterized as a
fixed fraction of the star formation rate for simplicity). Incorporating the gas inflow rate
required by the K-S law then allows a determination of the outflow rate. The metallicities
and gas fractions of the z 2 galaxies are best fit with an outflow rate approximately
equal to the star formation rate, and an inflow rate roughly equal to the combined outflow
and star formation rates (Erb 2008). This model is also a good fit to the metallicities
and gas fractions at z 3.5 (Mannucci et al. 2009).
While these values are in general agreement with outflow rates estimated from observations (Pettini et al. 2000) and theoretically predicted inflow rates (e.g. Keres et al. 2005),
significant uncertainties remain. In addition to the indirectly inferred gas fractions and
the systematic uncertainties associated with the metallicity measurements, these models require assumptions of the metallicity of the inflowing and outflowing gas, neither of
which can currently be measured at high redshift. The appropriate value of the true yield
y is also not well known. In short, measurements of metallicity and gas fraction at high
redshift provide a model for gas inflows and outflows consistent with observational data
and theoretical expectations, but because of the large number of free parameters this
model is not a unique solution. It does seem clear, however, that the canonical low redshift model of increasingly efficient removal of metals in low mass galaxies does not match
the observed data at high redshift (Erb et al. 2006a, Mannucci et al. 2009). Additional
observations of lower mass galaxies are needed to confirm this result.
153
simple model in which outflows are more effective at removing metals from galaxies at
lower masses. Rather, the high redshift mass-metallicity relation is best fit with combined inflow and outflow models, where the outflow rate is approximately equal to the
star formation rate and significant gas accretion is required. While this picture is in
general agreement with the (very limited) observational data and theoretical expectations regarding inflows and outflows of gas at high redshifts, many uncertainties and free
parameters remain.
The number of 1 . z . 3 galaxies with metallicity measurements will grow by leaps
and bounds in the very near future, as multi-object near-IR spectrographs are put into
place on 810 m telescopes. These will allow larger samples of larger numbers of emission
lines, and deep, higher S/N spectra and spectra of fainter objects as it becomes more
worthwhile to spend large amounts of time on a single observation. Less optimistically,
measurements beyond z 3 will remain extremely difficult, as the strong [O iii] lines and
H shift beyond the K-band and become inaccessible from the ground; a space-based
near-IR spectrograph such as that planned for JWST will be needed for robust measurements of metallicities in these galaxies. The other option is to turn to the rest-frame
UV, which remains accessible; however, although there are many metallicity-dependent
features here, their use as diagnostics has proven difficult due to the complicated origin of
the spectral features and the low S/N and resolution of the spectra. As these data are relatively easy to obtain and will dominate very high redshift observational efforts for some
time to come, improved calibrations of the metallicity dependence of the strongest features in the rest-frame UV may offer the best hope for constraining the abundances of the
most distant galaxies. Metallicity measurements from new, larger samples of rest-frame
optical emission lines at z 23 will provide the necessary data for these calibrations.
References
Ando, M., Ohta, K., Iwata, I., Akiyama, M., Aoki, K., & Tamura, N. 2007, PASJ, 59, 717
Bundy, K., Ellis, R. S., Conselice, C. J., Taylor, J. E., Cooper, M. C., Willmer, C. N. A., Weiner,
B. J., Coil, A. L., Noeske, K. G., & Eisenhardt, P. R. M. 2006, ApJ, 651, 120
Conroy, C., Shapley, A. E., Tinker, J. L., Santos, M. R., & Lemson, G. 2008, ApJ, 679, 1192
Dalcanton, J. J. 2007, ApJ, 658, 941
Denicol
o, G., Terlevich, R., & Terlevich, E. 2002, MNRAS, 330, 69
Edmunds, M. G. 1990, MNRAS, 246, 678
Erb, D. K. 2008, ApJ, 674, 151
Erb, D. K., Shapley, A. E., Pettini, M., Steidel, C. C., Reddy, N. A., & Adelberger, K. L. 2006a,
ApJ, 644, 813
Erb, D. K., Steidel, C. C., Shapley, A. E., Pettini, M., Reddy, N. A., & Adelberger, K. L. 2006b,
ApJ, 646, 107
F
orster Schreiber, N. M., Genzel, R., Lehnert, M. D., Bouche, N., Verma, A., Erb, D. K., Shapley,
A. E., Steidel, C. C., Davies, R., Lutz, D., Nesvadba, N., Tacconi, L. J., Eisenhauer, F.,
Abuter, R., Gilbert, A., Gillessen, S., & Sternberg, A. 2006, ApJ, 645, 1062
Halliday, C., Daddi, E., Cimatti, A., Kurk, J., Renzini, A., Mignoli, M., Bolzonella, M., Pozzetti,
L., Dickinson, M., Zamorani, G., Berta, S., Franceschini, A., Cassata, P., Rodighiero, G.,
& Rosati, P. 2008, A&A, 479, 417
Hayashi, M., Motohara, K., Shimasaku, K., Onodera, M., Uchimoto, Y. K., Kashikawa, N.,
Yoshida, M., Okamura, S., Ly, C., & Malkan, M. A. 2009, ApJ, 691, 140
Heckman, T. M., Robert, C., Leitherer, C., Garnett, D. R., & van der Rydt, F. 1998, ApJ j, 503,
646
Kennicutt, R. C. 1998, ApJ, 498, 541
Kennicutt, R. C., Bresolin, F., & Garnett, D. R. 2003, ApJ, 591, 801
Keres, D., Katz, N., Weinberg, D. H., & Dave, R. 2005, MNRAS, 363, 2
Kewley, L. J. & Ellison, S. L. 2008, ApJ, 681, 1183
154
Dawn K. Erb
1. Introduction
By using the on-line off-line technique in the [O III] 5007 and H lines, we have searched
for emission line nebulae in three different galaxies: NGC 3109, a dwarf irregular at the
Local Group edge; NGC 6822, a dwarf irregular of the Local Group and NGC 300, a
Scd-type nearby spiral galaxy. This technique allows us to detect HII regions, planetary
nebulae (PNe) and other emission line objects. To select PN candidates we employed the
criteria described in Pe
na et al. (2007a) and Hern
andez-Martnez et al. (2009a) which
have been very successful to distinguish PNe from compact HII regions. So far we have
detected 20 PNe in NGC 3109 (Pe
na et al. 2007a), 26 PNe in NGC 6822 (HernandezMartnez et al., 2009a) and more than 60 PN candidates in NGC 300 (Pe
na et al., in
preparation).
Follow-up spectroscopy of a sub-sample of emission line objects has been performed, to
verify their nature and to determine their chemical abundances. Data have been collected
with the VLT-FORS and Gemini GMOS multi-object spectrographs. Exposure times
were long enough to detect , with signal-to-noise better than 3, the diagnostic-lines [O III]
4363 for electron temperature and [SII] 6719,6731, for electron density. This is extremely
important in order to calculate reliable ionic abundances, based on electron temperature
and density measurements.
Abundances of He, O, N, Ne, Ar and S were calculated for most of our objects from
the ionic abundances determined and using the ionization correction factors suggested by
Kingsburgh & Barlow (1994). Results for NGC 3109 were presented in Pe
na et al. (2007b),
and for NGC 6822 in Hern
andez-Martnez et al. (2009b). A summary of these results are
presented in Table 1, where PN and HII region abundances are presented separately. The
analysis for NGC 300 objects is in preparation (Stasi
nska et al., in preparation). In the
following we discuss these results comparatively.
155
156
M. Pe
na
Table 1. Total abundances for PNe and HII regions in NGC 3109, NGC 6822 and the
Magellanic Clouds
object
12+log O/H N/O
Ne/O
Ar/H
references1
NGC3109-HII
7.75 0.10 1.28 0.05 0.88 0.06 2.01 0.15 PSR
NGC3109-PNe
8.14 0.20 0.80 0.20 0.80 0.15 2.45 0.25 PSR
NGC6822-HII
8.06 0.04 1.34 0.15 0.73 0.07 2.18 0.10 H-M et al.
NGC6822-young PNe 8.08 0.12 0.50 0.50 0.78 0.08 2.26 0.20 H-M et al.
NGC6822-old PNe
7.70 0.20 0.88 0.30 0.77 0.12 2.26 0.20 H-M et al.
LMC-HII
8.40
1.50
0.80
2.20
D89, G99
LMC-PNe2
8.33
0.88
0.79
2.40
LD06
SMC-HII
8.00
1.50
0.80
2.10
D89, G99
SMC-PNe2
8.09
0.98
0.93
2.58
LD06
1
Ref: PSR: Pe
na et al. (2007); H-M et al: Hernandez-Martinez et al. 2009b;
D89: Dennefeld (1989); G99: Garnett (1999); LD06: Leisy & Dennefeld (2006)
2
Type I PNe not included
157
Figure 1. Oxygen abundance as a function of the normalized galactocentric distance for PNe
(dots) and HII regions (open squares, data from Bresolin et al. 2009) in NGC 300. A similar
chemical gradient is evident.
Ar and S abundances were derived for most of the nebulae. The results are reported in
Hern
andez-Martnez et al. (2009b). Combining our results for the compact HII regions
with others from the literature, we have found that the present interstellar medium in
this galaxy is homogeneous at least in its three central parsecs, showing a value 12+log
O/H = 8.060.04. Other elements like N, Ne and Ar show also a homogeneous galaxy.
The situation is different for the PN sample where we detected two kind of objects. Six
PNe of the analyzed sample present abundance pattern similar to HII regions, with an
average of 12+log O/H = 8.080.12. This sample includes two Peimbert Type I PN,
extremely N-rich. We have identified these PNe as young objects. The other 5 PNe show
abundances a factor of 2 poorer than the young group, with a 12+log O/H average of
7.700.20. An age of about 6 Gyr can be attributed to these objects. Although metal
poor, these PNe do not show evidence of O-enrichment as their Ar/O abundance ratios
are similar to those in the young PNe and HII regions. Thus O dredge-up has not occurred
in their central stars. Therefore, apparently a metallicity lower than 12+log O/H=7.7 is
needed for significant O dredge-up in low-intermediate mass stars.
Chemical evolution models, to reproduce the chemical behavior of PNe and HII regions in this galaxy, have been computed. They will be presented elsewhere (HernandezMartnez et al., in preparation).
2.3. NGC 300, a Sc-type nearby spiral
The search for PN candidates and follow-up spectroscopy in this galaxy, were performed
with VLT-FORS 2 on August, 2006 (program ID 077.B-0430B). This galaxy extends over
an angle of about 25 arcmin in the sky and, as said in the introduction, two zones were
analyzed, a central field and an outskirts one, each one of 6.86.8 squared arcmin. In
total we have detected 60 PN candidates (about 40 in the central region and 20 in the
outskirts) and follow-up spectroscopic data were obtained for 22 of them. A few compact
HII regions were also analyzed. The 22 candidates were confirmed as PNe, which implies
that our method for choosing candidates is very reliable.
Preliminary chemical abundances have been derived for the PNe for which [O III]
electron temperatures were measured. Our results for O abundances are shown in Fig. 1
(kindly provided by Grazyna Stasi
nska), where 12+log O/H is plotted as a function of
the galactocentric distance normalized to R25 . In this figure we include the abundances
158
M. Pe
na
determined for a large sample of HII regions by Bresolin et al. (2009). Some of our
compact HII regions are also shown.
From this figure it is clear that PNe analyzed in NGC 300 show O abundances similar
to HII regions, all over the galaxy. In this sense, all our PNe behave as young objects.
No old under-abundant PN population has been detected. This is most probably a selection effect, because, to perform a reliable spectroscopical study, only the brightest (and
probably the youngest) PN candidates were observed. NGC 300, at a distance of 1.88 kpc
(Gieren et al. 2005) is much farther away than NGC 6822. The abundance gradient with
galactocentric distance, observed for HII regions, is also present in our PN sample, with
very similar value. This gradient has been computed to be 0.077 0.06 by Bresolin et
al. (2009)
3. Conclusions
We have analyzed the PNe chemical behavior in three different galaxies, two dwarf
irregulars and one spiral. Different behaviors are found. So far, a few conclusions that
can be drawn:
In the very low metallicity galaxy NGC 3109 (present metallicity 12+log O/H=7.750.10),
PNe analyzed appear 0.39 dex O-richer than HII regions, while Ar/H ratio is, in average,
0.15 dex poorer. We interpret this as an evidence of significant O dredge-up in these lowintermediate mass stars (LIMS), that were born in a very low metallicity environment
some Gyr ago.
In NGC 6822, with a present metallicity 12+log O/H=8.06, two PN populations were
found. Young PNe have similar abundances than HII regions. A population of old PNe
shows metallicity a factor of two lower. In this case no strong evidence for O dredge-up
in LIMS is found. Therefore, metallicities lower than 12+log O/H =7.7 are required for
an efficient O dredge-up.
Both irregular galaxies are chemically homogeneous.
From our preliminary analysis of the chemical composition in PNe in NGC 300 we
find that they show similar abundances than HII regions. Apparently, the PNe analyzed
belong to a young population. Clear abundance gradients with galactocentric distance,
are found for HII regions and for PNe. Both gradients are very similar.
This work received financial support from DGAPA-UNAM (ID 112708).
References
Bresolin, F Gieren, W., & Kudritzki, R.-P. 2009, ApJ, 700, 309
Evans, C. J., Bresolin, F., Urbaneja, M. A., et al. 2007, ApJ, 659, 1198
Gieren, W., Pietrzy
nski, G., Soszy
nski, I., Bresolin, F., et al. 2005, ApJ, 628, 695
Hern
andez-Martnez, L., & Pe
na, M. 2009a, A&A, 495, 447
Hern
andez-Martnez, L., Pe
na, M., Carigi,L., & Garca-Rojas, J. 2009b, A&A, 505, in press
Kingsburgh, R., & Barlow, M. 1994, MNRAS, 271, 257
Lee H., McCall M. L., Kingsburgh R. L., Ross R., Stevenson C. C. 2003, AJ, 125, 146
Kniazev, A. Y., Grebel, E. K., Pustilnik, S. A., Pramskij, A. G., & Zucker, D. B. 2005, AJ, 130,
1558
Leisy, P., & Dennefeld, M. 2006, A&A, 456, 451
Magrini, L., Leisy, P., Corradi, R. L. M., et al. 2005, A&A, 443, 115
Marigo, P. 2001, A&A, 370, 194
Pe
na, M., Richer, M. G., & Stasi
nska, G. 2007a, A&A, 466, 75
Pe
na, M., Stasi
nska, G., & Richer, M. G., 2007b, A&A, 476, 745
UFRJ - Observat
orio do Valongo, Ladeira Pedro Antonio 43, 20080-090 Rio de Janeiro, Brazil
email: denise@astro.ufrj.br
2
INAF - Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, I-50125 Firenze, Italy
email: laura@arcetri.astro.it
Abstract. The irregular dwarf galaxy IC10 is located within the Local Group (LG) at a distance
of 750 kpc. Although several studies have revealed the existence of stellar populations with a
broad range of ages, its star formation history (SFH) and age-metallicity (AM) relationship
remain quite unknown. In this contribution we present our spectroscopic investigation of 15 H ii
regions, 9 planetary nebulae (PNe) and 1 symbiotic star so far the farthest known symbiotic
binary. Our main goal is to reconstruct the SFH of IC10 and to constrain its AM relationship
using young and intermediate-age stars. The direct availability of the electron temperature in
our emission-line spectra allows an accurate determination of the IC10 metallicity map at two
different epochs. We find a non-homogeneous distribution of metals at both epochs, but similar
average abundances for the two analyzed populations. The derived AM relationship shows a little
global enrichment, which is interpreted as due to the loss of metals by supernovae winds and
to differential gas outflows. Our results bring strong observational constraints to the chemical
enrichment history of IC10, the formation of dwarf irregular galaxies and the evolution of the
LG as well.
Keywords. galaxies: abundances, galaxies: evolution, Local Group, PNe, HII regions.
160
abundances of a significant sample of LG dwarf galaxies using PNe. PNe are indeed
present in early- and late-type galaxies. In addition, their chemistry provides information
about the distant past of the galaxy ISM (old and intermediate-age stars), whereas the
simultaneous analysis of H ii regions give access to the present time ISM chemistry.
Together, these strong line emitters provide key constraints to the LG chemical evolution.
161
12+log(O/H) PNe
> 8.40
> 8.4
8.10 to 8.30
8.10 to 8.30
> 8.4
log(S/O)
log(Ne/O)
log(N/O)
12+log(O/H)
Figure 1. Metallicity maps of IC10. Left: from O/H of H ii regions (present-time ISM). Right:
from the old and intermediate-age population, PNe. Ellipses of different grey tones mark regions
of various oxygen abundances.
log(Ar/O)
12+log(O/H)
Mv
Figure 2. Left: Chemical abundance patterns for PNe (filled circles) and H ii regions (empty
circles), compared with the BCGs relations of Izotov et al. (2006) for the -element to oxygen
ratios, and their dispersion. Right: The dwarf galaxies metallicity-luminosity relation, which
show our own results and others from the literature.
dispersion of N/O and the different stellar origin of nitrogen and oxygen, the relationship
of this ratio with 12+log(O/H) is not plotted. In a starburst galaxy, as IC10, the N/O
ratio should be lower than in BCGs, due to the time delayed nitrogen production if
compared to the much faster production of oxygen. The reason why N/O in IC10 is
consistent with other BCGs reflects the fact that anything ejected by massive stars
whether oxygen by type II supernovae or nitrogen in the winds of WR stars, will not
be immediately incorporated into the nebular gas. This ejected matter is too hot, and
needs to cool before it can mix with the ISM and can be seen in the optical spectra of
nebular gas.
Fig. 2 also shows that the abundance ratios of BCGs and IC10 PNe agree very well.
Noting that PNe progenitors produce and dredge-up nitrogen during their lifetimes, the
agreement of N/O ratio would appear unexpected. However, as noted by Richer & McCall
(2007), the brightest PNe in starburst galaxies often behave in this way.
The similar abundance ratios of PNe and H ii regions is an evidence that IC10 had
small variations in the metal content from the PN progenitors formation to the present
time. There are two possible explanations. i) PNe in dIrrs are very young, and thus
162
they represent the same age as H ii regions. However this hypothesis is ruled out by the
absence of the HeII lines in their spectra, tracers of the most massive and youngest PNe.
ii) PNe and H ii regions belong to different epochs, and so their similar composition is a
probe of the metal loss in dIrrs by strong winds and outflows which remove the chemical
enriched material from the place they were formed. The winds mainly allow the loss of
-elements, through type II SN winds (MacLow & Ferrara 1999; Recchi et al. 2008).
School of Earth and Space Exploration, Arizona State University, P.O. Box 871404, Tempe,
AZ, 85287-1404
email: evan.scannapieco@asu.edu
2
Dept. of Physics, University of California, Santa Barbara, CA, 93106
Abstract.
Intergalactic metals are ubiquitous, but their sources remain unknown. A key constraint on
these sources is the spatial distributions of metals. Yet, the clustering of metals is difficult
to interpret along single lines-of-sight, because distance and velocity information are mixed in
redshift space. To overcome this situation we are carrying out detailed comparisons between the
line-of-sight and transverse distributions of metal line absorption systems observed in a large
sample of QSO pairs and simulations including a wide range of IGM-enrichment scenarios. The
degeneracy between distance and velocity is broken by the transverse information available in
pairs of sightlines, and thus these comparisons are providing unique new constraints on when
and where metals were ejected from galaxies.
Keywords. intergalactic medium, QSO absorption lines
1. Introduction
The intergalactic medium (IGM) spans the most remote and empty regions of the
universe. However, even this most rarified of environments is teeming with the byproducts
of stellar evolution. In fact, for over 15 years, quasar (QSO) absorption-line studies have
shown that the IGM is filled with heavy elements (e.g. Tytler et al. 1995; Songaila &
Cowie 1996), which are highly inhomogeneous (Rauch, Haehnelt & Steinmetz 1997), and
present even at high redshifts (Becker et al 2009; Ryan-Weber et al. 2009).
Yet despite these many observations, the sources of intergalactic metals remain unknown. There were numerous starburst-driven outflows at z 3 (e.g. Pettini et al. 2001)
but the evidence pointing toward them being the driving force behind the majority of
enrichment is unclear. Instead, there are strong theoretical arguments that suggest that
these galaxies may only represent the tail end of a larger number of smaller starbursts at
higher redshifts (Scannapieco, Ferrara, & Madau 2002) that may have been much more
important for IGM enrichment. Indeed, a number of studies suggest that the highestredshift metal-free stars may have been extremely massive, resulting in a large number
of supernovae that would have distributed metals at redshifts of z > 15. (e.g. Bromm
et al. 2001; Schneider et al.2002; Abel et al. 2002; Heger & Woosley 2002).
Regardless of which objects enriched the IGM, it is clear that they must have formed in
dense regions of space, which are far more clustered than the overall matter distribution.
This geometrical-biasing, is a systematic function of the masses of these structures
(Kaiser 1984). Thus, information from the large scale clustering of metal absorbers provides valuable insight into the objects from which metals were ejected.
Scannapieco et al. (2006) studied this spatial distribution in high-resolution, high
signal-to-noise VLT (UVES) spectra of absorbers at redshifts z = 2 3. They com163
164
puted a line-of-sight correlation function in redshift space, and found that it had a high
amplitude and a knee at 150 km/s. By comparing this with a suite of simulations,
they found that these measurements were consistent with bubbles of metals 2 comoving Mpc in radius, surrounding dark matter halos with masses 1012 M . However, the
clustering of metals is difficult to interpret along single lines of sight, because distance
and velocity information are mixed in redshift space.
2. Method
To overcome this degeneracy we made use of a sample of 29 quasar pairs, with redshifts
between 1.7-4.3 and separations from 0.1 to 2 comoving Mpc (Martin et al. 2009),
which were selected from the ongoing work of Hennawi et al. (2006,2009). Spectra for
each pair were taken with the ESI spectrograph on Keck II, which provided a resolution
of 60 km s1 , and continuous coverage from 4,000
A to 10,000
A. For this study we
focused on CIV absorbers, whose 1548.204, 1550.781 doublet is the best tracer of metals
in highly-ionized intergalactic gas.
The absorbers were identified by inspection, fitted by a pair of Gaussian line profiles
if they were well separated from other absorption lines, and fitted using the SPECFIT
program (Kriss 1994) in the case of blends. For most spectra, this allowed us to detect
CIV absorbers at the 5 level down to (1550.781) equivalent widths of 0.02
A. Our full
data set included 450 absorbers, but as we are interested in the general distribution of
intergalactic metals, we flagged absorbers within 5,000 km s1 of the QSO as associated
absorbers and excluded them from the analysis. This left a sample of 316 intervening
CIV absorbers, with a median redshift of 3.0.
To compare with this data set, we used the Gadget-2 code (Springel 2005) to perform
a large numerical simulation that contained 7003 dark matter particles, each with a mass
of 3.35 107M , in a periodic cubic region of the universe, 50 comoving Mpc/h on a side.
Initial conditions were generated using the Grafic code (Berschinger 2001), and evolved
to z = 1.7 assuming a CDM model with a Hubble constant of 70 km/s, and 0 = 0.3,
= 0.7, b = 0.046, and 8 = 0.9, where 0 , , and b are the total matter, vacuum,
and baryonic densities in units of the critical density, and 82 is the variance of linear
fluctuations on the 8h1 Mpc scale.
Outputs of the simulation were written once per light crossing time of the box, and
for each output we identified bound dark-matter groups with masses above 1010 M
using a friends-of-friend algorithm (Davis et al. 1985). Following the approach in Pichon
et al. (2003), we selected all groups above a threshold group mass Ms as markers of the
centers of enrichment events, about which we painted spheres of radius Rs which we
took to represent regions containing metals. All smaller groups within these regions were
associated with CIV absorbers, leading to modeled metal distributions as ilustated in
Figure 1, which shows the results of two choices of of RS with Ms fixed at 5 1011 M
For each pair of quasars we cast two sightlines through the simulation outputs, separated by the appropriate angular separation and covering the appropriate redshift range.
Each time a sightline passed near one of the groups associated with CIV absorbers, we
added an absorption line system to our sample, with an equivalent width which was set
by the impact parameter between the sightline and the absorber, such that the overall
column density distribution was the same as in observations. Finally, for each simulated
sightline pair, we imposed the same noise limits as in the real data, giving us simulated
data sets with the same properties as the observed data.
165
Figure 1. Distribution of intergalactic metals at z = 2.9 in two simulated models. In the left
panel, metal-line systems are relatively near (within 0.7 comoving Mpc) of 5 1011 M halos In
the right panel, metal-line systems are clustered at relatively large distances from 5 1011 M
halos (up to 2.5 comoving Mpc), as would be expected if most intergalactic metal enrichment
took place during the epoch of reionization.
3. Results
With these models in hand, we were able to directly compare our simulated measurements of both the line-of-sight correlation function and transverse correlation function.
Along the line-of-sight, this was computed by counting up pairs as a function separation, and dividing by the number of pairs expected for a random spatial distribution of
absorbers, subjected to the same noise limits as in the observations. For the transverse
correlation funciton, this was computed by matching up all pairs of absorbers within
1,000 km s1 of each other, and again dividing by the number of such pairs in a random
spatial distribution of absorbers subjected to the same noise limits as in the data.
While this analysis is still in progress, Figure 2 shows the results of a comparison
between the observations a model with Ms = 5 1011M and Rs = 2.5 Mpc. In general,
this large bubble model, which is similar to the one fit in Scannapieco et al. (2006)
provides a good match to both the line-of-sight and transverse correlation functions. On
the other hand, decreasing the bubble size, Rs or changing the mass of the central dark
matter halos, Ms , leads to a poorer fit to both the line-of-sight and transverse correlation
functions simultaneously.
Note this large bubble size is more easily accommodated as arising from the overlap of
numerous clustered sources at very high-redshifts (e.g. Scannpieco 2005) than it is from
single bubbles driven by outflows from z 3 Lyman break galaxies. Indeed, for winds to
reach these large distances within the 100 1, 000 Myr timescale over which a galaxy is
most likely to be identified by the drop-out technique (e.g. Ferguson et al. 2002; Shapley
et al. 2005) would require typical velocities between 1,000 and 10,000 km/s, which are
not realistic.
Furthermore, we considered a simulation model with ongoing winds, in which we applied a boost of velocity vboost pointed radially outward from the large Ms halos to each
of lower halos associated with CIV absorbers. In general introducing this additional parameter did not improve the fits between the models and the simulations, suggesting that
widespread outflows at z 6 4 may not play a role in determining the spatial and velocity distribution of most CIV systems. Further observations and simulations are ongoing,
which are quickly closing in on the sources of intergalactic metals.
166
Figure 2. Line-of-sight correlation function for our observed data set (blue points) and simulations (black points) with Ms = 5 1011 M and Rs = 2.5 Mpc. Transverse correlation function
from the observations divided up into 8 bins (red points) and 4 bins (blue points) with an equal
number of expected pairs, as compared to the transverse correlation function derived from the
simulations (green points)with Ms = 5 1011 M and Rs = 2.5 Mpc.
References
Abel, T., Bryan, G. L., & Norman, M. L. 2002, Science, 295, 93
Becker, G. D., Rauch, M., & Sargent, W. L. W. 2009, ApJ, 698, 1010
Bertschinger E. 2001, ApJS, 137, 1
Bromm V., Ferrara A., Coppi P. S., & Larson R. B. 2001, MNRAS, 328, 969
Cowie, L. L., Songaila, A., Kim T.-S., & Hu, E. M. 1995, AJ, 109, 1522
Davis, M., Efstathiou, G., Frenk, C. S., & White, S. D. M. 1985, ApJ, 292, 371
Feguson, H. C., Dickenson, M., & Papovich, C. 2002, ApJ, 569, 65
Heger, A. & Woosley, S. E. 2002, ApJ, 567, 532
Hennawi, J. F. et al. 2006, AJ, 131, 1
Hennawi, J. F. et al. 2009, ApJ, submitted (arXiv:09083907)
Kaiser N. 1984, ApJ, 284, L9
Kriss, G. 1994, A.S.P. Conference Series, 61, 437
Pettini M. et al. 2001, ApJ, 554, 981
Pichon, C. et al. 2003, ApJL, 597, 97
Martin, C. L. et al. 2009, in preperation
Rauch M., Haehnelt M. G., & Steinmetz M. 1997, ApJ, 481, 601
Rauch M., Haehnelt M. G., & Steinmetz M. 1997, ApJ, 481, 601
Scannapieco E., Ferrara A., & Madau P. 2002, ApJ, 574, 590
Scannapieco E. 2005, ApJL, 724, 1
Scannapieco E., et al. 2006, MNRAS, 365, 615
Schneider R., Ferrara A., Natarajan P., & Omukai K. 2002, ApJ, 571, 30
Shapley, A. E. et al. 2005, ApJ, 626, 698
Springel V. 2005, MNRAS, 364, 1105
Tytler, D., et al. 1995 in QSO Absorption LInes, ed. G. Meylan, 289
Institute of Astronomy, University of Latvia, Rainis blvd. 19, Riga, LV-1586, Latvia
email: dima@latnet.lv
2
Max Planck Institute for Astrophysics, Karl-Schwarzschild-Str. 1, Postfach 1317, 85741
Garching, Germany
email: sunyaev@mpa-garching.mpg.de
3
Space Research Institute, Russian Academy of Sciences, Profsoyuznaya 84/32, 117997
Moscow, Russia
Abstract. We discuss possibility of observations of the warm-hot intergalactic medium using the
hyperfine structure line of highly charged nitrogen ion 14 N VII (rest wavelength = 5.652 mm).
Observations of this line will allow to separate bulk and turbulent motions in the observed
target and will broaden the information about the gas ionization state, chemical and isotopic
composition.
Wavelength of this line is well-suited for ground-based observation of objects at z 0.15 0.6
when it is redshifted to the widely-used 6.5 9 mm spectral band, and, for example, for z > 1.3,
when the line can be observed in 1.3 cm band and at lower frequencies.
Keywords. galaxies: intergalactic medium - radio lines: general
1. Introduction
Hyperfine structure (HFS) lines of highly-charged ions may open a new window in
observations of hot astrophysical plasmas. Radio observations of these lines will allow
obtaining information about velocity field, mass, temperature and chemical abundance
distribution of hot, warm and highly photoionized gas. The first discussion of the astrophysical applications of HFS lines was given by Sunyaev & Churazov (1984), who
especially pointed out the millimeter lines of 14 N VII and 57 Fe XXIV.
The latter line was predicted to be bright in hot intracluster gas in clusters of galaxies
at temperatures around 2 107 K. This line should also be bright in spectra of young
supernova remnants, such as Cas A, if 57 Fe isotopic fraction in ejecta is not much less
than the terrestrial value.
The 14 N VII line was not considered interesting for galactic astronomy, as its rest
frequency of 53.041(15) GHz (Shabaev, Shabaeva & Tupitsyn (1995), Volotka et al. 2008)
is in the wing of atmospheric oxygen absorption band. Sunyaev & Churazov (1984)
proposed to use it for observations of objects having small positive redshifts, when the
observed frequency moves out of this strongly attenuated spectral band. Following this
idea, both theoretical studies (Sunyaev & Docenko (2007)) and observational searches
(Bregman & Irwin (2007)) have been performed recently, aiming at the line detection
from objects at redshifts z > 0.15. The 14 N VII HFS line is unique in the sense that
it allows to study from the ground astrophysical plasma at temperatures around 106 K,
that otherwise can be explored only by rocked-based and space ultraviolet and soft X-ray
missions.
Several theoretical studies have focused on specific environments in which the 14 N VII
167
168
HFS line might be detectable: surroundings of quasars (Docenko & Sunyaev (2007b)),
hot interstellar medium in elliptical galaxies (Docenko & Sunyaev (2007a)).
In this contribution, we concentrate on a possible application of the 14 N VII hyperfine
structure line for studies of intergalactic metals in the warm-hot intergalactic medium
(WHIM).
According to modern hydrodynamical simulations of the large-scale structure of the
Universe (e.g., Cen & Ostriker (1999)), the warm-hot intergalactic medium (rarefied
intergalactic gas heated to temperatures T 105 107 K) contains dominant fraction of
barions in the present Universe. However, it is almost unobserved till now.
It is clear that radio observations of HFS lines of ions abundant in these conditions will
provide additional information about the velocity field, mass, temperature and chemical
abundance distribution of the warm-hot intergalactic medium. Interestingly, computations by Hellsten, Gnedin & Miralda-Escude (1998), Cen & Ostriker (1999) show that
heavy element abundances in the WHIM rise sharply in regions of higher temperature
and density reaching values close to the solar ones.
2. The
14 N
While the solar abundance of 14 N relative to hydrogen is only 8105, the spontaneous
decay rate of its hyperfine transition is 7 104 times higher. However, the WHIM column
density is much lower than e.g. of the atomic gas in galaxies, and resulting typical optical
depth of the 14 N VII HFS line is very small.
Let us compare it with the soft X-ray lines of O VII and O VIII, that are widely
discussed as promising probes of the WHIM. Absorption cross-sections of these X-ray
transitions around 20
A are about three orders of magnitude larger than that of the
HFS transition. Optical depth of the HFS transition is additionally diminished due to
population of upper hyperfine sublevel in the field of the CMB radiation (see Sunyaev &
Docenko (2007)). Resulting rough estimate of 14 N VII HFS line optical depth corresponding to O VII or O VIII soft X-ray line (O) 1 is only about (14 NVII) (3 10) 105.
To assess frequency of occurrence of 14 N VII absorption lines in WHIM, we have used
the distribution function of O VIII X-ray absorption line equivalent width from Cen &
Fang (2006) and correspondence between the equivalent width and ionic column density
from Chen et al. (2003) cosmological simulations. As a first approximation, we also
assumed that in WHIM conditions the ionization equilibrium curve of O VIII atoms is
the same as of N VII atoms and their relative abundance N/O is solar. Then on average
in one sight line in the redshift intervals z = 0.15 0.30 and z = 0.3 0.6 we expect one
14
N VII HFS line with > 2 105 and > 3 105 , respectively.
Our estimates of the 14 N VII occurrence frequency are plotted on Figure 1 with dashed
lines, where they are compared with sensitivity of Green Bank Telescope (GBT ) and
planned performance of the Square Kilometer Array (SKA). The solid lines denote theoretical 3- detection limits of an absorption line in one velocity channel for several
brightest quasars in mm and cm bands.
The decrease with redshift of the expected 14 N VII absorber number having the same
optical depth is explained by combination of two effects: (a) increase of the CMB temperature and resulting line damping due to population of the upper HFS sublevel (Sunyaev
& Docenko (2007)), and (b) decrease of the WHIM mass at redshifts higher than about
one (Cen & Ostriker (1999)).
Although lines with optical depth of the order of 105 may seem too weak to be observable, emission lines of comparable line-to-background ratio have already been detected
on the GBT by Vanden Bout, Solomon & Maddalena (2004).
14
169
1e-04
Observing the source
0552+398
3C 454.3
3C 273
1
3
0.5
1
Redshift z
1.5
Figure 1. Expected 14 N VII line detection frequency on the sightlines to the brightest quasars.
Green lines denote expected HFS absorption line depth frequency inferred from cosmological
simulations. Red lines show 3- spectral line detection limits. (Upper panel ) Observations with
GBT, ten hours of on-source integration, 10 km/s spectral channel width; (Lower panel ) Observations with SKA, one hours of on-source integration, 30 km/s spectral channel width. In one
hour of integration, SKA will be able to detect five to ten lines at 3- level and about three
lines at 10- level in redshift range from one to two.
3. Conclusions
Thanks to relatively high abundance of 14 N in the WHIM, the line of 14 N VII is
the most prospective to be observed in absorption in spectra of brightest mm-band
extragalactic radio sources with z > 0.15. Typical optical depth predicted for WHIM
filaments reaches levels of several105 that is within the reach of the existing and
planned instruments.
170
Using GBT, a 3- low-z WHIM detection would take about 10 hours on-source. With
SKA, high-z detection may be achieved in less than an hour.
Acknowledgements
D.D. gratefully acknowledges the IAU travel grant.
References
Bregman, J.N. & Irwin, J.A. 2007, ApJ 666, 139
Cen, R. & Fang, T. 2006 ApJ 650, 573
Cen, R. and Ostriker, J. P., 1999 ApJ 514, 1
Cen, R. and Ostriker, J. P., 2006 ApJ 650, 560
Chen, X., Weinberg, D. H., Katz, N. & Dave, R. 2003 ApJ 594, 42
Docenko, D. & Sunyaev, R.A. 2007a, in: H. B
ohringer, G. W. Pratt, A. Finoguenov, &
P. Schuecker (eds.), Heating versus Cooling in Galaxies and Clusters of Galaxies, ESO
Astrophysics Symposia (Berlin, Heidelberg: Springer), p. 333
Docenko, D. & Sunyaev, R.A. 2007b, in: From Planets to Dark Energy: the Modern Radio
Universe, Published online at SISSA, Proceedings of Science, p.90
Hellsten, U., Gnedin, N. Y., Miralda-Escude, J. 1998 ApJ 509, 56
Shabaev, V. M., Shabaeva, M. B. & Tupitsyn, I. I. 1995, Phys. Rev. A 52, 3686
Sunyaev, R.A. & Churazov, E.M. 1984, Soviet Astron. Lett. 10, 201
Sunyaev, R.A. & Docenko, D.O. 2007, Astron. Lett. 33, 67
Vanden Bout, P. A., Solomon, P. M. & Maddalena, R. J. 2004 ApJ 614, L97
Volotka, A. V., Glazov, D. A., Tupitsyn, I. I., Oreshkina, N. S., Plunien, G., Shabaev, V. M.
2008 Phys. Rev. A 78, 062507
Abstract.
High-redshift quasars provide a unique glimpse into the early evolution of massive galaxies.
The physical processes that trigger major bursts of star formation in quasar host galaxies (mergers and interactions) probably also funnel gas into the central regions to grow the super-massive
black holes (SMBHs) and ignite the luminous quasar phenomenon. The globally dense environments where this occurs were probably also among the first to collapse and manufacture stars in
significant numbers after the big bang. Measurements of the elemental abundances near quasars
place important constraints on the nature, timing and extent of this star formation. A variety
of studies using independent emission and absorption line diagnostics have shown that quasar
environments have gas-phase metallicities that are typically a few times solar at all observed
redshifts. These results are consistent with galaxy evolution scenarios in which large amounts of
star formation (e.g., in the central regions) precede the visibly bright quasar phase. An observed
trend for higher metallicities in more luminmous quasars (powered by more massive SMBHs) is
probably tied to the well-known massmetallicity relation among ordinary galaxies. This correlation and the absence of a trend with redshift indicate that mass is a more important parameter
in the evolution than the time elapsed since the big bang.
Keywords. galaxies: formation, galaxies: high-redshift, quasars: general, quasars: absorption
lines, quasars: emission lines
1. Introduction
Quasars are powered by matter accreting onto super-massive black holes (SMBHs)
with masses of order 108 to 1010 M . During their brief lifetimes, just 107 to 108 yr,
quasar bolometric luminosities easily exceed the total energy output from stars in the
host galaxies. Most of this accretion-driven quasar activity occurred at high redshifts, in
the range z 13 but extending out to z > 6 (Silverman et al. 2005, Richards et al.
2006). This is also the era when massive elliptical galaxies were rapidly being assembled
and making most of their stars (Hopkins & Beacom 2006). We see the results of this early
evolution in the galaxies around us today. Giant ellipticals are gas poor with primarily old
and metal-rich stellar populations. Their mean stellar metallicities are typically above
solar, with the nuclear regions having hZi 23 Z (Trager et al. 2000, Spolaor et
al. 2009). The cores of giant spheroids also contain SMBHs that are mostly dark and
dormant because their gaseous fuel supply has essentially run out. However, at early
cosmic times, when the SMBHs were still growing and rapidly accreting matter, they
were the engines of quasars. Every giant elliptical today was a quasar host at earlier
times and higher redshifts.
The observed correlation between the masses of SMBHs and their surrounding spheroids
(Tremaine et al. 2002) implies that SMBH growth in galactic nuclei is somehow connected
to the same physical processes that make the galaxies. Remarkably, this physical connection occurs across spatial scales from <1 pc for the accreting SMBHs to >1 kpc in the
171
172
galaxy hosts. One possibility is that feedback from the quasar, in the form of powerful
outflows and perhaps ionizing radiation, directly regulates both star formation in the
host galaxy and the growth of the SMBH itself (Kauffmann & Haehnelt 2000).
A major goal of quasar research is to understand the evolutionary relationship between
quasars, SMBHs and their host galaxies. The most basic question involves simply the
relative timing of events: When does the quasar phase occur in relation to the major
episodes of star formation and mass assembly in the host galaxy? One common view is
that major mergers of gas-rich galaxies trigger both the growth of a central SMBH and
bursts of star formation in the host that appear observationally as ultra-luminous infrared
galaxies (ULIRGs, Sanders et al. 1988, Veilleux et al. 2009). Most of the SMBH growth
occurs in obscurity, behind the dusty veil of the star-bursting ULIRG, but eventually a
blowout of gas and dust reveals a visibly luminous quasar in the galactic center. In this
scenario, the short-lived quasar phase is part of (and maybe an active participant in)
the transition from the early stages of violent star formation to the more relaxed and
passively evolving galaxy that comes later (also Hopkins et al. 2005).
Measurements of the elemental abundances near quasars provide important tests of the
evolution. For example, the metallicities can tell us the extent to which star formation
and chemical enrichment occurred in different environments prior to the observed quasar
epoch. Specific abundance ratios, such as Fe/, can further constrain the nature and
timing of this star formation. The first review of quasar abundance studies appeared in
Hamann & Ferland (1999), which we updated most recently in Hamann et al. (2004 and
2007). Here we present a brief summary of the main results with further updates. Please
see our previous reviews for more extensive discussions of earlier work.
173
Figure 1. Mean spectra of quasars in three redshift intervals show the lack of evolution in the
broad emission line ratios and thus the inferred metallicities (from Juarez et al. (2009).
further simplification in the calculations is that the metal abundances are scaled from
the solar values, with nitrogen treated as a secondary element (such that N/O O/H).
Hamann et al. described the resulting strong metallicity dependence of several UV line
ratios involving nitrogen, such as N v/C iv, N v/(C iv+ O vi) and N iii]/O iii]. Nagao
et al. additionally promoted the use of other ratios, such as (Si iv+O iv])/C iv, whose
metallicity dependence appears to arise from a combination of saturation effects and
weaker lines carrying a greater share of the total cooling as Z increases.
The main result from these analyses is that quasar BEL regions have metallicities of
typically 25 times solar at all redshifts (see also our reviews, Jiang et al. 2007, Juarez et
al. 2009). It has been noted that different BEL ratios can indicate different metallicities
by up to a factor of 2 in the same quasar (Dietrich et al. 2003, Nagao et al. 2005).
The best strategy is to take the average metallicity inferred from several line ratios.
The worst case disparities between the line ratios might be a reasonable estimate of the
theoretical uncertainties in individual quasars. However, there are no obvious biases that
would push the derived metallicities higher, and therefore the general result for supersolar metallicities should be secure. Also, our ability to measure abundance trends in
quasar samples should be much better than the uncertainties in individual estimates.
Another important result is that there is no significant trend in the BEL metallicities
with redshift out to at least z 6 (Figure 1; Nagao et al. 2005, Dietrich et al. 2003).
There is, however, a significant trend with quasar luminosity that appears to have its
origins in an even stronger positive correlation between the metallicity and SMBH mass
(Warner et al. 2003 and in prep.). If we consider that more massive SMBHs reside in more
massive galaxies, then the quasar massmetallicity correlation is probably just another
manifestation of the well-known massmetallicity relation among galaxies.
In the first attempt to compare quasar metallicities directly to host galaxy properties,
Simon & Hamann (2009 and this proceedings) looked for trends with LF IR as a surrogate
for the star formation rate in the hosts. They found none. Visibly bright quasars are
The metallicities reported in these papers should be scaled downward by 30% to reflect the
most recent solar abundances (see Dhanda et al. 2007). The numbers here include this scaling.
174
generically metal-rich. The quasars associated with the most extreme star formation
rates, with LF IR in the range of ULIRGs, do not appear to be signifantly younger or
chemically less mature based on their BEL metallicities. This result is consistent with
evolutionary schemes that have most of the star formation in the (inner) host galaxies
occurring before the visible quasar epoch.
It should be noted that the BEL results apply strictly to the gas within 1 pc of
the central quasars. However, the star formation that produces the BEL metallicities
probably occurs on much larger scales. If BEL regions are continuously replenished by
gas from the accretion disk, then the total amount of metal-rich gas probed by the BELs
is at least equal to the amount that funnels through the disk during a quasars lifetime.
This is roughly the black hole mass or >109 M for luminous quasars. This amount of
gas with Z > 2 Z nominally requires enrichment by a stellar population with mass
>1010 M (assuming a normal Galactic IMF, see Hamann et al. 2007). Thus it appears
that at least Galactic-bulge-size stellar populations are already present or in their final
stages of assembly at the observed quasar epochs.
A more specific constraint on the ages of the stellar populations around quasars might
be gained from measurements of the Fe/ abundance. The age constraint comes from the
0.3 to 0.5 Gyr delay in the Fe enrichment from SN Ia. Attempts to measure this ratio in
quasars using their Fe ii/Mg ii BELs are poorly calibrated in an absolute sense and comparisons between quasars should probably be controlled for non-abundance parameters
like luminosity and perhaps Eddington ratio L/Ledd (Dong et al. 2009). Nonetheless, it
has been argued that the lack of an observed trend in this line ratio with redshift means
that SN Ias are contributing at all epochs and therefore the bursts of star formation that
enriched the BEL regions started at least 0.3 Gyr before the observed quasar redshifts
(Dietrich & Hamann 2008, Jiang et al. 2007 and refs. therein).
2.2. Narrow Emission Lines
Narrow emission lines probe much larger scales than the BELs, from roughly 100 to 104 pc
around luminous quasars. The metal abundances in these regions have been measured in
two ways. One approach is to use the usual H i and forbidden lines at visible wavelengths
that are also used to classify starbursts and assorted AGN. At high redshifts, these lines
present observational difficulties because they appear in the observed-frame infrared. In
one large study involving 23000 low-redshift Seyfert 2 galaxies, which have much lower
luminosities and than quasars, Groves et al. (2006) found the typical metallicities to be
24 Z , consistent with earlier work on similar objects.
A second approach is to use the same UV line ratios as the BEL studies. This is possible
only in sources where the BELs are weak or absent, e.g., in type 2 AGN. Nagao et al.
(2006) used this approach to derive metallicities from the narrow UV lines in a sample
of high-redshift type 2 quasars and radio galaxies. They found a larger range of results
than Groves et al., but super-solar metallicities still appear to be typical.
All of the narrow emission line studies so far rely on calculations that choose a particular gas density, even though the actual densities are poorly known. It would be helpful
to check these results against calculations that use a range of gas densities in cloud
distributions, similar to the BEL studies discussed above.
2.3. Broad Absorption Lines
Broad absorption lines (BALs) form in massive outflows from the inner accretion disk.
Figure 2 shows one example of a quasar BAL spectrum. Part of the motivation for
studying BAL outflows is that they might play a role in the kinetic energy feedback that
affects SMBH growth and the host galaxy evolution (1). Unfortunately, BAL studies
175
Figure 2. Rest-frame UV spectra of a redshift 2.4 quasar showing modest variability and
some blending effects in its broad absorption lines (from Capellupo et al. in prep.).
are plagued by large uncertianties in the derived column densities due to line blending.
In the best effort so far to measure the abundances in a BAL flow, Arav et al. (2001)
used constraints on many lines to derive a metallicity consistent with solar. However,
the uncertainties are still large and their finding for an enhanced phosphorus abundance
suggests that the actual metallicity is several times larger than solar (Hamann 1998).
2.4. Narrow Absorption Lines
Narrow absorption lines (NALs, with profile widths less than a few hundred km s1 )
can form in a wide range of quasar environments, from outflows like the BALs to much
more distant clouds in the outer host galaxies. They can also form in cosmologically gas
that has no relation to the quasars whatsoever. Figure 3 shows a mixture of NALs in
one quasar spectrum. The outflow lines (labeled A-E) were identified by their variable
strengths, line profiles that are broad and smooth compared to the thermal speed and
doublet ratios in C iv and O vi that reveal partial covering of the background continuum
source (<0.01 pc across, Hamann et al. in prep.).
The theoretical steps in the abundance analysis are straightforward. They involve only
an ionization correction applied to the measured ionic column densities. Even without an
estimate of the ionization (e.g., if just one metal ion is measured) we can still derive useful
lower limits on the metal/hydrogen abundance ratios (Hamann & Ferland 1999). The
main difficulties are usually more practical. Line saturation and possibly inhomogeneous
column density distributions across the background emission source(s) can affect the
reliability and interpretation of the derived column densities. However, these concerns
can be addressed by analyzing (or fitting) several lines together and making reasonable
assumptions about the spatial distribution of the absorbing gas (Hamann & Sabra 2004,
Arav et al. 2005, Gabel et al. 2005).
The particular outflow systems shown in Figure 3 indicate a metallicity of roughly
Z 2 Z with large uncertainties owing to poor constraints on the H i column. More
reliable analyses of multiple outflow NALs in two other quasars (Arav et al. 2007, Gabel
176
Figure 4. A complex of 9 C iv NALs (labeled above the spectrum) at velocity shifts from
roughly 1400 to 6200 km s1 in a quasar at redshift 3.4 (from Simon & Hamann 2009).
et al. 2006) yield Z 29 Z . Simon & Hamann (2009 and in prep.) are conducting a
survey of low-velocity NALs to derive metallicities and other properties of these absorbers
in quasars at redshifts from 2 to 4.7. Figure 4 shows the diverse C iv systems measured
in one particular quasar. This study emphasizes high redshifts to examine any redshift
effects and obtain accurate H i column densities using ground-based measurements of
several lines in the H i Lyman series. A significant fraction of the observed NAL systems
(e.g., in Fig. 4) clearly form in a quasar outflow based on their broad profiles and evidence
for partial covering. The best-measured systems in our survey so far have Z 24 Z ,
with a few outliers (that might be unrelated to the quasars) having sub-solar Z.
Overall, the NAL abundances agree with other quasar data. The metallicities tend to
be a few times solar in the confirmed or suspected NAL outflow systems, and there is no
evidence for a trend with redshift (see Hamann et al. 2007 and refs. therein).
177
178
al. 2007). The entire process, from the onset of the merger/starburst to the end of the
quasar phase, can take <0.5 Gyr.
The absence of a trend in the observed quasar metallicities with redshift means that
this rapid evolutionary sequence occurs in approximately the same way at all redshifts.
The positive correlation between SMBH mass and metallicity near quasars (2.1), akin
to the well-known massmetallicity relation among galaxies, suggests further that mass
is a more important parameter in this evolutionary sequence than the time elapsed since
the big bang.
We are grateful to Daniel Capellupo for help with Figure 2 and to NASA for financial
support through the HST guest observer and Chandra theory programs.
References
Arav, N., et al. 2001, ApJ, 561, 118
Arav, N., Kaastra, J., Kriss, G. A., Korista, K. T., Gabel, J., & Proga, D. 2005, ApJ, 620, 665
Arav, N., et al. 2007, ApJ, 658, 829
Beelen, A., et al. 2006, ApJ, 642, 694
Cox, P., et al. 2005, ESA-SP, 577, 115
Dhanda, N., Baldwin, J. A., Bentz, M. C., & Osmer, P. S. 2007, ApJ, 658, 804
Di Matteo, T., Croft, R. A. C., Springel, V., & Hernquist, L. 2004, ApJ, 610, 80
Dietrich, M., et al. 2003, ApJ, 589, 722
Dietrich, M., & Hamann, F. 2008, RMxAC, 32, 65
Dong, X., Wang, J., Wang, T., Wang, H., Fan, X., Zhou, H., & Yuan, W. 2009, arXiv:0903.5020
Gabel, J. R., Arav, N., & Kim, T.-S. 2006, ApJ, 646, 742
Gabel, J. R., et al. 2005, ApJ, 623, 85
Granato, G. L., De Zotti, G., Silva, L., Bressan, A., & Danese, L. 2004, ApJ, 600, 580
Groves, B. A., Heckman, T. M., & Kauffmann, G. 2006, MNRAS, 371, 1559
Hamann, F., Warner, C., Dietrich, M., & Ferland, G. 2007, ASP-CS, 373, 653
Hamann, F., Dietrich, M., Sabra, B. M., & Warner, C. 2004, COAS, 440
Hamann, F., Korista, K. T., Ferland, G. J., Warner, C., & Baldwin, J. 2002, ApJ, 564, 592
Hamann, F., & Ferland, G. 1999, ARAA, 37, 487
Hamann, F. 1998, ApJ, 500, 798
Hamann, F., & Sabra, B. 2004, ASP-CS, 311, 203
Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651, 142
Hopkins, P. F., et al. 2005, ApJ, 630, 705
Jiang, L., et al. 2007, AJ, 134, 1150
Juarez, Y., et al. 2009, AA, 494, L25
Kauffmann, G., & Haehnelt, M. 2000, MNRAS, 311, 576
Li, Y., et al. 2007, ApJ, 665, 187
Nagao, T., Maiolino, R., & Marconi, A. 2006, AA, 447, 863
Nagao, T., Marconi, A., & Maiolino, R. 2006, AA, 447, 157
Netzer, H., et al. 2007, ApJ, 666, 806
Richards, G. T., et al. 2006, AJ, 131, 2766
Rupke, D. S. N., Veilleux, S., & Baker, A. J. 2008, ApJ, 674, 172
Sanders, D. B., Soifer, B. T., Elias, J. H., Neugebauer, G., & Matthews, K. 1988, ApJ, 328, L35
Silverman, J. D., et al. 2005, ApJ, 624, 630
Simon, L. E., & Hamann, F. 2009, submitted to MNRAS
Spolaor, M., Proctor, R. N., Forbes, D. A., & Couch, W. J. 2009, ApJ, 691, L138
Trager, S. C., Faber, S. M., Worthey, G., & Gonz
alez, J. J. 2000, AJ, 120, 165
Tremaine, S., et al. 2002, ApJ, 574, 740
Veilleux, S., et al. 2009, ApJS, 182, 628
Wang, R., et al. 2008, ApJ, 687, 848
Warner, C., Hamann, F., & Dietrich, M. 2003, ApJ, 596, 72
Graduate School of Science and Engineering, Ehime University, 2-5 Bunkyo-cho, Matsuyama
790-8577, Japan
email: kenta@cosmos.phys.sci.ehime-u.ac.jp
Abstract. We investigate the metallicity of the narrow line regions (NLRs) of high-z radio
galaxies (HzRGs), using new deep optical spectra of 9 HzRGs obtained with FORS2 on VLT
and data from the literature. To estimate the metallicity of NLRs we focus on the Civ/Heii
and Ciii]/Civ flux ratios. Based on comparison between the observed emission-line flux ratios
and the prediction of our photoionization model calculations, we find no significant metallicity
evolution in NLRs of HzRGs, up to z 4. We discuss the possibility that massive galaxies had
almost completed the major epoch of the star formation in the very high-z universe (z > 5).
Keywords. galaxies: active, galaxies: evolution, quasars: emission lines, quasars: general
1. Introduction
Metallicity of galaxies is one of the most important aspects to understand the formation
and evolution of galaxies, since it is closely related with the past star formation history
of galaxies. The most straightforward way to investigate chemical evolution of galaxies is
by measuring the metallicity of galaxies at various redshifts and exploring the systematic
trends in the metallicity as a function of redshift (see Maiolino et al. 2008 and references
therein).
An approach to study the metallicity of galaxies in high-z universe is to focus on
active galactic nuclei (AGNs). AGNs generally show various emission lines in rest-frame
ultraviolet to infrared wavelengths. Here we focus on the rest-frame ultraviolet lines,
since we can easily measure the emission-line fluxes of high-z AGNs by means of optical
spectroscopic observations. Now we focus on the narrow-line region (NLR) of high-z radio
galaxies (HzRGs) which is a good tracer of chemical properties in the spatial scale of
their host galaxies.
Some studies of the NLR metallicity of HzRGs have already been carried out by studying the emission-line flux ratios of Nv/Civ and Nv/Heii. However, the emission-ine flux
of Nv in HzRGs is generally too faint, especially for metal-poor gas. Nagao et al. (2006)
proposed a new metallicity diagnostic diagram with the Civ, Heii, and Ciii] emission
lines, all of which are moderately strong in the rest-frame UV spectra of HzRGs, even
at low metallicities. They studied the NLR metallicity of HzRGs and reported that the
observational data do not show any evidence of significant evolution in the gas metallicity
of NLRs within the redshift range 1.2 < z < 3.8. We note, however, that their sample
included only 5 objects at z > 2.7. Thus, observing more HzRGs at this redshift range
is crucial to assessing the possible metallicity evolution.
We present new spectroscopic observations of 9 HzRGs at 2.7 < z < 3.5. By combining
the new data with the Nagao et al. (2006) database, we discuss the chemical evolution
of HzRGs in the whole z 1 4 redshift range.
179
180
K. Matsuoka et al.
3. Discussion
By comparing the observed flux ratios of HzRGs with the prediction of photoionization
models, we investigate the gas metallicity. In Fig. 1, we plot the flux ratios of HzRGs
on the diagnostic diagram with the calculated model grids. Although 9 new data are at
higher redshift than the data in Nagao et al. (2006) on average, there is no systematic
difference in the data distribution between our new observations and the data of Nagao
et al. (2006). This result naively suggests that there is no significant chemical evolution in
NLRs of HzRGs, even up to z 4. If the minimum timescale for a significant enrichment
of carbon ( 0.5 Gyr) is take into account, the major epoch of the star formation in
HzRGs may have occurred at z > 5. If the host galaxies of HzRGs are very massive, this
scenario implies that massive galaxies had almost completed the major epoch of the star
formation at z > 5 (see Matsuoka et al. 2009 for more details).
Figure 1. The flux ratios of HzRGs and calculated model grids plotted on the diagnostic
diagram of Civ/Heii versus Ciii]/Civ. Our new data are shown with red-filled circles, while the
data compiled by Nagao et al. (2006) are shown with blue-open circles and arrows. Constant
metallicity and constant ionization parameter sequences are denoted by solid and dashed lines,
respectively. These models are calculated by adopting nH = 104 cm3 (a typical NLR density).
References
Ferland, G. J., Korista, K. T., Verner, D. A., et al. 1998, PASP, 110, 761
Maiolino, R., Nagao, T., Grazian, A., et al. 2008, A&A, 488, 463
Matsuoka, K., Nagao, T., Maiolino, R., et al. 2009, A&A, 503, 721
Nagao, T., Maiolino, R., & Marconi, A. 2006, A&A, 447, 863
1. Introduction
The evolution of the metal content of the gas in galaxies is a useful tool to disentangle
between different scenarios of disk galaxy evolution. In particular the study of the stellar
mass-metallicity relation (M-Z) can help to constrain the contribution of the several
processes taking place during a galaxy lifetime : Star-formation history, outflow of gas by
SN and stellar winds, and infall of gas by merger or secular accretion. There are several
studies of the M-Z relation at different redshifts but the value evolution of the relation is
still in debate. See e.g previous work: Local galaxies (Tremonti et al. (2004) ), 0 < z < 1
Savaglio et al. (2005), Rodrigues et al. (2008) and z > 1 Liu et al. (2008), Erb et al.
(2006), Maier et al. (2006) and Maiolino et al. (2008).
182
M. Rodrigues et al.
Figure 1. The metallicity shift from the local relation as function of the loockback time. The 4
high-z sample and the two low-redshift sample are plotted as black open squares, the 3 redfhift
bins of the IMAGES sample as black square and the median of the 3 bins as red triangle. The
evolution found by Savaglio et al. (2005) in the frame of a close box model is indicated with
dashed line.
metals at early epoch is then expected to be very high and the gas is rapidly enriched by
metals. Due to the high initial star formation rate, the fraction of gas diminishes rapidly
and the star formation rate decreases exponentially with time. The gas is then slowly
enriched in metals at intermediate redshift. However, the current observation shows a
strong evolution of the metal content from intermediate redshift to local Universe. This
evolution is incompatible with the close box model in which galaxies evolve secularly. To
explain the observed evolution an external supply of gas is required. The input of gas
powers the star formation and dilutes metals at the same time. Combining our results
with the reported evolution of the Tully-Fisher relation (Puech et al. (2008)), we do find
that such metal content evolution requires that 30% of the stellar mass of local galaxies
have been formed through an external supply of gas.
References
Erb, D.K. et al. 2006, ApJ, 644, 813E
Lara-Lopez, M.A. et al. 2009, arXiv0907.0450
Liang, Y.C., Hammer, F., and Flores, H. 2006, A&A,447,113L
Liu, X. et al. 2008, ApJ, 678, 758L
Maier, C. et al. 2006, ApJ 639, 858M
Maiolino, R. et al 2008, A&A, 488, 463M
Puech, M. et al. 2008, A&A, 484, 173P
Rodrigues, M et al. 2008, A&A, 492, 371R
Shapley, A.E. et al. 2004, ApJ, 612, 108S
Savaglio, S. et al. 2005, ApJ, 635, 260S
Tremonti, C., et al. 2004,ApJ 613, 898T
High redshift quasars are thought to represent an early stage of galaxy evolution,
in which models by e.g., Dimatteo et al. (2008) and Hopkins et al. (2008) predict that
major mergers trigger violent star formation and the rapid growth of a central super
massive black hole. However, the timing of the quasar phase during a galaxys evolution
is less well understood: quasar feedback could quench, trigger, or have little effect on star
formation in host galaxies. The gas-phase metal abundance works as a fossil record for
the star formation in the host galaxy before the visible quasar epoch. Abundances have
been probed in the near-quasar environment using broad emission lines (BELs) out to
redshifts z > 6 by e.g., Juarez et al. (2009), Nagao et al. (2006) and Dietrich et al. (2003),
who consistently find metallicities near or above the solar value. This result suggests that
there is always significant star formation before the quasar becomes visible.
We compare the metallicities in high-redshift quasars to the ongoing star formation
rates (SFRs) in their host galaxies using measurements of BEL line flux ratios and farinfrared (far-IR) luminosities to constrain the late stages of galaxy-quasar evolution and
the possible effects of quasar feedback on star formation. We measure BEL flux ratios,
N V(1240) / C IV(1550) and Si IV(1397)+O IV](1402) / C IV(1550), for three
composite rest-frame ultraviolet spectra consisting of a sample of 29 2.3 < z < 4.6 Sloan
183
184
Digital Sky Survey quasars grouped by far-IR luminosity, with < L60 > = 1046.90 , 1046.52
and 6 1046.21 erg s1 for the three composites, corresponding to SFRs of 6740, 2810 and
6 1360 M yr1 if the far-IR is powered by star formation. We convert the flux ratios into
metallicities using the theoretical relationship in which secondary enrichment processes
increase the N / C ratio as metallicity increases, and determine the average metallicity
for each composite (e.g. Shields (1976)).
We find uniformly high (5-10 times solar) metallicities for the three composite spectra. No significant variations in metal enrichment exist among the three LF IR groups. By
the time a quasar becomes visible, most of the gas enriching star formation is complete
and any ongoing star formation does not contribute significantly to the enrichment (See
Simon & Hamann (2009)).
We investigate the galaxy-black hole formation relationship from a different perspective using the largest ever survey (others include Savaglio et al. (1994), Wampler et al.
(1996)) of narrow associated absorption line (AALs, vwidth < 500 km s1 , forming within
12,000 km s1 of the quasar emission velocity) metallicities and physical properties in a
sample of high-redshift (z >3) quasars with Keck high resolution spectra in a range of
near-quasar environments, including quasar outflows and host galaxy halos. We measure
C IV, and H I optical depths, covering fractions, widths, column densities and obtain
good ionization constraints, and determine well-constrained metallicities for five AAL
systems in one z>3 and two z>4 quasars, and determine lower limits for metallicities for
AALs in four other z4 quasars.
Both BELs and AALs are composed of metal rich gas at all redshifts although metallicities derived from emission lines tend to be higher than those from absorption lines.
The AALs have a broader range in metallicity, from only 100ths solar to up to 10 times
solar. One z 3.5 quasar has AALs that are supersolar in 7 out of 9 systems (See Simon
& Hamann (2010)). Smaller velocity shifts and broader lines are more likely to be associated directly with the quasar as outflows, but we see no strong trend in metallicity
with velocity shift or line width. The AALs with covering fraction less than one are more
likely to be metal rich.
These super-solar metallicity results are consistent with previous observations and
with evolutionary models, in which the quasar host galaxies experience significant star
formation before the appearance of the visible quasar (Dimatteo et al. (2008), Hopkins
et al. (2008)). High SFRs (comparable to ULIRGs) in the host galaxies are not clearly
linked to younger or chemically less mature quasar environments.
References
Dietrich M., Hamann, F., Shields, J. C., Constantin, A., Heidt, J., J
ager, K., Vestergaard, M.,
& Wagner, S. J. 2003, ApJ, 589, 722
DiMatteo, T., Colberg, J., Springel, V., Hernquist, L., & Sijacki, D. 2008, ApJ, 676, 33
Hamann, F., Korista, K. T., Ferland, G. J., Warner, C., & Baldwin, J. 2002, ApJ, 564, 592
Hopkins, P. F., Hernquist, L., Cox, T. J., & Keres, D. 2008, ApJS, 175, 356
Juarez, Y., Maiolino, R., Mujica, R., Pedani, M., Marinoni, S., Nagao, T., Marconi, A., & Oliva,
E. 2009, A&A, 494, L25
Nagao, T., Marconi, A., & Maiolino, R. 2006, A&A, 447, 157
Savaglio, S., DOdorico, S., & Mller, P. 1994, A&A, 281, 331
Shields, G. A. 1976, ApJ, 204, 330
Simon, L. E., & Hamann, F. 2009, in prep.
Simon, L. E., & Hamann, F. 2010, in prep.
Wampler, E. J., Williger, G. M., Baldwin, J. A., Carswell, R. F., Hazard, C., & McMahon, R. G.
1996, A&A, 316, 33
Session IV
1. Introduction
Since the first stellar abundances were derived from model atmospheres many decades
ago, the question has been repeatedly asked whether such models have been accurate
enough for the purpose. Model atmospheres may be used for various purposes: to describe and further explore the physics of the atmospheres themselves, to reproduce the
stellar spectra and other observables quantitatively with few (if any) free parameters,
or to solve the inverse spectral-analysis problem, i.e. to deduce the stellar fundamental parameters like Tef f , logg, and the array of elemental abundances from the spectra,
supplemented with additional observational information like colours, parallaxes and angular diameters. The question to answer in this latter case is then: how accurate will the
abundance determinations be, in view of the lack of realism in the models? And, with a
specified and required accuracy: what steps need to be taken in constructing the models
in order to ascertain this accuracy? In this short review I shall thus neglect the fact that
other circumstances, that may often seem more trivial but are sometimes nevertheless
quite difficult to deal with in practice, like continuum definition in noisy or crowded spectra, blends, or uncertain gf -values, may be sources of error as important as the model
atmospheres.
Since the modeling situation for stellar atmospheres and their spectra is widely varying
along the spectral sequence, I shall divide the discussion into several parts, starting with
solar-type stars and then move towards the cool end of the HR diagram. I shall not cover
the early-type stars here, due to lack of space and expertise. Before entering into this
discussion, I shall however make a few bibliographic comments.
187
188
Bengt Gustafsson
189
190
Bengt Gustafsson
tions in the upper layers which lead to expansion cooling. This corresponds to layers
above the unstable hydrogen ionization zone in an 1D model, where the standard MLT
does not allow convection at all. Although the correct calculation of the radiative energy
transfer in these upper layers of the 3D models may still be a problem (only a relatively
small number of frequency points can be afforded), the results seem to suggest that very
considerable adjustments of standard abundances for such stars have to be made. For
elements based on spectral lines from molecules (like CH, NH and OH) these abundance
corrections downwards may amount to more than 1 dex (Asplund & Garca Perez 2001,
Collet, Asplund and Trampedach 2007, see also Behara et al. 2009, and Hernandez et al.
2008). Even for abundances derived from low-excitation atomic lines, the effects may be
very considerable (see Bonifacio, Caffau & Ludwig 2009, who find effects of as much as
0.8 dex in the Cu abundances for Pop II dwarfs).
Very considerable progress has also taken place in the latest decade in the calculation
of stellar spectra with the assumption of LTE relaxed, and a great number of different
elements have now been studied with detailed statistical-equlibrium calculations in the
formation of solar-type stellar spectra, not a least as a function of stellar metallicity (e.g.,
from the last few years for Li: Shi et al. 2007, Lind et al. 2009; N: Caffau et al. 2009;
O: Caffau et al. 2008, Fabbian et al. 2009; Na: Gehren et al. 2004, 2006, Liu et al. 2007,
Andrievsky et al. 2007; Mg: Gehren et al. 2004, 2006, Liu et al. 2007, Sundqvist et al.
2008; Al: Gehren et al. 2004, 2006, Liu et al. 2007, Andrievsky et al. 2008; Si: Shi et al.
2008, K: Zhang et al. 2006; Ca: Mashonkina et al. 2007; Sc: Zhang, Gehren & Zhao 2008;
Mn: Mergemann & Gehren 2008; Fe: Collet, Asplund & Thevenin, 2005; Co: Bergemann
2008; Sr: Short & Hauschildt 2006; Ba: Short & Hauschildt 2006, Andrievsky et al. 2009;
Nd: Mashonkina et al. 2005. For earlier studies, see Asplund 2005). As mentioned above,
an important problem in these SE studies is the shortage of accurate cross sections for
atomic inelastic collisions with electrons and hydrogen atoms (note, however, the point
made by Gehren et al., 2006, that this in not always very problematic). One must in
general be critical concerning the classical or semi-classical recipes that are often used
for the collision cross sections in the absence of more adequate quantum-mechanical data.
Also, the semi-empirical derivation of collision cross sections from solar (or stellar) data,
by requiring observed solar line strengths to fit the calculated ones, is risky, since problems
with the solar model or model atom may be hidden in this fit, preventing correct cross
sections from being derived. However, proper quantum-mecanical calculations are getting
possible (current work by Belyaev, Barklem and others) and this seems to be the way
to go, with laboratory checks for measurable transitions. An interesting aspect which is
also illustrated by the work by the latter authors is the need to care about details in the
model atoms of the statistical-equilibrium calculations. Thus, Barklem (2007) found that
the correct treatment of the electron-impact excitation of OI from the triplet 3s to the
singlet 3s state causes very significant corrections of the oxygen abundances, as derived
from the OI IR triplet lines, and Barklem et al. (2003) and Lind et al. (2009) showed that
the standard semiclassical collision rates for H+Li collisions are highly exaggerated and
tend to lead to underestimated Li abundances for Pop II stars, but also that the charge
transfer Li + H Li+ +H has even more severe effects in the converse direction on
the abundances. In abundance analysis at least, the devil is in the details.
Another key factor to worry about is the need for realistic UV fluxes in these calculations, not the least for the proper estimation of photo-ionization rates. Here, the different
atomic species cannot always be treated individually; e.g. the departures from LTE for Fe
(leading to over-ionization which increases the UV-flux of the model) may couple to the
statistical equilibrium of other elements like Sr and Ba (see Short & Hauschildt 2006).
As yet, very few model atmospheres for late-type stars have been calculated with the
191
LTE assumption relaxed. The pioneering NLTE solar models made by Anderson (1989)
have now been replaced by those of Short and Hauschild (2005) in which 24 different
elements were consistently treated in SE, with up to 6 different ionization stages each,
and thousands of individual transitions for each species. The resulting 1D solar model
structure is a few hundred K hotter in the surface layers than the corresponding LTE
model. Its UV flux is also significantly higher. These effects are to be expected, the first
one as a result of the loosening of the radiative transfer from the local temperature (more
lines formed in scattering processes) and the over-ionization of primarily iron. For the K
giant Arcturus, Short and Hauschildt (2003, 2009) find again a significantly higher UV
flux but a conversely cooler SE model as compared with the corresponding LTE model.
This latter result is not understood in detail, but one may speculate that it is due to
the surface CO cooling, which gets more dominating for the K giants if the metal-line
opacity is decoupled from the local gas.
How far have we come in joining the 3D approaches with the statistical equilibrium
(SE) treatment of excitation and ionization of the gas and of radiative transfer? Some
diagnostic work has been done so far, with calculation of solar and stellar spectra in SE
from a 3D model, the latter, however, constructed under the assumption of LTE. This
work includes studies of Li by Asplund, Carlsson & Botnen (2003) and O by Asplund et
al. (2004), of Na and Ca by Uitenbroek (2006), and, although not with a complete 3D
treatment of radiative transfer, of Fe and O by Shchukina, Trujillo Bueno & Asplund
(2005) and of Sr by Trujillo Bueno & Shchukina (2007). We also note that Hansteen et
al. (2004, 2007) and Leenaarts et al. (2007, 2009) have developed magneto-hydrodynamic
models of the upper solar atmosphere with descriptions of the radiative transfer in the
most important transitions in considerable detail. The number of atomic levels one can
afford in problems of this character is a severe restriction, although Carlsson (2008) has
argued that atoms with typically 102 levels should presently be possible to handle if
the most efficient methods are optimized. The already available results are, however,
quite interesting and clearly demonstrate the complexity of the situation. One example
is the results for Fe I and Fe II of Shchukina, Trujillo Bueno & Asplund (2005) which
indicate that the Fe abundances for a subdwarf (Tef f /logg/[Fe/H] = 5700K/3.7/ 2.5)
are underestimated if LTE is assumed relative to SE by about 0.5 dex (due to overionization) while it is overestimated by 0.3 dex if 1D MLT LTE models instead of 3D
hydrodynamical LTE models are used (which essentially is due to the surface expansion
cooling in the 3D model). Here, one would expect that the combination with SE+3D
would lead to an effect in between, but it turns out instead to be an almost as large
positive effect as for the pure SE case. A naive adding of the two separate effects
would lead to an underestimate of [Fe/H] by about 0.25 dex. Also for Fe II, the two
effects combine in a clearly non-linear way.
Fully consistent and realistic hydrodynamic 3D models in SE have still not been constructed, and it is unclear whether such a project can be undertaken with existing algorithms and computers. In taking this on it would be necessary to find an adequate
treatment of the radiative transfer, e.g. by reducing the number of atomic levels involved
and the points in the frequency spectrum to a small number of representative levels and
frequencies, respectively, but still producing a realistic radiative field. Important steps
in developing such methods were taken by Nordlund (1982) in developing his opacitybinning method, by Skartlien et al. (2000) in including scattering in such treatments,
and by Trampedach (unpubl.) in optimizing the choice of frequency points further. It
seems, however, that decisive and final steps towards physical consistency, entailing both
3D and SE for photospheres will have to wait for further computer development.
An important question is then what errors in abundances one may expect as a result of
192
Bengt Gustafsson
this lack of consistency. The gradual development of SE and 3D models in recent decades
has along the way generated a number of estimates of systematic errors due to the neglect
of each of these complications in standard models. Looking back at such estimates it is
fair to say that they were often off, sometimes severely exaggerated, but not seldom also
severely underestimated. This certainly reflects the complexity of the phenomena. While
in the standard radiative-equilibrium model atmosphere (most MLT models are also in
radiative equilibrium in the upper layers) the local temperature is simply set by the
balance between heating by absorption of radiation from the deeper atmospheric layers,
and cooling by emission from the local gas, in the hydrodynamic case the compression and
expansion heating/cooling are also decisive. These dynamical effects affect the capacity
of the gas to absorb and emit radiation. So, the coupling between hydrodynamics and
radiation gets very intricate. With all this in mind, I would still dare to conjecture that
the abundance errors caused by the neglect of coupling SE and 3D hydrodynamics in the
models may well amount to 0.1 dex for numerous chemical elements in solar-type stars.
193
in general looks very similar for a very wide range of stellar surface gravity (Freytag &
Ludwig, 2007). The characteristic size of convection elements, or granulae, scales as Hp
which means (as was already predicted by Schwarzschild, 1975) that the surface of the
super-giant stars will be covered with a few giant convection elements, although finer
structures may also occur. All this is clearly seen in the 3D hydrodynamic star-in-a-box
simulations by Freytag (2003). These impressive simulations show a good agreement
with the interferometric observations of Chiavassa et al. (2009, see also Kervella et al.
2009) but the radiative transfer in them, basically assuming gray opacities and a coarse
grid of points in the atmosphere, is still not treated in the detail needed for high-quality
abundance work.
For atmospheres of supergiants and AGB stars, the lower boundary condition of the
model atmosphere is obviously crucial for the whole model structure. These stars of
almost always pulsating, more or less regularly. For several decades, such atmospheres
have been modeled by 1D dynamic models, set in pulsation by a lower piston. Here,
the piston amplitude and frequency are free parameters. Such models were pioneered by
Wood (1979) and Bowen (1988) and later developed by Fleischer, Gauger and Sedlmayr
(1992), H
ofner & Dorfi (1997) and subsequently by the groups in Berlin, Vienna and
Uppsala. The models also have a free outer boundary condition, and a basic aim is to
study stellar mass-loss: the gas is elevated and expanded by the pulsations, such that dust
can form and, driven by radiative forces, drag the gas along. More recently, the models
have included non-gray opacities (Gautschy-Loidl et al. 2004), dust formation with a
two-phase treatment of gas and dust (Sandin & H
ofner 2004, and references therein,
Sandin 2008), and extensive grids of models have been calculated (Mattsson et al. 2009).
Freytag & H
ofner (2008) have recently added possibilities for dust to form in a 3D
simulation for a carbon AGB star by Freytags program. Indeed, plumes of dusty gas
are expelled from the model (similar to those later found by Kervella at al. 2009 for
Betelgeuse), and when following the dust development further out (using a 1D hydrodynamics code) the authors find final wind velocities characteristic of carbon stars and in
fair agreement with 1D models.
Finally, it should be noted that the effects of magnetic fields on super-giant atmospheres were explored by MHD star-in-box simulations by Dorch (2004). He found
that local dynamos driven by the giant convection motions generated considerable
large-scale magnetic fields of up to 500 Gauss and that this field at the densities characteristic of the tenuous photospheres of these stars could have considerable effects on
their structures.
It is still premature to give figures on how these various improvements, and uncertainties, in the modeling of M and C stars will affect the abundances derived from their
photospheric spectra. No doubt, the progress basically demonstrated for the solar-type
stars as regards convection and departures from LTE will apply also in this case, although
as regards the statistical-equilibrium calculations the demands for physical data are different since molecular transitions are more at focus. Also, at least for the super-giants,
the convective dynamics will be more violent and partly supersonic. As compared with
the solar-type stars, in a sense more elementary demands (complete opacities) and
more advanced physics (dynamic lower boundary conditions, dust formation, magnetic fields) are also important. Another aggravating circumstance in the cool-star case
is that there is no nearby star like the Sun to compare the models directly with. Previous
studies of uncertainties from standard models tend to suggest abundance errors ranging
from 0.15 - 0.40 dex for M and C stars. Since the coupling between the various unknown
properties of the phenomena at play is so intricate, it would not be very astonishing if
errors of this order of magnitude will still remain after the implementation of 3D models,
194
Bengt Gustafsson
195
References
Anderson, L. S. 1989, ApJ, 339, 558
Andrievsky, S. M., Spite, M., Korotin, S. A., et al. 2007, A&A, 464, 1081
Andrievsky, S. M., Spite, M., Korotin, S. A., et al. 2008, A&A, 481,481
Andrievsky, S. M., Spite, M., Korotin, S. A., et al. 2009, A&A, 494, 1083
Asplund, M. 2005, ARA&A, 43, 481
Asplund, M., & Garca Perez 2001, A&A, 372, 601
Asplund, M., Carlsson, M., & Botnen, A. V. 2003, A&A, 399, L31
Asplund, M., Grevesse, N., Sauval. A. J., et al. 2004, A&A, 417, 751
Barklem, P., Belyaev, A. K., & Asplund, M. 2003, A&A, 409, L1
Barklem, P. 2007, A&A, 462, 781
Behara, N. T., Ludwig, H. -G., Bonifacio, P., et al. 2009, textitarXiv, 0909.1010
Bergemann, M. 2008, Physica Scripta, 133, 14013
Bergemann, M., & Gehren, T. 2008, A&A, 492, 823
Bonifacio, P., Caffau, E., & Ludwig, H. -G. 2009, Mem. S.A.It, 75, 282
Bowen, G. H. 1988, ApJ, 329, 299
Caffau, E., Ludwig, H. -G., Steffen, M., et al. 2008, A&A, 488, 1031
Caffau, E., Maiorca, E., Bonifacio, P., et al. 2009, A&A, 498, 877
Carlsson, M. 2008, Physica Scripta, 133, 4012
Chabrier, G., Baraffe, I., Allard, F., & Hauschildt, P. H. 2005, arXiv 0509798
Collet, R., Asplund, M., & Thevenin, F. 2005, A&A, 442, 643
Collet, R., Asplund, M., & Trampedach, R. 2007, A&A, 469, 687
Chiavassa, A., Plez, B., Josselin, E., & Freytag, B. 2009, A&A, 506, 1351
Dorch, S. B. F. 2004, A&A, 423, 1101
Fabbian, D., Asplund, M., Barklem, P., et al. 2009, A&A, 500, 1143
Fleischer, A. J., Gauger, A., & Sedlmayr, E. 1992, A&A, 266, 321
Freytag, B. 2003, ANS, 324, 67
Freytag, B. 2008, EAS Publ. Ser., 28, 9
Freytag, B., & Ludwig, H. -G. 2007, SF2A-2007: Proceedings of the Annual meeting of the French
Society of Astronomy and Astrophysics, eds. J Bouvier, A. Chalabaev, and C. Charbonnel,
481
Freytag, B., & H
ofner, S. 2008, A&A, 483, 571
Gautschy-Loidl, R., H
ofner, S., Jrgensen, U. G., & Hron, J. 2004, A&A, 422, 289
Gehren, T., Liang, Y. C., Shi, J. R., et al. 2004, A&A, 413, 1045
Gehren, T., Shi, J. R., Zhang, H. W., et al. 2006, A&A, 451, 1065
Hansteen, V. 2004, IAU Symp, 223, 385
Hansteen, V. H., de Pontieu, B., Carlsson, M., et al. 2007, PASJ, 59, 699
Hern
andez, J. I. G., Bonifacio, P., Ludwig, H. -G., et al. , 2008, AIP Conf. Proc., 990, 175
H
ofner, S., & Dorfi, E. A. 1997, A&A, 319, 648
Kervella, P., Verhoelst, T., & Ridgway, S. T. 2009, A&A, 504, 115
Korn, A., Grundahl, F., Richard, O. et al. 2006, Nature, 442, 657
Kucinskas, A., Dobrovolskas, V., Ivanauskas, A. et al. 2009, this symposium
Leenaarts, J., Carlsson, M., Hansteen, V., et al. , 2007, A&A, 473, 625
Leenaarts, J., Carlsson, M., Hansteen, V., et al. , 2009, ApJ, 694, L128
Lind, K., Asplund, M., & Barklem, P. 2009, A&A, 503, 541
Liu, Y. J., Zhao, G., Shi, J. R, et al. 2007, MNRAS, 382, 553
Ludwig, H. -G., Allard, F., & Hauschildt, P. H. 2002, A&A, 395, 99
Ludwig, H. -G., Allard, F., & Hauschildt, P. H. 2006, A&A, 459, 599
Ludwig, H. -G., & Kucinskas, A. 2005, ESA-SP, 560, 319
Ludwig, H. -G., & Steffen, M. 2008, in Precision Spectroscopy in Astrophysics, Proc.
ESO/Lisbon/Aveiro Conf., eds. N. C. Santos et al., p. 133
Mashonkina, L, Ryabchikova, T, & Ryabtsev, A. 2005, A&A, 441, 309
Mattsson, Wahlin, R., & H
ofner, S. 2009, arXiv, 0909.1513
Melendez, J., Asplund, M., Gustafsson, B., & Yong, D. 2009, ApJL, 704, L66
196
Bengt Gustafsson
Nordlund,
A 1982, A&A, 107, 1
Nordlund,
A, Stein, R. F., & Asplund, M. 2009, Living Rev. Solar Phys., 6, No. 2
Pereira, T., Asplund, M., & Kiselman, D. 2009, arXiv 0909.4121
Sandin, C., & H
ofner, S. 2004, A&A, 413, 789
Sandin, C. 2008, MNRAS, 385, 215
Schwarzschild, M. 1975, ApJ, 195, 137
Shchukina, N. G., Trujillo Bueno, J, & Asplund, M 2005, ApJ, 618, 939
Shi, J. R., Gehren, T., Zhang, H. W., et al. 2007, A&A, 465, 587
Shi, J. R., Gehren, T., Butler, K., et al. 2008, A&A, 486, 303
Short, & C. I., Hauschildt, P. H. 2003, ApJ, 596, 501
Short, & C. I., Hauschildt, P. H. 2005, ApJ, 618, 926
Short, & C. I., Hauschildt, P. H. 2006, ApJ, 641, 494
Short, & C. I., Hauschildt, P. H. 2009, ApJ, 691, 1634
Skartlien, R. 2000, ApJ, 536, 465
Sundqvist, J. O., Ryde, N., Harper, G. M., et al. 2008, A&A, 486, 985
Trujillo Bueno, J., & Shchukina, N. 2007, ApJ, 664, L135
Uitenbroek, H. 2006, ApJ, 639, 516
Wood, P. R. 1979, ApJ, 227, 220
Zhang, H. W., Gehren, T., Butler, K., et al. 2006, A&A, 457, 645
Zhang, H. W., Gehren, T., & Zhao, G. 2008, A&A, 481, 489
Abstract. Non-local thermodynamic equilibrium (NLTE) line formation for neutral and singlyionized iron is considered through a range of stellar parameters characteristic of cool stars. A
comprehensive model atom for Fe I and Fe II is presented. Our NLTE calculations support the
earlier conclusions that the statistical equilibrium (SE) of Fe I shows an underpopulation of Fe I
terms. However, the inclusion of the predicted high-excitation levels of Fe I in our model atom
leads to a substantial decrease in the departures from LTE. As a test and first application of the
Fe I/II model atom, iron abundances are determined for the Sun and four selected stars with
well determined stellar parameters and high-quality observed spectra. Within the error bars,
lines of Fe I and Fe II give consistent abundances for the Sun and two metal-poor stars when
inelastic collisions with hydrogen atoms are taken into account in the SE calculations. For the
close-to-solar metallicity stars Procyon and Vir, the difference (Fe II - Fe I) is about 0.1 dex
independent of the line formation model, either NLTE or LTE. We evaluate the influence of
departures from LTE on Fe abundance and surface gravity determination for cool stars.
Keywords. atomic data, line: formation, stars: atmospheres
1. Introduction
Iron plays an outstanding role in studies of cool stars thanks to quite numerous lines in
the visible spectrum, which are easy to detect even in very metal-poor stars. Iron serves
as a reference element for all astronomical research related to stellar nucleosynthesis and
chemical evolution of the Galaxy. Iron lines are used to determine the surface gravity,
log g, and the microturbulence of stellar atmospheres. In the atmosphere with Teff >
4500 K, neutral iron is a minority species. The ionization equilibrium between Fe I
and Fe II and the excitation equilibrium of Fe I easily deviate from thermodynamic
equilibrium. Since the beginning of the 1970s a number of studies attacked the problem
of non-local thermodynamic equilibrium (NLTE) for Fe (e.g., Athay & Lites (1972),
Thevenin & Idiart (1999), Gehren et al. (2001)). However, a consensus on the expected
magnitude of the NLTE effects was not reached.
In this study, we update the model atom of Fe I-II treated by Gehren et al. (2001)
197
198
L. Mashonkina et al.
(hereafter Paper I) and apply it to analysis of the Fe spectrum in the Sun and selected
cool stars with the aim of empirically constraining the role of inelastic collisions with
hydrogen atoms in the SE of Fe I-II.
3. Results
The coupled radiative transfer and statistical equilibrium equations are solved with
an improved version of the DETAIL program (Butler & Giddings (1985)) based on the
accelerated lambda iteration. All calculations are performed with plane-parallel, homogeneous, LTE, and blanketed model atmospheres computed with the MAFAGS-OS code
(Grupp et al. (2009)).
For comparison with observed data, a total of 43 lines of Fe I and 18 lines of Fe II
were chosen. For the Sun and HD 84937, the analysis was extended to a larger line list
including 271 lines of Fe I and 34 lines of Fe II. The Sun is also used as a reference
star for a line-by-line differential analysis of stellar spectra. Solar flux observations were
taken from the Kitt Peak Solar Atlas (Kurucz et al. (1984)). The absolute solar iron
abundances were determined using gf -values from OBrian et al. (1991) and Melendez
& Barbuy (2009) for Fe I and Fe II, respectively. We find that virtually all models of
line formation, whether LTE or NLTE with SH > 0.1, lead to acceptable solar ionization
equilibria within their 1 error bars. To show the maximal NLTE effect on abundance
199
Table 1. Stellar parameters and iron abundances obtained for selected stars
HD Teff
Sun
10700
61421
84937
102870
5777
5377
6510
6350
6060
log g
4.44
4.53
3.960.02
4.000.12
4.110.01
Vmic ,
km s1
0.9
0.8
1.8
1.7
1.4
[Fe/H]I
NLTE0
LTE
7.630.08 7.490.10
0.430.04 0.490.02
0.100.06 0.140.05
1.940.06 2.160.07
0.040.03 0.040.03
[Fe/H]II
NLTE0
LTE
7.440.06 7.450.06
0.530.05 0.520.05
0.040.03 0.040.03
2.080.04 2.110.04
0.130.04 0.120.04
determination, Table 1 presents the average abundances for both ionization stages derived
from the NLTE with SH = 0 (denoted as NLTE0 ) and LTE calculations.
Four stars with effective temperature and surface gravity measured from the modelindependent methods were chosen to investigate the ionization equilibrium between Fe I
and Fe II for various SH values. They are listed in Table 1 together with the Teff and
log g values taken from Di Folco et al. (2004) for HD 10700 ( Cet), Allende Prieto et al.
(2002) for HD 61421 (Procyon), Korn et al. (2003) for HD 84937, North et al. (2009) for
HD 102870 ( Vir). Observational data were obtained with the FOCES spectrograph at
the 2.2m telescope of the Calar Alto Observatory during a number of observing runs
between 1997 and 2005, with a spectral resolution of R 60 000 and a signal-to-noise
ratio S/N > 200.
The NLTE, SH = 0 and LTE abundances obtained from the lines of Fe I (denoted
as [Fe/H]I ) and Fe II ([Fe/H]II ) are presented in Table 1. It is worth noting that, with
the updated model atom of Fe I-II, the departures from LTE are substantially smaller
compared to those from the previous studies. For example, with SH = 0, we obtain
an average NLTE abundance correction NLTE = log NLTE log LTE = 0.22 dex for
the Fe I lines in HD 84937, while the corresponding value amounts 0.40 dex in Korn
et al. (2003). Figure 1 displays the abundance difference between Fe I and Fe II for
various assumptions for the hydrogen collisions. We find that NLTE with pure electronic
collisions (SH = 0) is not acceptable for HD 84937 and Cet. This indicates the need
Figure 1. The difference in abundance between Fe I and Fe II in selected stars from calculations
with various line formation models. For each star, the error bars is indicated in the upper part
of panel.
200
L. Mashonkina et al.
for thermalizing processes not involving electrons in the atmosphere of metal-poor stars.
For each object, the NLTE effect on abundance determination is small (within the error
bars) when hydrogen collisions are included with SH > 1. For Procyon and Vir, the
mean Fe abundance from Fe I lines is about 0.1 dex lower compared to that from Fe II
lines. The origin of such a discrepancy will be investigated in a forthcoming paper. For
the Fe I/Fe II ionization equilibrium in two metal-poor stars, LTE seems to be as good
as NLTE with SH > 1.
The NLTE calculations were performed with SH = 1 for the small grid of model
atmospheres with Teff = 5000 and 6000 K, [M/H] = 1 and 3, and log g ranging
between 2 and 4 in order to inspect the departures from LTE depending on stellar
parameters. Negligible NLTE effects were obtained for Fe II. Fe I is subject to significant
NLTE effects for low gravity (log g < 3) and very metal-poor models. An important
consequence is that surface gravities of giants and very metal-poor stars derived by LTE
analysis are in error with a magnitude strongly depending on log g/[Fe/H]. For example,
LTE leads to a 0.26 dex lower gravity for Teff = 5000 K, log g = 2, and [Fe/H] = 3.
4. Conclusions
Completeness of model atom for Fe I is important for a correct calculation of the
Fe I/Fe II ionization equilibrium in the atmosphere of cool stars.
Thermalizing processes not involving electron collisions have to be included in the
SE calculations for Fe I-II. Collisions with hydrogen atoms could be good candidates for
such processes.
Fe I is affected by significant NLTE effects for giants and very metal-poor stars.
Only minor departures from LTE are obtained for Fe II.
Acknowledgements. L.M. acknowledges a partial support from the International Astronomical Union, the Russian Foundation for Basic Research (08-02-92203-GFEN), and
the Russian Federal Agency on Science and Innovation (02.740.11.0247) of the participation at the IAU XXVII General Assembly. This study is supported by the Deutsche
Forschungsgemeinschaft (GE 490/34.1). A.K. acknowledges support by the Swedish Research Council (VR).
References
Allende Prieto, C., Asplund, M., Garsia Lopez, R.J. & Lambert, D. 2002, ApJ, 567, 544
Athay, R.G. & Lites, B.W. 1972, ApJ, 176, 809
Bautista, M.A. 1997, A&AS, 122, 167
Butler, K. & Giddings, J. 1985, Newsletter on the analysis of astronomical spectra, No. 9, University of London
Di Folco, E., Thvenin, F., Kervella, P,; et al. 2004, A&A, 426, 601
Gehren, T., Butler, K., Mashonkina, L., Reetz, J., & Shi, J. 2001, A&A, 366, 981 (Paper I)
Grupp, F., Kurucz, R. L., & Tan, K. 2009, A&A, 503, 177
Korn, A., Shi, J. & Gehren, T. 2003 A&A, 407, 691
Kurucz, R. 2007, http://kurucz.harvard.edu
Kurucz, R.L., Furenlid, I., Brault, J. & Testerman, L. 1984, Solar Flux Atlas from 296 to 1300
nm. Nat. Solar Obs., Sunspot, New Mexico
Melendez, J. & Barbuy, B. 2009, A&A, 497, 611
Nave, G., Johansson, S., Learner, R.C.M., Thorne, A.P., & Brault, J.W. 1994, ApJS, 94, 221
North, J. R., Davis, J., Robertson, J. G., et al. 2009, MNRAS, 393, 245
OBrian, T.R., Wickliffe, M.E., Lawler, J.E., et al. 1991, J. Opt. Soc. Am. B, 8, 1185
Steenbock, W. & Holweger, H. 1984, A&A, 130, 319
Thevenin, F. & Idiart, T.P. 1999, ApJ, 521, 753
Abstract. We present solar photospheric abundances for 12 elements from optical and nearinfrared spectroscopy. The abundance analysis was conducted employing 3D hydrodynamical
(CO5 BOLD) as well as standard 1D hydrostatic model atmospheres. We compare our results
to others with emphasis on discrepancies and still lingering problems, in particular exemplified
by the pivotal abundance of oxygen. We argue that the thermal structure of the lower solar
photosphere is very well represented by our 3D model. We obtain an excellent match of the
observed center-to-limb variation of the line-blanketed continuum intensity, also at wavelengths
shortward of the Balmer jump..
1. Motivation
In recent years several solar abundances experienced a significant downward revision,
among them major contributors to the overall solar metalicity (Asplund et al. 2005). In
part, the downward revision was attributed to the application of 3D model atmospheres.
Due to the importance of the solar composition as a fundamental yardstick in astronomy, the CIFIST Team and its collaborators started an independent investigation of
the solar abundances applying its self-developed analysis tools, in particular its own 3D
model atmosphere code dubbed CO5 BOLD (Freytag et al. 2002, Wedemeyer et al 2004).
Table 1 summarizes the result for 12 elements in comparison to other works. Considering
the latest compilation of Asplund and collaborators one can note a convergence towards
a unique abundance set. However, there are still sizable differences present, in particular concerning the abundant element oxygen. Formally, the overlapping error bars could
be taken to basically signal correspondence. However, one must keep in mind that certain systematics (observed spectra, oscillator strength, analysis methodology) are shared
among all groups, and from that perspective differences are still on a rather high level. In
this contribution we want to comment on a few of the lingering problems when it comes
to the spectroscopic determination of solar abundances.
201
202
Table 1. Abundances derived by the CO5 BOLD group in comparison to other compilations:
AG89 Anders & Grevesse (1989); GS98 Grevesse & Sauval (1998); AGS05 Asplund et al. (2005);
AGSS09 Asplund et al. (2009). El denotes the element, N the number of spectral lines used in
our analysis. The last two rows give the total mass fraction of metals, and metals relative to
hydrogen. Values set in italics refer to meteoritic abundances.
El
GS98
AGS05
AGSS09
Li
C
N
O
P
S
Eu
Hf
Th
K
Fe
Os
1
43
12
10
5
9
5
4
1
6
15
3
Z
Z/X
1.03 0.03
8.50 0.06
7.86 0.12
8.76 0.07
5.46 0.04
7.16 0.05
0.52 0.03
0.87 0.04
0.08 0.03
5.11 0.09
7.52 0.06
1.36 0.19
1.16 0.10
8.56 0.04
8.05 0.04
8.93 0.035
5.45 0.04
7.21 0.06
0.51 0.08
0.88 0.08
0.12 0.06
5.12 0.13
7.67 0.03
1.45 0.10
1.10 0.10
8.52 0.06
7.92 0.06
8.83 0.06
5.45 0.04
7.33 0.11
0.51 0.08
0.88 0.08
0 .09 0 .02
5.12 0.13
7.50 0.05
1.45 0.10
1.05 0.10
8.39 0.05
7.78 0.06
8.66 0.05
5.36 0.04
7.14 0.05
0.52 0.06
0.88 0.08
0 .06 0 .05
5.08 0.07
7.45 0.05
1.45 0.10
1.05 0.10
8.43 0.05
7.83 0.05
8.69 0.05
5.41 0.03
7.12 0.03
0.52 0.04
0.85 0.04
0.02 0.10
5.03 0.09
7.50 0.04
1.25 0.07
0.0153
0.0209
0.0189
0.0267
0.0171
0.0234
0.0122
0.0165
0.0134
0.0183
accurate are our measurements of the lines equivalent width? The ever-lasting problem
of the continuum placement constitutes a difficult to overcome limit to the achievable
precision. Line shapes fitted to observations can mitigate but not eliminate this precision
bottleneck. iii) How good are our model atmospheres, in particular 3D models? There
have been long-lasting arguments about insufficient spatial resolution, and wavelength
resolution when representing the energy exchange between gas and radiation field. iv)
How great are the departures from local thermodynamic equilibrium? In particular, the
poorly constraint efficiency of collisions with neutral hydrogen atoms in the calculation
of the statistical equilibrium established a limit to which one can determine abundances
from some lines. Prominent examples are the infrared triplet lines of neutral oxygen. v)
Which solar spectrum is the solar spectrum? There are surprising differences among high
quality solar atlases which need to be better understood or even better overcome by a
newer generation of atlases.
3. 3D model properties
In this section we want to demonstrate that 3D models atmosphere have reached a
high level of realism when it comes to the thermal structure of the lower photosphere
including temperature inhomogeneities due to granulation. Figure 1 illustrates the
exceptional performance of our standard solar CO5 BOLD model representing the centerto-limb variation of the solar radiation field on a spatial scale where granulation is not
resolved. The calculation was done for a time series of 19 snapshots of the 3D flow
field, whose intensity pattern was subsequently horizontally and temporally averaged.
In the spectral synthesis calculations line blocking was accounted for by applying an
ATLAS (Kurucz 2005) Opacity Distribution Function with 1200 wavelength intervals,
and 12 sub-intervals each. Fig. 1 shows the emergent intensity averaged over the 12
sub-bins. The same calculation was repeated for the 1D semi-empirical Holweger-M
uller
atmosphere (Holweger & M
uller 1974, HM). The overall match to the observations by
the 3D model is remarkable, including the wavelength range in the Balmer continuum
203
Figure 1. Center-to-limb variation with line blocking using ATLAS ODFs: for four
heliocentric angles the intensity relative to disk-center is depicted as a function of wavelength.
suffering from heavy line blocking. The precision is challenging the available observations
and the performance of the HM model which was at least in part constructed to match
the solar center-to-limb variation.
204
6.5
6.4
A(Ni)
6.3
5.7
4.8
5.2
4.3
Reetz99(HM)
Reetz99(GRS88)
Caffau08(3D,HM)
6.2
6.1
Centeno08
Ayres08
Caffau09(HM)
Allende01
Melendez04
6.0
Caffau09(3D)
flux
intensity
5.9
8.55
8.60
8.65
8.70
A(O)
8.75
8.80
8.85
Figure 2. Oxygen and nickel abundances obtained by various groups from the 630 nm feature.
flux refers to disk-integrated, intensity to disk-center spectra. The solid and dashed curves
delineate the relation between O and Ni abundance at fixed total equivalent width of the feature
(labels in m
A). The grey bar indicates the currently accepted range of the Ni abundance from
other Ni lines. Further details see text.
the two components is far from unique. Here, the recent 3D based result of Caffau et al.
(2009) indicates a particularly low Ni abundance. While it is difficult to reconcile with the
presently accepted Ni abundance, it provides a striking illustration of the still lingering
problems in the determination of solar photospheric abundances from spectroscopy.
Acknowledgements
HGL, EC, BF, and PB acknowledge support from EU contract MEXT-CT-2004-014265.
References
Allende Prieto, C., Lambert, D.L., Asplund, M. 2001, A&A, 556, 63
Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
Asplund, M., Grevesse, N., & Sauval, A. J. 2005, ASP Conf. Ser, 336, 25
Asplund, M., Grevesse, N., Sauval, J., & Scott, P. 2009, ARAA, 47, 481
Ayres, T. R., 2008, ApJ, 686, 731
Caffau, E., Ludwig, H.-G., Steffen, M., Ayres, T. R., Bonifacio, P., Cayrel, R., Freytag, B., &
Plez, B. 2008, A&A, 488, 1031
Caffau, E., Ludwig, H.-G., Steffen, M., Livingston, W., Bonifacio, P., & Cayrel, R. 2009, A&A,
submitted
Centeno, R., & Socas-Navarro, H. 2008, ApJ, 682, L61
Freytag, B., Steffen, M., & Dorch, B. 2002, AN, 323, 213
Grevesse, N., & Sauval, A. J. 1998, Space Science Reviews, 85, 161
Holweger, H., & M
uller, E. A. 1974, Solar Physics, 39, 19
Kurucz, R. L. 2005, MSAIS, 8, 14
Melendez, J. 2004, ApJ, 615, 1042
Neckel H., & Labs D. 1994, Solar Physics, 153, 91
Reetz, J. 1999, PhD thesis, Ludwig-Maximilians University, Munich
Wedemeyer, S., Freytag, B., Steffen, M., Ludwig, H.-G., & Holweger, H. 2004, A&A, 414, 1121
1. Introduction
The classical theory of stellar evolution predicts a very simple Galactic destiny to 3 He,
dominated by a large production of this isotope by low-mass stars (Iben 1967; Rood
1972; Rood et al. 1976; Dearborn et al. 1996; Weiss et al. 1996). As a consequence, one
expects a large increase of 3 He with time in the Galaxy with respect to its primordial
abundance (e.g., Tosi 1996). However, the 3 He content of Galactic HII regions (Balser et
al. 1994, 1999a; Bania et al. 1997, 2002) is similar to that of the Sun and the solar system
(Geiss & Reeves 1972; Geiss 1993, Mahaffy et al. 1998), and very close to the BBN value
(Coc et al. 2004; Cyburt 2004; Serpico et al. 2004). This is the so-called 3 He problem
that could be resolved if only 10 % or less of the low-mass stars were releasing 3 He
as predicted by classical stellar theory (Tosi 1998, 2000; Palla et al. 2000; Romano et
al. 2003).
In 2007, Charbonnel & Zahn showed that thermohaline mixing drastically reduces
the 3 He production in low-mass, low-metallicity stars. Simultaneously, this mechanism
changes the surface carbon isotopic ratio as well as the abundances of lithium, carbon
and nitrogen.
206
Figure 1. Li and Be abundances in IC 4651 main sequence and turnoff stars (black points
and triangles for actual determinations and upper limits respectively). Open circles, squares,
and triangles show model predictions for initial rotation velocities of 50, 80, and 110 km s1
respectively. On the cool side of the Li and Be dip the model with Teff6250 K is from Talon
& Charbonnel (2005) and takes into account additional transport of angular momentum by
internal gravity waves. Figures from Smiljanic et al. (2009)
accounted for with the complete formalism developed by Zahn(1992) and Maeder & Zahn
(1998) that takes into account advection by meridional circulation and diffusion by shear
turbulence (see Palacios et al. 2003, 2006 and Decressin et al. 2009 for a description of
the implementation in STAREVOL). Typical initial (i.e., ZAMS) surface rotation velocities are chosen for all the models depending on the stellar mass. We assume magnetic
braking on the early main sequence for the stars with Teff on the ZAMS lower than
6900 K that have relatively thick convective envelopes. The adopted braking law follows
the description of Kawaler (1988). Rotational velocity further decreases when the stars
evolve on the subgiant branch due to radius expansion. Atomic diffusion is included in
the form of gravitational settling as well as that related to thermal gradients, using the
formulation of Paquette et al. (1986).
207
Figure 2. Evolution of the surface 12 C/13 C value as a function of stellar luminosity for the 1.25
and 2 M models (left and right respectively). Different tracks correspond to different initial
rotation velocities (50, 80, and 110 km.s1 for the 1.25 M , and 0, 110, and 250 km.s1 for the
2 M ). Observations for M67 stars by Gilroy & Brown (1991) are also shown (triangle, squares,
and circles for subgiant, RGB, and clump stars respectively; turnoff mass of M67 1.2 M ).
Figures from Lagarde & Charbonnel (in preparation)
208
Figure 3. Evolution of the surface abundance of 3 He (in mass fraction) for solar metallicity
stars of various initial masses. Figure from Lagarde & Charbonnel (in preparation)
References
Balser, D.A., Bania, T.M., Brockway, C.J., Rood, R.T., Wilson, T.L., 1994, ApJ, 430, 667
Balser, D.A., Bania, T.M., Rood, R.T., Wilson, T.L., 1999a, ApJ, 510, 759
Bania, T.M., Balser, D.A., Rood, R.T., Wilson, T.L., Wilson, T.J., 1997, ApJS, 113, 353
Bania, T.M., Rood, R.T., Balser, D.A., 2002, Nature, 415, 54
Charbonnel, C. 1995, ApJ, 453, L41
Charbonnel, C. 2002, Nature, 415, 27
Charbonnel, C. & Zahn, J.P. 2007, A&A Letters, 467, L15
Coc, A., Vangioni-Flam, E., Descouvemont, et al. 2004, ApJ, 600, 544
Cyburt, R.H. 2004, Phys.Rev.D, 70, 023 505
Dearborn, D.S.P., Steigman, G., Tosi, M., 1996, ApJ, 465, 887
Decressin, T., Mathis, S., Palacios, A., et al. 2009, A&A, 495, 271
Eggleton, P.P., Dearborn, D.S.P., Lattanzio, J.C 2006 Science, 314, 5805, 1580
Geiss, J. & Reeves, H., 1972, A&A 18, 126
Geiss, J., 1993, in Origin and evolution of the elements, eds. N.Prantzos et al., p.89
Gilroy, K.K., & Brown, J.A. 1991, ApJ, 371, 578
Iben, I., 1967, ApJ, 143, 642
Kawaler, S.D., 1988, ApJ, 333, 236
Krishnamurti, R. 2003, J. Fluid Mech., 483, 287
Maeder, A., & Zahn, J.P. 1998, A&A, 334, 1000
Mahaffy, P.R., Donahue, T.M., Atreya, S.K.,et al., 1998, Space Sci. Rev., 84, 251
Palacios, A., Charbonnel, C., Talon, S., & Forestini, M. 2003, A&A, 399, 603
Palacios, A., Charbonnel, C., Talon, S., & Siess, L. 2006, A&A, 453, 261
Palla, F., Bachiller, R., Stanghellini, L., Tosi, M., Galli, D., 2000, A&A, 355, 69
Paquette, C., Pelletier, C., Fontaine, G., & Michaud, G., 1986, ApJS, 61,177
Rood, R.T., 1972, ApJ, 177, 681
Rood, R.T., Steigman, G., Tinsley, B.M., 1976, ApJ, 207, L57
Rood, R.T., Bania, T.W., Wilson, T.L. 1984, ApJ, 280, 629
Romano, D., Tosi, M., Matteucci, F., Chiappini, C. 2003, MNRAS, 346, 295
Serpico, P.D., Esposito, S., Iocco, F., et al. 2004, JCAP, 12, 10S
Smiljanic, R., Pasquini, L., Charbonnel, C., Lagarde, N. 2009, A&A, in press, astro-ph 0910.4399
Talon, S. & Charbonnel, C 2005, A&A, 440, 981
Tosi, M., 1996, ASP Conference Series, Vol. 98, 299
Tosi, M., 1998, Space Science Reviews, Vol.84, 207
Tosi, M., 2000, IAUS 198, 525
Weiss, A., Wagenhuber, J., Denissenkov, P.A., 1996, A&A, 313, 581
Zahn, J.P. 1992, A&A, 265, 115
Institute of Theoretical Physics and Astronomy, Gostauto 12, Vilnius LT-01108, Lithuania
email: ak@itpa.lt
2
210
A. Kucinskas et al.
0.08
0.0
= 4000
= 4000
0.00
-0.2
- log
= 8500
-0.2
-0.4
(X )
3D
-0.08
-0.16
0.08
= 16000
0.00
log
- log
3D
0.00
(X )
log
0.0
= 8500
(X )
-0.16
0.08
1D
-0.4
(X )
1D
-0.08
0.0
-0.2
-0.08
= 16000
-0.4
Fe I
-0.16
0
Si I
Ni I
Mg I
Ti I
Ca I
Fe II
Zn I
4
, eV
-0.6
Si II
Ba II
Ni II
Mg II
Zn II
Ti II
Ca II
O I
N I
10
, eV
Figure 1. The 3D1D abundance differences plotted versus excitation potential, , and shown
at several wavelengths for neutral (left panel) and ionized species (plus N I and O I, right panel)
in a late-type giant characterized by Teff = 3640 K, log g = 1.0, [M/H]= 0.0.
1. Observation material
High-resolution UV spectra of 24 galactic OB supergiants were obtained with IUE
in the wavelength range 1200-1900
A. In this work we have selected the following nonresonance absorption photospheric lines: C III 1247
A, N IV 1718
A, N III 1748
A, N III
1752
A and He II 1640
A. Sample stars: HD 207198, HD 30614, HD 209975, HD 167264,
HD 37128, HD 204172, HD 38771, HD 192422, HD 213087, HD 2905, HD 13854, HD
24398, HD 193183, HD 13841, HD 190603, HD 14818, HD 206165, HD 41117, HD 42087,
HD 198478, HD 53138, HD 225094, HD 164353, HD 58350. Spectral type and effective
temperatures are adopted from McErlean, Lennon & Dufton (1999).
212
in stars with large vsini, which leads to overestimate of equivalent widths and underestimate of the continuum level. Moreover, the models do not include all lines that exist
in real objects, the latter showing much stronger absorption features relative to the synthetic spectra. Further, the lines considered to be photospheric might be possible affected
by neighboring resonant wind-contaminated line which are not included in hydrostatic
Tlusty models. As the synthetic spectra were convolved with averaged rotational and
macroturbulent velocities, large deviations for real objects may exist.
Figure 1. log W vs. Teff from observed (symbols) and synthetic (lines) spectra. The CN-processed model assumes a He abundance increased to He/H=0.2 by number, N abundance increased by a factor of 5, and a halved C abundance.
3. Conclusion
Differential analysis is not sufficiently reliable in determination of abundances from
UV spectra of B supergiants. In order to estimate the contribution of individual lines in
the observed profile, exact modeling by use of adequate codes is needed. It is a question
whether the present models are able at all to correctly produce strong UV lines. For
example, the wind model CMFGEN has failed in this attempt (Searle et al. 2008). The
non-wind Tlusty code models purely photospheric lines and can be successfully applied
on weak optical lines. We have shown that use of Tlusty grids and convolution with
rotational and macroturbulent broadening cannot reproduce the photospheric UV lines
appropriately.
References
Cassatella, A. & Martin, T. 1982, Report to IUE Three-Agency Coordinatiom Meeting
Dufton, P.L., Ryans, R.S.I., Simon-Diaz, S., Trundle, C. & Lennon, D.J. 2006, A & A, 451, 603
Lanz, T. & Hubeny, I. 2007, ApJS, 169, 83
Lanz, T. & Hubeny, I. 2003, ApJS, 146, 417
McErlean, N.D., Lennon, D.J. & Dufton, P.L. 1999, A & A, 349, 553
Ryans, R.S.I., Dufton, P.L., Rolleston, W.J.R., Lennon, D.J., Keenan, F.P., Smoker, V.J. &
Lambert, D.J. 2002, MNRAS, 336, 577
Searle, S.C., Prinja, R.P., Massa, D. & Ryans, R. 2008, A & A , 481, 777S
Turnrose, B.E. & Thompson, R.W. 1984, Conctract NAS 5-27295 Task Assignment 401
1. Overview of Project
Historically the atmospheres of B-type main-sequence band stars have provided perhaps the most accurate information on the present-day abundances of the elements in a
galaxy and its chemical evolution. We are currently determining the abundances of light
and Fe group elements in a subset of early B stars that display exceptionally sharp lines
and investigating the extent to which contemporary NLTE and LTE models represent
their atmospheres. The abundances of the Fe group elements are of particular interest,
as they serve as a tracer of earlier supernova activity across a galaxy. High resolution
spectral data from HST, FUSE, and the KPNO Coude Feed Telescope are compared
with computations from the NLTE codes TLUSTY/SYNSPEC (Hubeny 1988, Hubeny
& Lanz 1995) and the LTE programs ATLAS9/SYNTHE (Kurucz 1993ab). The atmospheric parameters Teff , log g are determined in standard ways from the SED, Balmer
line profiles, and the ionization balance criterion, while the microturbulence, turb , is
assessed by the method of examining the abundances versus the EWs of the lines of O ii
and other ions or from the SED. Some results for Her (B3V) are presented below.
2. Herculis
Her is an ultrasharp-lined B3V field star that has long been considered an abundance
standard. To determine the abundances of the Fe group elements, we obtained FUV/NUV
spectra of Her with the HST/STIS E140M/E230M echelle gratings as part of a HST
SNAPSHOT program. Other than for a few weak multiplets of Fe iii in the optical region,
213
214
Figure 1. The observed UV spectrum of Her compared with model spectra computed from
TLUSTY/SYNSPEC and ATLAS9/SYNTHE. Left panel: Fits in a region that contains numerous lines of Ti iii, Cr iii, and Mn iii suggest abundances that are 0.3-0.7 dex below solar values.
Right panel: A solar Fe abundance is implied by fits to Fe iii lines in the 1900
A region. There
is good agreement between NLTE & LTE calculations. Model parameters are Teff = 17500 K,
log g = 3.75, turb = 0 km s1 , and V sin i = 0 km s1 .
the measurable lines from the Fe group are typically found in the UV. Model spectra
were computed with the Hubeny/Lanz NLTE and the Kurucz LTE codes and compared
with the continuum-normalized observations. In Fig. 1 we show sample spectral fits in
the 12861290
A and 18921899
A regions that are rich in lines from the Fe group. The
grid of B star models by Lanz & Hubeny (2007) were used for the NLTE calculations.
Fe was treated in NLTE. The calculations for the other Fe group elements were made
in the LTE approximation. Most fits to the lines of Ti iii, Cr iii, and Mn iii (cf. Fig. 1)
suggest abundances that are in the range of 0.3-0.7 dex below the solar values given
in Asplund et al. (2005). The numerous Fe iii lines in the 18802000
A region were
fit well with a solar abundance (Fig. 1), and there was good agreement between the
NLTE and LTE calculations. But other lines of this species (e.g. near 1400
A ) were
computed to be too strong for both NLTE and LTE treatments. Besides its sharp lines,
Her makes an excellent abundance standard because its microturbulence appears to be
near zero. Typically UV lines in early B stars suggest a smaller value of turb than the
optical O ii features, and this causes most of the prominent UV lines to suggest lower
abundances, but this is not a problem here. However it should be mentioned that the
lines in Her are much sharper than the spectral resolution of the E140M and E230M
gratings ( 7 10 km s1 ) and observations with the E140H and E230H echelle gratings,
which have a resolution of 3 km s1 , are warranted to investigate photospheric velocity
fields and improve the abundance determinations.
The authors appreciate support from NASA grants NAG5-11802, NAG5-12239, NAG513212, and STScI grants GO-09848 & GO-06709.
References
Asplund, M., Grevesse, N., & Sauval, A. J. 2005, in F. N. Bash & T. G. Barnes (eds.), Cosmic
Abundances as Records of Stellar Evolution and Nucleosynthesis (ASP Conf. Ser. 336,San
Francisco: ASP), p. 25
Hubeny, I. 1988,Comput. Phys. Commun. 52, 103
Hubeny I, & Lanz, T. 1995 ApJ, 439, 875
Kurucz, R. L. 1993a,Kurucz CD-ROM No. 13, (Cambridge:SAO)
Kurucz, R. L. 1993b,Kurucz CD-ROM No. 18, (Cambridge:SAO)
Lanz T., & Hubeny, I. 2007ApJS, 169, 83
Figure 1. The data used to determine the bolometric flux of Cen A+B . IUE and ground-based
spectrophotometric data are shown with thin lines and filled symbols are broad-band fluxes.
216
Hans Bruntt
Figure 2. Left panel: Comparison of Teff , R and M from direct and indirect methods. Right
panel: H-R diagram with ASTEC tracks. Solid circles are stars with measured angular diameters.
only the limb-darkening is from models. (2) We made a classical spectroscopic analysis
of 100s of Fe i lines requiring that lines with a range of different excitation potentials and
line strengths yield the same abundance. We use the VWA tool (Bruntt et al. 2008, Bruntt
2009) employing 1D LTE MARCS atmospheric models (Gustafsson et al. 2008; spectra
are from HARPS@ESO except Boo observed with FIES@NOT). As shown in Fig. 2
(left top panel) the mean difference is Teff = 48 49 K (rms scatter), and there is no
significant correlation with Teff . We thus claim that after correcting for the Teff offset
we can determine Teff from a high-quality spectrum to 50 K in the spectral range from
Procyon A (F5) to Cen B (K1). The stars are shown in the H-R diagram in Fig. 2.
Radius with 3% accuracy without interferometry
We determined the radii of the stars using a direct and an indirect method: (1) We
combined the measured angular diameters from the literature with the updated parallaxes
from van Leeuwen (2007). (2) We combined the spectroscopic Teff with the luminosity
through L/L = (R/R )2 (Teff /Teff; )4 . The luminosities were determined from the V
magnitude, BC from Girardi et al. (2002), and the parallax from van Leeuwen (2007).
The comparison of the two methods to determine R is shown in Fig. 2 in the left middle
panel. The agreement is good and shows that radii can be determined from indirect
methods to 3% (1- uncertainty).
Mass with 4% accuracy for single stars
Three of the stars are members of binary systems and have well-determined masses
(< 2%). For all stars the radius has been measured (< 2%) from interferometry and the
mean density is inferred from asteroseismic data (large separation) by scaling from the
Sun. Combining the radius and density we get the mass to 4%. The left lower panel in
Fig. 2 compares the mass for the two methods (Procyon A and Cen A+B). Although
the number of stars is small it is reassuring that the agreement is good (2% rms scatter).
References
Bruntt, H., 2009, A&A accepted, arXiv 0907.1198
Bruntt, H., De Cat, P., Aerts, C., 2008, A&A 478, 487
Girardi, L., et al., 2002, A&A 391, 195
Gustafsson, B., et al., 2008, A&A 486, 951
van Leeuwen, F., 2007, A&A 474, 653
Session IV
2 - Dwarf galaxies
Abstract.
As the least massive galaxies we know, dwarf spheroidal galaxies (dSph) allow to probe
chemical enrichement on the smallest scales, and perhaps in its simplest expression. Particularly
interesting are the issues concerning the efficency with which metals are retained or lost in these
shallow potential wells (supernovae feedback), and the effect of this on star formation itself.
Another fundamental issue concerns the earliest epochs of star formation: are first stars formed
in similar ways and proportions in all halos ? Finally, as the smallest galaxies know, dSph have
been suggested to be the surviving cousins of galaxy building blocs that (in -CDM) assemble
to make larger galaxies. This parenthood would not necessarily hold at all late times, when
survivors have lived their own differentiated life, but is expected at least at the earliest epochs.
I review here the chemical abundances of individual stars in the nearest dwarf spheroidal
galaxies, that have become available in increasing numbers (sample size and galaxies probed) in
the last decade. Special emphasis is given to: a) recent results obtain with FLAMES on VLT,
highlighting the power of detailed chemical abundance patterns of large samples of stars to
unravel the various evolutionnary paths followed by dSph; b) the oldest and most metal-poor
populations in dSph.
Keywords. stars: abundances, galaxies: abundances, galaxies: evolution, galaxies: formation,
Local Group
1. Introduction
The detailed chemical abundance patterns in individual stars of a stellar population
provide a fossil record of chemical enrichment over different timescales. As generations of
stars form and evolve, stars of various masses contribute different elements to the system,
on timescales directly linked to their mass. These studies require precise measurements
of elemental abundances in individual stars, only achievable with high-resolution and
reasonably high signal-to-noise spectra. It is only very recently that this has become possible beyond our Galaxy. It is efficient high-resolution spectrographs on 810m telescopes
that have made it possible to obtain high resolution (R>40000) spectra of RGB stars
in nearby swarf spheroidals (dSphs) and O, B and A super-giants in more distant dwarf
irregulars (dIs). These stars typically have magnitudes in the range V=17 19. After the
pioneering works of Shetrone et al. (1998, 2001) and Bonifacio et al. (2000) in dSphs, or
Venn et al. (2003) and Kaufer et al. (2004) in dIs, samples have slowly started to grow. In
the following, I will concentrate on the case of classical dSph galaxies, in which samples
now reach for the first time samples sizes within in a given galaxy large enough to actually follow its chemical enrichment. Further considerations about dIs and the faintest
dSph (known as ultra-faint dwarf spheroidal galaxies) can be found in Tolstoy, Hill &
Tosi (2009).
219
220
Vanessa Hill
2. Alpha elements
/Fe, is commonly used to trace the star-formation timescale in a system, because it is
sensitive to the ratio of SNII (massive stars) to SNIa (intermediate mass binary systems
with mass transfer) that have occurred in the past. SNIa progenitors have a longer lifetime
than SNII and as soon as they start to contribute they dominate the iron enrichment and
[/Fe] decreases. This is seen as a knee in a plot of [Fe/H] vs. [/Fe], see Fig. 1. The knee
position indicates the metal-enrichment achieved by a system by the time SNIa start to
contribute to the chemical evolution, i.e. a galaxy that efficiently produces and retains
metals will reach a higher metallicity by the time SNIa start to contribute than a galaxy
which either loses significant metals in a galactic wind, or simply does not have a very
high astration. The position of this knee is expected to be different for different dSphs
because of the wide variety of star formation histories. The level of /Fe in metal-poor
stars (before the knee) is on the other hand sensitive of the high-mass IMF and selective
galactic winds.
0.5
-0.5
0.5
0
-0.5
-2
-1
[Fe/H]
Figure 1. Abundances of two -elements Mg and Ca in dSPhs (large coloured symbols), compared to the Milky-Way disc and halo (small black dots Venn et al. 2004). Sgr (red triangles:
Sbordone et al. 2007, Monaco et al. 2005, McWilliam & Smecher-Hane 2005), Fnx (blue circles:
Letarte et al. 2009, Shetrone et al. 2003), Scl (green squares: Hill et al. in prep, Shetrone et al.
2003, Geisler et al. 2005) and Carina (magenta stars: Koch et al. 2008a, Shetrone et al. 2003).
The knee : The apparent paucity of -elements (relative to iron) in dSph galaxies
compared to the MW disk or halo was first noted by (Shetrone et al. 1998, 2001, 2003,
Tolstoy et al. 2003) from small samples. Fig. 1 shows this convincingly over most of the
metallicity range in each system. However, it also appears that each of these dSphs starts,
at low [Fe/H], with [/Fe] ratios similar to those in the MW halo at low metallicities,
thereby defining the location of the knee. At present the available data only cover the
knee with sufficient statistics to quantify the position in the Sculptor (Scl) dSph, a
system which stopped forming stars 10 Gyr ago, and the knee occurs at [Fe/H] 1.8.
This is the same break-point as the two kinematically distinct populations in this galaxy
Tolstoy et al. (2004). This means that the metal-poor population has formed before any
SNIa enrichment took place, which means on a timescale shorter than 1 Gyr. In other
dSphs the knee is not well defined due to a lack of data, but limits can be established
that suggest that not all dSphs have a knee at the same position. The Carina dSph has
had an unusually complex SFH, with at least three separate bursts of star formation
(Hurley & Tout 1998). The abundance measurements in Carina are presently too scarce
to confidently detect these episodes in the chemical enrichment pattern, but Carina
221
appears to possesses [/Fe] poor stars between [Fe/H] = 1.7 and 2.0, which suggests
that the knee occurs at lower [Fe/H] than in Scl. As shown by Cohen & Huang (2009),
Draco may be an even more drastic example where the knee may appear as low as
[Fe/H] 2.8. At the other extreme, Sagittarius (Sgr) has enhanced [/Fe] up to [Fe/H]
1.0, which is significantly more metal-rich than the position of the knee in Scl. This
is consistent with what we know of the SFH of Sgr, which has steadily formed stars over
a period of 810 Gyrs, and only stopped forming stars about 23 Gyr ago (e.g., Dophin
& Kennicutt 2002). In the Fornax (Fnx) dSph, another galaxy with a complex SFH, the
sample does not include a sufficient number of metal-poor stars to determine even an
approximate position of the knee. From this (small) sample of dSph galaxies, it appears
that the position of the knee correlates with the total luminosity of the galaxy, and the
mean metallicity of the galaxy, suggesting that the presently most luminous galaxies are
those that must have formed more stars at the earliest times and/or retained metals
more efficiently than the less luminous systems.
Low-metallicity stars The abundance ratios observed in all dSphs for stars on the
metal-poor side of the knee, tend to be indistinguishable from those in the Milky-Way
halo. It therefore seems that the first billion year(s) of chemical enrichment gave rise to
similar enrichment patterns in small dwarf galaxies and in the Milky-Way halo, at least
in systems as luminous as Scl, Fnx or Sgr.. This is comsistent with a constant IMF, with
no need to resort to incomplete IMF sampling as was proposed early on (Tolstoy et al.
2003, Carigi & Hernandez 2008). Furhermore, this opens up a window for merging dSphlike Milky-Way halo building blocks, at a time when their internal chemical evolution
was not yet visible.
On the other hand, there is now a hint that the slightly less luminous Sextans (Mv=9.5) could display a scatter in the /Fe ratios at the lowest metallicities, including [/Fe]
close to solar Aoki et al. (2009). Such a scatter is so far observed only in this purely old
system, and suggests a very inhomogeneous metal-enrichement in this system that presumably never retained much of the metal it produced. The true extent of this scatter
in Sextans remains to be investigated, and extension to other similar sytems is needed
before general conclusions can be reached on the mechanisms leading to the chemical
homogeneity -or not- of dwarf galaxies. Note that based on hydrodynamical Nbody simulations, Revaz et al. (2009) suggest that the degree of inhomogeneity is linked to the
galaxy total mass (through low-level star formation stochasticity), and that the least
massive systems are expected to be inhomogeneous.
222
Vanessa Hill
dSph galaxies and in the Milky-Way. In the MW, the Ba and Y are dominated by the
r-process for [Fe/H] 6 2.0 (e.g., Travaglio et al 2004), while the s-process dominates at
higher metallicities.
2
1
0
-1
1.5
1
0.5
0
-0.5
-2
-1
[Fe/H]
Figure 2. Abundances of two neutron-capture elements Ba (s- and r- processes) and Eu (r-process element) in dSphs (large coloured symbols), compared to the Milky-Way disc and halo.
Symbols and references as in Fig.1
At early times (at [Fe/H] < 1) there seems to be little difference between the various
dSphs, and the MW halo in Fig 2. The scatter and downturn of [Ba/Fe] at and below
[Fe/H] = 3 is a well known feature in the MW halo and references therein]Francois07,
Barklem05. In the dwarfs however, there is a hint that this may occur already at higher
metallicities ([Fe/H] < 1.8). This hint is confirmed in Fig. 5 which includes very metalpoor stars in other dSphs, although from much smaller samples (see figure caption). This
should still be regarded as a tentative hint, because of the low number statistics for dSph
low-metallicity stars. But if confirmed with larger samples, these low r-process values
at higher [Fe/H] than in the Galactic halo would either mean that the dwarf galaxies
enriched faster than the halo at the earliest times or that the site for the r-process is less
common (or less efficient) in dSphs.
The Ba/Eu ratio, shown in Fig. 3 (left panel), indicates the fraction of Ba produced
by the s-process to that produced by the r-process. In dSph, as in the MW, the early
evolution of all neutron-capture elements is dominated by the r-process (as was already
noted by Shetrone et al. 2001, 2003). In each system, however, the low and intermediate
mass AGB stars contribute s-process elements, that soon start to dominate the Ba (and
many other neutron capture elements) production. The metallicity at which this switch
from r- to s-process is only somewhat constrained in the Scl dSph. It is found to occur
at [Fe/H] 1.8, which is the same as the [/Fe] knee. This turnover needs to be better
constrained in Scl and even more so in other galaxies to provide timing constraints on
the chemical enrichment rate. It could reveal the metallicity reached by the system at
the time when the s-process produced in AGBs starts to contribute.
For the more metal-rich stars ([Fe/H] > 1) there is also a distinctive behaviour of
[Ba/Fe] in dSphs (Fig. 2). In the Scl dSph the [Ba/Fe] values never leave the MW trend,
but this galaxy also has almost no stars more metal-rich than [Fe/H]< 1. Fnx, on the
other hand, and to a lesser extent Sgr, display large excesses of barium for [Fe/H]> 1.
This is now barium produced by the s-process, and it shows the clear dominance of the sprocess at late times in dSphs. Furthermore, Fig. 3 (left panel) shows that, in the domain
where the neutron-capture enrichment is dominated by the s-process, [Y/Ba] in dSphs are
223
-1
[Fe/H]
Figure 3. (Left panel) : Ba/Y and Ba/Eu abundance ratios in dSphs (large coloured symbols),
compared to the Milky-Way disc and halo. Symbols and references as in Fig.1. (Right panel) :
Eu/H run versus Ba/H in Fornax (large circles, Letarte et al. 2009), compared to simple models
assuming, starting at [Ba/H] = 0.8 [Ba/H] = 0.5 (where the Ba/Fe rise occurs in Fnx,
see Fig. 2): a pure r-process enrichment (solid red line), a 90% s-process and 10% r-process
contribution (long dashed green line), a pure s-process enrichment using respectively solar and
1.14 metallicity AGB Ba/Eu yields (Cristallo et al. 2009; dotted black lines)
224
Vanessa Hill
Hamburg ESO Survey (HES, Christlieb et al. 2004). Indeed, when considering the whole
sample of stars in the four dSph galaxies Fornax, Sculptor, Sextans and Carina, a distinctive lack of stars with metalicities [Fe/H] < 3 was found. This conclusion was based
on metallicities obtained from low resolution spectra at the infrared Ca II triplet, empirically calibrated using globular clusters and high-resolution metallicities of stars in
Sculptor and Fornax (Battaglia et al. 2008). In the meantime, the low metallicity tail
of the galactic halo distribution has been revised downwards, taking into account biases arising from the design of the HES to find the most metal-poor stars, and bringing
the halo closer to the dSph metallicity distributions (Sch
ork et al. 2009). More importantly still, we have undertaken a systematic study of the low-metallicity tail of the four
abovementioned galaxies by :
(i) building up a new calibration of the Ca II triplet based on synthetic spectra rather
than observed globular clusters stars (Starkenburg et al. in prepartion). Indeed, because
of the Ca II triplet lines change regime, from strongly wing dominated in giants of
metallicities > 2 to core dominated in very metal-poor giants (< 3), the empirical
simple linear relation between the equivalent width of the Ca II triplet lines (reduced,
by folding in a luminosity-dependent factor) and the metallicity is not expected to hold
down to the lowest metallicities. The new synthetic spectra based calibration overlays
nicely the empirical relation down to metallicities of 2.0, but deviates strongly at
the lowest metallicities: as illustrated in Fig.4 left panel, the slope of the equivalent
widths versus luminosities differs from the empirical one, and the metallicity dependency
becomes shallower. Both these effects concur to produce a bias in the Ca II metallicity
estimates based on the empirical calibration at the lowest metallicities.
Figure 4. (left panel): empirical (dotted lines) and synthetic spectra based (symbols according
to metallicity) relations between the Ca II triplet equivalent widths (sum of the two strongest
lines) and luminosity for various metallicities. The two relations agree for metallicities > 2; at
low metallicities, the empirical relation is too steep and lines of iso-metallicities are too far appart
compared to the synthetic spectra expectations. (middle and right panels): metallicities deduced
from Ca II triplet ([Fe/H]LR ) using the empirical (middle, Battaglia et al. 2008) and synthetic
spectra based (right) calibrations are compared to metallicities measured in high-resolution
spectra of DART samples ([Fe/H]HR ). The clear bias observed at the lowest metallicities with
the empirical relation disapears once the true (synthetic spectra) behaviour of the Ca II lines is
taken into account in our new calibration.
(ii) endertaking a high resolution follow-up of the most promissing candidates (Taflemeyer et al. in preparation, Aoki et al. 2009 and Venn et al. in preparation). This has
225
allowed us to firmly establish the presence of very low metallicity stars in dSph (down
to 4 in Sculptor, Tafelmeyer et al. in prepartion), and to build a comparison sample to
validate the new Ca II calibration (Fig. 4).
Using the new calibration, we revisit the metallicity distribution of dSph (Starkenburg
in prep.). Fig. 5 (left panel) shows that while the metallicity distribution is globally
unchanged, the shape of the metal-weak tail is altered, the new calibration providing a
more extended tail. It is thus now clearly established that dSph contain, albeit in small
numbers, extremely metal poor stars, at least down to 4 metallicity. There does not
seem to be a clear difference between classical dSph galaxies and the ultra-faint dwarfs
in this respect anymore (Kirby et al. 2008).
Thanks to the high-resolution follow-up of the extremely metal-poor candidates ([Fe/H] <
3) from these galaxies (by the DART collaboration, but also other groups, Fulbright,
Rich & Castro 2004, Cohen & Huan 2009, Frebel et al. 2009, Frebel et al. 2009, this conference), not only the metallicity, but also the detailed chemical composition extremely
metal poor stars can also be probed, and Fig. 5 shows Mg/Fe and Ba/Fe for a compilation
of known stars in classical and ultra-faint dwarf galaxies below [Fe/H] < 2.3. There is
a global agreement in the abundance ratios of extremely metal poor stars ([Fe/H] < 3)
in dwarfs galaxies and the Milky-Way halo.
1
0.5
0
-0.5
0
Sgr
Fnx
Scl
-1
Sext
Car
Dra
Ufdwarfs
-2
-3
-4
-3
-2
-1
[Fe/H]
Figure 5. (Left panel) Metallicity distributions of Sculptor, Fornax, Carina and Sextans, with
the Battaglia et al. (2008) (black) and new Starkenburg et al. (in preparation) (red/grey) calibrations. The metallicity distribution now clearly extends down to [Fe/H] = 3 and below.
(Right panel) Abundance ratios Mg/Fe and Ba/Fe as a function of metallicity for stars with
[Fe/H] < 2.3 in classical (large squares: DART follow-ups, Tafelmayer et al. in preparation
(Scl, Fnx, Sext), Venn et al. in preparation (Car), Aoki et al. 2009 (Sext); filled circles: see refs
in Fig 1 and Cohen et & Huang 2009 (Draco)) and ultrafaint (triangles: Frebel et al. 2009, Koch
et al. 2008b). The mean trends (running average on 10 points) and dispersion of Scl, Fnx and
Sgr (as of Figs. 1 and 2) are indicated as lines.
Acknowledgements
This paper is mostly based on the work of the DART Dwarf Abundances and Radial
velocity Team collaboration, and on the review paper of Tolstoy, Hill & Tosi (2009)
References
Aoki W., Arimoto N., Sadakane K., Tolstoy E., Battaglia G., Jablonka P., Shetrone M., Letarte
B., Irwin M., Hill V., Francois P., Venn K., Primas F., Helmi A., Kaufer A., Tafelmeyer
M., Szeifert T., Babusiaux C., 2009, A&A502, 569
226
Vanessa Hill
Battaglia G., Irwin M., Tolstoy E., Hill V., Helmi A., Letarte, B., Jablonka P., 2008, MNRAS
383, 183
Bonifacio P., Hill V., Molaro P., Pasquini L., Di Marcantonio P., Santin P., 2000, A&A, 359,
633
Busso, M., Gallino, R., & Wasserburg, G.J. 1999, ARAA, 37, 239
Carigi L., Hernandez X., 2008, MNRAS, 390, 582
Cristallo S., Straniero O., Gallino R., Piersanti L., Domnguez I., Lederer M.T., 2009, ApJ, 696,
797
Christlieb N., Reimers D., Wisotzki L., 2004, The Messenger, 117, 40
Cohen J., Huang W., 2009, ApJ, 701, 1053
Dolphin A., Kennicutt R.C., 2002, AJ, 123, 3154
Frebel A., Simon J., Geha M., Willman B., 2009, arXiv:0902.2395v1
Fulbright J., Rich R.M., Castro S., 2004, ApJ 612, 447
Geisler D., Smith V., Wallerstein G., Gonzalez G., Charbonnel C., 2005, AJ, 129, 1428
Helmi A., Irwin M., Tolstoy E., Battaglia G., Hill V., Jablonka P., Venn K., Shetrone M., Letarte
B., Arimoto N., Abel T., Francois P., Kaufer A., Primas F., Sadakane K., Szeifert T. 2006,
ApJ 651, L121
Hurley J. Tout C., 1998, MNRAS, 300, 977
Kaufer A., Venn K., Tolstoy E., Pinte C., Kudritzki R.-P., 2004, AJ 127, 2723
Koch A., Grebel E., Gilmore G., Wyse R., Kleyna J., Harbeck D., Wilkinson M., Wyn E., 2008,
AJ 135,1580
Koch A., McWilliam A., Grebel E., Zucker D., Belokurov V., 2008, ApJ 688, 13
Lanfranchi G., Matteucci F., Cescutti G., 2008, A&A, 481, 635
Letarte B., Hill V., Tolstoy E., Jablonka P., Shetrone M., Venn K., Spite M., Irwin M., Battaglia
G., Helmi A., Primas F., Francois P., Kaufer A., Szeifert T., Arimoto N., Sadakane K., 2009,
A&A submitted
McWilliam A., & Smecker-Hane T., 2005, in: T.G. Barnes & F.N. Bash (eds), Astronomical
Society of the Pacific Conference Series 336, 221
Monaco L., Bellazzini M., Ferraro F., Pancino E.,, 2005, MNRAS 356, 1396
Revaz Y., Jablonka P., Sawala T., Hill V., Letarte B., Irwin M., Battaglia G., Helmi A., Shetrone
M. D., Tolstoy E., Venn K., 2009, A&A 501, 189
Sbordone L., Bonifacio P., Buonanno R., Marconi G., Monaco L., Zaggia S., 2007, A&A 465,
815
Sch
orck T., Christlieb N., Cohen J.G., Beers T.C., Shectman S., Thompson I., McWilliam
A., Bessell M.S., Norris J.E., Melendez J., Ramirez S., Haynes D., Cass P., Hartley M., Russell K., Watson F., Zickgraf F.-J., et al. 2009, A&A preprint doi
http://dx.doi.org/10.1051/0004-6361/200810925
Shetrone M., Bolte M., Stetson P., 1998, AJ 115, 1888
Shetrone M., C
ote P., Sargent W.L.W., 2001, ApJ 548, 592
Shetrone M., Venn K., Tolstoy E., Primas F., Hill V., Kaufer A., 2003, ApJ 125, 648
Smecker-Hane T.A., & McWilliam A. 2002 (ApJ submitted) arXiv:astro-ph/0205411
Sneden C., Cowan J., Gallino R., 2008, ARAA 46, 241
Tolstoy E., Venn K., Shetrone M., Primas F., Hill V., Kaufer A., Szeifert T., 2003, AJ 125, 707
Tolstoy E., Irwin M., Helmi A., Battaglia G., Jablonka P., Hill V., Venn K., Shetrone M., Letarte
B., Cole A., Primas F., Francois P., Arimoto N., Sadakane K., Kaufer A., Szeifert T., Abel
T., 2004, ApJ 617, L119
Tolstoy E., Hill V., Tosi M., 2009, ARA&A 47, 371
Travaglio C., Gallino R., Arnone E., Cowan J., Jordan F., Sneden C., 2004, ApJ 601, 864
Venn K., Irwin M., Shetrone M., Tout C., Hill V., Tolstoy E., 2004, AJ, 128, 1177
Venn K., Tolstoy E., Kaufer A., Skillman E., Clarkson S., Smartt S., Lennon D., Kudritzki R.,
2003, AJ 126, 1326
1. Introduction
Dwarf spheroidal (dSph) galaxies are intriguing for a plentitude of reasons: Owing to
their very low luminosities (MV >14 mag) they have been characterized as faint systems
ever since their first discovery. They further have low total masses of only a few 107 M
and a puzzling deficiency of gas (e.g., Grebel et al. 2003; Bailin & Ford 2007; Gilmore et
al. 2007). Over the past three years, the faint end of the galaxy luminosity function has
been traced even further down, towards the ultrafaint regime. These ultrafaint dSphs,
discovered in large number in sky surveys such as the SDSS, are now the faintest galaxies
known to exist in the Universe, with absolute magnitudes above MV > 6, and stellar
masses of up to a mere few ten thousand Solar masses (e.g., Martin et al. 2008).
Furthermore, the dSphs are fairly metal-poor systems, with mean metallicities starting
at 1 to 2 dex and decreasing. The range of metallicity in a given dSph is normally
large and of up to 0.5 dex. Typically these spreads greatly exceed the measurement
errors. Despite deep photometric studies and complementing spectroscopy for selected
stars, the detailed properties of these ultrafaint galaxies remain poorly investigated until
now. Interestingly, the ultrafaint dSphs are more metal-poor on average than their more
luminous counterparts; their mean metallicities reach as low as about 2.5 dex (e.g.,
Simon & Geha 2007). While no star more metal-poor than [Fe/H]< 3 dex had been
found in any of the classical dSphs until recently (e.g., Koch et al. 2006; Helmi 2006;
227
228
A. Koch et al.
cf. Cohen & Huang 2009), several such metal-poor stars, down to 3.3 dex have been
detected in the ultrafaint galaxies (Kirby et al. 2008; Norris et al. 2008; Frebel et al.
2009).
Although cosmological simulations like CDM predict a wealth of small-scale substructures that hierarchically merge into larger structures like the present-day Milky
Way (MW), a number of arguments against such a simplistic view has arisen over the
years (Moore et al. 1999). Those comprise the oft-cited missing satellite problem, which
is, however, nowadays much alleviated (e.g., Robertson et a. 2005; Simon & Geha 2007).
Another problem is the discrepancy between the chemical abundances of the dSph stars
compared to the halo stars (Sect. 2). Furthermore, the aforementioned apparent lack of
very metal-poor stars in the dSphs was long considered a major contradiction to the large
number of such stars, below 3 dex, found in the Galactic halo. This leaves us with the
question of how and when the (ultrafaint) dSphs formed and evolved, and how they fit
into the cosmological CDM models. In particular, what fraction of dSph-like systems
contributed to the build-up of the stellar halo of the MW?
0.6
[Ca/Fe]
0.4
0.2
0
0.2
0.4
3.5
2.5
1.5
0.5
[Fe/H]
Figure 1. Currently available abundance data for Galactic stars (black dots) and dSphs
(squares). See Koch (2009) for a detailed review of the sources for these data. Red symbols
indicate the outer halo GCs Pal 3 (Koch et al. 2009) and Pal 4 (Koch & C
ote in prep.)
observations by Shetrone et al. (2001) that the dSph stars are systematically depleted
in those elements relative to halo stars at the same metallicity (as already predicted by
the models of Unavane et al. 1996), the abundance data have now vastly grown, with up
to several tens of stars in a few of the more luminous systems (e.g., Koch et al. 2008a,
2009; Tolstoy et al. 2009) and progressive measurements in the fainter ones (e.g., Koch
et al. 2008b; Frebel et al. 2009). The first thing to note is that this picture of -depletion
remains valid also when taking the new data into account. However, the new picture
that slowly emerges is that there are in fact metal-poor stars below 3 dex found in
the classical dSphs (Cohen & Huang 2009; Frebel et al., this meeting [S265-o:18]) and in
particular the ultrafaint dSphs appear to host a large number of those stars (Kirby et
al. 2008; Norris et al. 2008; Frebel et al. 2009). For many of those stars, high-resolution
229
abundance data are currently being gathered. The overlap of these metal-poor stars
abundances with those of the Galactic metal-poor halo plateau at [/Fe]+0.4 dex then
underlines the picture in which the accretion and disruption of dSph-like systems had a
major contribution to the metal-poor halo at most.
1.2
1.6
1.4
0.8
1.2
0.6
0.4
0.8
[Co/Cr]
[Mg/Ca]
In Koch et al. (2008b) we showed that the elemental abundance patterns in Her are
peculiar. Firstly, neither of the only two red giants analyzed in the literature to date
show any evidence for heavy elements (e.g., Ba, Eu) above the noise in the spectra.
Secondly, we found remarkably high abundance ratios of the hydrostatic (O, Mg) to the
explosive -elements (Si, Ca, Ti), see Fig. 2. This suggests that very massive stars were
the main drivers for the chemical enrichment in Her. For instance, the high [Mg/Ca] of
0.61.0 dex in the Her stars suggests that stars of at least 30 M governed the enrichment
(Heger & Woosley 2008). Thus we may be seeing the results of stochastical enrichment
in terms of an incompletely sampled IMF. Such abnormally high [Mg/Ca] ratios are also
found in the low-mass dSph Boo I (Feltzing et al. 2009).
0.2
0.6
0.4
0.2
0.2
0.4
0.6
0.2
0.8
4
3.5
2.5
2
1.5
[Fe/H]
0.5
0.5
0.4
4.5
3.5
2.5
Fe/H]
1.5
Figure 2. Left panel: [Mg/Ca] ratios for the same MW and dSph stars as in Fig. 1. Highlighted
as red squares are the two stars in Her. The right panel shows the Her [Co/Cr] abundance ratio
in comparison to the metal-poor Galactic halo stars of McWilliam et al. (1995).
Another peculiarity is the unusally high [Co/Cr] ratio (Fig. 2): at [Co/Cr]0.58 dex,
the two moderately metal-poor Her stars ([Fe/H] of 2 dex) rather resemble the metalpoor Galactic halo below 2.5 dex. This is explicable, if we assume that Her was
enriched towards its observed higher Fe-abundances through standard SNe Ia contributions according to its IMF, but it experienced an early enrichment from a first generation
of metal-free, very massive Population III stars, the models of which reproduce the high
[Co/Fe] and low [Cr/Fe] ratios very well. If this scenario is supported by further observations in other (fainter) stars in Her and other dSphs, it would mean that the ultrafaint
dwarfs may in fact be the site of the very first generations of stars in the Universe.
230
A. Koch et al.
3.2. Clean member selection
One difficulty in the interpretation of resolved galaxy properties from low- or medium
resolution data is the inevitable contamination with Galactic foreground stars. While the
member selection in some dSphs is straightforward, thanks to their high radial velocities
of (positive or negative) several hundred km s1 (e.g., Draco, Carina) and/or high Galactic latitudes, pure color-magnitude and radial velocity criteria in low-resolution mode fail
for those cases where the dSphs are deeply embedded in the foreground. With a systemic
mean velocity of 45 km s1 , Hercules is in fact strongly affected.
In Aden et al. (2009a) we showed, however, that the dwarf contamination can be
efficiently identified using Str
omgren photometry: this filter system is able to discern the
evolutionary stages of stars, based on a set of gravity sensitive index definitions (e.g.,
Faria et al. 2007). Using this we removed all the contaminating foreground dwarf stars.
As a result, we could isolate a bona fide member sample of 45 red giants, AGB, RHB, and
BHB stars. As it turned out, about five of the stars that overlap with previous studies
(Simon & Geha 2007; Kirby et al. 2008) are likely foreground stars. The cleaned member
candidate sample then yields a lower velocity dispersion and thus significantly lower mass
by a factor of 3 (Aden et al. 2009b) compared to the literature (e.g., Simon & Geha
2007; Strigari et al. 2008). Furthermore, our sample is slightly more metal-rich than the
previous estimates. Both from our calcium triplet spectroscopy and from the calibration
of the Str
omgren photometry onto metallicity, we find a mean [Fe/H] of 2.35 dex with
a 1 scatter of 0.31 dex. Currently, there are four objects known that have systemic
radial velocities comparable to the Galactic foreground around 050 km s1 , viz. CVn I,
Willman I, Leo T, and Her. Therefore we emphasize the importance of a clear member
selection for all future studies, by means of sophisticated photometric and spectroscopic
techniques. High-resolution follow-up is vital for assessing the true stellar properties and
sampling the true, full metallicity (preferably, iron) ranges of the dSphs.
231
outer halo cluster NGC 7492 (Cohen & Melendez 2005a,b). On the other hand, Pal 3
does not resemble dSph field stars in any regard: the enhanced abundances and most
other abundance patterns are incompatible with, e.g., the -depletions seen in the dSph
stars. The situation is, however, slightly different for the GC system of the Fornax dSph
(Letarte et al. 2006). Its abundance patterns are unlike the surrounding dSph field stars
and rather appear to be similar to the patterns found in the Galactic halo and its GCs
and a possible resemblance with Pal 3 cannot be refuted by the present data. However,
the observed very small scatter of the Pal 3 stars advocates against the large abundance
spreads detected in all of the Local Group dSphs analyzed to date and we conclude that
Pal 3 (and Pal 4, Koch & C
ote, in prep.) did likely not contribute any major fraction to
the outermost halo, but rather are part of an underlying, genuine halo population.
There is yet one peculiarity found in Pal 3 that deserves notion: All the heavy elements
in this cluster are fully compatible with a pure r-process origin without any need to invoke
any significant contribution from the s-process. Such a behaviour has so far only been
observed in very metal-poor field stars (e.g., Honda et al. 2007) and in the GC M 15
(Sneden et al. 2000) and its surrounding stellar stream (Roederer 2009, this meeting
[S265-p:8]). Contrary to the massive progenitors that dominated the enrichment in the
low-mass environment of the Her dSph, the n-capture patterns in Pal 3 do not require any
such enrichment by massive stars, but are explicable by a standard r-process in 8 M
SNe II (e.g., Qian & Wasserburg 2003). A detailed interpretation of this loss of s-process
material is beyond the scope of this work. We just remark that the large spatial extent
of this loose cluster and the corresponding shallow potential could have favored an early
loss of the ejecta of a first generation of AGB stars, followed by efficient star formation
with a high SNe II rate leading to strong r-enrichment through this second generation of
cluster stars (cf. DAntona & Caloi 2008). While it is clearly unwarranted to discern any
trend of the [r/Fe] and [r/s] ratios with structural parameters in Fig. 3, it is certainly
noteworthy that Pal 3, with its large radius, also exhibits the highest [Eu/Fe] (likewise
[Dy/Fe]), lowest [Ba/Eu] ratios, respectively. A complete and homogeneous survey of the
n-capture elements within the Galactic GC system is clearly necessary.
0.9
0.2
0.8
0.7
[Ba / Eu]
[Eu / Fe]
0.6
0.5
0.4
0.2
0.4
0.3
0.2
0.6
0.1
0
0
r [pc]
10
15
0.8
0
r [pc]
10
15
Figure 3. Abundance ratios of a sample of Galactic GCs (blue points) as function of their
core radii. Pal 3 is denoted as a red star symbol.
References
Aden, D., et al. 2009a, A&A, in press (astro-ph/0908.3489)
Aden, D., et al. 2009b, ApJL, submitted
232
A. Koch et al.
Discussion
Nomoto: Your abundance patterns in Hercules could also be explained by enrichment
from faint Supernovae with progenitor masses of 25 M [this meeting; S265-i:26].
Koch: That is an interesting possibility and one should look into fits to the entire
abundance distribution of the stars. In any case, 25 M still leaves us in the high-mass
regime for Hers enrichment.
Sarajedini: Are you saying from your data in only two GCs that the outer and inner
halo are coeval? There is evidence that parts of the Galactic GCs were accreted from
dSphs (e.g., the Sagittarius clusters).
Koch: Our two clusters appear to have experienced a similar chemical evolution as the
inner halo clusters. Furthermore, as opposed to the Sgr clusters, Pal 3 is not a member
of any currently known stream.
1. Introduction
The spectral analysis of H ii regions has been an invaluable tool in astrophysics for the
past few decades, providing a straightforward means to measure present-day chemical
abundances in a variety of galactic environments, which has led, for example, to the
study of radial abundance gradients in spiral galaxies (Vila-Costas & Edmunds 1992,
Zaritsky et al. 1994). It has recently become possible to extend the nebular techniques
to star-forming galaxies at high redshift (Pettini et al. 2001), allowing the investigation
of the cosmic chemical enrichment (Savaglio et al. 2005, Maier et al. 2006) in connection
with fundamental properties of galaxies such as the mass-metallicity relation (Tremonti
et al. 2004, Maiolino et al. 2008). In the local Universe, the knowledge of abundance
gradients in spiral galaxies and their evolution with time provides the necessary observational constraints to the parameters that drive models of the chemical evolution of
galaxies, such as the radial dependence of accretion and star formation rate in galactic
disks (Matteucci & Francois 1989, Chiappini et al. 2001).
The work on young massive stars to obtain reliable metallicities in nearby galaxies is
less advanced, due to the need for a sophisticated NLTE treatment of the physical processes involving millions of metal lines in expanding atmospheres (Hillier & Miller 1998,
Pauldrach et al. 2001, Puls et al. 2005), and the requirement for telescopes with large
collecting areas to secure spectra of individual stars located a few Mpc away (Bresolin
et al. 2001, Bresolin et al. 2002). While it is possible to measure stellar metallicities
from the integrated spectra of young star clusters (Larsen et al. 2006, 2008) or for starforming galaxies at high redshift (Rix et al. 2004, Halliday et al. 2008), more stringent
tests that compare the chemical compositions of galaxies as obtained from H ii regions
and massive stars should be carried out in nearby, well-resolved systems. In doing so,
the comparison is limited to chemical elements that are measurable in both types of objects and that, contrary to nitrogen, are largely unaffected by rotational mixing (Maeder
233
234
Fabio Bresolin
& Meynet 2000, Hunter et al. 2007). For these reasons the abundances of oxygen can
be directly compared between ionized nebulae and young stars, mostly early-B dwarfs
(within the Milky Way and in the Magellanic Clouds), and brighter A and B supergiants
(in more distant galaxies). The expectation is that, once evolutionary effects in stars are
properly accounted for, the present-day abundances derived from young massive stars
and H ii regions agree within the uncertainties of the measurements and of the modeling.
In low-metallicity and generally chemically homogeneous galaxies, such as the Magellanic Clouds and a small number of dwarf irregulars of the Local Group, the agreement
found between H ii region and young star chemical abundances is satisfactory (e.g. Trundle et al. 2005, Bresolin et al 2006, Lee et al. 2006). The possibilities for comparison
are even fewer in the case of spiral galaxies, as only data for the Milky Way (Rolleston
etal. 2000, Deharveng et al. 2000, Daflon & Cunha 2004) and M33 (Vilchez et al. 1988,
Urbaneja et al. 2005a, Magrini et al. 2007, Rosolowsky & Simon 2008) have insofar allowed meaningful comparisons. Yet, these are perhaps more interesting cases, because the
metallicity in spiral galaxies can span a wide range, from the metal-rich nuclear regions
to the metal-poor outskirts, of up to 1 dex (as in the case of M101, Kennicutt etal. 2003).
235
to rely on gas-phase chemical abundances derived from strong-line methods (which are
subject to the uncertainties mentioned above), we obtained a sample of 28 H ii regions for
which we detected different auroral lines ([O iii] 4363, [S iii] 6312, [N ii] 5755), which
we used to derive electron temperatures and O, S, N, Ar and Ne abundances. This procedure yielded typical errors in O/H of less than 0.1 dex. The main conclusion from our
new work is that stellar and direct nebular abundances agree very well, in the range
spanned by our objects, 12 + log(O/H) approximately between 8.1 and 8.5, i.e. from the
metallicity of the Small Magellanic Clouds to an intermediate value between that of the
Large Magellanic Cloud and the Sun (Fig. 2).
References
Bono, G., Caputo, F., Fiorentino, G., Marconi, M., & Musella, I. 2008, ApJ, 684, 102
Bresolin, F., Kudritzki, R.-P., Mendez, R. H., & Przybilla, N. 2001, ApJ, 548, L159
Bresolin, F., Kudritzki, R.-P., Najarro, F., Gieren, W., & Pietrzynski, G. 2002, ApJ, 577, L107
Bresolin, F., Garnett, D. R., & Kennicutt, R. C. 2004, ApJ, 615, 228
Bresolin, F., et al. 2006, ApJ, 648, 1007
Bresolin, F. et al. 2009, ApJ, 700, 309
Chiappini, C., Matteucci, F., & Romano, D. 2001, ApJ, 554, 1044
Daflon, S., & Cunha, K. 2004, ApJ, 617, 1115
Deharveng, L., et al. 1988, A&AS, 73, 407
Halliday, C., et al. 2008, A&A, 479, 417
Hillier, D. J., & Miller, D. L. 1998, ApJ, 496, 407
Kennicutt, R. C., Bresolin, F., & Garnett, D. R. 2003 ApJ, 591, 801
Kewley, L. J., & Ellison, S. L. 2008, ApJ, 681, 1183
Kudritzki, R.-P., et al. 2008, ApJ, 681, 269
Larsen, S. S., Origlia, L., Brodie, J., & Gallagher, J. S. 2008, MNRAS, 383, 263
Larsen, S. S., Origlia, L., Brodie, J. P., & Gallagher, J. S. 2006, MNRAS, 368, L10
Lee, H., Skillman, E. D., & Venn, K. A. 2006, ApJ, 642, 813
236
Fabio Bresolin
Figure 2. The radial oxygen abundance gradient in NGC 300 obtained from H ii regions (circles)
and blue supergiants (star symbols: B supergiants; open squares: A supergiants) by Bresolin
et al. (2009), in terms of the isophotal radius R25 . The weighted regression line to the H ii
region data is shown by the continuous (green) line. The dotted lines show the 95% confidence
level interval for the regression line. The dashed line represents the weighted regression to the
BA supergiant star data. For reference, we include the oxygen abundances of the Magellanic
Clouds and the solar photosphere. The central abundance of NGC 300, obtained from a linear
extrapolation of the H ii region data, is sub-solar, 12 + log(O/H) = 8.57 0.02, and the slope of
the radial abundance gradient is 0.077 0.006 dex kpc1 . The metallicities of the B and A
supergiants analyzed by us are fully consistent with the nebular results.
Maeder, A., & Meynet, G. 2000, ARA&A, 38, 143
Magrini, L., Vilchez, J. M., Mampaso, A., Corradi, R. L. M., & Leisy, P. 2007, A&A, 470, 865
Maier, C., et al.. 2006, ApJ, 639, 858
Maiolino, R., et al. 2008, A&A, 488, 463
Matteucci, F., & Francois, P. 1989, MNRAS, 239, 885
McGaugh, S. S. 1991, ApJ, 380, 140
Modjaz, M., et al. 2008, AJ, 135, 1136
Pauldrach, A. W. A., Hoffmann, T. L., & Lennon, M. 2001, A&A, 375, 161
Pettini, M., et al. 2001, ApJ, 554, 981
Pettini, M., & Pagel, B. E. J. 2004, MNRAS, 348, L59
Puls, J., et al. 2005, A&A, 435, 669
Rix, S. A., et al. 2004, ApJ, 615, 98
Rolleston, W. R. J., Smartt, S. J., Dufton, P. L., & Ryans, R. S. I. 2000, A&A, 363, 537
Rosolowsky, E., & Simon, J. D. 2008, ApJ, 675, 1213
Savaglio, S., et al. 2005, ApJ, 635, 260
Tremonti, C. A., et al. 2004, ApJ, 613, 898
Trundle, C., & Lennon, D. J. 2005, A&A, 434, 677
Urbaneja, M. A., et al. 2005a, ApJ, 635, 311
Urbaneja, M. A., et al. 2005b, ApJ, 622, 862
Vila-Costas, M. B., & Edmunds, M. G. 1992, MNRAS, 259, 121
Vilchez, J. M., et al. 1988, MNRAS, 235, 633
Zaritsky, D., Kennicutt, Jr., R. C., & Huchra, J. P. 1994, ApJ, 420, 87
Abstract. We present Keck/HIRES spectra of six metal-poor stars in two of the ultra-faint
dwarf galaxies orbiting the Milky Way, Ursa Major II and Coma Berenices, and a Magellan/MIKE spectrum of a star in the classical dwarf spheroidal galaxy (dSph) Sculptor. Our
data include the first high-resolution spectroscopic observations of extremely metal-poor stars
([Fe/H] < 3.0) not belonging to the Milky Way (MW) stellar halo field population. We obtain
abundance measurements and upper limits for up to 26 elements between carbon and europium.
The stars span a range of 3.8 < [Fe/H] < 2.3, with the ultra-faints having large spreads
in Fe. A comparison with MW halo stars of similar metallicity reveals substantial agreement
between the abundance patterns of the ultra-faint dwarf galaxies and Sculptor and the MW halo
for the light, and iron-peak elements (C to Zn). This agreement contrasts with the results of
earlier studies of more metal-rich stars (2.5 . [Fe/H] . 1.0) in more luminous dwarfs, which
found significant abundance discrepancies with respect to the MW halo data. The abundances of
neutron-capture elements (Sr to Eu) in all three galaxies are extremely low, consistent with the
most metal-poor halo stars, but not with the typical halo abundance pattern at [Fe/H] & 3.0.
Our results are broadly consistent with a galaxy formation model which predicts that massive
dwarf galaxies are the source of the metal-rich component ([Fe/H] > 2.5) of the MW inner
halo, but we propose that dwarf galaxies similar to the dSphs are the primary contributors to
the metal-poor end of the metallicity distribution of the MW outer halo.
Keywords. stars: abundances, stars: Population II, Galaxy: formation, Galaxy: halo, galaxies:
dwarf, (galaxies:) Local Group, galaxies: stellar content, (cosmology:) early universe
1. Introduction
According to prevailing theories of the early Universe, the recombination of the plasma
of electrons and light nuclei (mostly H and He, with trace amounts of Li) emerging from
the Big Bang was followed by the Dark Ages, during which space was filled with gas and
radiation but no sources of light. This era ended with the formation and ignition of the
very first (Pop III) stars. These presumably very massive ( 100 M ; Bromm & Larson
2004) objects soon died as energetic supernovae. The freshly synthesized metals expelled
by those explosions provided the first effective coolants for the interstellar medium. This
pollution set the stage for the second generation of stars (Pop II), which included stars
over a wide range of masses. The low-mass (M 6 0.8 M ) early Pop II stars, which
237
238
A. Frebel et al.
formed from gas enriched by the first stars, live longer than a Hubble time and are thus
still observable. In their atmospheres, these stars preserve the characteristic chmemical
signature of the gas at the time of their birth prior to the formation of the Galaxy.
Metal-poor stars represent the local equivalent of the high-redshift Universe because
they supply us with direct information on conditions shortly after the Dark Ages (stellar
archaeology). We are thus provided with a unique window on early star formation and
the onset of galaxy formation. Detailed studies of these oldest stars can therefore address
a broad range of astrophysical issues, including the Pop III initial mass function, the
formation of the first low-mass stars, the yields of the first supernovae and relevant
nucleosynthesis processes and sites, the relation between chemistry and kinematics, and
whether the Milky Ways stellar halo was built from accreted satellites.
239
Figure 1. Abundance ratios ([X/Fe]) as a function of metallicity ([Fe/H]) for light and iron-peak
elements in comparison with those of Cayrel et al. (2004). Red squares indicate UMa II stars,
blue circles show ComBer stars, open black circles are the Cayrel et al. halo sample. The halo
star HD122563 is shown by a black diamond. (Taken from Frebel et al. (2009b).)
240
A. Frebel et al.
formed in much larger systems such as the Magellanic Clouds. Our results now support
a scenario where the ultra-faint dwarf galaxies contributed some individual metal-poor
stars that are now found primarily in the outer Galactic halo (although not exclusively).
However, these systems may not have been suffciently numerous to acount for the entire
metal-poor end of the Fe metallicity distribution of the Milky Way halo. Since the classical
dSphs contain more stellar mass and have been shown to also contain at least some of
the most metal-poor stars, they could have been a major source of the lowest-metallicity
halo content.
3. Conclusions
It appears that the hunt for the most metal-poor stars may have just begun since these
dwarfs host a large fraction of low-metallicity stars, perhaps even much higher than what
has so far been inferred for the Galactic halo (Sch
orck et al. 2008). The currently available
stars already provide important clues on early star and galaxy formation, the nature and
sites of the first stars, the onset of chemical enrichment and the relevant nucleosynthesis
processes. The detailed abundance patterns of the stars in UMa II, ComBer and Sculptor
are strikingly similar to that of the Milky Way stellar halo, thus renewing support for
dwarf galaxies as the building blocks of the halo. Future discoveries of additional faint
dwarf galaxies (for example with Skymapper) will enable the identification of many more
metal-poor stars in systems similar to UMa II and ComBer. But also the brighter dSphs
have to be revisited for their metal-poor content (Kirby et al. 2009). More stars at the
lowest metallicities are clearly desired to better quantify the emerging chemical signatures
and to solidify our understanding of the early Galaxy assembly process. Only in this way
can the hierarchical merging paradigm for the formation of the Milky Way be put on
firm observational ground. The next-generation optical telescopes will facilitate a great
advance in this field since the vast majority of the stars in any dwarf galaxy are too
faint to be followed-up with high-resolution spectroscopy on current 6-10 m telescopes.
These observations, however, are mandatory for investigating how the surviving dwarfs
are linked to their early counterparts that built the halo.
References
Bromm, V. & Larson, R. B. 2004, ARAA, 42, 79
Cayrel, R. et al. 2004, A&A, 416, 1117
Cohen, J. G. & Huang, W. 2009, ApJ, 701, 1053
Frebel, A. et al. 2005, Nature, 434, 871
Frebel, A., Kirby, E., & Simon, J. D. 2009a, submitted
Frebel, A., Simon, J. D., Geha, M., & Willman, B. 2009b, ApJ submitted, astro-ph/0902.2395
Fulbright, J. P., Rich, R. M., & Castro, S. 2004, ApJ, 612, 447
Helmi, A. et al. 2006, ApJ, 651, L121
Johnston, K. V. et al. 2008, ApJ, 689, 936
Kirby, E. N. et al. 2009, astro-ph/0909.3092
Kirby, E. N., Simon, J. D., Geha, M., Guhathakurta, P., & Frebel, A. 2008, ApJ, 685, L43
Koch, A., McWilliam, A., Grebel, E. K., Zucker, D. B., & Belokurov, V. 2008, ApJ, 688, L13
Norris, J. E. et al. 2008, ApJ, 689, L113
Robertson, B., Bullock, J. S., Font, A. S., Johnston, K. V. & Hernquist, L. 2005, ApJ, 632, 872
Sch
orck, T. et al. 2008, astro- ph/0809.1172, A&A in press
Searle, L. & Zinn, R. 1978, ApJ, 225, 357
Simon, J. D. & Geha, M. 2007, ApJ, 670, 313
Tolstoy, E., Hill, V. & Tosi, M. 2009, ARA&A, 47, 371
242
Our abundance analysis shows that Feh-Duf has enhanced carbon and heavy s-element
abundances, [hs/Fe]= +0.88. The yttrium abundance is low, [Y/Fe]= 0.07. Comparison of the oxygen-to-iron ratio in Feh-Duf ([O/Fe]= +0.10) and in Galactic halo stars
indicates that the [O/Fe] ratio in Feh-Duf is about 0.3 dex below the corresponding value
for the stars of the same metallicity in the Galaxy. Other -elements, such as Ca and Ti,
also have lower abundances ([Mg+Ca+Ti/3Fe] = 0.23).
Assuming the distance to the star to be d = 5.9 kpc we calculated Galactic spacevelocity components (U, V, W ) (Johnson & Soderblom 1987). Proper motions were taken
from NOMAD catalog (Zacharias et al. 2004). The obtained heliocentric space velocities
are (U, V, W ) = (116, 491, 106) km s1 . We transformed the V component of the
space velocity of Feh-Duf to the Galactic Reference Frame using (U, V, W ) = (9, 232, 7)
(Venn et al. 2004) which results in VGRF = 259 km s1 . This value of VGRF shows that
Feh-Duf has extreme retrograde motion which may be a sign that this star was accreted
by the Milky Way from a dwarf satellite galaxy.
3. Conclusions
Our analysis of the chemical abundances and kinematic properties of Feh-Duf showed
that:
Feh-Duf is a CH star. Even though its luminosity is compatible with the early AGB
phase, the binary nature of the carbon and s-element enrichment is preferable since the
envelope mass of a 0.8 M star is too small for the third dredge up phenomenon to occur
(Straniero et al. 2006). Feh-Duf is also a lead star since its lead-to-cerium ratio is high,
[Pb/Ce]= +0.70.
The extreme retrograde motion (VGRF = 259 km s1 ) may be a sign that this
star was accreted by the Milky Way from a dwarf satellite galaxy. Also as pointed out by
Marsakov & Borkova (2006), a star born in a monotonically collapsing single protogalactic
cloud could not be in a retrograde orbit.
The [/Fe] ratio in the photosphere of Feh-Duf is lower than typical ratio for halo
stars. Recently, Venn et al. (2004) carried out an analysis of the chemical abundances of
the stars in the Galaxy and in the Milky Way dwarf spheroidal (dSph) satellite galaxies.
They confirmed that the [/Fe] ratios of most stars in the dSph galaxies are lower than
Galactic stars of similar metallicity.
The large retrograde velocity and the unusual abundance ratios suggest that Feh-Duf
may have originated within a dSph galaxy that experienced a different nucleosynthetic
chemical evolution history and has been captured by the Milky Way.
References
Anders, E., & Grevesse, N. 1989, Geochim. et Cosmochim. Acta, 53, 197
Drake, N.A., & Pereira, C.B., 2008, AJ, 135, 1070.
Fehrenbach, Ch., & Duflot, M. 1981, A&A, 101, 226
Johnson, D.R.H., & Soderblom, D.R. 1987, AJ, 93, 864
Kurucz, R. L. 1993, CD-ROM 13, Atlas9 Stellar Atmosphere Programs and 2 km/s Grid (Cambridge: Smithsonian Astrophys. Obs)
Marsakov, V.A., & Borkova, T.V. 2006, Astron. Letters, 32, 545
Pereira, C.B., & Drake, N.A. 2009, A&A, 496, 791
Sneden, C. 1973, Ph.D. Thesis, Univ. of Texas
Straniero, O., Gallino, R., & Cristallo, S. 2006 Nuclear Physics A, 777, 311
Venn, K.A., Irwin, M., Shetrone, M.D., Tout, C.A., Hill, V., & Tolstoy, E. 2004, AJ, 128, 1177
Zacharias, N., Monet, D.G., Levine, S.E., et al. 2004, AAS, 205, 4815
Ver
onica Firpo1 , Guillermo Bosch1 , Guillermo H
agele1,2 , Angeles
I.
2
3
Daz , and Nidia Morrell
1
Facultad de Ciencias Astron
omicas y Geofsicas, Universidad Nacional de La Plata,
Paseo del Bosque s/n, B1900FWA, La Plata, Argentina.email: vfirpo@fcaglp.unlp.edu.ar
2
Departamento de Fsica Te
orica,C-XI, Universidad Aut
onoma de Madrid, Spain.
3
Las Campanas Observatory, Carnegie Observatories, La Serena, Chile.
Abstract. We present a detailed study of the physical properties of the nebular material in
multiple knots of the blue compact dwarf galaxy Haro 15. Using long slit and echelle spectroscopy, obtained at Las Campanas Observatory, we study the physical conditions (electron
density and temperature), ionic and total chemical abundances of several atoms, reddening and
ionization structure. The latter was derived by comparing the oxygen and sulphur ionic ratios
to their corresponding observed emission line ratios (the and plots) in different regions of
the galaxy. Applying direct and empirical methods for abundance determination, we perform a
comparative analysis between these regions.
Keywords. (ISM:) H ii Regions, starburst, Haro 15, abundances.
Observations
We obtained high dispersion (Echelle Spectrograph; = 0.148
A px1 at 5400
A equiv1
1
alent to 6.72 kms px ) and long-slit low resolution spectra (WFCCD; = 4.2
A px1
at 5400A) of five luminous knots in the BCD galaxy Haro 15 ( Fig. 1) at the du Pont
Telescope. We have obtained a good flux calibration in both groups of data and we can
confirm that the two different data groups are comparable in knot B and C.
Electron density and temperatures
We have followed the same procedure described in detail in H
agele et al. (2006), (2008),
H
agele (2008) to derive the physical properties of the nebular material in the observed
regions. The [OIII] 4363
A auroral emission line is only detected in knot B (longslit and
echelle) and knot C (longslit). It was only possible to derive T[OIII], T[SIII], T[OII]
and T[SII] from direct measurements for knot B. For the other knots, for which we are
not able to measure some lines such as [OII] 3727
A, [SIII] 9069,9532
A or auroral
lines, we have resorted to models that predict relationships between emission lines for
different temperatures, for example the relation between T[OII] and T[OIII] found by
Perez-Montero & Daz (2003). In the case of densities, the knots show density values
in the low density limit (ne < 100cm3 ), typical for of this kind of objects, with the
exception of knot E (ne 280cm3).
Chemical Abundances: O, S, N, Ar, Ne
Oxygen abundances and their uncertainties were derived for each observed knot using
the direct method, where it could be applied, or several empirical methods using the
strong emission lines present in the spectra. We notice a difference in the O/H ratio
between knots A and B. This difference was suggested by L
opez-S
anchez & Esteban
(2009) as the two objects might have had a different chemical evolution. Our results
support these differences between the two regions. Knot C shows oxygen abundance
similar to that of knot B, while the oxygen abundance derived for knot E is closer to the
243
244
V. Firpo et al.
Figure 1. Left panel: HST archive image showing the location of the observed regions. Right
panel: Contour maps of the continuum-subtracted H image (Cair
os et al. 2001). In this image
we show the slits distribution. The long-slit spectroscopy (1 wide) can be seen running through
Haro15-B and Haro15-C. Five pointings were observed with the echelle mode, shown as dark
rectangular (1x4) slits.
abundance calculated for knot A. The S/N in our knot D spectra is not as good as for
the other knots, and the quantities derived for this region should be used with caution.
In the cases that we can use the direct method, the total abundances have been derived
taking into account the unseen ionization stages of each element, resorting to the most
widely accepted ionization correction factors (ICF) for each species [X/H=ICF(X+i ) *
(X+i /H+ )] (see Perez-Montero et al. 2007, H
agele et al. 2008). The N and Ar abundances
are higher in knot A than in knot B, although the S abundance derived for knot A lies
between those derived from echelle and longslit of knot B.
Ionization Structure
The ratio between O+ /O2+ and S+ /S2+ denoted by is intrinsically related to the
shape of the ionizing continuum and depends on nebula geometry only slightly (Vlchez &
Pagel, 1988). We can include knot B in this diagram, lying in the highest excitation region.
The purely observational counterpart, the diagram (=[([OII]/[OIII])/([SII]/[SIII])),
where and are related through the electron temperature but very weakly. The position
of knot B in both diagrams shows a compatible ionization structure.
References
H
agele, G. F., Perez-Montero, E., Daz, A. I., Terlevich, E. and Terlevich, R.2006, M.N.R.A.S.,
372, 293-312
H
agele, G. F., Daz, A. I., Terlevich, E., Terlevich, R., Perez-Montero, E. and Cardaci, M. V.
2008, M.N.R.A.S., 383, 209-229
H
agele, G. F. 2008, PhD Thesis at Universidad Aut
onoma de Madrid, arXiv, 0908.4285v1
L
opez-S
anchez A. R.& Esteban C., 2009 arXiv, 0910.1578
Perez-Montero E. & Daz A. I., 2003 M.N.R.A.S., 346, 105
I., 2007 M.N.R.A.S., 381, 125
Perez-Montero, E., H
agele, G. F., Contini, T. and Daz, A.
Vlchez J. M. & Pagel B. E. J., 1988 M.N.R.A.S., 231, 257
Cair
os, L. M. and Caon, N.,Vlchez, J. M., Gonz
alez-Perez, J. N. and Mu
noz-Tu
n
on, C., 2001
ApJS, 136, 393
Abstract. The evolution of Fe-peak elements of several Local Group Dwarf Spheroidal Galaxies
are discussed based on the comparison between a chemical evolution model and obsevations. In
our scenario, the evolution of these galaxies are mainly controlled by a low star formation
efficiency coupled with very intense galactic winds. The low star formation rate gives rise to
the observed low metallicities and to [alpha/Fe] and [s/Fe] ratios below solar, whereas the
intense galactic winds are responsible for the sharp decrease observed in several abundance
ratios. The shape of the stellar metallicity distributions are defined by both parameters and
the observed data cannot be reproduced without evoking galactic winds. The same scenario
applied to a standard model fits very well several Fe-peak elements, with different nucleosynthesis
prescriptions for each set of elements.
Keywords. stars: abundances, galaxies: dwarf, Local Group
1. The model
We use the chemical evolution model for Dwarf Spheroidal (dSph) galaxies as described
in Lanfranchi & Matteucci (2004). Each galaxy model is specified by the prescriptions of
the star formation (SF) and by the galactic wind (GW) efficiency chosen to reproduce the
main features of these galaxies. The main characteristics of the models are the following:
one zone with instantaneous and complete mixing of gas inside this zone; the stellar
lifetimes are taken into account; the SFR is proportional to the gas mass and given by its
efficiency (); the wind efficiency (wi ) is proportional to the SFR. In the nucleosynthesis
prescriptions we adopted the yields of van den Hoeck & Groenewegen (1997) for low and
intermediate mass stars, Nomoto et al. (1997) for type Ia supernovae and Woosley &
Weaver (1995) for massive stars.
2. Results
The abundance ratios of Fe-peak elements observed in dSph galaxies are compared
to a standard model for these galaxies (Lanfranchi, Matteucci and Cescutti 2008). The
main aim of such model is to draw a general scenario, without taking into account the
particularities of each system, by putting all the data together in the comparison with
the predictions of the model. A model capable of reproducing several other observational
constraints, can also shed some light into the nucleosynthesis of Fe-peak elements. The
SF proceeds in one long episode of activity (8 Gyr) with low efficiency ( = 0.3 Gyr1 ).
When the thermal energy of the gas equals or exceeds the binding energy of the galaxy,
245
246
[Cr/Fe]
[Co/Fe]
[Mn/Fe]
-3
-2
-1
[Fe/H]
Figure 1. [Cr,Co,Mn/Fe] vs. [Fe/H] observed in dSphs compared to the predictions of the
standard model.
an intense galactic wind (wi =10) develops, removing a large fraction of the gas content
of the galaxy. The SFR decreases then substantially, almost halting the formation of
new alpha and r-process elements and of stars with metallicities higher than the one of
the ISM when the wind began. This scenario allows the model to reproduce very well
a few observational constraints: [alpha/Fe], [Fe-peak/Fe], and [s-r/Fe] ratios, the total
mass, the gas mass and the stellar metallicity distributions. In Figure 1 the predictions
of the standard model for Cr, Co, and Mn are compared to the observations (Shetrone
et al. 2001, 2003, Sadakane et al. 2004, Geisler et al. 2005) with a very good agreement.
The same model, which reproduces very well the evolution of alpha and neutron capture
elements for the set of dSph as a whole (LM03, LMC08), fits the fe-peak elements, however
with different yields for different elements. Whereas for V, Cr, and Mn, we adopted the
metallicity dependent yields of WW95, for Ti, Sc, Co, Zn, and Ni the WW95 yields with
solar metallicity are used. The minor offsets in the plots are caused by the particularities
in the SFHs and SFRs of each galaxy. Because of the differences in the evolution of each
system, one should interprete these plots not as evolutionary diagrams: they only suggest
a trend in the evolution of these elements instead of tracing the evolution.
References
Geisler D., Smith V.V., Wallerstein G., Gonzalez G.,Charbonnel C., 2005, AJ, 129,1428
Lanfranchi, G., Matteucci, F., 2003, MNRAS, 345, 71
Lanfranchi, G., Matteucci, F., 2004, MNRAS, 351, 1338
Lanfranchi, G., Matteucci, F., Cescutti, 2008,A&A, 481, 635
Nomoto, K., et al., 1997, Nucl. Phys. A, 616, 79
Sadakane, K., et al., 2004, PASJ, 56, 1041
Shetrone, M., Cote, P., Sargent, W.L.W., 2001, ApJ, 548,59
Shetrone, M., Venn, K.A., Tolstoy, E., Primas, F., 2003, AJ,125, 684
van den Hoeck, L.B., Groenwegen, M.A.T., 1997, A&AS, 123, 305
Woosley, S.E., Weaver, T.A., 1995, ApJS, 101, 181
1. Introduction
The study of the chemical evolution of nearby spiral galaxies is very important to
improve our knowledge about the main ingredients used in chemical evolution models
and test the basic assumptions made for modelling our Galaxy. M31 and M33 are other
spiral members of the Local Group of galaxies and during the last years many observational studies have been made to investigate the chemical and dynamical properties
of these neighbour systems. New surveys (Braun et al. 2009, Magrini et al. 2007,2009)
contributed to the analysis of different stellar populations and provided more accurate
data to constrain the chemical evolution models.
The disks of M31 and M33 are similar to the Milky Way disk but some observational
constraints, such as the present day gas distribution, can only be explained by assuming different star formation histories for these galaxies. The star formation rate (SFR)
is one of the most important parameters regulating the chemical evolution of galaxies
(Matteucci & Recchi 2001, Boissier et al. 2003) together with the initial mass function.
Another important mechanism is the inside-out disk formation that is very important
to reproduce the radial abundance gradients (see Colavitti et al. 2008 for the most recent
247
248
paper on the subject). A faster formation of the inner disk relative to the outer disk
was proposed by Matteucci & Francois (1989) and supported in the following years by
Boissier & Prantzos (1999) and Chiappini et al. (2001).
NASA Ames Research Center, M.S. 245-6, Moffett Field, CA 94035-1000, USA
email: rubin@cygnus.arc.nasa.gov
2
Orion Enterprises
3
Kavli Institute for Astronomy & Astrophysics, Peking University
4
Rice University
5
University Munich
Abstract. We observed several H ii regions in the dwarf irregular galaxy NGC 6822 using the
infrared spectrograph on the Spitzer Space Telescope. Our aim is twofold: first, to examine the
neon to sulfur abundance ratio in order to determine how much it may vary and whether or not,
it is fairly universal; second, to discriminate and test the predicted ionizing spectral energy
distribution between various stellar atmosphere models by comparing with our derivation of the
ratio of fractional ionizations involving neon and sulfur. This work extends our previous similar
studies of H ii regions in M83 and M33 to lower metallicities.
Keywords. ISM: abundances, H ii regions, stars: atmospheres, galaxies: individual (NGC 6822)
1. Synopsis
We have observed emission lines of [S iv] 10.51, [Ne ii] 12.81, [Ne iii] 15.56, and [S iii]
18.71 m in a number of extragalactic H ii regions with the Spitzer Space Telescope.
Previous papers (Rubin et al. 2007 and 2008) presented our data and analysis for the
substantially face-on spiral galaxies M83 and M33. We undertook a similar program
for H ii regions in the dwarf irregular galaxy NGC 6822. The observations were made
with the Infrared Spectrograph with the short wavelength, high resolution module. The
above set of four lines is observed cospatially, thus permitting a reliable comparison of
the fluxes. From the measured fluxes, we determine the ionic abundance ratios including
Ne++ /Ne+ , S3+ /S++ , and S++ /Ne+ . By sampling the dominant ionization states of Ne
(Ne+ , Ne++ ) and S (S++ , S3+ ) for H ii regions, we can estimate the Ne/S ratios. From
our own Cycle 4 data, we had clear detections of all 4 program lines in only 5 sources
after subtracting a background off. We also measured these 4 ionic lines in four large
H ii regions (Hu I, III, V, X) in NGC 6822 previously observed by the SINGS team.
The results from the analysis of these new data have been added to those we published
previously. There is no variation in the Ne/S ratio with RG . Because of the low metallicity
and high ionization of H ii regions in NGC 6822, the Ne/S ratios we derive are likely a
robust estimate of the true value. The median Ne/S ratio derived for the 9 H ii regions
in NGC 6822 is 10.8. We compare this with previous values. Figure 1 presents these new
data along with the previous data sets. We continue to conclude that the Ne/S ratio is
significantly higher than the controversial solar value.
Our observations may also be used to test the predicted ionizing spectral energy distribution of various stellar atmosphere models. We compare the ratio of fractional ioniza249
250
tions <Ne++ >/<S++ >, <Ne++ >/<S3+ >, and <Ne++ >/<Ne+ > vs. <S3+ >/<S++ >
with predictions made from our photoionization models using several of the state-of-theart stellar atmosphere model grids. The trends of the ionic ratios established from the
prior M83 and M33 studies continue to higher ionizations with the present NGC 6822
objects, and are remarkably similar.
M83
M33
NGC 6822
Wu et al. (2008)
Lebouteiller et al. (2008)
NGC 3603
30 Dor
N 66
M33 (22 sources)
Wu (9 sources)
Lebouteiller (23 sources)
40
30
HuIII
20
18
10
10
28e
HuI
HuX
HuV
50
30
HuIII
20
10
5e
M83
M33
NGC 6822
Wu et al. (2008)
Lebouteiller et al. (2008)
NGC 3603
30 Dor
N 66
M33 (22 sources)
Wu (9 sources)
Lebouteiller (23 sources)
40
(Ne+ + Ne++)/(S++ + S3+)
50
HuI
HuX
HuV
18
10
28e
2e
4
6
Ne++/Ne+
10
0.0
2e
5e
0.2
0.4
0.6 0.8
S3+/S++
1.0
1.2
1.4
Figure 1. Colour version in on-line edition only. Plot of Ne/S vs. Ne++ /Ne+ [left panel] and
S3+ /S++ [right panel]. Our prior M33 results are shown as black stars for the 22 sources where
we detected all four lines. There are various linear least-squares fits to the data, all described
in Rubin et al. (2008). Space precludes a repeat here. The results from our prior M83 study
are shown as circles. These data demonstrate a huge variation in the inferred Ne/S ratio at low
ionization. The orange squares show the Wu et al. (2008) data for blue compact dwarf galaxies,
as reanalyzed with our program. We show only 9 points, those objects where they actually
detected all four lines: [S iv], [Ne ii], [Ne iii], and [S iii]. The median Ne/S for the 9 galaxies is
14.0 very close to the Orion value of 14.3 (Simpson et al. 2004) shown as the dashed line. The
Lebouteiller et al. (2008) data were also reanalyzed and are presented as follows: NGC 3603
(red asterisks), 30 Dor (green triangles), and N 66 (blue diamonds). The median Ne/S ratios
for each are 14.6, 11.4, and 10.1, respectively, possibly indicating a decreasing trend with lower
metallicity. The 4-sided violet stars show our new Spitzer preliminary results for NGC 6822.
Only five H ii regions labeled in black with their KD number (Killen & Dufour 1982), have all
four lines measured. Our remeasure of the four SINGS giant H ii regions are labeled in violet.
Acknowledgements
Support is from Spitzer Space Telescope Cycle 4 program 40910.
References
Killen, R. M., & Dufour, R. J. 1982, PASP 94, 444
Lebouteiller, V., Bernard-Salas, J., Brandl, B. R., Whelan, D. G., Wu, Y., Charmandaris, V.,
Devost, D., & Houck, J. R. 2008, ApJ 680, 398
Rubin, R. H., et al. 2007, MNRAS 377, 1407
Rubin, R. H., et al. 2008, MNRAS 387, 45
Simpson, J.P., Rubin, R.H., Colgan S.W.J., Erickson, E.F., & Haas, M.R. 2004, ApJ 611, 338
Wu, Y., Bernard-Salas, J., Charmandaris, V., Lebouteiller, V., Hao, L., Brandl, B. R., & Houck,
J. R. 2008, ApJ 673, 193
Abstract. The corotation radius in a spiral galaxy is the place where the spiral pattern speed
has the same velocity of the rotation curve. By compiling results coming from the literature for
20 spiral galaxies we verified a strong correlation between the radius of the minima or inflections
of the metallicity distribution and the corotation radius.
Keywords. galaxies: abundances, dynamics, spiral, corotation, metallicity gradient.
1. Introduction
The gradients of metallicity observed in disk galaxies provide important constraints to
the models of the chemical evolution of these objects. It is frequently observed, in a first
approximation, that spiral galaxies have a declining gradient of metallicity. However, it is
not rare to observe galaxies with a clear change in the slope of their gradient of metallicity
(eg. Zaritsky et al., 1994). Considering that spiral arms are the main contributors to the
formation of massive stars, which in turn are responsible for the metal enrichment of
the interstellar medium, Mishurov et al.(2002) proposed a simple model for the chemical
evolution of spiral galaxies that effectively introduces the role of the spiral arms in the
star formation rate. This model is able to explain the plateau observed in the metallicity
distribution of our galaxy, which would be a consequence of the minimum expected in the
metallicity distribution produced at corotation. In this work we compiled a set of spiral
galaxies for which the rotation curves, the gradients of metallicity and the corotation
radii (or the spiral pattern speeds) have been evaluated in the literature, to verify if the
variations in their metallicity gradient are also related to the corotation.
252
determined using the derivatives of the fitted curve. The uncertainties were calculated
by propagating the errors in the fitted parameters.
90
Typical Uncertainty
Typical Uncertainty
20
NGC0628
75
30
15
NGC5236
NGC5457
= (0.87 0.08)
NGC5033
+ (2.2 1.4)
Number
Number
NGC0598
NG NGC1365
C7
47
dZ
9
NGC2903
NGC6946
NGC2543
NGC4321
Milky Way
NGC1232
NGC2403
Milky Way
NGC5194
45
NGC3031
dZ
RdZ [kpc]
10
NGC4254
NGC3319
NGC5055
NGC0224
IC0342
[km/s/kpc]
60
15
4
2
0
4
2
0
10
15
RCR [kpc]
20
25
15
30
45
60
[km/s/kpc]
75
90
Figure 1. Correlation between the minima or inflection radii in the metallicity distribution
and the corotation radii (left) and the equivalent graph for the angular velocities associated to
breaks in the gradients of metallicity and the velocities of spiral pattern speed of all sampled
galaxies (right). The Sun symbol represents the results for our galaxy. The histograms at the
bottom of each graph show the statistics (number of points per bin) of the data.
In this figure the continuous lines are the linear regressions using robust statistics methods, which show the strong correlation between the radii of the breaks in the metallicity
gradients and of corotation. The fitted lines are very near to the one-to-one correspondence represented by the dashed lines. Concerning other interesting relations discovered
in the present study, the p distribution shows a peak around 20 km/s/kpc, revealing
that our galaxy is a typical one in this aspect. A surprising correlation between the corotation radius and the Elmegreen spiral arm class was also found, and when it is applied
to our Galaxy, it results in a spiral pattern similar to that revealed by recent observations
from the Spitzer telescope (Churchwell et al., 2009).
References
Churchwell, E. et al. 2009, PASP, 121, 213
Mishurov, Y.N., Lepine, J.R.D. & Acharova, I.A. 2002, ApJ, 571L, 113M
Scarano Jr, S. 2008, PhD thesis, Universidade de Sao Paulo.
Zaritsky, D., Kennicutt, Jr., R.C. & Huchra, J.P. 1994, ApJ, 420, 87Z
Session IV
1. Introduction
Eggen, Lynden-Bell & Sandage (1962; hereafter ELS) provided a first coherent picture
for the formation of our Galaxy as the outcome of the dissipative collapse of a selfgravitating perturbed gas. Stars formed during the collapse would show a quite spherical
distribution and the remaining gas would settle in an equilibrium disk configuration
where star formation would proceed in a more steady pace. In spite of its simplicity,
this scenario has some serious drawbacks, since its halo and disk components do not
properly resemble the real halo and disk(s) inferred from the observations. For instance,
it is presently known that the age range in the halo globular clusters is larger than
the expected time scale for dissipative collapse formation of the halo (Sarajedini et al.
1997; De Angeli et al. 2005; Marn-Franch et al. 2009). Moreover, some clear and smooth
chemokinematical trends found by ELS for the Galaxy are likely to have been formed by
the particular local mixture of the main Galactic populations contributing to the stellar
content of the solar neighborhood (Carollo et al. 2007, 2009).
CDM models aimed at explaining the observed large scale structure of the distribution of galaxies have a much better performance at issues that are seen as failures of the
monolithic dissipative collapse model. These give a quite intuitive explanation for the
formation of the Galaxy as a hierarchical process, driven by the successive mergers of
smaller subgallactic units into larger ones, making room for the existence of a population of faint satellites in the outskirts of the present-day bright gallaxies. Nevertheless,
the predicted number of these satellites in the Local Group is a order of magnitude or
more larger than the presently known number of its satellite dwarfs (Moore et al. 1999;
Kravtsov et al. 2004; Strigari et al. 2007; see also Tollerud et al. 2008 and Koposov et al.
2009). Several attempts have been made to solve this missing satellite problem, ranging
from deep observations which could increase the present census of Local Group faint
satellites (Willman et al. 2005a,b; Martin et al. 2006; Zucker et al. 2006a,b; Walsh et al.
2007; McConnachie et al. 2008; Liu et al. 2008) to reassessment of events in the early
Universe that could hinder the lighten up of these, mainly dark, satellites (Simon & Geha
2007; Donghia & Lake 2008; Madau et al. 2008; Bovill & Ricotti 2009; Yang et al. 2009).
Despite this inconsistency, it is clear that the Galaxy has experienced a number of
merger events in its quite recent past, on account of the presence of nearby satellite
galaxies. Their number should have been higher in the past and the outcome of their
255
256
Helio J. Rocha-Pinto
interactions with the Milky Way (MW) can probably still be found as statistically significant substructures somewhat detached from the smooth stellar population distributions
of the typical, canonical Galactic components. Halo substructures are also found in other
spiral galaxies (e.g., Shang et al. 1998, Ibata et al. 2007, 2009; Herrmann et al. 2009;
Martnez-Delgado et al. 2009a,b), and may be a quite common phenomenon attesting
the tumultuous hierarchical formation of these systems.
It happens that much of the definition of what is a canonical Galactic component is
somewhat biased towards what we expect from an ELS-type galaxy, especially in the
case of the Galactic halo, which is expected to be an old, well-mixed, kinematically hot
and non (or slightly) rotating component. Any deviation from this particular type of
halo is a possible signature of past merger events between the MW and its satellite
companions. This can lead to conflicting interpretations: the Virgo stellar overdensity
(Juric et al. 2008), the second largest MW halo substructure in angular size in the sky,
can be the signature of a triaxial halo instead of the cloudy debris of a former satellite
galaxy (Newberg et al. 2007; but see Bell et al. 2007).
Halo tidal substructures are diverse in shape and properties, as shown by Johnston
et al. (2008) in a study of the morphology of tidal debris in merger-built haloes. Three
main halo groups as well as transition groups can be distinguished, according to
the distribution of the debris system around the galaxy and in phase space: the mixed,
cloudy and great circles morphology. Each of these halo morphologies is produced by a
particular accretion history such that the characterization the phase-space and chemical
content of the halo substructure allows the uncovering of the MW accretion history.
257
if the stellar halo is triaxial (Newberg et al. 2007) or towards regions having selective
reddening RV very different from the RV of their symmetric fields.
3. Sagittarius et alli
Figure 1 shows the approximate sky distribution of most of the presently known halo
substructure in an aitoff projection using Galactic coordinates. By far, the largest debris
feature in the Galactic halo is created by the disruption of the Sgr dSph (Ibata et al.
1994, Majewski et al. 2003, 2004, Newberg et al. 2002). Debris from Sgr are distributed
according to a mixgreat circle transition morphology with at least two clearly visible
wraps (Belokurov et al. 2006). The two wraps of the Sgr tidal stream are displaced
because of the precession of the Sgr orbital plane, and this puts a strong constraint to
the shape of the Galactic dark halo being spherical or slightly prolate (Fellhauer et al.
2006), although the analysis of other parts of this debris system suggest other halo shapes
(Helmi 2004a,b; Law et al. 2003; Johnston et al. 2005). Recently, Law et al. (2009) have
suggested that this conflict can be solved the dark matter halo is triaxial.
In spite of being a dwarf spheroidal galaxy, Sgr had a complex chemical enrichment
history. This can be attested by the 1.58 < [Fe/H] < 0.71 range shown by its stars
(Marconi et al. 1998) and the 2.0 < [O/H] < 0.2 range found in its planetary nebulae
Figure 1. Aitoff projection showing the distribution of some of the presently known halo substructure in Galactic coordinates: Sgr tidal stream (Majewski et al. 2003), candidate Monoceros
fields (Rocha-Pinto et al. 2003), Monoceros Northern and Southern arcs (Martin et al. 2004;
Rocha-Pinto et al. 2006), Canis Major (Martin et al. 2004), Argo (Rocha-Pinto et al. 2006),
TriAnd 1 and 2 clouds (Rocha-Pinto et al. 2004; Martin et al. 2007), Virgo OD (Juric et al.
2008), Virgo SS (Keller et al. 2009), Pisces (Watkins et al. 2009), Hercules-Aquila cloud (Belokurov et al. 2007a), Orphan stream (Belokurov et al. 2006), GD-1 stream (Grillmair & Dionatos
2006), Cetus stream (Newberg et al. 2009), Anticenter stream tributaries and Eastern Band
Structure (EBS; Grillmair 2006b). Acheron, Cocytos, Lethe and Styx (Grillmair 2009) not
shown in this plot are located near the region where the Orphan and GD-1 cross each other.
See the eletronic version of this paper for a colored version of this plot.
258
Helio J. Rocha-Pinto
259
Gyr ago. In Pe
narrubia et al. (2005)s model TriAnd is explained as a past wrap of the
Monoceros debris system. However, Majewski et al. (2010; in this proceedings) argues
that TriAnd is chemically very distinct from Mon.
A large set of halo substructures was unveiled in the analysis of the SDSS stellar
content, ranging from several cold streams to new ultrafaint satellite galaxies. Among
them, a large debris system is the Virgo Stellar Overdensity (VOD; Juric et al. 2008),
whose existence has been suspected since the early 2000 from an overdensity of RR Lyrae
in the QUEST survey (Vivas et al. 2001). It has the morphology of a star cloud, like
TriAnd, with no apparent center. VOD could be part of the Sgr leading tail (MartnezDelgado et al. 2007), but Newberg et al. (2007) and Yanny et al. (2009) argue that the
Sgr tail passes at a different distance along the same line of sight. Just like for Monoceros
and TriAnd, smaller substructures (including the Virgo Stellar Stream, a.k.a VSS) seem
to share the VOD distance range and have possibly an independent origin (Duffau et al.
2006; Prior et al. 2009; Keller et al. 2009).
Other cloudy SDSS overdensities have been reported in the literature: Hercules-Aquila
(HerAql; Belokurov et al. 2007a) and Pisces (Sesar et al. 2007; Watkins et al. 2009;
Kollmeier et al. 2009). HerAql occupies an angular size of 80 50 and can be seen
above and below the Galactic plane (Belokurov et al. 2007a), lying between 10 to 20
kpc from the Sun toward l 40 . Belokurov et al. (2007a) proposes that it could be
a new structural component of the inner halo. However, the southern regions of this
overdensity (not covered by the presently available SDSS data) still need to be properly
mapped before confirming this hypothesis. The Pisces overdensity has been initially seen
by Sesar et al. (2007) as an overdensity of RR Lyrae stars in SDSS Stripe 82. Later,
Watkins et al. (2009) independently found it and christened it that way. It is centered
at (l, b) (80 , 55 ), at a distance of 80 kpc. Watkins et al. (2009) estimates its
total mass as a few 104 M . Kollmeier et al. (2009) have spectroscopically confirmed the
existence of the overdensity and suggests that Pisces can be a new Milky Way satellite
dwarf galaxy possibly in the process of disruption.
A great deal of attention has been given to the discovery of cold streams in SDSS
on account of the deepness of the survey which allows reaching the distant turnoff stars
of these narrow substructures. The most well studied of these are the Orphan Stream
(Belokurov et al. 2007b; Grillmair 2006a) and the 63 -long Grillmair-Dionatos 1 Stream
(Grillmair & Dionatos 2006). Other reported discoveries are Acheron, Cocytos, Lethe and
Styx (Grillmair 2009) a creative pattern of designations to avoid the boring repetition
of constellation names in several different astronomical objects and Cetus (Newberg
et al. 2009). Structures like these can persist over timescales of 2-4 Gyr (Younger et al.
Figure 2. Distance and metallicity ranges for some of the presently known halo substructures.
260
Helio J. Rocha-Pinto
2008) and are particularly interesting because they trace the orbit of their parental body
(e.g., Willett et al. 2009) and allow the constraining of the Galactic potential (Koposov
et al. 2009; Eyre & Binney 2009).
It is not very easy to find halo structures by exclusively looking for correlations between
chemical anomalies and spatial positions. In spite of it, Roederer (2008) has showed that
the kinematically diverse outer halo population is also chemically diverse, suggesting
that they were formed in regions where chemical enrichment was dominated by local
SN events. Also, there are some chemically distinct overdensities in the SDSS data as
Monoceros, Virgo and an unnamed significantly overabundant region in Figure 9a from
Ivezic et al. (2009) at 2-3 kpc above the Galactic plane.
261
262
Helio J. Rocha-Pinto
Martnez-Delgado, D., Pe
narrubia, J., Gabany, R. J., et al. 2009a, ApJ, 689, 184
2003, ApJ, 660, 1264
Martnez-Delgado, D., Pe
narrubia, J., Juric, M., Alfaro, E. J., Ivezic, Z.
Martnez-Delgado, D., Pohlen, M., Gabany, R. J., et al. 2009b, ApJ, 692, 955
Mateu, C., Vivas, A. K., Zinn, R., Miller, L., & Abad, C. 2009, AJ, 137, 4412
McConnachie, A. W., Huxor, A., Martin, N. F., et al. 2008, ApJ, 688, 1009
Moitinho, A., V
azquez, R. A., Carraro, G., et al. 2006, MNRAS, 368, L77
Momany, Y., Zaggia, S., Bonifacio, P., et al. 2004, A&A, 421, L29
Momany, Y., Zaggia, S., Gilmore, G., et al. 2006, A&A, 451, 515
Moore, B., Ghigna, S., Governato, F., Lake, G., et al. 1999, ApJ, 524, L19
Newberg, H. J., Yanny, B., Cole, N., et al. 2007, ApJ, 668, 221
Newberg, H. J., Yanny, B., Rockosi, C., et al. 2002, ApJ, 569, 245
Newberg, H. J., Yanny, B., & Willett, B. A. 2009, ApJ, 700, L61
Pe
narrubia, J., Martnez-Delgado, D., Rix, H.-W., et al. 2005, ApJ, 626, 128
Prior, S. L., Da Costa, G. S., Keller, S. C., Murphy, S. J. 2009, ApJ, 691, 306
Robin, A. C., Reyle, C., Derri`ere, S., & Picaud, S. 2003, A&A, 409, 523
Rocha-Pinto, H. J., Majewski, S. R., Skrutskie, M. F., Crane, J. D. 2003, ApJ, 594, L115
Rocha-Pinto, H. J., Majewski, S. R., Skrutskie, M. F., et al. 2004, ApJ, 615, 732
Rocha-Pinto, H. J., Majewski, S. R., Skrutskie, M. F., et al. 2006a, ApJ, 640, L147
Rockosi, C. M., Odenkirchen, M., Grebel, E. K., et al. 2002, AJ, 124, 349
Roederer, I. U. 2009, AJ, 137, 272
Sales, L. V., Helmi, A., Starkenburg, E., et al. 2008, MNRAS, 389, 1391
Sarajedini, A., Chaboyer, B., & Demarque, P. 1997, PASP, 109, 1321
Shang, Z., Zheng, Z., Brinks, E., et al. 1998, ApJ, 504, L23
Lupton, R. H., et al. 2007, AJ, 134, 2236
Sesar, B., Ivezic, Z.,
Simon, J. D., & Geha, M. 2007, ApJ, 670, 313
Starkenburg, E., Helmi, A., Morrison, H. L., et al. 2009, ApJ, 698, 567
Strigari, L. E., Bullock, J. S., Kaplinghat, M., et al. 2007, ApJ, 669, 676
Tollerud, E. J., Bullock, J. S., Strigari, L. E., Willman, B. 2008, ApJ, 688, 277
Vivas, A. K., Zinn, R., Andrews, P., et al. 2001, ApJ, 554, L33
Walsh, S. M., Jerjen, H., Willman, B. 2007, ApJ, 662, L83
Walsh, S. M., Willman, B., Jerjen, H. 2009, AJ, 137, 450
Watkins, L. L., Evans, N. W., Belokurov, V., et al. 2009, MNRAS, 398, 1757
Willett, B. A., Newberg, H. J., Zhang, H., Yanny, B., & Beers, T. C. 2009, ApJ, 697, 207
Willman, B., Dalcanton, J. J., Martnez-Delgado, D., et al. 2005, ApJ, 626, L85
Willman, B., Blanton, M. R., West, A. A., Dalcanton, J. J., Hogg, D. W., Schneider, D. P.,
Wherry, N., Yanny, B., & Brinkmann, J. 2005, AJ, 129, 2692
Yang, X., Mo, H. J., van den Bosch, F. C. 2009, ApJ, 693, 830
Yanny, B., Newberg, H. J., Johnson, J. A., et al. 2009, ApJ, 700, 1282
Younger, J. D., Besla, G., Cox, T. J., Hernquist, L., Robertson, B., Willman, B. 2008, ApJ, 676,
L21
Zucker, D. B., Belokurov, V., Evans, N. W., et al. 2006a, ApJ, 643, L103
Zucker, D. B., Belokurov, V., Evans, N. W., et al. 2006a, ApJ, 650, L41
Abstract. This paper presents a detailed kinematic and chemical analysis of 16 members of
the Kapteyn moving group. The group does not appear to be chemically homogenous. However,
the kinematics and the chemical abundance patterns seen in 14 of the stars in this group are
similar to those observed in the well-studied cluster, Centauri. Some members of this moving
group may be remnants of the tidal debris of Cen, left in the Galactic disk during the merger
event which deposited Cen into the Milky Way. A more detailed version of this work can be
found in Wylie de-Boer et al. 2009.
Keywords. stars: abundances, stars: kinematics, globular clusters: individual( Centauri),
Galaxy: structure
1. Introduction
The most massive Galactic globular cluster, Cen, has several unique physical properties which suggest that there are very significant differences in star formation histories,
enrichment processes and structure formation between Cen and other normal globular clusters Bekki & Freeman (2003). A commonly accepted scenario of formation of
Cen is that it is the surviving nucleus of an ancient dwarf galaxy, the outer envelope of
which was entirely removed by tidal stripping as it was accreted by the Galaxy (Bekki
& Freeman (2003) and references therein). Dinescu (2002) also provided a theoretical
prediction of where the Cen group would appear kinematically using the metal-poor
star sample of Beers et al. (2000). This region covers a large interval of energy over an
angular momentum range from about Lz = -200 to -600 kpc km s1 .
We have recently investigated the chemical properties of stars in the Kapteyn moving
group, first introduced by Eggen (1962). The stars of the Kapteyn group, as most recently
tabulated by Eggen (1996), are mostly metal-poor and in retrograde galactic orbits, so
they were identified as a halo moving group. It turned out that some of the group stars
show chemical peculiarities similar to those seen in Cen, and we will argue that the
Kapteyn group may be part of the Cen debris.
2. Kinematics Results
The Lindblad diagram is shown in Figure 1 for the Kapteyn group stars and the
Gratton peak sample from Gratton et al. (2003). The large open circles represent the
Gratton peak stars. The large solid circles are for the retrograde Kapteyn group stars and
the elongated cloud of smaller coloured points around each solid circle show the outcome
263
264
of the Monte Carlo simulations on uncertainties, and outline the 1 confidence level
for each Kapteyn group star. As expected, the errors in Lz and E are highly correlated.
The open star is for Cen. The region in which Dinescu (2002) suggested that candidate
stars from Omega Cens host galaxy could lie is outlined by the dashed line. This range
depends on the adopted velocity of the Local Standard of Rest (taken here to be 220
km s1 ). The dotted lines indicate the Dinescu range using recent extreme values of the
LSR velocity.
-2000
-1000
1000
2000
Figure 1. Lindblad diagram for Kapteyn group and Gratton peak. The black filled circles are
the 16 Kapteyn values and the open star indicates the position of Centauri. The black open
circles are stars from the Gratton Cen peak. The smaller dots represent the error distribution
in E,Lz for the Kapteyn group stars, as discussed in the text. The different colors used are
intended only to provide clarity between the distributions for different stars.
The Kapteyn group stars in Figure 1 appear to lie mainly in two bands. The appearance
of these bands is accentuated by the correlated errors on E and Lz . Figure 1 shows that
nine of the Kapteyn stars lie within the box covering the Dinescu et al. region of Cen
candidates. Within the 1- uncertainties, this number increases to 12 members of the
Kapteyn group. There are four prograde stars in the Kaptyen sample. Two of these stars
lie such that the ends of their 1- distribution fall near the extreme of the Cen region,
leaving only two stars (HD13979 and CD -30 1121) which are clearly separate from the
Cen candidate region.
In the following analysis, the 16 Kapteyn stars are partitioned into two groups. From
Figure 1, fourteen stars are taken as kinematically associated with Cen, while two of
the stars are likely to have no kinematic connection to Cen debris.
3. Abundance Results
First we compare the abundances in Cen and in the field halo stars. Then we compare
the Kapteyn group abundances with those in the cluster and the field halo, in order to
265
ascertain whether the Kapteyn group stars show the abundance patterns of either of these
two populations. The main point of interest here is that the Kapteyn stars which appear
to be kinematically related to Cen share distinct abundance similarities with those stars
from previous Cen studies. The abundances for the potential Cen candidates among
the Kapteyn stars could be drawn from the same abundance distribution as the Cen
stars, although we note that there is much more scatter in the abundance distributions
for most elements in the Cen stars. Whatever the reason for this large scatter in the
cluster star abundances, it seems clear that the Kapteyn candidates show abundance
anomalies relative to the field halo which are similar to those shown by the stars of
Cen. The similarities between the Kapteyn candidates and the Cen stars, relative to
the field halo, are clearest in Na, Mg and the s-elements.
Although the copper abundance was derived for only five of the stars in our sample,
the results are worth mentioning. Cen is known for its unique copper signature. At a
constant value of [Cu/Fe] 0.6, it is distinct from the halo stars in the same metallicity
range, which show an increase with metallicity up to [Cu/Fe] = -0.3 as discussed by Smith
et al. (2000). Of the four stars for which copper was measured in our sample, all show
deficiencies, with [Cu/Fe] values between -0.47 and -0.60. This adds weight to the case
that this group of stars come from the same population as Cen.
Statistical tests were undertaken on the abundance distributions to determine the
likelihood that the three groups of stars (Kapteyn group, Cen and the halo) come from
the same population. The Cen group consisted of stars from the Norris & Da Costa
(1995) and Smith et al. (2000) studies and includes the stars from Gratton et al. (2003)
only for elements in common with this study, Na, Mg and Ca. The halo group is taken
from the Venn et al. (2004) compilation and is defined as being any halo population star
that falls within the plausible range of [Fe/H] values for Cen of [Fe/H] = -0.5 to -2.5.
Figure 2 shows the mean and standard deviation of the mean, mean , graphically for the
five elements studied, with the Kapteyn group shown as filled red squares, the Cen
stars as filled magenta triangles and the field halo stars as open green circles. It is evident
from this figure that in all elements, except Ca and to a lesser extent Mg, the field halo
stars are chemically distinct from Cen. It is also clear that, except for Ca, the Kapteyn
group and Cen populations overlap in [X/Fe].
266
Figure 2. The mean abundance and standard deviation of the mean abundance of the Kapteyn
group stars (filled squares), Cen stars (filled triangles)and halo stars (open circles) for the five
different elements studied: a) [Na/Fe], b) [Mg/Fe], c) [Ca/Fe], d) [ls/Fe] and e) [hs/Fe]. It is easy
to see the mean abundance and distributions of each population and compare to the others. The
y axis simply spreads the three different groups out for clarity and ease of comparison.
massive in order for dynamical friction to have the required effect, and was also relatively
dense in order to survive the Galactic tidal stresses in to the current orbital radius of
Cen.
References
Beers, T. C., Chiba, M., Yoshii, Y., Platais, I., Hanson, R. B., Fuchs, B., & Rossi, S. 2000, AJ,
119, 2866
Bekki, K., & Freeman, K. C. 2003, MNRAS, 346, L11
Dinescu, D. I. 2002, in Astronomical Society of the Pacific Conference Series, Vol. 265, Omega
Centauri, A Unique Window into Astrophysics, ed. F. van Leeuwen, J. D. Hughes, &
G. Piotto, 365+
Eggen, O. J. 1996, AJ, 112, 1595
Eggen, O. J., Lynden-Bell, D., & Sandage, A. R. 1962, ApJ, 136, 748
Gratton, R. G., Carretta, E., Desidera, S., Lucatello, S., Mazzei, P., & Barbieri, M. 2003, A&A,
406, 131
Meza, A., Navarro, J. F., Abadi, M. G., & Steinmetz, M. 2005, MNRAS, 359, 93
Norris, J. E., & Da Costa, G. S. 1995, ApJ, 447, 680
Smith, V. V., Suntzeff, N. B., Cunha, K., Gallino, R., Busso, M., Lambert, D. L., & Straniero,
O. 2000, AJ, 119, 1239
Venn, K. A., Irwin, M., Shetrone, M. D., Tout, C. A., Hill, V., & Tolstoy, E. 2004, AJ, 128,
1177
Wylie-de Boer, E. C., Freeman, K. C., Williams, M. 2009, AJ, in press
268
D. Carollo et al.
present in the local calibration-star sample, concluding that the Galactocentric rotational
velocity, V , and Zmax (the maximum vertical distance of a stellar orbit above or below
the Galactic plane) are superior to available alternatives. The subtle, but important,
question of how many components are required to account for the observed kinematics of
halo stars in the local sample was then considered. On the basis of an objective clustering
approach, applied to the low-metallicity sub-sample of local stars with [Fe/H] < 2.0,
we demonstrated that (a) a single-population halo is incompatible with the observed
kinematics, that (b) a dual-component halo, comprising populations we associate with
the inner and outer halo, is sufficient to accommodate the observed kinematics, and that
(c) although additional components (of unspecified physical origin) can be added, they
are not required (and in any case, provide no statistically significant improvement in the
kinematic fits). We concluded that a dual-halo model is preferred on the grounds of its
simplicity (see Carollo et al. 2009 for details).
We have developed a flexible Maximum Likelihood (ML) approach for further analysis of the kinematics of Galactic components. We used a low-metallicity sub-sample of
local stars in order to obtain estimates of the mean rotational velocity and dispersion
of the outer-halo population, obtaining values of hV i = 80 13 km s1 , slightly
more retrograde than that obtained by Carollo et al. 2007 ( 70 km s 1 ) from the
DR5 sample of local calibration stars. For the dispersion, we obtained V = 165 9
km s1 , a substantially larger value than reported by previous authors who considered
only a single-component halo. The ML approach was then used in order to estimate the
rotation and dispersion of the inner-halo component, based on an examination of cuts
on Zmax for the low-metallicity sub-sample. We obtained, for stars with Zmax < 10 kpc
(where the inner-halo component dominates) values of hV i = 3 4 km s1 , and V =
100 2 km s1 for the rotational velocity and dispersion, respectively. The existence of
a gradient in the mean rotational velocity for the inner-halo component, as claimed previously by Chiba & Beers 2000, is confirmed. We obtained a value of 4hV i/4|Z| = 28
9 km s1 kpc1 , for stars located within 2 kpc of the Galactic plane. Such a gradient may represent the signature of a dissipatively-formed flattened inner halo. Clearly,
in many aspects our inner-halo population is essentially kinematically identical to the
halo population studied by Chiba & Beers 2000, and many others.
The mean rotational velocity and dispersion of the thick disk were then considered,
based on inspection of a metal-rich (0.8 < [Fe/H] < 0.6) sub-sample of stars close to
the Galactic plane. We obtained values hV i = 182 2 km s1 and V = 51 2 km s1
for stars in the range 1 < |Z| < 2 kpc, where the thick disk is espected to dominate, and
contamination from thin-disk stars should be negligible. The gradient in the asymmetric
drift of the thick-disk component as a function of height above the plane, noted by
previous authors, is very clear in our data as well; we derived 4hV i/4|Z| = 36
1 km s1 kpc1 , in excellent agreement with the rotational velocity gradients for disk
stars obtained by Chiba & Beers 2000, Girard et al. 2006, and Ivezic et al. 2008. The
ML approach was then applied to the full range of metallicities in our sample of local
calibration stars, in order to determine the fractional contribution of the three primary
components in our model as a function of Zmax (fixing as inputs the values of mean
rotational velocities and dispersions derived previously for each component). This exercise
indicated that, within 5 kpc from the plane, the thick-disk and inner-halo components
contribute roughly equally (Figure 1). Beyond 5 kpc, the thick-disk component is absent,
as expected. The inner-halo population dominates between 5 and 10 kpc. Beyond 10
kpc, the outer halo increases in importance, is present in equal proportion to the inner
halo between 15 and 20 kpc, and dominates beyond 20 kpc. The inversion point in the
dominance of the inner/outer halo is located in the range Zmax = 15-20 kpc.
269
Figure 1. Derived stellar fractions, as a function of Zmax , for the thick disk (circle),
inner-halo (triangle), and outer-halo (square) components.
We then used our extensive local dataset to examine the question of whether an independent MWTD component may be required in order to account for the rotational
properties of stars close to the Galactic plane. The clustering analysis approach was applied to sub-samples of stars selected in metallicity and vertical distance |Z| (Carollo et
al. 2009). We have demonstrated that, in the region closer to the Galactic plane, an independent MWTD component with hV i = 100-150 km s1 and V = 35-45 km s1 may
be required in order to account for the rotational behavior of the stars in our local sample.
We were also able to demonstrate that the metallicity range covered by the MWTD is
1.8 < [Fe/H] < 0.8, and possibly up to [Fe/H] 0.7. Finally, we found that the as
observed MDF of stars in our sample changes with distance from the Galactic plane,
exhibiting the anticipated shift from dominance by thick-disk and metal-weak thick-disk
populations, with peaks at [Fe/H] = 0.6 and 1.3, respectively, to dominance by the
inner-halo ([Fe/H] = 1.6) and outer-halo ([Fe/H] = 2.2) populations (see Beerss
article, this volume, for more details).
1.1. Velocity Ellipsoids
We derived estimates of the velocity ellipsoids for the thick disk, inner halo, and outer
halo; an approximate ellipsoid for the MWTD was also derived. In the case of the thickdisk and outer-halo components, we examined the same sub-samples of stars used for
determination of the rotational properties of these components. For the inner-halo, in
the radial and vertical directions (where the strong overlap with multiple additional
components complicates a mixture-model analysis), the velocities and dispersions were
obtained by adopting the mean velocity and dispersion of a sub-sample of stars in a more
restricted range of metallicity (2.0 < [Fe/H] < 1.5), where the inner halo is expected
to be dominant. The velocity ellipsoid for the canonical thick disk, the inner- and outerhalo, as well as for the MWTD, are summarized in Table 1. The values obtained for the
canonical thick disk match those of Chiba & Beers 2000. The inner halo is essentially
non-rotating; its velocity differs from that reported by Chiba & Beers 2000, hV i 3050 kms1 . Our velocity ellipsoid values are consistent, within the reported errors. This
agreement has been obtained even though the Chiba & Beers 2000 analysis adopted a
one-component halo, from which we conclude that their sample (and others) did not
include significant numbers of outer-halo stars. For the outer halo, which exhibits a
large net retrograde rotation, we have obtained a more tangentially anisotropic velocity
270
D. Carollo et al.
hVR i
km s1
hV i
km s1
hVZ i
VR
km s1 km s1
Thick Disk
MWTD
Inner Halo
Outer Halo
3 2 182 2 0 1
-13 5 125 4 -14 5
32
3 4
31
-9 6 -80 13 2 4
53 2
59 3
150 2
159 4
V
km s1
VZ
km s1
51 1
40 3
100 2
165 9
35 1
44 3
85 1
116 3
Abstract. The Galactic bulge is the central spheroid of our Galaxy, containing about one
quarter of the total stellar mass of the Milky Way (Mbulge = 1.8 1010 M ; Sofue, Honma &
Omodaka 2009). Being older than the disk, it is the first massive component of the Galaxy to
have collapsed into stars. Understanding its structure, and the properties of its stellar population,
is therefore of great relevance for galaxy formation models. I will review our current knowledge
of the bulge properties, with special emphasis on chemical abundances, recently measured for
several hundred stars.
Keywords. Galaxy: bulge, abundances, stellar content, formation
272
Manuela Zoccali
Figure 1. The CMD of a clean sample of bulge stars, selected on the basis of their proper
motions. Also shown are isochrones for different ages and metallicities. It is evident that bulge
stars follow the old isochrones, with the precise age depending on the adopted metallicity.
The residual stars brighter than the old turnoff do not follow a younger isochrone, and indeed
Clarkson et al. (2009) argues that they are likely to be all blue stragglers. Figure adapted from
Clarkson et al. (2008).
dances, resembling more those of galaxy disks rather than those of classical spheroids
(Kormendy & Kennicutt 2004).
In contrast with that, Ortolani et al. (1995) first demonstrated that the stellar population of the bulge, in Baades Window at (l, b) = (0, 4), is as old as the stars in
the globular cluster 47 Tucanae. This result was later confirmed by Feltzing & Gilmore
(2000), Kuijken & Rich (2002), Zoccali et al. (2003), and more recently by Clarkson et al.
(2008) who excluded disk stars from the color magnitude diagram (CMD) on the basis of
their proper motions (Fig 1). It should be noted, however, that all the above works studied the bulge stellar population in small fields very close to the minor axis. Dynamical
simulation of bulge formation in the bar-driven scenario suggest that the vertical heating
is significantly larger at the two ends of the bar, indicating these positions as the places
where the intermediate age components should be found (e.g., Debattista et al. 2004; and
references therein). Due to the larger interstellar extinction, deep photometry reaching
the main sequence turnoff away from the minor axis is not yet available in the literature.
Some data, acquired with the near IR camera HAWKI@VLT, are under analysis (Valenti
et al. 2010). Brown et al. (2009) present a new photometric system employing five WFC3
bands spanning the UV, optical, and near IR, especially designed to break the degeneracy between reddening, temperature and metallicity, for the forthcoming WF3 Galactic
bulge Treasury Program. This program will allow us to derive the bulge star formation
history in four fields, including one at the far edge of the bar.
273
274
Manuela Zoccali
Figure 2. The bulge MDF in three fields along the minor axis, at latitudes listed on the top
right corners. Mean metallicities and dispersions are labeled.
neglected and would obviously change the slope of the gradient. However it cannot erase
the presence of a gradient (see Zoccali et al. 2008 for a discussion).
The bulge MDF in Baades Window obtained by Zoccali et al. (2008) is centered at
the same mean metallicity than the one by Fulbright et al. (2006). However, the former is
significantly narrower than the latter, likely due to the smaller errors on the metallicity
of individual stars. Recently Rangwala & Williams derived the bulge MDF in a few fields
including Baades Window, by means of Fabry-Perot photometry across the Ca II triplet
line at 8542
A. Their MDF is compatible with both the Fullbright et al. (2006) and
Zoccali et al. (2008) results, once the latter two are convolved with the larger error of
the Fabry-Perot measurements.
The presence of a radial metallicity gradient along the bulge minor axis favors a bulge
formation through dissipational collapse, against the bar-driven scenario. In fact, the
latter is a dynamical process that should not, in principle, segregate metallicities. The
independent evidence of the absence of a radial gradient in the inner region (< 600
pc) cannot be contradicted here, thus one should keep in mind the possibility of a two
component bulge, uniform in metallicity in its inner part, and with radially decreasing
metallicity in the outer part.
Further evidence in the direction of a two component bulge comes from a detailed
analysis of the MDF in Baades Window. The observed MDF, once deconvolved from
the estimated errors, appears clearly bimodal (Hill et al. 2009). Furthermore, the two
components seems to have different kinematics (see below).
It is worth mentioning that in the past few years several high resolution spectroscopic
analysis of bulge dwarf stars were carried on during microlensing events that temporarily
magnifyied their brightness (Johnson et al. 2007, 2008; Cohen et al. 2008, 2009; Bensby
et al. 2009a, 2009b). The results of these analysis were extremely surprising, because the
mean [Fe/H] of microlensed dwarfs was too high to be compatible with a random (if
275
small) sampling of the bulge MDF obtained from giants. Newer results, however, now
including a total of 13 bulge microlensed dwarfs, demonstrate that the initial discrepancy
has almost completely disappeared, and it was very likely due to small number statistics
(Bensby et al. 2009c, these proceedings).
276
Manuela Zoccali
but have smaller statistical errors. The comparison with the thick disk giants observed by
Melendez et al. (2008) confirmed the chemical similarity between the two components.
It might be worth mentioning that high resolution near IR spectra have recently been
obtained for a sample of red supergiant in the Galactic center (< 50 pc). All the available
studies (Cunha et al. 2007; Davies et al. 2009; Najarro et al. 2009) agree on a [Fe/H]
distribution sharply peaked around 0.1; and on [O/Fe] ratios evenly spread between
+0.05 and +0.45. The reason for alpha element enhancement in this case is rather unclear. The independent evidence of intense star formation in this region (e.g., An et al.
2009) implies recent formation of massive stars thus core collapse SNe shortly after
naturally enriching the inter stellar medium of oxygen and other alphas. Alternatively
or simultaneously it might also be that the recent star forming activity in the center
was fueled by gas coming from the bulge/bar (known to produce gas inflows), hence already enriched by different kind of SNe and stellar winds, explaining the unusual element
ratios.
6. Conclusions
The nature of the Galactic bulge is somehow puzzling. Its stellar population is old
(10 12 Gyr) and it has a metallicity distribution compatible with chemical enrichment
277
models assuming a fast star formation. The abundance ratio of alpha elements over
iron also supports a short star formation timescale, certainly more rapid than that of
the thin disk, and possibly more rapid than that of the thick disk. The presence of a
radial metallicity gradient, at least outside 600 pc, favors a formation scenario via
dissipational collapse, rather than secular evolution of the disk. Nevertheless, the bulge
has the shape of a bar (perhaps including some X-shape feature in the outer part) and a
cylindrical rotation velocity, both characteristics of a pseudobulge formed via dynamical
heating of a bar, resulting from disk secular evolution.
A possible solution to these conflicting results may come from the confirmation of a
double component bulge, as already suggested by several studies (e.g., Soto et al. 2007,
Hill et al. 2009, Babusiaux et al. 2009) and seen in several bulges of external galaxies
(Peletier et al. 2007).
In any case, it is now clear from several independent evidences that the Galactic bulge
is a complex structure. There is a metallicity gradient in the outer region that seems
not to be present in the inner region. There are indications that the MDF in Baades
Window is bimodal, with each of the two component having different kinematics. Stellar
ages are predicted to be different along the minor axis, compared to the edges of the
bar. Even the morphology itself does not seem to be simply that of a bar, but rather
something like an X-shape. The properties of the stellar population in Baades Window
cannot be considered as representative of the whole bulge: larger area photometric and
spectroscopic maps are needed in order to understand the bulge structure and origin.
The VVV survey, and its spectroscopic followups, will certainly reserve many surprises
in this sense.
Acknowledgements
This work was supported by the Fondap Center for Astrophysics 15010003, CATA
PFB-06, and Fondecyt Regular #1085278.
References
Alard, C. 2001, A&A, 379, L44
Alcock, C. et al., 2001, ApJ, 562, 337
An, D. et al. 2009, ApJ, 702, L128
Athanassoula, E. 2005, MNRAS, 358, 1477
Athanassoula, E., & Misiriotis, A. 2002, MNRAS, 330, 35
Babusiaux, C. & Gilmore, G. 2005 MNRAS, 358, 1309
Babusiaux, C. et al. 2009, A&A, in preparation
Binney, J. 1978, MNRAS, 183, 501
Ballero, S.K., Matteucci, F., Origlia, L., & Rich, R.M. 2007, A&A, 467, 123
Bensby, T., Feltzing, S. & Lundstr
om, I. 2004, A&A, 421, 155
Bensby, T., Johnson, J. A., Cohen, J. G., et al. 2009a, A&A, 499, 737
Bensby, T., Feltzing, S., Johnson, J. A. et al. 2009b, ApJL, 69, L174
Brown, T. M. et al. 2009, AJ, 137, 3172
Cescutti, G., Matteucci, F., McWilliam, A., & Chiappini, C. 2009, A&A, 505, 605
Clarkson, W. et al. 2008, ApJ, 684, 1110
Clarkson, W. et al. 2009, ApJ, submitted
Cohen, J. G., Huang, W., Udalski, A., Gould, A. & Johnson, J. A. 2008, ApJ, 682, 1029
Cohen, J. G., Thompson, I. B., Sumi, T. et al 2009, ApJ, 699, 66
Combes, F., Debbasch, F., Friedli, D. & Pfenniger, D. 1990, A&A, 233, 82
Combes, F. & Sanders, R. H. 1981, A&A, 96, 164
Cunha, K. & Smith, V. V., 2006, ApJ, 651, 491
Cunha, K. et al., 2007, ApJ, 669, 1011
278
Manuela Zoccali
Abstract. 1. McWilliam and Zoccali (2009) show the existence of two Red Clump populations
towards the Galactic bulge, based on 2MASS data. 2.Measured [Mg/Fe], [Al/Fe], and [La/Eu]
ratios in the bulge are consistent with a rapid formation timescale (<1Gyr), which also requires a
slightly top-heavy IMF to reproduce the mean bulge metallicity. The [C/O] and [O/Fe] ratios are
consistent if their predicted metal-dependent yields from massive stars with winds are considered.
The decline in explosive [/Fe] (Si, Ca, and Ti) can only be understood if their yields also decline
with metallicity above [Fe/H]1.
Keywords. stars: abundances, Galaxy: abundances, Galaxy: bulge, nucleosynthesis.
2. Introduction
Detailed abundance ratios can constrain the bulge formation timescale, SFR, IMF as
well as the role of accretion in bulge history, and stellar nucleosynthesis yields. Enhanced
-element abundances (O, Mg, Si, Ca, Ti), relative to Fe, in the Galactic halo (e.g.
Wallerstein 1962; Conti et al. 1967) are thought to arise from the time delay between Fe
produced by Type II and Type Ia supernovae (henceforth SNII and SNIa); see Tinsley
(1979). Matteucci & Brocato (1990) showed that the star formation rate (SFR) determines the [Fe/H] where the [/Fe] ratio declines (the knee) due to the late addition of
SNIa iron. Smecker-Hane & Wyse (1992) found a mean SNIa timescale of 1Gyr. Because
bulges are thought to have formation timescales shorter than SNIa progenitor lifetimes
they should possess enhanced [/Fe] even at solar metallicity. Wyse & Gilmore (1993)
showed that a top heavy massive star initial mass function (IMF) leads to larger mean
enhancement of [O/Fe] before the knee. This work is motivated by a simple question:
What was the formation timescale of the Galactic bulge?
279
280
Figure 1. a)Left: Explosive alpha-elements in the bulge and disk (see FMR07) b)Right:
Revised [O/Fe] in the bulge and disk (see McWilliam, Fulbright & Rich 2009, in preparation)
281
Figure 2. a) Left:
b)Right:
[Al/Mg]
onset of the knee suggests that bulge evolution occurred over a long timescale greater
than 1Gyr, perhaps significantly longer. On the other hand, the 0.2 dex enhancement
of [<SiCaTiI>/Fe] at solar metallicity might indicate a shorter evolution timescale than
for the Galactic thin disk, and the overall higher [/Fe] values may be understood with
an IMF slightly more top-heavy than for the thin disk.
4.2. Hydrostatic Alphas
The original FMR07 bulge star [O/Fe] trend was slightly enhanced relative to the solar neighborhood, but declined by 0.7 dex from [Fe/H]=1.5 and +0.5 dex. FMR07
employed Fe II lines in the ratio for robust O/Fe ratio measurements; however, the C
abundance required to account for CO formation was adopted assuming the solar C/Fe
ratio, adjusted for RGB dredge-up. The FMR07 oxygen abundances were recently revised by McWilliam, Fulbright and Rich (2009, in preparation, MFR09). In this case we
employed the C/Fe abundances determined from the near-infrared Fe and CO molecular
lines by Melendez et al. (2008). Because well over 90% of the atmospheric carbon is in
the form of CO in RGB stars, total carbon abundances derived from CO lines are more
reliable than from other species. Figure 1b shows the new [O/Fe] trend for Galactic bulge
stars from MFR09, which appears very similar to the trend seen in the thick disk; certainly, there is no compelling evidence that the trend differs significantly from the thick
disk. This conclusion supports that of Melendez et al. (2008), who employed OH lines,
rather than the [O I] forbidden line at 6300
A used by FMR07 and MFR09.
Figure 1b indicates a steeper decline of [O/Fe] than seen in the explosive alpha-elements
of Figure 1a. In the Tinsley paradigm the decline is due to iron from SNIa, thus indicating a long formation timescale for the bulge and thick disk, greater than 1 Gyr. SNIa
nucleosynthesis models (e.g. Nomoto, Thielemann, & Yokoi 1984) indicate production of
the explosive alpha elements, but an insignificant [O/Fe] ratio, which might explain the
steeper decline in [O/Fe] than [<SiCaTi>/Fe] in the bulge.
The consistency of the picture presented so far is destroyed by the Mg and Al abundances in the FMR07 bulge stars. Figure 2a shows the [Mg/Fe] trend in the bulge
compared to thin and thick disk stars. At solar metallicity [Mg/Fe] is 0.4 dex; even
the most metal-rich bulge stars, near [Fe/H]=+0.5, have [Mg/Fe] ratios of 0.3 dex. A
282
slight decline in [Mg/Fe] from [Fe/H]=1.5 to +0.5 might be due to Fe from SNIa, but
this is smaller than the decline in [O/Fe] by 0.5 dex, and smaller than the decline in
[<SiCaTi>/Fe] by 0.15 dex. It is not possible to explain the decline in [/Fe] of the
explosive -elements, O and Mg with only the addition of iron from SNIa. The relatively
small decline in Mg/Fe is the most difficult to comprehend in light of a time-delay scenario for oxygen and the explosive -elements. Indeed, from a strict delay scenario the
[Mg/Fe] ratios indicate that the bulge formed very rapidly, in less than 1Gyr, while the
[O/Fe] ratios suggest slow formation, on the timescale of the Galactic disk, significantly
longer than 1 Gyr. This is particularly surprising because O and Mg are both thought to
be formed in the hydrostatic phase of core-collapse SN progenitors, while neither element
is predicted to have significant production from SNIa.
Because the implied formation timescale for the bulge is very dependent on the Mg
abundances FMR07 considered additional Mg I lines, but found the same result as initially obtained. Enhanced Mg abundances are also supported by the results of Lecureur
et al. (2007) and McWilliam & Rich (1994, 2004). Finally, FMR07 employed Al as a
proxy for Mg. Arnett (1971) noted that in massive stars the odd/even yield ratios, as
indicated by [Al/Mg], should depend on the neutron excess, i.e., proportional to [Fe/H].
Figure 2b shows the FMR07 [Al/Mg] versus [Fe/H] for the bulge compared with the thin
and thick disks; the trends appear to be identical. If FMR07 had erroneously overestimated Mg then the Al abundances must also be overestimated by the same amount,
which seems unlikely. Therefore, the Mg and Al abundances are consistent and enhanced
in the bulge RGB stars, even at [Fe/H]=+0.5. In the Tinsley time-delay scenario these
Mg and Al enhancements indicate a rapid bulge formation timescale. A serious problem is that abundances of lensed bulge dwarf stars, reported by Bensby at this meeting,
disagree with the [Mg/Fe] trends found for unlensed bulge RGB stars.
McWilliam & Rich (2004) proposed that the different bulge trends for [O/Fe] and
[Mg/Fe] resulted from reduced oxygen yields from massive stars from stripping of their
envelopes by metal-dependent winds, as expected in the formation of Wolf-Rayet stars.
Bulge chemical evolution models by McWilliam et al. (2008) and Cescutti et al. (2009) using metal-dependent oxygen yields from massive stars with mass-loss, from Maeder (1992,
M92) and Meynet & Maeder (2002, MM02), and semi-empirical Mg yields, found qualitative agreement with the observed, universal, decline in [O/Mg] versus [Mg/H]. Thus,
it appears that the decline in [O/Mg] with [Mg/H] is consistent with metal-dependent
oxygen yields from massive stars, and indicates that the decline in [O/Fe] is not due to
the addition of iron from SNIa.
If metallicity-dependent winds are responsible for reducing the oxygen yields from
massive stars then the [C/O] yield ratio should increase with metallicity. Cescutti et
al. (2009) investigated chemical evolution models for [C/O] ratios using massive star C
and O yields, compared with the revised [C/O] ratios from MFR09. The bulge chemical
evolution model adopted a rapid formation timescale and top-heavy IMF, consistent
with Ballero et al. (2007). The MFR09 abundance measurements showed a 0.7 dex
increase in [C/O] with [O/H]. WW95 element yields without mass-loss failed to match
the observations, while the models with mass-loss closely matched the observed trend,
and bracketed the observations. It appears that a mass-loss rate higher than the MM02
but lower than M92 would best fit the [C/O] observations. One inconsistency is the metalpoor bulge stars, which show a higher [C/O] plateau than expected; Cescutti et al. (2009)
suggested that the required source of carbon may be from a population of low-metallicity
Wolf-Rayet stars formed via binary Roche-lobe stripping of the envelopes.
The conclusion is that the observed [O/Fe], [C/O] ratios, and the Mg and Al abundances are consistent with the high SFR, short bulge formation timescale and top-heavy
283
Figure 3. a)Left: [La/Eu] in the bulge and disk b)Right: [Eu/Fe] in the bulge and disk.
Blue and red crosses: thin and thick disk stars. (see Fulbright et al. 2009, in progress)
IMF obtained by Ballero et al. (2007). If correct, this conclusion means that the decline
of the [<SiCaTi>/Fe] ratios in Galactic bulge stars cannot be due to increased iron production by SNIa. FMR07 proposed a solution to this puzzle: that the yields for explosive
alpha elements decline with increasing metallicity in SNII.
5. Neutron-Capture Elements
Figure 3a compares the [La/Eu] ratios in the halo, bulge, thick disk and thin disk.
This ratio is sensitive to the relative importance of the s- and r-processes; higher values
indicate a greater s-process fraction. Since the main s-process is thought to be produced
by relatively low-mass AGB stars an enhanced s-process fraction (i.e., high [La/Eu]) indicates long timescales. The r-process is thought to be produced in core-collapse supernova
events (i.e. massive stars on short timescales) with low [La/Eu] ratios. Figure 3a also
indicates the solar-system pure r-process [La/Eu] ratio as determined by Burris et al.
(2000). It is clear that our bulge [La/Eu] ratios (Fulbright et al. 2009, in progress) are
halo-like and closer to the pure r-process value than the thin disk trend. This indicates
a rapid bulge formation timescale, qualitatively consistent with our conclusions from
[Mg/Fe] and [Al/Mg]. The [La/Eu] ratio certainly indicates that the bulge formed more
rapidly than the thin disk. Much of the thick disk shares the halo-like [La/Eu] ratios, but
some of the thick-disk points indicate higher s-process contributions. Therefore, either
the thick disk evolved more slowly than the bulge, but faster than the thin disk, or some
of the thick disk stars are mis-identified thin disk interlopers.
One difficulty with the s- to r-process ratio is that the r-process astrophysical site is
not known; even the neutron source for the r-process is unconstrained. Figure 3b shows
the [Eu/Fe] in the bulge compared to the thick disk and thin disk, which measures the
r-process/Fe ratio. If the r-process is made only by core-collapse SNe, then, naively, the
alpha-like decline in [Eu/Fe] with [Fe/H] in the disk, which follows the [/Fe] trend, is
due to a declining SNII/SNIa ratio as in the Tinsley (1979) scenario. Therefore, Figure 3b
indicates that the disk-like trend of the bulge [Eu/Fe] ratio is due to the addition of iron
from SNIa ejecta, and suggests that the bulge formed on a long timescale, contrary to
our conclusion from the bulge [La/Eu] ratios.
284
Abstract. The carbon, nitrogen, and oxygen abundances and trends in the bulge are discussed
in the context of our recent analysis of these elements in an on-going project based on near-IR
spectra (Ryde et al. 2009). We obtained these using the CRIRES spectrometer on the VLT.
The formation and evolution of the Milky Way bulge can be constrained by studying elemental
abundances of bulge stars. Due to the large and variable visual extinction in the line-of-sight
towards the bulge, an analysis in the near-IR is preferred.
Keywords. stars: abundances, Galaxy: bulge, infrared: stars
It is strange, but we actually do not know the origin and the classification of the Milky
Way bulge. There are two main avenues for the formation of bulges, one leading the the
pseudobulges and one to the classical ones (see, e.g. Kormendy & Kennicutt, 2004). The
former are formed through secular, dynamical evolution of the disks over time, driven
by the development of a bar. One would then expect an age spread of its stars. The
dynamics and morphology of the Milky Way bulge places it as pseudobulge. Its shapes
and structures (boxy, peanut-shaped bar) indicates this. Indeed, Binney (2009) writes
that ...there is every indication that our bulge is a pseudobulge. For this reason alone it
would be connected to the disc by history. The second avenue leads to the formation of
the classical bulges, which are formed by merger-driven star-bursts during a short phase.
The stars of the Milky Way bulge are old, the bulk being 11 Gyrs (Ortolani et al
1995, Clarksson et al. 2009), even though there is a younger population at the centre
(see e.g. Figer et al. (2004)). Furthermore, the stars are metal-rich and are -element
enhanced. This indicates that the bulge underwent rapid chemical enrichment as the
result of starbursts, as in the case of a classical bulge. Thus the formation of the Milky
Way bulge is clearly not well understood and its classification inconclusive. Classical
bulges are most often found in early-type spirals (Sa and Sb) whereas the pseudobulge
are found in late-type spirals (Sc and Sd). The Milky Way is classified as a Sbc and is
morphologically on the borderline. It is therefore maybe not a surprise that the origin
of it is under debate. Maybe the bulge formed from a secular, but fast evolution of the
early disk, resulting in an early formation (Genzel et al. 2008).
The different formation scenarios can be constrained by abundance surveys. From stellar populations and abundance analyses it can be investigated which process dominated
the star formation: whether stars formed rapidly a long time ago during hierarchical
clustering of galaxies (as a classical bulge), or through secular evolution (such as defines
a pseudo-bulge). The -element composition (e.g. O, Mg, Ca, etc.) relative to iron as a
function of the metallicity, [Fe/H], can infer star-formation rates (SFR) and initial-mass
functions (IMF). A shallower IMF will increase the number -element producing stars
thus leading to higher [/Fe] values. A faster enrichment due to a high star-formation rate
will keep the over-abundance of the elements at a high value also at higher metallicity.
Different populations may show different behaviours. However, the chemical properties
and evolution of the MW bulge are poorly constrained, mainly due to the large distance
285
286
Nils Ryde
and large and variable visual extinction toward the bulge. We, therefore need more stellar abundance studies of many more bulge stars of various bulge fields, especially in the
near-IR, to constrain bulge models.
The Bulge is more accessible in the IR than in the optical for multiple reasons (Ryde
et al. 2005). The most important is of course the smaller interstellar extinction in the
IR (AK 0.1 AV ; Cardelli et al. (1989)). The near-IR is also preferred for analysis
of abundances, due to the fact that the absorption spectra are less crowded with lines,
that fewer lines are blended, and that it is easier to find portions of the spectrum which
can be used to define a continuum compared to wavelength regions in the optical spectral window. Furthermore, in the Rayleigh-Jeans regime, the intensity is less sensitive to
temperature variations. This means that the effects of, for example, effective-temperature
uncertainties or surface inhomogeneities on line strengths should be smaller in the IR.
Note, however, that 3D effects on the atmosphere can have an impact though the sensitivness of the molecular equilibria to temperature. A general drawback of a spectral
analysis in the near-IR is that existing HR spectrometers are still much less effective than
optical ones, one of the main reasons being the lack of cross-dispersion. Also, determining
the stellar parameters based only on near-IR spectra is difficult. The main advantage of
a high spectral resolution (HR, R > 50, 000) is the smaller line-blending, reduction of
background noise, and the fact that telluric lines can be taken care of more easily.
Especially, the CNO abundances are of interest and infrared spectroscopy at high spectral resolution can provide the most secure measurements of the CNO elements. Oxygen
provides information on the SFR and IMF through O/Fe and C/O ratios. The C+N abundances give an indication of the evolutionary stage of the red giants investigated, and N is
needed to estimate the primordial C abundance of the red giants. The C abundance can
give information on its origin in the bulge. To get the CNO abundances, normally one
needs to measure them simultaneously due to the molecular equilibria. Recently, a few
studies of elemental abundances of bulge stars using near-IR spectra at high resolution
have been done, see for instance Ryde et al. (2009), Melendez et al. (2008), Cunha et al.
(2007), Cunha & Smith (2006), and Melendez et al. (2003).
We are engaged in an on-going VLT project (080.D-0675), in which we are analysing
near-IR spectra recorded at a spectral resolution of R = 60, 000 with the CRIRES spectrometer (Moorwood 2005; K
aufl et al. 2006) on the Very Large Telescope, VLT, see
also Ryde et al. (2007, 2009). We have measured the abundances of Fe, C, N, and O
in addition to the elements Si, S, and Ti for a well-chosen sample of 11 stars, sampling different stages of the chemical enrichment history and different parts of the bulge.
The stars are chosen from the optical work by Lecureur et al. (2007) and Zoccali et al.
(2006, 2008). The stars lie in three fields at b = 3 (NGC 6553 field), 4 (BW), and 6
and have H magnitudes between 10.3 and 12.0. The spectra are analysed with tailored
MARCS models and fully consistent synthetic spectra. The temperatures lie between
3900 4300 K, 1.25 <[Fe/H]< 0.0, 1.0 < log g < 2.2. The quality of the spectra are
very high and numerous molecular lines of CO, CN, and OH can be identified, see Ryde
et al. (2009). We have also identified 10 Si, 5 Ti, 4 S, 5 Ni, one Cr, and many Fe lines in
the spectra.
We find a high [O/Fe] vs. [Fe/H] trend of 0.4 up to metallicities of [Fe/H= 0.3
whereafter the [O/Fe] declines. The data suggest that the Milky Way bulge experienced
a rapid and early star-formation history compared to the thin disk. Our data agrees
well with other determinations of the oxygen abundances in the bulge both from near-IR
B. Gustafsson (PI), B. Edvardsson, J. Melendez, A. Alves Brito, M. Asplund, B. Barbuy,
V. Hill, H.-U. K
aufl, D. Minniti, S. Ortolani, A. Renzini, N. Ryde, and M. Zoccali
287
lines (Melendez et al.(2008), Cunha & Smith (2006), and Rich & Origlia (2005)) and from
optical lines (Fulbright et al. (2007) and Lecureur et al. (2007) and Zoccali et al. (2006)).
Our stars are a subsample of the latter. Our and Melendez et al.s data confirm the result
of Zoccali et al. When comparing with thick and thin disks giants from Melendez et al.
(2008) we find that the thick disk and bulge trend are similar. Thus, both the bulge and
thick disk must have formed rapidly (Melendez et al. 2008). A key point is that one can
reduce the systematic uncertainties by homogeneously comparing disk giants with bulge
giants on the same scale (Melendez et al. 2008).
Recently, Cescutti et al. (2009) showed in a plot of primordial [C/O] vs. [O/H] for
thin and thick disk dwarfs and turnoff stars and bulge giants a similar similarity between
the bulge and thick disk relationships. The primordial C abundance is estimated by
calculating C+N Fe(N/Fe) . In Figure 1 we show our relationship in the same
[C/O] vs. [O/H] diagram. The difference is that we compare with the Melendez et al.
(2008) thin and thick disk giants which are on the same scale. We also find a similar
relationship and an upturn, but our data show a larger scatter.
To conclude, we have analysed near-IR spectra from the CRIRES spectrometer on the
VLT with tailored MARCS model atmospheres and consistent synthetic spectra. We find
that the stellar parameters are important for the derivation of the CNO abundances from
molecular lines. We find [O/Fe] vs. [Fe/H] values enhanced up to metallicities of [Fe/H]=
0.3, implying a rapid and early star formation in the bulge, as for a classical bulge.
0.2
[C/O]
0.0
-0.2
-0.4
-0.6
-0.8
-0.6
-0.4
-0.2
[O/H]
0.0
0.2
0.4
Figure 1. [C/O] vs. [O/H] for our bulge giants (dimonds) and those from Melendez et al. 2008
(triangles) shown with blue symbols. The thin and thick disk giants from Melendez et al. (2008)
are shown with red and green circles, respectively. The thick disk giants and the bulge giants
show a similar behavior.
288
Nils Ryde
The [C+N/Fe] is nearly constant with metallicity implying that the oxygen abundance
we measure is the primordial, unprocessed values, and that our target stars are indeed
first ascent giants or clump giants. We show that the [O/Fe] vs. [Fe/H] trends in the
literature all corroborate each other within uncertainties. The bulge and thick disk seem
similar in [O/Fe] vs. [Fe/H] and [C/O] vs. [O/H], implying that both formed rapidly.
Such a similarity could suggest that the bulge has a pseudobulge origin. Note, that
the metallicity distributions are not the same. We do not see an increased [C/Fe] but
corroborate an increase in the [C/O] vs. [O/H] plot for [O/H]> 0.1.
Stellar surface abundances in Bulge stars, especially those of the C, N, and O elements,
can extensively be studied in the near-IR, due to lower extinction. It will be very important to extend the analysis to other regions of the Galactic bulge, such as in the galactic
plane, in order to get a proper handle on its formation and evolution. In these regions
the optical extinction is large which only permits observations in the near-IR. Near-IR,
high-spectral-resolution spectroscopy offers a promising methodology to study the whole
bulge to give clues to its formation and evolution.
References
Binney, J., 2009, in: J. Andersen, J. Bland-Hawthorn, & B. Nordstr
om (eds.), The Galaxy Disk
in Cosmological Context, Proc. IAU Symposium No. 254 (CUP), p. 145
Cardelli, J. A., Clayton, G. C., Mathis, J. S., 1989, ApJ 345, 245
Cescutti, G., Matteucci, F., McWilliam, A., Chiappini, C., 2009, arXiv:0907.4308v1
Clarkson, W., Sahu, K., Anderson, J., et al., 2008, ApJ 684, 1110
Cunha, K., Sellgren, K., Smith, V. V., et al., 2007, ApJ 669, 1011
Cunha, K., Smith, V. V., 2006, ApJ 651, 491
Figer, D. F., Rich, R. M., Kim, S. S., Morris, M., Serabyn, E., 2004, ApJ 601, 319
Fulbright, J. P., McWilliam, A., Rich, R. M., 2006, ApJ 636, 821
Fulbright, J. P., McWilliam, A., Rich, R. M., 2007, ApJ 661, 1152
Genzel, R., Burkert, A., Bouche, N., et al., 2008, ApJ 687, 59
Hinkle, K. H., Blum, R. D., Joyce, R. R., et al., 2003, in: P. Guhathakurta (ed.), Discoveries
and Research Prospects from 6- to 10-Meter-Class Telescopes II. , SPIE, Vol. 4834 p. 353
Hinkle, K. H., Cuberly, R. W., Gaughan, N. A., et al., 1998, SPIE 3354, 810
K
aufl, H. U., Amico, P., Ballester, P., et al., 2006, The Messenger 126, 32
Kormendy, J., Kennicutt, Jr., R. C., 2004, ARAA 42, 603
Lecureur, A., Hill, V., Zoccali, M., et al., 2007, A&A 465, 799
Melendez, J., Asplund, M., Alves-Brito, A., et al., 2008, A&A 484, L21
Melendez, J., Barbuy, B., Bica, E., et al., 2003, A&A 411, 417
Moorwood, A., 2005, in: H. U. K
aufl, R. Siebenmorgen, & A. F. M. Moorwood (eds.), High Resolution Infrared Spectroscopy in Astronomy, Proc. of an ESO Workshop (ESO Astrophysics
Symposia: Springer Berlin / Heidelberg), p. 15
Ortolani, S., Renzini, A., Gilmozzi, R., et al., 1995, Nature 377, 701
Rich, R. M., Origlia, L., 2005, ApJ 634, 1293
Ryde, N., Edvardsson, B., Gustafsson, B., K
aufl, H.-U., 2007, in: A. Vazdekis & R. F. Peletier
(eds.), Stellar Populations as Building Blocks of Galaxies, Proc. IAU Symposium No. 241
(CUP), p. 260
Ryde, N., Edvardsson, B., Gustafsson, B., et al., 2009, A&A 469, 701
Ryde, N., Gustafsson, B., Eriksson, K., Wahlin, R., 2005, in: H. U. K
aufl, R. Siebenmorgen, &
A. F. M. Moorwood (eds.), High Resolution IR Spectroscopy in Astronomy, p. 365
Zoccali, M., Hill, V., Lecureur, A., et al., 2008, A&A 486, 177
Zoccali, M., Lecureur, A., Barbuy, B., et al., 2006, A&A 457, L1
Zoccali, M., Renzini, A., Ortolani, S., et al., 2003, A&A 399, 931
1. Introduction
Deciphering the history of the birth and the growth of the Milky Way Galaxy is one
of the major outstanding astrophysical problems. Detailed studies of our Galaxy would
help to gain insights into the formation and evolution of other galaxies, and thereby large
scale structure formation in the universe. As inhabitants, we have much better access to
our Galaxy to resolve its building blocks: the stars. The properties of stars that constitute
the Galaxy are clues to the processes involved in the making of the Milky Way Galaxy
we see today. The observational data on stellar motions and photospheric abundances
played decisive roles in the progression of our understanding of the Galaxy. The major,
perhaps the first, quantitative study of stellar kinematic and chemical properties of a
large sample of nearby dwarfs by Eggen, Lynden-Bell and Sandage (1962), known as
ELS, laid the foundation for the Galaxy formation models. In this seminal paper ELS
showed that the eccentricity of stellar orbits and the photometrically derived UV excess
( (U-B)), an indicator of metallicity, are correlated: stars with the largest excess (metalpoor) are found to be moving in highly elliptical orbits, where as the stars with little or
no excess (solar metallicity stars) are found to have circular orbits. ELS attributed this to
the rapid collapse of a proto-galactic cloud forming the halo quickly, and the rotationally
supported disk component thereafter. Further studies, based on improved observations
of globular clusters in the Galaxy, led to the fact that mergers play significant role in the
evolution of the Galaxy (Searle & Zinn 1978). Thanks to large scale surveys of both the
kinematic and photometric properties of stars, we now know that the Galaxy consists of
four major components: the halo, the thick disk, the thin disk and the bulge. There are
excellent reviews which trace the progressive understanding of the subject (e.g; Gilmore
1989; Majewski 1993; Freeman & Bland-Hawthorn2002). In this review, we primarily
concentrate on the thick disk component and its relation to the Milky Way galaxy.
289
290
Bacham E. Reddy
3. Global properties
Identification of Galactic stellar populations into distinct components relies, primarily, on five basic differences: location, structural properties, kinematic motions, age and
chemical make-up. Until late 90s, studies mostly focused on defining the global properties of the thick disk like scale height, scale length, local normalization, metallicity
distribution, kinematic properties and to some extent ages relative to the other components. Thanks to the Hipparcos space mission for accurate stellar astrometry, and the
major ground based surveys like Radial Velocity Experiment (RAVE), Sloan Digital Sky
Survey (SDSS), and 2 Micron Sky Survey (2MASS) it became abundantly clear that the
thick disk exists and it has distinct global properties which are different from that of the
thin disk, halo and the bulge. Since the halo and the bulge populations are known to
differ both in morphology and kinematics from the disk populations, we concentrate on
distinguishing two distinct populations within the Galactic disk: the thin disk and the
thick disk.
Recently measured values of the thick disk population stars are summarized in Table 1.
The thick disk has scale height in the range of 500pc to 1100 pc which is two to four times
of the thin disk (300 pc),and its metallicity peaks at [Fe/H] 0.6 dex compared to the
thin disk which peaks at [Fe/H] 0.2. Thick disk stars are older (8-12 Gyrs) compared
to their counter parts in the thin disk stars (<10 Gyrs). Local normalization of the thick
disk population relative to the thin disk population varies from survey to survey and
which is in the range of 2% to 13%. That is there are 2 to 13 thick disk stars for every
100 thin disk stars in the solar neighborhood. Volume limited (stars within the radius
of 25 pc from Sun) survey by Fuhrman (2008) suggests even much larger proportion
(20%) of thick disk stars in the local neighborhood implying a massive thick disk. The
population of the thick disk stars increases as one moves away from the Galactic plane
291
Vlag (kms
Velocity Ellipsoid
Metallicity peak
1
Carollo et al.
(2009)
Juric et al.
(2008)
Siegel et al.
(2002)
0.510.04
2.200.35
157%
900 pc
3.6 kpc
...
940 pc
...
8.5%
Carollo et al.
Soubiran et al.
382
515
20
(532,511, 351) (636,394, 394) (464,504, 353)
0.6
0.48
in the z-direction. The thick disk dominates other populations beyond 1.0 kpc. If the
thick disk is a separate component one would also expect distinct kinematic motion (U ,
V , W ) for the thick disk population which is different from the other two components:
the halo and the thin disk.
One of the potential discriminators of the Galactic components is their average rotational velocity (Vrot ) or asymmetric drift velocity (Vlag ) with respect to the Galactic
rotation of 220 km s1 . This has been known for a while (see Majewski 1993). As shown
in Table 1, values of Vlag for the thick disk range anywhere between 20 km s1 (Chiba
& Beers 2000) to 51 km s1 (Soubiran et al. 2003). Large range in the measured Vlag is,
perhaps, due to the sample selection in the respective studies. However, Vlag of 40 - 50
km s1 is generally in good agreement with the other recent studies (e.g; Robin et al.
2003). Compare this with the halo population which has Vlag 220 km s1 (or statistically stationary w.r.t to the Galactic rotation) and with the thin disk of Vlag 5-10 km
s1 (Robin et al. 2003). The measured velocity ellipsoid (u , v , w ) of the thick disk is
about twice that of the thin disk and half of the halo values. In all of the measurements,
Gaussian distribution is assumed, and the non-Gaussian velocity distributions cant be
ruled out.
In summary, as far as the Global properties of the thick disk are concerned, measurements based on different data sets and different methodologies agree with each other
suggesting a level of overall confidence in characterizing the thick disk. Discontinuity in
the luminosity function, metallicity distribution and the difference in rotational velocity
are some key evidences for the distinct nature of the thick disk population in the disk of
the Galaxy.
292
Bacham E. Reddy
Table 2. Summary of a few key abundances ratios for the kinematically selected stars the Halo,
thick disk, MWTD, and thin disk. Data is adopted from a series of papers by Reddy et al. group.
[X/Fe]
[Mg/Fe]
[Si/Fe]
[Ni/Fe]
[Eu/Fe]
Halo
MWTD
Thick Disk
Thin disk
0.310.06 0.320.07
0.310.09
0.090.05
0.230.06 0.270.09
0.230.15
0.070.04
0.020.04 0.080.09 0.080.11 0.020.02
0.380.13 0.44 0.24 0.35 0.15
0.120.10
making it difficult to analyze, and hence larger uncertainties. Another important consideration is the selection of elements. The -elements like O, Mg, Si, Ca or Ti are thought
to be predominantly produced in the short lived massive star explosions (SNII) and elements like Fe, Ni etc. are the main products from the long lived low mass star explosions
(SNIa). SNII are more frequent in the early epochs and SNIa starts contributing much
later. Thus, ratio of [/Fe] is a tracer of the chemical evolution and star formation rates
in the Galaxy as well as a indicator of relative frequency of SNII to SNIa explosions.
The abundance study of the Galactic disk stars of Edvardsson et al. (1993) is the first
major work in this direction. They have analyzed high resolution spectra of 189 F- and Gdwarfs and measured abundances of 12 elements, which provided important constraints
for the Galactic chemical evolutionary models. They have noticed significant scatter in
[/Fe] versus [Fe/H], particularly, below [Fe/H] -0.4. Stars with relatively higher [/Fe]
ratios found to be preferentially more eccentric and closer to the Galactic center. This was
attributed to an efficient star formation rate in the inner disk. Surprisingly, the possibility
of the thick disk population to the observed scatter in the abundances was never raised
in the entire lengthy paper. By the late 1990s, it became increasingly clear that the thick
disk harbours a stellar population whose chemical history is distinct from the rest of the
components in the Galaxy, particularly the thin disk. Fuhrman (1998) study of carefully
selected and unbiased, but volume limited, sample of F and G dwarfs indicated that
stars with intermediate kinematic properties show higher [Mg/Fe] ratio compared to the
thin disk stars in the overlapping metallicities. This was confirmed from the detailed
study of many more elements but for a small kinematically pre-selected sample of 10
stars (Prochaska et al. 2000). Abundance ratios of [/Fe] found to be higher compared
to the thin disk stars but very similar to halo and bulge stars. Contrary to the above
results, Chen et al. (2000) study of pre-selected samples of thick disk showed continuous
chemical evolution from the thick to the thin disk populations.
293
et al. These studies are complimentary to each other as the sample stars of Bensby et al.
are drawn from the southern hemispheres whereas the samples in the studies of Reddy
et al. are from the northern hemisphere. Common stars among the two studies are very
few. However, the agreement between the two results,in most cases, is extremely good
indicating the maturity of the different model atmospheres and the analysis techniques.
In the studies of Reddy et al. (2003, 2006) a carefully pre-selected sample, based on
kinematic properties, of about 400 nearby dwarfs belong to the thin disk, the thick disk,
or the halo were subjected to high resolution spectroscopy. Stars were grouped into the
thin disk, thick disk and the halo populations based on their kinematic definitions and
relative fraction of each of these populations in the solar neighbourhood. Quantitative
abundances of 22 to 26 elements were determined with uncertainties as low as 0.04 0.06 dex for most of the key elements. The mean [X/Fe] ratios are given in Table 2 for
the thick disk, thin disk and the halo stars. Results for a few key elements are shown in
Figure 1. Similar studies based on similar number of stars were carried out in a series
of papers by Bensby et al. group. In general, results from both the studies are in good
agreement. Some of the Key conclusions from these results are: a) Abundance ratios of
elements that are known to be the dominant yields from SNII are clearly larger than their
counter parts in the thin disk in the overlapping metallicity (0.3 - 0.8 dex) implying
0.4
0.2
0
-0.2
0.4
0.2
0
-0.2
0.2
0
-0.2
-0.4
-0.6
-0.8
-2
-1.5
-1
-0.5
[Fe/H]
Figure 1. Abundances ratios [Mg/Fe], [Ti/Fe] and [Mn/Fe] versus [Fe/H] for samples of the
thick disk (red circles), the thin disk (black crosses) and the halo stars (blue circles). Light
blue and green circles are shown for stars of metal-weak and metal-rich thick disks, respetively.
TKTA stars are shown as red circles embedded in blue circles.
294
Bacham E. Reddy
a distinct chemical history for the thick disk from that of the thin disk. However, the
mean abundance ratios of elements of thick disk stars are indistinguishable from those of
MWTD or halo stars and even the bulge stars ( e.g; Melendez et al. 2008). The similarity
of the thick disk abundance ratios with that of the halo and bulge may suggest that the
three components are similarly old and had a rapid star formation. b) The run of [/Fe]
with [Fe/H] for stars of the thick disk shows little or no slope against [Fe/H] suggesting
rapid formation of the thick disk. Slight slope towards higher metallicities is attributed
to -element yield dependence on the initial metallicities of SNII. c) Another important
outcome of the analysis is that the star-to-star scatter in the trends which is comparable
to the scatter expected from the measurements indicating stars both in the thin as well
as in the thick disk formed from the well mixed gas, and d) Age estimates indicate that
the stars of the thick disk are older on average (8-12 Gyrs) compared to the thin disk
stars (t < 10 Gyrs). Age results show that 2-3 Gyrs elapsed between the first stars of the
thin disk and the first stars of the thick disk suggesting sufficient time for the SNIa role
in the chemical history of the thick disk.
295
there seems to be multiple channels that are connecting thick disk trends with that of
the thin disk.
296
Bacham E. Reddy
kinematic and spacial overlapping with the halo and the canonical thick disk, and the
expected large uncertainties in derived astrometry will require additional studies to fully
resolve the status of MWTD. Accurate astrometry for large samples and followed by
accurate abundance study would certainly help to unravel the origin thick disk over its
full metallicity range.
Figure 2. Abundances ratio [Fe/] versus [/H] for samples of the thick disk, the thin disk
and the halo stars. Colour and symbol coding are same as those for in the Figure 1.
297
either through a rapid or dissipative collapse of protogalactic clouds. In this scenario halo
forms first, then the thick disk, and then the thin disk. Top down models, by their nature,
predict gradients both in abundances and kinematics, and do not predict distinct chemical
abundances of the thick disk from the thin disk over the overlapping metallicities. Top
down models fail to meet many of the well defined observational parameters listed above.
For an exhaustive list of different models and their full description, readers are advised
to refer to Majewski (1993) and references therein.
In the case of bottom up models thin disk already exists, and the thick disk forms
later either through satellite mergers, accretion of debris or through radial mixing. These
models are successful in predicting many of the observed properties of the thick disk.
It may not be exaggerating to say that all the three possibilities may have contributed
to the formation of the thick disk. Simulations of galaxy formation through hierarchical
build up based on the CDM universe are regularly reported in the literature. Many
of these simulations predict emergence of thick disk in the galactic disk, and some of
its distinct properties from the thin disk and the halo. Quinn, Hernquist & Fullagar
(1993) proposed that satellite mergers heat the preexisting thin disk but not the gas.
Thick disk stars are primarily the fossils of the thin disk. These models predict some
of the distinct global properties like scale height and kinematic properties of thick disk
stars. Later thin disk forms from the reformed gas from the merged satellites. Models
predict only mild reduction in the rotational velocity, and do not account for the chemical
enrichment of the thick disk from the delayed SNIa. Thin disk heating by mergers was
probed further ( Kazantzidis et al. 2008) by including massive satellites to heat the thin
disk sufficiently to form the thick disk. Massive satellites were more frequent at redshifts
above z1 which corresponds to the age of the thick disk in the CDM. Alternatively,
Abadi et al. (2003) simulations of Galaxy formation in CDM universe, resulted a thick
disk consisting mostly of stars from the satellite debris and a few through accretion. A
clear distinction between dynamically cold disk of stars on nearly circular orbits and
a thicker disk with orbital parameters in between the thin disk and the halo emerged.
Results help to explain the existence of non-negligible fraction of thin disk stars which
pre-date the last major merger history. However, the chemical evolution of the thick disk
in the simulations in this framework need to be probed.
In a series of papers Brook et al. (2007 and references therein) proposed the formation
of thick disk through heating of the thin disk by gas rich mergers and starburst during the
merger process. Starburst during the merger accounted the observed thick disk properties
like kinematic dispersions, higher [/Fe], and rapid star formation. Rapid star formation
ensures higher [/Fe] even at higher metallicities with significantly much older age compared to thin disk stars with similar metallicities but much younger. The models also
predict very low velocities for the merged stars and some of them have counter rotation.
Though, simulations predict trend of decreasing of [/Fe] with [Fe/H] but stops at much
higher [/Fe] 0.2 dex even at solar metallicity. Simulations do not predict observed
drop in [/Fe] which is the signature of SNIa to the thick disk chemical enrichment (see
Figures 1 & 2).
Radically different from the merger/accretion scenarios, recent studies focused on creating the thick disk out of the Galactic disk without accretion from outside. Radial
migration of stars from inner to outer radii and the scattering of stars by spiral structure
and the molecular clouds could explain many of the observed morphological, kinematical
and chemical properties of the thick disk (e.g; Haywood 2008, Schonrich & Binney 2009).
Though, this is an interesting exercise, there are many assumptions like rate of infall and
its distribution, and scattering time scales that need to be constrained further. Models
298
Bacham E. Reddy
based on radial mixing may find it difficult to explain the evidence of counter rotating
thick disk stars that are seen in some of the spiral galaxies.
9. Conclusions
Significant progress has been made, over the last two decades, towards establishing
various observational properties. However, we still lack information on the metal-poor
side of the thick disk. This has implications to favour a particular merging model. The
issue of understanding the MWTD is related to accurate astrometry for a large number of
stars at relatively farther distances. Hopefully, this would be resolved with the proposed
GAIA space mission data expected to be available in 2015/16. Systematic studies of thick
disk in large number of spiral galaxies with varying mass would help to constrain the mass
of the merging satellites in the simulations of a particular galaxy with a particular thick
disk to thin disk mass ratio. As far as decoding the disk components in other galaxies,
we require more powerful tools like next generation large aperture telescopes which are
at the horizon. As shown in Figure 2, thick disk structure is much more complex and
observations point to multiple process for this structure: mergers, accretion of stars from
the debris, heating of the thin disk through scattering, and radial mixing. Hopefully, in
the near future, an unified model will be developed that would match all the well defined
observed parameters.
Acknowledgements
I thank SOC for inviting and giving me this opportunity to review the thick disk and
LOC for their superb arrangements at the meeting. I thank INSA, IAU and IIA for
providing the funds to attend the meeting.
References
Abadi, M.G., Navarro, J., Steinmetz, M., Eke, V.R. 2003, ApJ, 597, 21
Arifyanto, M. I., Fuchs, B., Jahrei, H., Wielen, R. 2005, AA, 433, 911
Bensby, T., Feltzing, S., Lundstr
om, I. 2003, A&A, 410, 527
Bensby, T., Feltzing, S., Lundstr
om, I., Ilyin, I. 2005, A&A, 433, 185
Bensby, T., Zenn, A.R., Oey, M.S., Feltzing, S. 2007, ApJl, 663, 13
Brook, C., Richard, S., Kawata, D., Martel, H., Gibson, B.K. 2007, ApJ, 658, 60
Burstein, D. 1979, ApJ, 234, 829
Carollo et al. 2009, arXiv, 0909.3019
Chen, Y. Q., Nissen, P. E., Zhao, G., Zhang, H. W., Benoni, T. 2000, A&AS, 141, 491
Chiba, M., Beers, T. C. 2000, AJ, 119, 2843
Edvardsson, B., Andersen, J., Gustafsson, B., Lambert, D.L., Nissen, P. E., Tomkin, J. 1993,
A&A, 275, 101
Eggen, O.J., Lynden-Bell, D., Sandage, A. R. 1962, ApJ, 136, 748
Freeman, K., & Bland-Hawthorn, J. 2002, ARAA, 40, 487
Fuhrmann, K. 1998, A&A, 338, 161
Fuhrmann, K. 2008, MNRAS, 384, 173
Gilmore, G., & Reid, N. 1983, MNRAS, 202, 1025
Gilmore, G., Wyse, R.F.G., Kuijen, K. 1989, ARAA, 27, 555
Gilmore, G., Wyse, R.F.G., Jones, J. B. 1995, AJ, 109, 1095
Haywood, M. 2008, MNRAS, 388, 1175
Juric et al. 2008, ApJ, 673, 864
Kazantizidis, S., Bullock, J. S., Zenter, A.R., Kractsov, A.V., Mosutakas, L.A. 2008, ApJ, 688,
254
Majewski, S. R. 1993, ARAA, 31, 575
299
Abstract. We have obtained high-resolution spectra and carried out a detailed elemental abundance analysis for a new sample of 899 F and G dwarf stars in the Solar neighbourhood. The
results allow us to, in a multi-dimensional space consisting of stellar ages, detailed elemental
abundances, and full kinematic information for the stars, study and trace their respective origins. Here we briefly address selection criteria and discuss how to define a thick disc star. The
results are discussed in the context of galaxy formation.
Keywords. stars: abundances, stars: kinematics, Galaxy: abundances, Galaxy: disk
1. Introduction
The study of the Milky Way is important in the context of galaxy formation. The
hierarchical build-up of galaxies within CDM, the currently most successful theory for
the formation of large-scale structure in the Universe (e.g., Springel et al. 2006), implies
that a galaxy like the Milky Way has suffered a bombardment of merging blocks over its
life time. Stellar discs are very fragile to changes in the gravitational potential, hence the
existence of spiral galaxies like the Milky Way is a challenge for CDM. However, recent
studies by, e.g., Koda et al. (2009), Stewart et al. (2009), and Scannapieco et al. (2009),
show that these effects might not be as severe as previously envisioned.
Evidence of the presence of a Galactic thick disc was first presented by Gilmore &
Reid (1983), showing that the density of stars as a function of distance from the Galactic
plane towards the Galactic north pole could not be explained by a single exponential, but
rather two exponentials with different scale heights and with different number densities
in the plane. Subsequent observational studies have established that, (compared to the
thin disc) the thick disc stellar population is a kinematically hotter population, it rotates
more slowly, it is mainly an old stellar population (as old as the stellar halo), and that it
has different chemical properties, particularly in the sense that it possesses higher [/Fe]
ratios at a given metallicity (e.g., Bensby et al. 2003, 2005, 2007b; Reddy et al. 2006;
Fuhrmann2008).
In the context of galaxy formation and as tests of models of galaxy formation, it
is important to establish the properties of a disc system of the type observed in the
Milky Way. Therefore, to fully understand why the Milky Way has two different disc
populations, how they formed, how they evolved, and clarify their relationships to the
stellar halo and the Galactic bulge we have carried out an extensive study of long-lived
F and G dwarf stars in the Solar neighbourhood.
The Galactic thin and thick disks in the context of galaxy formation
301
SOFIN, and MIKE; see Bensby et al. 2010, in prep. for details). Among the things we
targeted during our observations are: (i) the oxygen trends in the metal-rich thin disc
(Bensby et al. 2004); (ii) chemical differences between the thin and thick discs (Bensby
et al. 2003, 2005); (iii) metal-rich and old stars; (iv) the metal-rich limit of the thick disc
(first results in Bensby et al. 2007b); (v) the origin of the Hercules stream (first results in
Bensby et al. 2007a); (vi) the metal-poor limit of the thin disc; (vii) the metal-poor limit
of the thick disc; (viii) the stars with intermediate kinematics that can not be classified
as neither thin disc nor thick disk stars (using current kinematical criteria, see Bensby et
al. 2003, 2005); (ix) the metal-rich halo and its interface to the thick disc. Compared to
our previous published thin and thick disc stellar sample that contained 102 dwarf stars
(Bensby et al. 2003, 2005), the current sample is a factor 8 larger, and will allow us to
put firmer observational constraints on the Galactic discs.
A Toomre diagram, which is a representation of the stars combined vertical and radial
kinetic energies as a function of the stars rotational energy is shown in Fig. 1. Low2
2
2
velocity stars, constrained within vtot (ULSR
+ VLSR
+ WLSR
)1/2 . 50 km s1 are to a
first approximation mainly thin disc stars, and stars with vtot greater than 70 km s1 ,
but less than 200 km s1 , are likely to be thick disc stars (e.g., Nissen 2004). Stars with
higher velocities are halo stars. The Toomre diagram shows that our study sample the
regions occupied by the thin and thick disc stellar populations well. There is also an excess
of stars with velocities around VLSR 50 km s1 and (U 2 + W 2 )1/2 50 100 km s1 ,
the velocity space occupied by the Hercules stream (e.g., Bensby et al. 2007a).
302
Figure 2. Thin disc (right-hand side) and thick disc (left-hand side) abundance trends using two
different selection criteria. The stars in the top panels have been selected using the kinematical
criteria used in Bensby et al. (2003, 2005), i.e., that a star has to be at least two times likely to
belong to a stellar population than to any of the other populations. In the middle two panels
the stars have been divided according to their ages: one sample with lower age estimates greater
than 8 Gyr, and one sample with upper age estimates less than 7 Gyr. The two bottom panels
show the Toomre diagrams for the old and young samples in the middle panels.
sample of F and G dwarf stars currently limits us to the stars of the immediate Solar
neighbourhood, typically the sphere covered by the Hipparcos mission, we have to rely
on kinematical criteria when selecting our stellar sample. The top panels of Fig. 2, show
the [Fe/Ti] [Ti/H] abundance trends for stars that are at least two times more likely
of being thick disc stars than thin disc stars (top left panel), and vice versa in the top
right panel. It is clear that these two sub-samples, only being different in terms of their
kinematic properties, show quite different abundance trends. It is at the same time also
clear that in each of these plots, there exist a number of stars that deviate from the
large majority, and that actually would fit better with the other kinematic sample. This
is most likely an effect of the fact that the velocity ellipsoids of the thin a thick discs
partially overlap. Hence, when using kinematical criteria to pick either thin or thick disc
stars we a likely to pick a few high-velocity thin disc stars in the thick disc sample, and
a few low-velocity thick disc stars in the thin disc sample.
As most investigations indicate that the thick disc mainly is an old stellar population
and that the thin disc is considerably younger, it might be illustrative to try and separate
The Galactic thin and thick disks in the context of galaxy formation
303
the sample by age criteria instead. This is what we have done in the middle panels of
Fig. 2: one sample of stars whose ages have a lower age estimate of 8 Gyr (left panel) and
one sample of stars whose ages have an upper age estimate lower than 7 Gyr (right panel).
Now we see two abundance trends very alike the ones we obtained by using kinematical
criteria only. However, the number of outliers have been reduced and the abundance
trends are cleaner. The bottom panels in Fig. 2 show Toomre diagrams for these two
age-selected samples and we see that the old sample mainly is a kinematically hot sample.
The young sample, on the other hand, contains a lot of kinematically hot stars as well,
stars that, when using kinematical criteria, would be selected as thick disc stars. This
kinematic confusion demonstrates that one has to be careful when using kinematical
criteria to select thin and thick disc stars, and that better criteria to separate the two
populations are needed. However, Fig. 2 shows that, and if focus is not put on individual
stars, kinematic criteria still can be used to obtain stellar samples of the Galactic stellar
populations in order to trace their evolution.
4. Discussion
The mostly regarded scenario for the formation of the Galactic thick disc is that it is
a result of an ancient merger event between the Milky Way and another (dwarf) galaxy
(see, e.g., Quinn et al. 1993; Villalobos & Helmi 2009). However, recent simulations by
Sch
onrich & Binney (2009) that show that it is possible to form a thick disc in the
Galaxy, without a merger event, through radial mixing of gas and stars. Their model is
able to reproduce all of the current observables (abundances, stellar ages, dichotomy of
the Galactic disc, etc). So, currently, it appears as if the observable that would serve in
favour of any these two formation scenarios is if there is a hiatus in the star formation
history between the two discs. If present, the merger scenario would be favoured, as the
model by Sch
onrich & Binney (2009) implies a continuous star formation history.
Acknowledgements
Sofia Feltzing is a Royal Swedish Academy Research Fellow supported by a grant from
the Knut and Alice Wallenberg Foundation.
References
Bensby, T., Feltzing, S., & Lundstr
om, I. 2003, A&A, 410, 527
Bensby, T., Feltzing, S., Lundstr
om, I., & Ilyin, I. 2005, A&A, 433, 185
Bensby, T., Oey, M. S., Feltzing, S., & Gustafsson, B. 2007a, ApJ, 655, L89
Bensby, T., Zenn, A. R., Oey, M. S., & Feltzing, S. 2007b, ApJ, 663, L13
Fuhrmann, K. 2008, MNRAS, 384, 173
Gilmore, G. & Reid, N. 1983, MNRAS, 202, 1025
Koda, J., Milosavljevic, M., & Shapiro, P. R. 2009, ApJ, 696, 254
Nissen, P. E. 2004, in Origin and Evolution of the Elements, Carnegie Observatories Astrophysics
Series, Vol. 4, (Eds.) A. McWilliam and M. Rauch, Pasadena: Carnegie Observatories, 156
Quinn, P. J., Hernquist, L., & Fullagar, D. P. 1993, ApJ, 403, 74
Read, J. I., Lake, G., Agertz, O., & Debattista, V. P. 2008, MNRAS, 389, 1041
Reddy, B. E., Lambert, D. L., & Allende Prieto, C. 2006, MNRAS, 367, 1329
Scannapieco, C., White, S. D. M., Springel, V., & Tissera, P. B. 2009, MNRAS, 396, 696
Sch
onrich, R. & Binney, J. 2009, arXiv:0907.1899v1 [astro-ph.GA]
Springel, V., Frenk, C. S., & White, S. D. M. 2006, Nature, 440, 1137
Stewart, K. R., Bullock, J. S., Wechsler, R. H., & Maller, A. H. 2009, ApJ, 702, 307
& Helmi, A. 2009, arXiv:0902.1624v1 [astro-ph.GA]
Villalobos, A.
Abstract.
We discuss recent observations of stars located close to the symmetry plane of the Milky Way,
and examine them in the context of theories of Galaxy formation and evolution. The kinematics,
ages, and compositions of thin disk stars in the solar neighborhood display complex patterns, and
interesting correlations. The Galactic disk does not seem to pose any unsurmountable obstacles
to hierarchical galaxy formation theories, but a model of the Milky Way able to reproduce the
complexity found in the data will likely require a meticulous study of a significant fraction of
the stars in the Galaxy. Making such an observational effort seems necessary in order to make a
physics laboratory out of our own galaxy, and ultimately ensure that the most relevant processes
are properly understood.
Keywords. Keyword1, keyword2, keyword3, etc.
1. Introduction
Large galaxies naturally produce disks. Radiative cooling of the gas and angular momentum conservation lead the early evolution of galaxies through dissipative collapse and
disk formation. Disks are frequently observed in galaxies, even at high redshift (F
orster
Schreiber et al. 2006), and the Milky Way does not seem unique at all in showing a dual
disk, with a distinct thin and thick components (Dalcanton & Bernstein 2002).
Early attempts to place the Galactic thin disk in the context of a CDM universe
exposed a number of problems. The number of observed surviving satellites appeared far
too small compared to simulations. This problem is somewhat alleviated after the Sloan
Digital Sky Survey (SDSS) has identified many new low-surface brightness galaxies in
the immediate Galactic neighborhood (but see Koposov et al. 2008). It was also deemed
hard for the disk to survive for as long as the observations suggested, 810 Gyr, and in
particular to stay thin (T
oth & Ostriker 1992, Kauffmann & White 1993). More recent
appraisals, however, indicate that as many as 85% of disk galaxies have not been involved
in a merger with a mass ratio larger than 0.5 since redshift 1 1.5, or approximately
in the last 8 Gyr (Koda et al. 2009). A higher gas fraction in the accreted building blocks
seems to favor disk survival (Hopkins et al. 2009).
The structure of galaxy disks is usually studied measuring surface brightness distributions. Anomalies such as spiral arms and HII regions are smoothed out, taking deprojected azimuthal averages in nearly face-on galaxies, modeling the radial dependence
of the light distribution with exponential profiles (see, e.g., Aguerri et al. 2000; Prieto et
al. 2001). Edge-on galaxies are, in turn, used to study the light distribution perpendicular
to the plane. In the Milky Way, the spatial distribution of stars is studied using deep
imaging surveys counting stars and exploiting photometric calibrations to estimate the
luminosity of the main-sequence as a function of color (e.g. Juric et al. 2008).
Present address: Instituto de Astrofsica de Canarias, Va L
actea S/N, La Laguna 38205,
Tenerife, Spain.
304
305
Being insiders to the Galaxy provides some advantages; for example, we can measure
in detail the properties of individual stars using spectroscopy. Modern surveys employ
cameras with a very broad dynamical range, and massive multiplexing capabilities for
spectroscopy (see, e.g., Gunn et al. 1998, 2005; Onaka et al. 2008), making it feasible to
obtain large data sets fast.
306
ratio (disk/total, about 0.35 for the Milky Way), and adopting Vrot ' 220 km s1 ,
h
h
0.35 + 1.10
.
2 ' 24200
R
R
This indicates that a range in between 8 and 30 km s1 , as observed for stars between 1
and 10 Gyr should be matched by an excursion in the scale height between 60 and 600 pc.
Measuring directly scale heights from astrometry will have to wait for Gaia (Lindegren &
de Bruijne 2005; Lindegren et al. 2009), as the Hipparcos parallaxes for late-type dwarfs
are limited to 100 pc.
307
Figure 1. Velocity and metallicity histograms for 2427 stars with metallicities from the catalog
of Cayrel de Strobel, Soubiran & Ralite (2001), Hipparcos astrometry, and radial velocities from
the compilations by Malaroda, Levato & Galliani (2001) and/or Barbier-Brossat & Figon (2000),
showing velocities V > 40 and 80 < W < +50 km s1 . The expected contamination by thick
disk and halo populations is very small, and therefore we identify the observed distributions with
the thin disk. The smooth solid lines are Gaussian curves fitted to the histograms. Note the peak
associated with the Hyades at (U, V, W ) ' (43, 18, 2) km s1 . Adopted from Allende Prieto
et al. (2004).
al. (1999) that different dating techniques are complementary. For example, isochrones
are most useful for turn-off stars, and hence more likely applicable to intermediate mass
stars, while activity and rotation can provide ages for low mass stars that stay on the
main sequence longer than a Hubble time.
Thin disk stars show a wide range of ages. Reddy et al. (2006) estimated ages between
1 and 9 Gyr, although predominantly < 5 Gyr. Holmberg et al. (2009) and Haywood
(2008), using larger samples, found wider ranges reaching up to 1314 Gyr, although
most concentrated again at < 4 5 Gyr. Very old ages for thin-disk stars may be at
odd with the upper limit to the age of the disk derived from the analysis of the white
dwarf cooling sequence. For example, Leggett al. (1998) estimated 8 1.5 Gyr, although
a critical assessment of the literature by Fontaine, Brassard & Bergeron (2001) led them
to propose a plausible range between 8.5 to 11 Gyr.
An inspection of the literature in the last decade shows that there is now consensus
regarding the metallicity distribution of the thin disk. Most authors find it is reasonably
Gaussian, with a standard deviation of about 0.2 dex (see, e.g. Allende Prieto et al. 2004;
Holmberg et al. 2007). More polemic is the exact mean of the distribution, which some
argue could be as low as [Fe/H]=0.10 dex, while others push for a value much close
to solar (hence around 0.00; see Luck & Heiter 2007; Haywood 2001; Taylor & Croxall
2005; Fuhrmann 2008).
As noted some years ago, it is interesting that blindly adopting the metallicities compiled in the Cayrel de Strobel et al. catalog, and just by simply cleaning thick disk stars
308
with slow Galactic rotation (V < 50 km s1 ), one recovers again a [Fe/H] distribution
centered at 0.1 dex with ' 0.20 dex remarkably close to those found from the analysis of much more homogeneous data sets (see Fig. 1). The implication is that despite
this compilation includes high-resolution determinations from many studies, the systematics across samples and analysis protocols are not large enough to widen the derived
distribution.
An additional complication should be noted. Although the metallicity distribution of
the thin disk may be well determined, this is likely not an ideal quantity to compare with
chemical evolution models. Selection effects such as mass biases due to different lifetimes
need to be considered, and so does the role of radial migration, which could be bringing
significant numbers of stars formed at different galactocentric distances, and hence with
different compositions even if formed at exactly the same time (see, e.g., Haywood et al.
2008).
6. Abundance ratios
There have been a multitude of studies performing high-resolution spectroscopy of
nearby GFK stars (e.g., Edvardsson et al. 1993; Feltzing & Gustafsson 1998; Chen et al.
2000; Nissen et al. 2000; Fulbright 2002; Reddy et al. 2003, 2006; Takeda 2007; Ecuvillon
et al. 2004; Gilli et al. 2006; Bensby et al. 2003; Ramrez et al. 2007; Fuhrmann 1998,
2004, 2008). Most of these studies found a remarkable uniformity in the abundance ratios
for thin disk stars at any given [Fe/H].
Reddy et al. (2003) looked for and failed to find a cosmic scatter in the abundance
ratios. Assuming [Fe/H] is a reliable clock, the interstellar medium where these samples
formed was very well mixed. Many works encounter non-solar ratios at solar [Fe/H] for
some elements. This puzzling result, which might fuel the idea of the Sun being somehow
special, was later traced to systematic errors in the abundances associated with using the
Sun as a reference for non-solar type stars (Allende Prieto 2008). There is no doubt that
highly homogeneous samples, in particular those restricted to a narrow range in effective
temperature (isothermal samples, if you will), can dramatically reduce systematic errors
still present in the analyses. Melendez et al., in these proceedings, show an extreme
example of exploiting such a trick.
309
-0.5
[Fe/H]
Figure 2. Average of the abundance ratios of Mg, Si, Ca and Ti to Fe for stars kinematically
assigned to the thin (crosses) and thick disk (filled circles). The filled circles surrounded with
an additional circumference are the so-called TKTA stars; they show thin-disk abundances but
thick-disk kinematics. Adopted from Reddy et al. (2006).
(2006). Some argue there exists genuine transition objects, but not all studies find them
in their samples.
Galactic disks can become thinner by dissipative collapse while they are still rich in
gas. Such straightforward connection between the thick and thin disks does not seem
viable in the light of the distinct chemical patterns that separate the two Milky Way
disks. Stellar disks can also become thicker with time, due to internal (scattering and
other dynamical interactions in the disk) or external (satellite accretion) mechanisms.
But again, such simple path does not match the distribution of ages.
Two scenarios recently proposed in the literature appear feasible. Modelers have argued for some time that mergers could have been responsible to produce the thick disk,
perhaps disrupting a previously existing disk, but allowing the thin disk to form and
evolve independently afterwards. While the accretion of dry (gasless) system(s) may lead
to features that clash with existing observations (think rings of stars and thick-disk characteristics that vary with galactocentric distance), the acquisition of gas rich systems
and on-the-fly formation of stars seems to work well (Brook et al. 2005, 2007). The second scenario is based on new, relatively simple, models which indicate that the observed
distributions of metallicity, age, and /Fe ratios could emerge as a result of the natural
density gradient and the associated variation of the star formation rate with galactocentric distance, when coupled to radial mixing of stellar orbits (Sch
onrich & Binney 2008,
2009).
310
are found for multiple elements using different tracers (see Fig. 13 in Allende Prieto et
al. 2006).
Both Cepheids (Andrievsky et al. 2002, 2004) and giants in open clusters (Yong et al.
2006) allow tracing the abundances of many elements at large galactocentric distances. A
remarkable outcome of these studies is that the well-known thin disk abundance gradient
(see the review by Maciel in these proceedings) may flatten out at R > 12 kpc (but see
Sale et al. 2009 a for discrepant voice). This has been suggested to indicate a flat density
profile in the inner stellar halo (Cescutti et al. 2007). Interestingly enough, these studies
also show that the /F e ratios increase with galactocentric distance and so do the
ratios of lanthanum (an sprocess tracer) as well as europium (an rprocess tracer) to
iron.
The disk is by no means flat, and a better understanding of its structure is needed,
in particular the flare and warp traced by stars (L
opez-Corredoira et al. 2002; Momany
et al. 2006) gas (Kalberla et al. 2007; Levine et al. 2006), and dust (Drimmel & Spergel
2001). Infrared spectroscopic observations of vast numbers of red giants across the disk
should provide much insight. APOGEE, part of SDSS-III, plans to obtain high-resolution
H-band spectra for 105 stars with a signal-to-noise ratio approaching 100 between 2011
and 2014 (Allende Prieto et al. 2008). Preliminary studies suggest that more than 15
chemical elements can be sampled within the H band, where dust obscuration is 5 times
less than in V .
9. Closing remarks
The thin disk of the Galaxy likely fits in the overall bottom-up galaxy formation
scenario in a CDM universe, but a detailed picture of its formation is still missing. The
stellar population of the thin disk is rich in kinematic structure, but appears chemically
well-mixed. The two statements in the previous sentence need not be in contradiction, as
fine structure is likely missed due to limited abundance precision (currently 0.05 dex),
and especially if most of the structure has a dynamical origin, excited by resonances
and/or (modest) accretion.
The solar neighborhood needs to be placed in the context of the whole Galactic disk.
Massive surveys of faint stars will do that, and they will happen over the next 510 years.
Among the most pressing questions in this field, we could single out: What is has been
the star formation history of the solar neighborhood? (Must consider radial mixing!).
Which process(es) are mainly responsible for the stellar clustering in phase space and
the disk heating? What is the connection between the thin and thick disks?
The tools to address these questions are already in place: global astrometry methods,
accurate spectroscopy from efficient instruments, detailed chemical analysis techniques
(isothermal samples, larger samples, refined analyses), chemo-dynamical modeling, improved statistical techniques, and last but not least, data, to be provided by SDSS, RAVE,
Gaia, and other supporting facilities.
Acknowledgements
I am grateful to the IAU, a Sociedade Astron
omica Brasileira, and UCL graduate school
for their generous support. Thanks go also to Katia Cunha for her warm hospitality, to
Vivien Reuter for answering inquiries swiftly and with a smile, and to Martn L
opez
Corredoira for stimulating discussions and comments on a draft.
311
References
Abazajian, K. N., et al. 2009, ApJS, 182, 543
Aguerri, J. A. L., Varela, A. M., Prieto, M., & Mu
noz-Tu
no
n, C. 2000, AJ, 119, 1638
Allende Prieto, C. 2008, The Metal-Rich Universe, 30
Allende Prieto, C., Beers, T. C., Wilhelm, R., Newberg, H. J., Rockosi, C. M., Yanny, B., &
Lee, Y. S. 2006, ApJ, 636, 804
Allende Prieto, C., Barklem, P. S., Lambert, D. L., & Cunha, K. 2004, A&A, 420, 183
Allende Prieto, C., et al. 2008, Astronomische Nachrichten, 329, 1018
Andrievsky, S. M., Luck, R. E., Martin, P., & Lepine, J. R. D. 2004, A&A, 413, 159
Andrievsky, S. M., Kovtyukh, V. V., Luck, R. E., Lepine, J. R. D., Maciel, W. J., & Beletsky,
Y. V. 2002, A&A, 392, 491
Arifyanto, M. I., & Fuchs, B. 2006, A&A, 449, 533
Aumer, M., & Binney, J. J. 2009, MNRAS, 397, 1286
Baranne, A., Mayor, M., & Poncet, J. L. 1979, Vistas in Astronomy, 23, 279
Barbier-Brossat, M., & Figon, P. 2000, A&AS, 142, 217
Bensby, T., Feltzing, S., & Lundstr
om, I. 2003, A&A, 410, 527
Bensby, T., Oey, M. S., Feltzing, S., & Gustafsson, B. 2007, ApJ, 655, L89
Bertelli, G., & Nasi, E. 2001, AJ, 121, 1013
Bilir, S., Cabrera-Lavers, A., Karaali, S., Ak, S., Yaz, E., & L
opez-Corredoira, M. 2008, Publications of the Astronomical Society of Australia, 25, 69
Brook, C., Richard, S., Kawata, D., Martel, H., & Gibson, B. K. 2007, ApJ, 658, 60
Brook, C. B., Gibson, B. K., Martel, H., & Kawata, D. 2005, ApJ, 630, 298
Cabrera-Lavers, A., Bilir, S., Ak, S., Yaz, E., & L
opez-Corredoira, M. 2007,A&A, 464, 565
Cayrel de Strobel, G., Soubiran, C., & Ralite, N. 2001, A&A, 373, 159
Cescutti, G., Matteucci, F., Francois, P., & Chiappini, C. 2007, A&A, 462, 943
Chakrabarty, D. 2007, A&A, 467, 145
Chen, Y. Q., Nissen, P. E., Zhao, G., Zhang, H. W., & Benoni, T. 2000, A&AS, 141, 491
Dalcanton, J. J., & Bernstein, R. A. 2002, AJ, 124, 1328
De Silva, G. M., Freeman, K. C., Bland-Hawthorn, J., Asplund, M., & Bessell, M. S. 2007, AJ,
133, 694
De Simone, R., Wu, X., & Tremaine, S. 2004, MNRAS, 350, 627
Drimmel, R., & Spergel, D. N. 2001, ApJ, 556, 181
Ecuvillon, A., Israelian, G., Santos, N. C., Mayor, M., Villar, V., & Bihain, G. 2004, A&A, 426,
619
Edvardsson, B., Andersen, J., Gustafsson, B., Lambert, D. L., Nissen, P. E., & Tomkin, J. 1993,
A&A, 275, 101
Eggen, O. J. 1992, AJ, 104, 2141
Famaey, B., Jorissen, A., Luri, X., Mayor, M., Udry, S., Dejonghe, H., & Turon, C. 2005, A&A,
430, 165
Famaey, B., Siebert, A., & Jorissen, A. 2008, A&A, 483, 453
Feltzing, S., & Gustafsson, B. 1998, A&AS, 129, 237
Fontaine, G., Brassard, P., & Bergeron, P. 2001, PASP, 113, 409
F
orster Schreiber, N. M., et al. 2006,ApJ, 645, 1062
Fuhrmann, K. 2004, Astronomische Nachrichten, 325, 3
Fuhrmann, K. 1998, A&A, 338, 161
Fuhrmann, K. 2008, MNRAS, 384, 173
Fulbright, J. P. 2002, AJ, 123, 404
Gilli, G., Israelian, G., Ecuvillon, A., Santos, N. C., & Mayor, M. 2006, A&A, 449, 723
Gunn, J. E., et al. 1998, AJ, 116, 3040
Gunn, J. E., et al. 2006, AJ, 131, 2332
Haywood, M. 2001, MNRAS, 325, 1365
Haywood, M. 2008, MNRAS, 388, 1175
Hernandez, X., Valls-Gabaud, D., & Gilmore, G. 2000, MNRAS, 316, 605
Holmberg, J., Nordstr
om, B., & Andersen, J. 2009, A&A, 501, 941
312
Abstract. The modern observations of the planetary nebulae (PNe) were used to create a new
catalogue of He, C, N, O, Ne, S and some other element abundances for more than 300 Galactic
and extragalactic PNe. This catalogue was a start point for studying the star formation history
in the Galactic disk. We have found that He, C, N and O abundances in the thin disk PNe are
very close to those in bulge PNe. We also established that the abundances of the same elements
in the PNe belonging the halo of the Milky Way and the Magellanic Clouds are similar. We
have estimated the mean ages of the progenitors of PNe in the thin and thick disks and in the
bulge. The evidences in the favor of the suggestion that the formation of the intermediate-mass
star began in bulge and only then continued in the thin disk were found.
Keywords. galaxies: evolution, planetary nebulae: general, Magellanic Clouds
1. Introduction
The PN phenomenon is an intermediate phase between the stars and interstellar
medium. So the research of the properties and chemical abundances in the ensemble
of Galactic and Magellanic Clouds planetary nebulae open us a possibility to study
of star formation history and the chemical evilution of the Milky Way. The following
problems are considered:
The comparison of the chemical abundance of the PNe ensemble with the synthetic
models of the LIMS evolution.
An evolution status of the PNe with the extremely low element abundances.
A comparison of the element abundances in the different galactic subsystems.
A comparison of the abundances for nebulae in our Galaxy and in the Magellanic
Clouds.
Determination of abundance gradients for the Milky Way PNe ensembles both for
all galactic PNe and for nebulae with the different ages of their progenitors.
Investigation of the evolution of the element abundances in the different regions of
the Milky Way in a comparison with the models of the Galaxy chemical evolution.
314
Fig. 1 the calculated and observed UV line intensities in spectra of planetary nebulae to
demonstrate a good quality of our model fit.
11.5
11.5
11
11
11
10.5
10
9.5
10
10
9.5
9.5
log(X/H)+12
log(X/H)+12
lg(X/H)+12
10.5
10.5
9
8.5
9
8.5
7.5
7.5
6.5
6.5
8.5
He
He
He
Figure 2. Comparison of the He, C, N, and O abundances in PNe of various Galactic subsystems:
(a) Galactic type IIa PNe (dark columns) and bulge nebulae (light columns); (b) the same as
(a) for PNe of the Galactic halo and the LMC; (c) the same as (b) for SMC nebulae.
In figure 2 (b,c) we compare the He, C, N, and O abundances in PNe of the Galactic
halo and the Magellanic Clouds. A similarity between the elemental abundances in these
objects can be seen. It can be the indicator of their close evolutionary status.
315
disk are similar to bulge objects we compare the He, C, N, and O abundances in bulge
and disk nebulae.
9.1
9.2
9.1
8.9
8.8
[O]=lg(O/H)+12
[O]=lg(O/H)+12
9.3
8.9
8.8
8.7
8.7
8.6
8.5
8.6
8.4
8.5
8.3
8.2
8.4
8.3
10
8.1
5
R
10
Figure 3. (a) Averaged oxygen abundances in type IIa (asterisks) and bulge (filled square) nebulae vz. their mean Galactocentric distance. The solid line gives a linear fit to this dependence,
the dotted lines shows one standard deviation from the fit. (b) The same as Fig. (a), but for all
type II nebulae.
The He, C, N and O abundances in the thin disk PNe (type IIa in the modified
Peimberts classification given by Quireza et al., 2007) appeared to be close to those for
bulge PNe (see Fig. 2 (a)).
In this case the dependence of elemental abundances in PNe on the galactocentric
distance R should must be continuous from thin disk to bulge. To test this assumption,
we analyzed the radial dependence of the O abundance, which is almost constant during
the evolution of intermediate-mass stars. As can be seen in Fig. 3 the mean oxygen
abundances for bulge PNe continue a dependence of ratio [O/H] for thin disk PNe.
The oxygen abundance gradient d[O/H]/dR = 0.017 0.01 dex kpc1 , when only the
type IIa nebulae are included in the sample. It is slightly higher than a value 0.012 dex
kpc1 obtained by Kholtygin and Milanova (2006) from the analysis of the O abundance
for all Galactic disk PNe. The value of d[O/H]/dR = 0.031 0.014 dex kpc1 For all
type II PNe. The type IIa nebulae are thin disk objects and they have an intermediate
age of their progenitor stars, which is equal to 4-6 Gyr from the present epoch (Quireza et
al. 2007). We propose that the bulge is formed on the early stage of the Galaxy evolution.
After that the stellar formation begins in the disk starting from the regions close to bulge.
We compare the mean radial abundance gradients for disks of spiral galaxies in the
table 1. We can see from the table that our values for the Milky Way disk are in a good
agreement with those for galaxy M33, while our abundance gradient for thin Galactic
disk PNe is similar to those for the Andromeda galaxy.
316
-0.03
0.012 0.011
-0.046
-0.08
0.028 0.01
0.102 0.009
0.016 0.017
-
-0.012
-0.019
-
-0.031
Present work
4.5
20
18
16
3.5
Z=0.008
Heburnng
models
14
2.5
sun
12
lg(L/L
N([He/H])
Garnett (1997)
Crockett et al., 2006
Garnett (1997)
Garnett (1997)
Cedres et al., 2004
Garnett (1997)
Lunyova &
Kholtygin (2002)
10
8
2
1.5
0.95M
1M
1.5M
2M
2.5M
Lowabund PNe
1
6
0.5
0.5
0
0.5
0
[He/H]
0.5
1
5.4
5.2
4.8
4.6
4.4
lgTeff
4.2
3.8
3.6
Figure 4. Left panel: distribution of [He/H] abundance ratio for PNe with low abundances.
Right panel: low abundances PNe central stars (asterisks) on a H-R diagram. Tracks are taken
from He-burning star evolutionary model by Vassilidias & Wood (1994) with initial metallicity
Z = 0.008. Initial masses of stars (from bottom to top) are equal to 0.95, 1.0, 1.5, 2.0 and 2.5
solar masses.
the low abundant PNe. We see the concentration of the central stars of the considered
PNe in the region of the low masses of the progenitors.
This study was supported by a grant from the President of Russia (NSh-1318.2008.2) for
Support of Leading Scientific Schools.
References
Cedres, B., Urbaneja, M.A., Cepa, J., 2004, A&A, 422, 511
Crockett, N.R., Garnett, D.R., Massey, P., Jacoby, G., 2006, Ap.J., 637(2), 741
Garnett, D.R., Shields, G.A., Skillman, G.A., et al., 1997, Ap.J., 489, 63
Golovatyj V.V., Sapar A., Feklistova T., Kholtygin A.F., 1997, Catalogue of atomic data for
low-density astrophysical plasma, Astron. Astroph. Trans., 12, 85
Kholtygin, A.F., 1998, A&A, 329, 691
Lunyova, Yu.V., Kholtygin, A.F., 2002, Astrophysics (English translation of Astrofizika), 45(3),
370
Milanova, Yu.V. & Kholtygin, A.F. 2006, Astronomy Letters, 32, 557
Milanova, Yu.V. & Kholtygin, A.F. 2009, Astronomy Letters, 35, 563
Quireza, C., Rocha-Pinto H.J.,& Maciel, W.J. 2007, A&A, 475, 217
Vassilidias, V., Wood, P.R. 1994, Ap.J. Suppl. Ser., 92, 125
1. Introduction
Radial metallicity gradients are observed in the disks of many galaxies, including the
Milky Way, galaxies of the Local Group and other objects. Current research topics include
the determination of (i) the magnitude of the gradients, (ii) any space variations along
the disk and (iii) possible time variations during the evolution of the host galaxy. In this
paper, we present some recent observational evidences of radial abundance gradients.
Our main focus is the Milky Way, but it will be shown that the analysis of some objects
in the Local Group, particularly M33, is useful in order to study the main properties of
the gradients in our own Galaxy. A brief discussion of some recent theoretical models
is also given, in an effort to highlight their predictions concerning the radial abundance
gradients and their variations. Some recent reviews and general papers on abundance
gradients include: Freeman (2009), Maciel & Costa (2009), Rudolph et al. (2006), and
Stasi
nska (2004). Theoretical models are discussed by a number of people, including
Fu et al. (2009), Magrini et al. (2009a), Cescutti et al. (2007), Moll
a and Daz (2005),
Chiappini et al. (2003, 2001, 1997), Hou et al. (2000), and Gensler (these proceedings).
Recent discussions on azimuthal and vertical gradients, not treated here, can be found
in Davies et al. (2009) and Ivezic et al. (2008).
318
behaviour of the abundances along the galactic disk. The main reason is that cepheids
are bright enough to be observed at large distances, so that accurate distances and
spectroscopic abundances of several elements can be obtained, which in principle allow
the determination of radial variations of the gradients better than any other indicator.
The ages of these objects are generally under 200 Myr (cf. Maciel et al. 2005), so that
the measured gradients can be safely considered as present-day gradients. Recent work
on Cepheids include Lemasle et al. (2008), Pedicelli et al. (2009), and Romaniello et al.
(2008). Lemasle et al. (2008) obtained high-resolution spectroscopic iron abundances for
galactic cepheids for which accurate distances were determined based on near-infrared
photometry. The abundances were determined within 0.12 dex, while the distances, based
on a near-infrared period-luminosity relation, are expected to be accurate within 0.5 kpc
in average. In order to improve the sampling process, additional objects were included.
The average gradient obtained in the range 517 kpc is d[Fe/H]/dR = 0.052 0.003
dex/kpc. A better solution proposed by the authors includes a change of slope, in the
sense that the gradient is steeper in the inner galaxy, with a flattened gradient of 0.012
0.014 dex/kpc in the outer galaxy. In this region the abundances show an increased
spread, as compared with the inner galaxy. The change of slope occurs at about R ' 10
kpc, farther away than the solar radius, located at R = 8.5 kpc. These results have
been largely confirmed by the recent work of Pedicelli et al. (2009), in which again a
large sample from different sources was considered with a new photometric metallicity
calibration. These results suggest a gradient of 0.051 0.004 dex/kpc for the whole
sample, or 0.130 0.015 dex/kpc for the inner sample (R < 8 kpc) and 0.042 0.004
dex/kpc for the outer Galaxy.
2.2. HII regions
Concerning HII regions, several determinations have been presented in the last few years,
which include Deharveng et al. (2000), Esteban et al. (2005), Rudolph et al. (2006) and
Quireza et al. (2006). Deharveng et al. (2000) obtained an oxygen gradient of 0.04
dex/kpc in the range 515 kpc, with no indication of flattening further out in the disk.
Pilyugin et al. (2003) compiled spectra of 13 HII regions in the range 714 kpc with
available [OIII]4363 measurements, and recomputed the oxygen abundances from the
data by Shaver et al. (1983), obtaining a gradient of 0.051 dex/kpc. Esteban et al. (2005)
obtained echelle spectrophotometry of 8 HII regions in the range 610 kpc, and derived
carbon and oxygen abundances from recombination lines. The oxygen gradient obtained
is 0.044 0.010 dex/kpc. More recently, Rudolph et al. (2006) used both infrared and
optical data to study abundance gradients in a sample of 117 HII regions. The data
include both new results and a reanalysis of previous material. The best fit to optical
data corresponds to a gradient of 0.060 dex/kpc. For the infrared data a gradient of
0.041 dex/kpc was obtained. Quireza et al. (2006), determined the electron temperature
gradient from high-precision radio recombination line and continuum measurements for
over a hundred HII regions, calibrated in terms of the oxygen abundance gradient. The
derived slope obtained using a relation between the electron temperature and the oxygen
abundances by Shaver et al. (1983) is 0.043 0.007 dex/kpc, in good agreement with
the previous work by Deharveng et al. (2000) and Esteban et al. (2005).
From these results it is difficult to establish whether or not the gradients flatten out
at large galactocentric distances. However, studies of HII regions in spiral galaxies are
consistent with essentially constant abundances in the outer parts of these objects, as
for example in the recent work of Bresolin et al. (2009a) on the extended disk of M83, in
which a flat oxygen gradient was obtained beyond the R25 isophotal radius, regardless
of the abundance indicator used.
319
320
Figure 1. The O/H radial gradient from PN. Top: Distances from the IAG/USP group, Bottom:
Distances from Stanghellini et al. (2008). Left: CSPN with ages in the range 210 Gyr; Right:
CSPN with ages in the range 46 Gyr.
clusters, ranging from about 1 Gyr to about 8 Gyr (cf. Maciel et al. 2009), so that any
time variation of the gradients would have to be taken into account. On the other hand,
abundances of elements such as O, Ne, S, and Ar can be obtained within about 0.2 dex
in average, which is probably lower than the abundance spread at a given galactocentric
radius. The results presented in the last couple of years have been rather contradictory. Pottasch and collaborators (cf. Pottasch and Bernard-Salas 2006) have presented
accurate abundance data based both on optical and infrared measurements from ISO,
which do not need the consideration of the usually uncertain ionization correction factors
(ICF), since more ionized species can be observed. As a result, the derived abundances
are expected to be more accurate than in the case of the traditional plasma diagnostic
method. The results suggest the presence of a strong negative gradient similar to the
ones observed in HII regions and early type stars. For the O/H ratio a slope of 0.085
dex/kpc was found. For most of the elements considered, which are O, Ne, S, and Ar, the
predicted abundances at the solar radius obtained by taking into account the observed
gradients match exactly the solar abundances. In a more recent work (Gutenkunst et
al. 2008), a larger sample was considered, including bulge nebulae. It can be concluded
that the gradient flattens out near the galactic bulge, a result also obtained by Cavichia
et al. (2009). In contrast with these results, Stanghellini et al. (2006) studied a sample
of galactic PN with abundances derived from the traditional optical plasma diagnostic
method, obtaining a flat gradient for oxygen and neon, based on a simple linear fit. However, from what we have seen in the previous sections, it seems clear that the gradients
flatten out in the outer Galaxy, so that using a single fit for the whole disk may be misleading, especially if the inner Galaxy, where the gradient is steeper, is undersampled as
is the case here. Moreover, the data in this sample located in the region around the solar
321
Figure 2. Time variation of the radial abundance gradient (Maciel & Costa 2009).
circle clearly show some evidence of a steeper gradient, so that a flat gradient for the
whole disk is probably incorrect. A more detailed analysis was made by Perinotto and
Morbidelli (2006), who considered a larger and more complete sample of PN for which
the abundances were recalculated in a homogeneous way. The results also suggest relatively flat gradients (< 0.04 dex/kpc) for oxygen, but a careful analysis of their data
shows that the uncertainties in the distances, coupled with a possible time variation of
the gradients, may wash out the gradients, so that a careful selection of the objects must
be made. This can be seen by considering the oxygen gradients for PN of sets A and B,
defined as follows: set A includes 131 objects whose abundances are considered by the
authors as the most reliable, and set B is a control sample, containing all PN abundances
published between 2000 and 2005, with about 200 objects. In order to avoid any bias
due to the adopted distances, Perinotto and Morbidelli (2006) considered four different
statistical scales. Considering for example the results corresponding to the distances by
Cahn et al. (1992), a gradient is apparent from set A, while set B presents a flat distribution, suggesting that the uncertainties in the abundances contribute to erase any existing
gradients. Some indication of the existence of a time variation of the gradients, in the
sense that the present-day gradient is flatter than in the past can also be observed from
the results by Perinotto and Morbidelli (2006). They have analyzed separately the PN
according to the Peimbert types, in which Type I are expected to be younger objects,
while Type III are older nebulae, generally located at higher distances from the galactic
plane and with a larger peculiar velocity. Type II are intermediate age objects in this
scheme. According to Perinotto and Morbidelli (2006), Type I objects do not show any
gradients in both sets, while Type II and III show measurable, albeit low, gradients. Also,
for the distance scales with a meaningful sample, the gradients of Type III objects are
larger than for Type II PN.
A different approach has been taken by the IAG/USP group (Maciel et al. 2003, 2005,
2006, 2009, Maciel & Costa 2009). Here an effort has been made to divide the PN sample
into age groups, as was done for the open clusters by Magrini et al. (2009a), so that
any time variation of the gradients could be appropriately taken into account. Several
methods have been developed to obtain the age distribution of the PN central stars.
Fig. 1 shows the O/H gradient for disk objects with ages of 210 Gyr (left) and 46 Gyr
(right), using the distance scale adopted by our IAG Basic Sample (top), or the distance
scale by Stanghellini et al. (2008) (bottom). The effect of restricting the age interval is
322
Figure 3. Left: The [OIII] electron temperature gradient from planetary nebulae. Right: The
correlation between the electron temperatures and oxygen abundances (Maciel et al. 2007).
clear in both figures. Taking into account the age distribution of the PN progenitor stars
(Maciel et al. 2009), it is possible to separate the PN sample according to their ages
(young, intermediate, old), so that an estimate of the time variation of the gradients can
be obtained. This is shown in Fig. 2 (Maciel & Costa 2009), where other objects are also
considered, namely open clusters, cepheids, OB stars and HII regions. As a conclusion,
data from planetary nebulae support the flattening of the gradients near the bulge-disk
interface and at large galactocentric distances. However, anticentre nebulae are difficult
to observe, and the problem of the distances is still a complicating factor. A considerable
improvement is expected with the advent of GAIA. As for the time variation of the
gradients, a conservative conclusion at this stage is that either they have not changed
very much in the last 6 Gyr approximately, or they may have flattened out by a small
amount. These conclusions are supported by some recent theoretical models by Fu et al.
(2009), which take into account infall, star formation based on the Kennicutt law and a
delayed disk formation.
323
5. Conclusions
From all objects considered, some tentative conclusions may be drawn: (1) Average
abundance gradients are generally between 0.03 dex/kpc and 0.10 dex/kpc, but a
single value for the whole disk may be misleading. (2) Most evidences point to a flattening
out of the gradients at large galactocentric distances. (3) There are some clear evidences
of a flattening of the gradients near the galactic bulge. (4) The change of slope in the outer
Galaxy occurs in the region around R ' 10 kpc. (5) Any further change of the slope needs
better data than presently available. Cepheids may be an exception. (6) There are no
evidences of a steepening of the gradients at large galactocentric distances, as suggested
by some theoretical models. (7) Either the gradients do not change appreciably during
galactic evolution, or they flatten out at a moderate rate. (8) There are no clear evidences
of a steepening of the gradients with time, as suggested by some theoretical models.
Acknowledgements. This work was partially supported by FAPESP and CNPq.
324
References
Andrievsky S.M., Bersier D., Kovtyukh V.V., et al. 2002b, A&A, 384, 140
Andrievsky S.M., Kovtyukh V.V., Luck R.E., et al. 2002a, A&A, 381, 32
Andrievsky, S.M., Kovtyukh, V.V., Luck, R.E., et al. 2002c, A&A, 392, 491
Andrievsky, S.M., Luck, R.E., Martin, P., Lepine, J.R.D. 2004, A&A, 413, 159
Bresolin, F., Gieren, W., Kudritzki, R.P., et al. 2009b, ApJ, 700, 309
Bresolin, F., Ryan-Weber, E., et al. 2009a, ApJ, 695, 580
Cahn, J. H., Kaler, J. B., Stanghellini, L. 1992, A&AS, 94, 399
Cavichia, O., Costa, R. D. D., Maciel, W. J. 2009, in preparation
Cescutti, G., Matteucci, F., Francois, P., Chiappini, C. 2007, A&A, 462, 943
Chiappini, C., Matteucci, F., Gratton, R. 1997, ApJ, 477, 765
Chiappini, C., Matteucci, F., Romano, D. 2001, ApJ, 554, 1044
Chiappini, C., Romano, D., Matteucci, F. 2003, MNRAS, 339, 63
Cioni, M. R. 2009, A&A, in press (astro-ph/0904.3136)
Colavitti, E., Cescutti, G., Matteucci, F., Murante, G. 2009, A&A, 496, 429
Davies, B., Origlia, L., Kudritzki, R., et al. 2009, ApJ, 696, 2014
Deharveng, L., Pe
na, M., Caplan, J., Costero, R. 2000, MNRAS, 311, 329
Esteban, C., Garcia-Rojas, J., Peimbert, M., et al. 2005, ApJ, 618, L95
Freeman, K. C. 2009, IAU Symp. 254, ed. J. Andersen et al., CUP, 111
Friel, E. D., Janes, K. A., Tavarez, M., et al. 2002, AJ, 124, 2693
Fu, J., Hou, J. L., Hin, J., Chang, R. X. 2009, ApJ, 696, 668
Gutenkunst, S., Bernard-Salas, J., Pottasch, S. R., et al. 2008, ApJ, 680, 120
Hou, J. L., Prantzos, N., Boissier, S. 2000, A&A, 362, 921
Ivezic, Z., Sesar, B., Juric, M., et al. 2008, ApJ, 684, 2871
Lemasle, B., Francois, P., Piersimoni, A., et al.. 2008, A&A, 490, 613
Luck, R.E., Gieren, W. P., Andrievsky, S.M., et al. 2003, A&A, 401, 939
Maciel, W. J., Costa, R. D. D. 2009, IAU Symp. 254, ed. J. Andersen et al., CUP,
electronic publication (astro-ph/0806.3443)
Maciel, W. J., Costa, R. D. D., Idiart, T. E. P. 2009, IAU Symp. 258, ed. E.E. Mamajek et al.,
electronic publication (http://abstracts.stsci.edu/IAU258/Maciel.pdf)
Maciel, W. J., Costa, R. D. D., Uchida, M. M. M. 2003, A&A, 397, 667
Maciel, W. J., Lago, L. G., Costa, R. D. D., 2005, A&A, 433, 127
Maciel, W. J., Lago, L. G., Costa, R. D. D., 2006, A&A, 453, 587
Maciel, W. J., Quireza, C., Costa, R. D. D., 2007, A&A, 463, L13
Magrini, L., Corbelli, E., Galli, D. 2007b, A&A, 470, 843
Magrini, L., Sestito, P., Randich, S., Galli, D. 2009a, A&A, 494, 95
Magrini, L., Stanghellini, L., Villaver, E. 2009b, ApJ, 696, 729
Magrini, L., Vilchez, J. M., Mampaso, A., et al., 2007a, A&A, 470, 865
Moll
a, M., Daz, A. I. 2005, MNRAS, 358, 521
Pedicelli, S., Bono, G., Lemasle, B., et al. 2009, A&A, in press (astro-ph/0906.3140)
Perinotto, M., Morbidelli, L. 2006, MNRAS, 372, 45
Pilyugin, L. S., Ferrini, F., Shkvarun, R. V. 2003, A&A, 401, 557
Pottasch, S., R., Bernard-Salas, J. 2006, A&A, 457, 189
Quireza, C., Rood, R. T., Bania, T. M., Balser, D. S., Maciel, W. J. 2006, ApJ, 653, 1226
Romaniello, M., Primas, F., Mottini, M. 2008, A&A, 488, 731
Rubin, R. H., Simpson, J. P., Colgan, S. W. J., et al. 2008, MNRAS, 387, 45
Rudolph, A. L., Fich, M., Bell, G. R., et al. 2006, ApJS, 162, 346
Sestito, P., Bragaglia, A., Randich, S., et al. 2008, A&A, 488, 943
Shaver, P. A., McGee, R. X., Newton, L. M., et al. 1983, MNRAS, 204, 53
Stanghellini, L., Guerrero, M. A., Cunha, K., et al. 2006, ApJ, 651, 898
Stanghellini, L., Shaw, R. A., & Villaver, E. 2008, ApJ, 689, 194
Stasi
nska, G. 2004, In: Cosmochemistry. The melting pot of the elements. XIII Canary Islands
Winter School of Astrophysics, Ed. C. Esteban, R. J. Garca L
opez, et al., CUP, 115
1. Introduction
The evolution of the Universe is clearly manifested by its structure formation and the
consumption of gas to star formation, the latter leading to the enrichment of chemical
elements heavier than those stemming from the primordial nuclesynthesis. Knowing four
facts, the star-formation (SF) rate at any time and its integral of the Hubble time, the
initial stellar mass function (IMF), the production and yield of each element for each
particular stellar mass, in each stage of the stellar life, respectively, would well define the
interstellar enrichment and with some SF delay also the stellar, so that an observational
trace-back of the chemical evolution (CE) of the Universe would allow to derive one of
these parameters definitely and for different specific structures.
Assuming that a structure like e.g. a galaxy in total or also any galactic region under
consideration behaves like an isolated volume, called a closed box, the longer-living lowmass stars lead to increase the stellar mass fraction continuously as well as the stellar
remnants lock-up mass while as a consequence of SF the gas fraction diminishs. The basic
set of analytical equations for the chemical evolution of galaxies has been formulated
by numerous authors (e.g. Talbot & Arnett (1971), Pagel & Patchett (1975), Tinsley
(1980), Pagel (1997) and a lot more since then: see references in the recent papers by
Prantzos (2008a) and Recchi et al. (2008)). In this simple temporal relation for the
metallicity as Zi (t) = yi [ln()] follows where = Mg (t)/M(g,0) is the temporal gas mass
fraction and yi the yield of element i, i.e. the metallicity release per stellar population
(see left panel of fig.1).
Unfortunately, nature is not as simple and our present knowledge has a lot of uncertainties and even gaps because of models that are reasonably and necessarily much
simpler than nature and some processes are not yet well understood. A strong simplification is implied to the analytical solution by the lifetimes of the stellar factories and
thus the various delay times of specific element releases. The discrepancy of the so-called
instantaneous recycling approximation becomes obvious after almost 2-3 Gyrs as demonstrated in fig. 12 of Prantzos (2008a). A trace-back of element ratios through long-living
325
326
Figure 1. Evolution in arbitrary time units of an element Zi (t) (green dots), of total mass m
(blue rhombi), gas mass fractions g (red squares), and stellar mass fraction s (black triangles)
with respect to t=0 for a closed-box model (left) and for the extreme inflow of primordial gas
(right) necessary to get the gas mass g constant.
stars is clearly representing this delay, the most prominent signature being e.g. in the
[/Fe]-[Fe/H] trend of the solar vicinity.
Another simplification can be relaxed by advancing the closed box to open boundaries
and allowing for gas infall or outflow. Fig. 1 (right panel) demonstrates the extreme case
of gas infall which is required to balance the gas comsumption. By this, also the stellar
Z enhancement is reduced by dilution with primordial gas. As a further limitation the
released gas is assumed to mix instantaneously with the local environment.
Another effect also enters into the element distribution, namely, the existence of different gas phases produced by means of various heating processes and the splitting of
stellar metal injecta to these phases. While the SF is limited to the cool gas alone, cooling
timescales of dilute hot gas are extremely long, by this, allowing to refrain characteristic elements from their immediate incorporation into newly formed stars. The ignorance
concerning such small-scale mixing timescales between the different gas phases until now
allows only an order-of-magnitude approach for chemical models. Mixing processes are
not only caused by diffusion in an otherwise static gas but triggered by dynamical effects.
This means that last but not least basic ingredients to understand and to model the CE
is gasdynamics.
Despite of the enormous activities all over the field of CE but because of limited space
and time in this paper, we cannot cover all aspects and objects under investigation of CE
and also not refer to the whole bunch of literature, but have to pick out a few papers as
representative and review the present state of modelling the CE of some cosmic structures
only from the early universe to the local. We therefore apologize for the reasonable
incompleteness. Starting from the solar vicinity the paper focusses on some hot topics
as e.g. the galactic halo and its dwarf spheroidal (dSphs) satellite system finishing with
dwarf gas-rich galaxies.
Since present-day chemical models are preferentially performed by numerical simulations a further aspect on the quality and reliability of models should focus on the applied
numerical methods. Since it is not the aim of this review to also address and discuss the
validity of numerical methods of galaxy evolution, the interested reader is referred to a
dedicated recent review by Hensler (2008).
327
the galactic CE are mostly executed for the different galactic components in separation
only and to some extent implying dynamical processes thru analytical terms, not selfconsistently but instead tuning them corresponding to a best fit of the observations.
2.1. The Milky Way disk
One of the problems within the solar neighbourhood realized at earliest already by Beatrice Tinsley (1974) was the so-called G-dwarf problem, the lack of low-metallicity G
dwarfs. Although this problem can be simply solved by exponentially declining infall
(but also outflow or both), Pagel (1987) has already emphasized that additional deviations from a constant slope in the ln()Z diagram, the yield y, point towards dynamical
influences. Moreover, Prantzos (2008a) (his fig. 14) shows that a metal pre-enrichment
leads to a better fit to the stellar Z ln() distribution.
Therefore, non-dynamical CE models have been conducted with a temporal and radial
function of a gas infall rate (see e.g. Chiappini et al. (1997), Portinari & Chiosi (1999)).
Because also the SF rate is related to the gas surface density g as (r, t) kg (r, t),
where represents a SF efficiency, the infall steers the SF. The exponent k can be set
to unity or adapted to the original Kennicutt (1989) relation 1.4. Additional factors are
also included, as e.g. the total mass density (see e.g. Chiappini et al. (1997), eq.4).
The most cited model of this kind by Chiappini et al. (1997) can fit the G-dwarf
problem at best by two infall episodes, a short one with 1 Gyrs timescale only and a
long one of 8 Gyrs. Astrophysically speaking this so-called two-infall model resembles
nothing exceptional or unexpected than a short initial collapse phase and the general
gas assembly with long-term decreasing infall rate. Nevertheless, also the SF efficiency
as another free parameter is changed between both infall eposides. Depending on the
stellar sample used for the G-dwarf problem the models fit more or less (Chiappini et al.
1997), but a quantitative comparison of the model abundances at present-day with solar
is desillusionary and not well surportive for the model since it provides not more than
a qualitative tendency. The same is discernible for the stellar age-metallicity relation
(AMR) in Chiappini et al. (1997).
On short timescales, however, the gas accretion rate should vary locally and affect the
SF (Pflamm-Altenburg & Hensler 2010) and the CE (Hensler et al. 2004a) as well as the
abundances (Koeppen & Hensler 2005) what was also invoked from observations of the
solar neighbourhood by Knauth et al. (2006).
How uncertain the deduction of relations can be and to some extent biased by the
sample taken for studies can also be documented by a comparison of the original AMR
of stars found by Twarog (1980) with that derived by Nordstroem et al. (2004) from
a well-defined and complete sample of solar vicinity stars. While the traditional results
revealed the expected tendency but with very strong early enrichment, already Tadross
(2003) did not find any relation at all from open cluster analyses, while Nordstroem et
al. (2004) data yield a trend. Showing a large hardly understandable scatter [Fe/H]
by almost 0.5, even for the average AMR tendency a CE model is hardly constructable
without any metal pre-enrichment.
The evolution of radial abundance gradients is reasonable, but values, ranges, and
the temporal behaviour are also uncertain (see Maciel & Costa, this volume). Galactic
abundance gradients that are measured from Hii regions as usually by the standard
method of collisionally excited emission line (CEL, Shaver et al. 1983) or recently used
recombination lines (RL) in the infrared (Esteban et al. 2005) can only yield the presentday abundances in the ISM. Over years of determinations the slope has not only been
reduced, e.g. for log(O/H) from -0.8 per kpc (Shaver et al. 1983) to -0.44 (Esteban et al.
2005), but also a flattening of the gradients in the innermost radii and also outside the
328
solar circle (Vilchez & Esteban 1996) became available from RLs. From massive stars
one can derive the same: the chemical abundances of the ISM from which they are born
most recently.
While Maciel et al. (2003) could derive a temporal flattening of the oxygen gradient
over the past 9 Gyrs from -0.11 to -0.06 dex/kpc unsing Planetary Nebulae, from stars
of different ages Nordstroem et al. (2004) could even attribute a positiv slope to disk
stars older than 10 Gyrs, and again a slight flattening for younger stars from almost -0.1
dex/kpc to -0.077 since 1.5 Gyrs.
While the present-day abundance gradients from Hii regions hold for all the determined
elements, i.e. C, N, O, and elements, their different radial slopes need explanation.
Interestingly, the abundance ratio C/O shows also a radial decline what means that C is
more abundant in O-rich regions. As explanation Carigi et al. (2005) proposed that this
behaviour is caused by a metal-dependent C yield. On the other hand, also the O yield
is metal dependent in the sense that at lower metallicities a weaker wind by massive
stars allow a longer shell-burning production of O (Maeder 1992). Cescutti et al. (2009)
modelled the chemical evolution of bulge and disk stars with metal-dependent yields by
Meynet & Maeder (2002b) and could by this also reproduce the enhancement of C/O in
metal-rich regions like the bulge.
Figure 2. Evolution of a 85 M star until its explosion as supernova typeII. Left panel:
Wind mass loss adapted to evolutionary models by Schaller et al. (1992). Right panel:
Time-dependent abundances of 12 C, 14 N, and 16 O the warm Hii region gas after the onset
of the WN phase at 2.83 Myrs. See Kroeger et al. (2006).
For the determination of C and O in Hii regions, another effect has to be studied
seriously: self-enrichment by Wolf-Rayet stars (Kroeger et al. 2006). At the end of its
lifetime at t = 3.22 Myr a 85 M star has supplied 0.28 M of 14 N, 13.76 M of 12 C, and
11.12 M of 16 O, which are contained in the combined stellar wind bubble/Hii region.
Since N was released at first in the WN stage it increases slightly in this period and is
thereafter diluted by the N-poor gas feed. These facts are discernible in the hot gas by a
first rise and a subsequent decrease of the N abundance (Kroeger et al. 2006) after the
transition to the WC stage when C and O are released. The C content in the hot gas
increases steeply and reaches an overabundance of 38 times solar while the enrichment
with O is weaker. Due to turbulent mixing of the hot gas with the photo-evaporated shell,
the warm Hii gas is also enriched but at most for C by not more than a factor of 1.22
(Fig. 2, right panel). Since the stellar mass range of the Wolf-Rayet phase increases with
metallicity, the negative slope of the C/O ratio within the galactic disk can be attributed
to this effect.
329
In a recent study Colavitti et al. (2009) found that the chemical properties of the
galactic disk are at best reproduced by models with an infall law derived from cosmological but purely Dark Matter simulations. In reality, the disk settling timescale is not
a simple stellar dynamical one but inherently determined by its thermal and turbulent
energy budget causing a delayed thin-disk formation following the thick disk (Burkert,
Truran, & Hensler 1992).
2.2. Multi-zone models
As already implied as advancement of simple CE models one can even account for gas exchanges through the boundaries of limited regions under consideration by simply adding
a corresponding time-dependent (source or sink) term to the gas-mass and metallity
equations. This was e.g. done in several above-mentioned models with gas infall to the
galactic disk. Nonetheless, no simple model exists that also account for outflow from the
disk. A first step was the exploration of the effect of the galactic fountain model on radial
metallicity gradients by Spitoni et al. (2009). The idea is, that the disk gas pushed-up by
supernovae (SNe) and redistributed while falling back on ballistic paths has to mix with
the disk gas within the site of impinge and affects the abundance gradient of the galactic
disk. As general results they found that neither the ballistic delay nor the cooling delay of
this fountain-driven mixing process alter the metallicity and the radial oxygen gradient.
As well as in this approach also e.g. for infall models, authors preferably avoid the
complexity of a proper hydrodynamical treatment combined with the chemical and thermal descriptions by means of an analytical ansatz. From gas infall and expected radial
gas flows, the above mentioned models mimic the first process by radius-dependent functions, but drop the latter effect. In order to differentiate between radial disk regions
and to follow radial flows, multi-zone models were developed e.g. that devide the galactic
disk into rings and allow for a radial gas flux, are artificially parametrized due to positive
flows characterizing the inside-out evolution of the disk (Portinari & Chiosi 2000) or are
consistently determined by the gas pressure gradient.
Different approaches were performed by coupling structural parts of the Milky Way,
e.g. halo and disk, and dealing with mass exchange between the regions (Ferrini et al.
1992, 1994). By fine-tuning not only the coefficients for mass exchange arbitrarily, but
also further parameters, as e.g. SF efficiency, etc., rough quantitative agreements with
gas and stellar mass, abundances, and their radial gradients in the disk could be achieved.
2.3. Chemo-dynamics
For all these simple-type CE models it is inherent that their results in comparison with
observational facts can provide a first insight into amount and timescales on which nonlocal effects have played a role during the evolution. Nonetheless, they are all devoid
of self-consistency according to the acting processes and the resulting dynamics. Two
major sources exist, driving the gas dynamics, namely, gravitation and stellar energy
release. Plasmaphysical processes according to heating and cooling lead to different gas
phases that have not only different energetics but also behave dynamically different and
interact by multiple processes as e.g. by drag forces, shock waves, and turbulence as well
as energeticly e.g. by heat conduction. Moreover, this means that also SF and thermal
state of the ISM have to be treated consistently.
A first-order approach into this direction is provided by multi-phase CE models (also
implied in Ferrini et al. 1992, 1994) where, however, also the interaction terms in a set
of equations are not self-consistently determined but arbitrarily applied.
How crucial an appropriate representation of the ISM and its processes is, will be
demonstrated in a few of the next sections. Two main ingredients can already be estab-
330
lished here: 1) Cool/warm and hot gas phases are coupled energeticly by heat conduction,
leading to a self-regulation of SF (Koeppen, Theis, & Hensler 1998), local mixture of hot
and cool gas and, by this, local metal enrichment of cool gas with SN elements. 2) The
dynamical coupling of both gas phases by drag and mass loading due to evaporation
hampers the outflow and enhances the cooling of the hot gas.
Figure 3. Metallicity evolution of three different components, cool/warm gas (CM, dark
hatched), hot intercloud medium (ICM, light hatched), and stars (S, grey) on average of a
Milky Way-type chemo-dynamical model (Samland, Hensler & Theis 1997) at 10 kpc galactocentric distance within the disk. The right panel is a cut-out of the left one demonstrating how
the metallicity Z is affected by different processes: 1) Due to infall of primordial gas, ZCM of the
decreases. 2) ZICM increases due to supernova-expelled metals and is always higher than ZCM .
3) The ZICM decreases therefore, when CM evaporates by heat conduction. 4) Condensation of
hot gas onto cool clouds increases their Z. The stellar Z represents an average over stars of all
ages and is therefore smaller than ZCM . (from Samland (1994)).
Up to now as the optimal grid code one can consider the further development of the
chemo-dynamical scheme by Samland, Hensler & Theis (1997) to 3D and with the stellar
dynamics for the stars. In addition, a cosmologically growing DM halo is included into the
simulations by Samland & Gerhard (2003). These models contain all the crucial processes
of SF self-regulation by stellar feedback, multi-phase ISM, and temporally resolved stellar
components according to the chemo-dynamical prescription (Hensler 2003). They cannot
only trace the formation and evolution of the disk galaxies components but also of
characteristic chemical abundances and are until now the best self-consistent evolutionary
models of disk galaxies because the disk formation is included into the global temporal
galaxy evolution.
2.4. The Milky Way halo
Not before half a decade ago only poor knowledge on the element abundances of the Milky
Way halo stars was available. With the wealth of new class 8-10m telescopes, spectroscopy
of faint and distant halo stars has changed our picture fundamentally. While in the era
of Sandage & Fouts (1987) a halo metallicity [Fe/H] not below -4 could be found and
only a few below -3, raising doubts on the first population of stars in the halo and their
chemical fingerprints, recent detections of hyper metal-poor stars (Beers & Christlieb
2005) in the galactic halo and their peculiar element abundances (see e.g. Frebel et al.
2005) opened a wide field of new activities, namely, modelling towards understanding the
zero metallicity nucleosynthesis and studying the formation of the halo. This is necessary
in order to decide between the two preferred formation scenarios of the Milky Way: the
monolithic collapse model (Eggen et al. 1962) and the accretion model (Searle & Zinn
1978). A general word of clarification of the two terms should be done here: Even if the
331
accretion model has formed an inhomogeneous halo (Argast et al. 2000, Oey 2000) its
mixing of infalling gas with stellar ejecta lead to a settling of more or less homogenised
gas with low angular momentum into the bulge and with larger one into a disk. Due to
the higher gas column density in the innermost disk due to its smaller area the mostly
invoked inside-out evolution in the disk is quite plausible.
Although a dichotomy of halo stars, the inner halo vs. the outer halo (Carollo et al.
2005), makes a dissipative formation of the inner halo probable but invokes a dissipationless accretion process for the outer halo stars, it is still under debate whether part of the
present-day dwarf spheroidal galaxies (dSphs) could have served as its building blocks.
Although most of the dSphs have formed their stellar population in an early short, but
active SF epoch, the chemical signatures of most of the dSph stars still alive are distinct
from the stars of each kinematic component of the Milky Way (Venn et al. 2004).
Even the most recent investigation by Prantzos (2008b) which is sold as a clear proof
that dSphs have served at least partly as building blocks of the Milky Way halo, lacks of
consistency: SF history, baryonic mass content and metallicity have been assumed from
the 4 most prominent present-day dSphs, and the accretion timescale was set to half a
billion years only.
Since even the CE of the dSphs is not yet conclusively understood, in general, models
have to be advanced for this morphological galaxy type.
3. Dwarf galaxies
3.1. The Milky Way dwarf spheroidal satellite system
Since the origin and the Dark Matter content of the galactic system of dSph satellites
is basically questioned by some authors (Metz et al. 2009), general problems are: Can
low-mass systems survive an intense SF epoch (as expected at the formation)?
From compilations by Dekel & Woo (2003) and Grebel et al. (2003) dSphs follow
a mass-metallicity relation. As an extreme case Hensler et al. (2004b) have performed
spherical low-mass galaxies by means of chemo-dynamical simulations in order to study
the galaxy survival, SF rates, gas loss, and (final) metallicity. They could demonstrate
that due to the SF self-regulation only short but vehement initial SF epochs occur and
lead to mass-dependent gas loss. Nonetheless, the DGs remain gravitationally bound with
the further issue that more cool gas survives than it is observed, but it forms a halo around
the visual body. Although the stellar energetic feedback is the driving mechanism to expel
the gas, its effect is not as dramatic as obtained in semi-analytic models (Salvadori et al.
2008) and the amount of unbound mass is considerably lower. To get lost, this gas has to
be stripped off additionally (Grebel et al. 2003) what probably happens because of the
ram pressure of the galactic halo gas or of tidal stripping (Read et al. 2006). Otherwise
it can bounce back to the DG and produce subsequent events, from a second SF epoch
even up to SF oscillations. The external gas reservoir around the Scl dSph (Bouchard
et al. 2003) might whitness this effect. While the models make some intermediate-age
stars, most stars are made in less than 1 Gyr after formation. The dependence of stellar
metallicity on the galaxy mass is a quite well confirmed observational relation.
Because of its simplifying one-dimensionality and the neglection of environment effects
like e.g. tidal effects, external gas pressure, gas inflow, etc. more complex models are required. The fascinating wealth of data and their precision on stellar ages and kinematics,
on their chemical abundances and tidal tails of dSphs (for most recent reviews see e.g.
Koch (2009) and Tolstoy et al. (2009)) have triggered numerous numerical models. Although all of them up to date are advanced by 3D hydrodynamics (see e.g. Marcolini et
332
al. (2006) and Revaz et al. (2009)), they still lack of the same aforementioned environmental agents. The most comprehensive paper by Revaz et al. (2009) e.g. simulates a
bunch of DG models with the method of smooth-particle hydronamics (SPH), but again
all of them in isolation. In their models mainly also sufficient gas mass is retained and
can fuel further SF epochs if it would not be stripped of by ram pressure or tidal forces,
as the authors mention. Those models that fit the presently best studied dSphs Fnx, Car,
Scl, and Sex at best, are than assumed as test cases for further exploration. Although
their results do not deviate too much from the further observational data, in addition
to the already mentioned neglections 3 further caveats must be mentioned: 1) If models
are selected according to agreement with one or two observed structural parameters, it is
not surprising if also other values would not deviate significantly. 2) The numerical mass
resolution of the SPH particles is too low to allow quantitative issues of galactic winds,
heating and cooling, etc. 3) Because of the single gas-phase description released metals
are too rapidly mixed with the cool gas and the metal-enrichment happens too efficiently.
Despite these facts, with appropriate initial conditions always models in agreement to
observations can be found.
Although the advancement a two-phase ISM by an SPH code is not trivial and inserts
various numerical problems, but is not impossible (Berczik, Hensler, et al. 2003, Harfst,
Theis & Hensler 2006, Scannapieco et al. 2006), such treatment would be absolutely
necessary in order to achieve reliable results. In addition the chemo-dynamical interaction
proceeses must be implied.
3.2. Dwarf Irregular Galaxies
Dwarf irregular galaxies (dIrrs) are characterized by large gas fractions, often an ongoing
SF and low metallicities. Until 10 years ago it was believed that at least some of them
were forming stars now for the very first time. Nowadays it is evident that even the most
metal-poor ones (like IZw18) contain stars at least 1 Gyr old (Momany et al. 2005). This
means that SF should have proceeded for at least a few Gyr in dwarf irregular (dIrr)
galaxies, albeit at a low intensity. The low intensity of the star formation in dwarf galaxies
is the best way to explain their chemical characteristics, like for instance the low [/Fe]
ratio. As explained in the introduction, the [/Fe] vs. [Fe/H] plot is representative of the
different delay in production of -elements (mostly produced by the short-living massive
stars) and the iron (2/3 of which comes from longer-living binary systems originating
Type Ia SNe). If the SF duration in a galaxy is very short, Type Ia SNe do not have
enough time to restore Fe into the ISM and most of the stars will be over-abundant in
-elements compared to Fe. The low average [/Fe] ratios in dIrrs (compared to large
galaxies) is a hint of a long-lasting (presumably very mild) SF in these galaxies (Ikuta &
Arimoto 2002; Lanfranchi & Matteucci 2004). The same trend of [/Fe] (decreasing as we
move towards dwarf galaxies) can be obtained by varying the IMF (Recchi et al. 2009),
but the hypothesis of a low SF efficiency in dwarf galaxies is important also to explain
the downsizing (Cowie et al. 1996), namely the trend of having older stellar populations
as we move towards more massive galaxies, indicative of a shorter (but more intense)
duration of the SF in large galaxies.
On the other hand, most people believe that a fundamental role in the chemical evolution of dIrrs is played by galactic winds. In fact, these systems are characterized by
shallow potential wells and less energy is required to extract gas from them. Galactic
winds have the effect of reducing the metallicity of a galaxy in comparison with the one
predicted by the closed box model. However, detailed numerical simulations (DErcole
& Brighenti 1999; Recchi et al. 2006) have shown that galactic winds are not very effective in removing gas from a galaxy. In fact, galactic winds develop vertically (along
333
the direction of steepest pressure gradient) whereas the horizontal transport along the
disk (where most of the gas lie) is very limited. On the other hand, the metals freshly
produced during a starburst can be easily carried out of the galaxies through a galactic
wind (which will be therefore metal-enhanced) and this, too, contributes to keep the
metallicity of the galaxy low. Although this overall picture is almost widely accepted,
the effect of galactic winds depends very sensibly on parameters like galaxy structure
and ISM properties. In particular, the presence of clouds can hamper the development of
galactic winds while in the meantime, through evaporation of metal-poor gas, reducing
the average metallicity of the galaxy. Detailed numerical simulations (Recchi & Hensler
2007) show that the leakage of metals is not prevented by the presence of clouds and
that the final metallicity is a few tenths of dex lower than in models without clouds.
4. Concluding Remarks
Because of its importance for understanding the evolution of baryonic structures in the
Universe, their chemical evolution is one of the main focusses of galaxy investigations.
For simplicity, numerous analytical approaches have been performed for different morphological units. They can provide a first insight into the temporal and spatial element
enrichment by stellar nucleosythesis products.
Since dynamical effects of different gas phases affect the element enrichment differentially leading to a redistributions of those elements, but because also plasmaphysical
processes as e.g. heat conduction and turbulence allow for a mixing of gas, more complex dynamically and chemically coupled descriptions have to be developed for advanced
models. The development and application of multi-phase so-called chemo-dynamical descriptions are on the way.
How important the implimentation of the multi-phase character of the ISM is, can
be simply understood by the following basic considerations of its inherent nature: In
a low-mass galaxy the stellar energy release can overcome the binding energy of the
whole gas reservoir and the gas is totally lost - and by this also the freshly produced
metals. Semi-analytical (Salvadori et al. 2008) or single-phase hydrodynamical (MacLow
& Ferrara 1999) models largely overestimate the gas loss due to the instantaneous energy
mixture, but to some extent also because of an arbitrarily high SF rate. Although the
first-order 1D chemo-dynamical approach (Hensler et al. 2004b) even overestimates the
momentum transfer by the drag exerted from hot gas flows to the cool clouds and leads
to an overly more efficient gas removal than in 2D or 3D chemo-dynamical models e.g.
of dIrrs (Hensler et al. 1999), the model description is still closer to reality because it
allows the outflow of hot gas, while hampered by mass loading and drag of cool infalling
intergalactic clouds. The amount of expelled metals is reduced (Recchi & Hensler 2007).
On massive galaxy scales, as e.g. the evolution of giant ellipticals (gE) (Pipino et al.
2006), however, this picture can turn. When hot SNeII gas of 106 107 K mixes with
a sufficiently high amount of 104 K and cooler gas, a rough estimate gives that with a
SF efficiency of 10% and a Salpeter IMF about 10% of the stellar mass are returned by
massive star ejecta, i.e. winds + SNeII, what is 1% of the initial SF gas. Accordingly,
the temperature of the mixed gas then amounts to 2.1 104 K - 1.2 105 K, i.e. always
in the range of shortest cooling time, and, by this, remains always bound. A collapsing
massive galaxy is not stopped and metal-enriched gas is completely following the collapse.
Lateron, during the evolution this changes due to the gas consumption so that the SN gas
dominates and gas at above 106 K has reasonably to re-expand (Pipino et al. 2006). It is
not surprising that out of a model grid any model will be found to fit the observations
at best, but nonetheless the validity of its issued parameters must be questioned. Again
334
the simple 1D chemo-dynamical model (Theis, Burkert, & Hensler 1992) demonstrates
that a metallicity of 1-2 times solar can be reached within 1 Re of a gE, while the hot
gas with 3-4 times solar metallicity extends out to 10 Re .
The here discussed considerations aim at sensibilizing the reader to the validity of CE
models. Finally, we wish to mention that the considerations of CE of tidal-tail DGs, of
the intra-cluster medium, of Ly-alpha structures in the early Universe, and furthermore,
are also in the focus of crucially important studies and provide further details and to
some extent initial and boundary conditions for the CE of galaxies.
Moreover, one should also keep in mind that the stellar yields applied to CE models
are still uncertain and change due to advanced stellar evolutionary models, taking into
account stellar rotation (see e.g. Chiappini et al. (2003), Hirschi et al. (2005), Meynet et
al. (2006)), binarity (De Donder & Vanbeveren 2004), stellar winds, etc.
Acknowledgement: The authors wish to thank A. Hren for providing some figures,
the symposium organizers for their invitation to this review, and the University of Vienna
for travel support.
References
Argast, D., Samland, M., Gerhard, O.E., & Thielemann, F.-K. 2000, A&A, 356, 873
Beers, T.C. & Christlieb, N. 2005, ARAA, 43, 531
Berczik, P., Hensler, G., Theis, C., & Spurzem, R. 2003, Ap Space Sci., 284, 465
Bouchard, A., Carignan, C., & Mashenko, S. 2003, AJ, 126, 1295
Burkert, A., Truran, J.S.W., & Hensler, G. 1992, ApJ, 391, 651
Carigi, L., Peimbert, M., Esteban, C.,& Garcia-Rojas, J. 2005, ApJ, 623, 213
Carollo, D., Beers, T.C., Lee, Y.S., et al. 2005, Nature, 450, 1020
Cescutti, G., Matteucci, F., McWilliams, A., & Chiappini, C., 2009 A&A, 505, 605
Chiappini, C., Matteucci, F., & Gratton, R. 1997, ApJ, 477, 765
Chiappini, C., Matteucci, F., & Meyner, G. 2003, A&A, 410, 257
Colavitti, E., Cescutti, G., Matteucci, F., & Murante, G. 2009, A&A, 496, 429
Cowie, L.L., Songalia, A., Hu, E.M., & Cohen, J.G. 1996, AJ, 112, 839
Dekel, A. & Woo, J. 2003, MNRAS, 344, 1131
De Donder, E. & Vanbeveren, D. 2004, New Astron. Rev., 48, 861
DErcole, A., & Brighenti, F. 1999, MNRAS, 309, 941
Eggen, O. J., Lynden-Bell, D., & Sandage, A. 1962, ApJ, 136, 748
Esteban, C., Garca-Rojas, J., Peimbert, M., et al. 2005, ApJ (Letters), 618, L95
Ferrini, F., Matteucci, F., Pardi, M.C., & Penco, U. 1992, ApJ, 387, 138
Ferrini, F., Molla, M., Pardi, M.C., & Diaz, A.I. 1994, ApJ, 427, 745
Frebel, A., Aoki, W., Christlieb, N., et al. 2005, Nature, 434, 871
Grebel, E.K., Gallagher, J.S., & Harbeck, D. 2003, AJ, 125, 1966
Harfst, S., Theis, C., & Hensler, G. 2006, A&A, 499, 509
Hensler, G. 2003, in: C. Charbonnel et al. (eds.), ASP Conf. Ser. Vol., 304, 371
Hensler, G. 2008, in: J. Andersen, J. Bland-Hawthorn, & B. Nordstroem (eds.), The Galaxy
Disk in Cosmological Context, Proc. IAU Symp., 254, p. 269
Hensler, G., Rieschick A., & Koeppen, J. 1999, in: J. Beckman & T.J. Mahoney (eds.), The
Evolution of galaxies on Cosmological Timescales, ASP Conf. Ser., 187, 214
Hensler, G., Koeppen, J., Pflamm, J., & Rieschick, A. 2004a, in: P.-A. Duc, J. Braine, & E.
Brinks (eds.) Recycling intergalactic and interstellar matter, Proc. IAU Symp., 217, 178
Hensler, G., Theis, C., & Gallagher, J.S., III. 2004b, A&A, 426, 25
Hirschi, R., Meynet, G., & Maeder, A. 2005, A&A, 433, 1013
Ikuta, C., & Arimoto, N. 2002, A&A, 391, 55
Kennicutt, R.C. 1989, ApJ, 344, 685
Knauth, D.C., Meyer, D.M., & Lauroesch, J.T. 2006, ApJ (Letters), 647, L115
Koch, A. 2009, AN, 330, 675
335
The Australian National University, Mt. Stromlo Observatory, Cotter Rd., Weston ACT
2611, email: chiaki@mso.anu.edu.au
Abstract. We simulate the formation and evolution of galaxies with a self-consistent 3D hydrodynamical model including star formation, supernova feedback, and chemical enrichment.
Hypernova feedback plays an essential role not only in solving the [Zn/Fe] problem, but also
reproducing the cosmic star formation rate history and the mass-metallicity relations. In the
Milky-Way type galaxy, the star formation history, and thus the kinematics and chemical abundances are different in bulge, disk, and thick disk.
1. Introduction
While the evolution of the dark matter is reasonably well understood, the evolution of
the baryonic component is much less certain because of the complexity of the relevant
physical processes, such as star formation and feedback. With the commonly employed,
schematic star formation criteria alone, the predicted star formation rates (SFRs) are
higher than what is compatible with the observed luminosity density. Thus feedback
mechanisms are in general invoked to reheat gas and suppress star formation. Supernovae inject not only thermal energy but also heavy elements into the interstellar medium,
which can enhance star formation. Chemical enrichment must be solved as well as energy
feedback. Feedback is also important for solving the angular momentum problem and
the missing satellite problem, and for explaining the existence of heavy elements in intracluster medium and intergalactic medium, and the mass-metallicity relation of galaxies
(Kobayashi et al. 2007, hereafter K07).
In the next decade, high-resolution multi-object spectroscopy (HERMES) and space
astrometry mission (GAIA) will provide kinematics and chemical abundances of a million
stars in the Local Group. Since different heavy elements are produced from different supernovae with different timescales, elemental abundance ratios can provide independent
information on age. Therefore, stars in a galaxy are fossils to tell the history of the
galaxy. The galactic archeology technique can be used to study the galaxy formation and
evolution in general. Metallicities are measured in various objects with different galaxy
mass scale and as a function of redshift/time. The internal structure of galaxies have
being observed with integral field spectrographs (e.g., the SAURON project, SINFONI
on VLT). Chemodynamical simulations can provide useful predictions and physical interpretation of these observations.
337
Figure 1. Evolutions of heavy element abundance ratios [X/Fe] against [Fe/H] for one-zone
models with only SNe II (long-dashed line), with our new yields and SN Ia model (short-dashed
line), with the double-degenerate scenario of SNe Ia (dotted line), and with AGB stars (solid
line). The dots are observational data (see K06 for the references).
from carbon to zinc are in good agreement with observations in the solar neighborhood,
bulge, halo, and thick disk (Kobayashi et al. 2006, hereafter K06).
Figure 1 shows the evolutions of heavy element abundance ratios [X/Fe] against [Fe/H]
with our new yields (short-dashed lines), and with only SNe II (long-dashed lines, Nomoto
et al. (1997)s yields adopted). The star formation history and total number of SNe Ia
are determined to meet the observed metallicity distribution function. In the early stage
of galaxy formation, only SNe II explode, and [/Fe] stays constant. Because of the
delayed Fe production by SNe Ia, [/Fe] decreases toward 0. -elements, O, Mg, Si, S,
and Ca, show the plateau at [Fe/H]
< 1. Ti is underabundant overall, which could be
solved with 2D calculation. The observed decrease in the odd-Z elements (Na, Al, and
Cu) toward low [Fe/H] is reproduced by the metallicity effect on nucleosynthesis. The
iron-peak elements (Cr, Mn, Co, and Ni) are consistent with the observed mean values at
2.5
< [Fe/H]
< 1, and the observed trend at the lower metallicity can be explained
by the energy effect under the assumption of inhomogeneous enrichment. Note that Cr
II observations are plotted.
The most important improvement is in Zn. The observed abundance of Zn ([Zn/Fe]
0) can be explained only by such large contribution of HNe. Since the observed [Zn/Fe]
shows a increase toward lower metallicity (Primas et al. 2000; Nisesn et al. 2007), the
HN fraction may be larger in the earlier stage of galaxy formation. Since neutron-rich
isotopes 6670 Zn are produced at high metallicity, the HN fraction can be as small as
1%. In the following chemodynamical simulations, we adopt HN = 0.5, 0.5, 0.4, 0.01 for
Z = 0, 0.001, 0.004, 0.02. Pair-instability supernovae, which produce much more Fe, more
[S/Fe], and less [Zn/Fe], should not to contribute galactic chemical evolution.
Type Ia Supernovae (SNe Ia) The progenitors of the majority of SNe Ia are
most likely the Chandrasekhar (Ch) mass white dwarfs (WDs). For the evolution of
accreting C+O WDs toward the Ch mass, two scenarios have been proposed: One is the
double-degenerate (DD) scenario, i.e., merging of double C+O WDs with a combined
mass surpassing the Ch mass limit, although it has been theoretically suggested that it
leads to accretion-induced collapse rather than SNe Ia. The other is our single-degenerate
(SD) scenario, i.e., the WD mass grows by accretion of hydrogen-rich matter via mass
transfer from a binary companion.
338
Chiaki Kobayashi
We construct a new model of SNe Ia, based on the SD scenario, taking account of
the metallicity dependences of the WD wind (Kobayashi et al. 1998) and the massstripping effect on the binary companion star (Kobayashi & Nomoto 2009). Our model
naturally predicts that the SN Ia lifetime distribution spans a range of 0.1 20 Gyr
with the double peaks at 0.1 and 1 Gyr. While the present SN Ia rate in elliptical
galaxies can be reproduced with the old population of the red-giants+WD systems, the
large SN Ia rate in radio galaxies could be explained with the young population of the
main-sequence+WD systems. Because of the metallicity effect, i.e., because of the lack of
winds from WDs in the binary systems, the SN Ia rate in the systems with [Fe/H]
< 1,
e.g., high-z spiral galaxies, is supposed to be very small. We succeed in reproducing the
galactic supernova rates with their dependence on the morphological type of galaxies,
and the cosmic SN Ia rate history with a peak at z 1. At z
> 1, the predicted SN Ia
rate decreases toward higher redshifts and SNe Ia will be observed only in the systems
that have evolved with a short timescale of chemical enrichment. This suggests that the
evolution effect in the supernova cosmology can be small.
From [Fe/H] 1, SNe Ia start to occur producing more Fe than -elements, and
thus [/Fe] decreases toward the solar abundance. The decreasing [Fe/H] depends on the
SN Ia progenitor model. Our SN Ia model can give better reproduction of the [(, Mn,
Zn)/Fe]-[Fe/H] relations in the solar neighborhood than other models such as the DD
scenario (dotted lines). With the DD scenario, the typical lifetimes of SNe Ia are 0.1
Gyr, which results in the too early decrease in [/Fe] at [Fe/H] 2. Even with our SD
model, if we do not include the metallicity effect, [/Fe] decreases too early because of
the shortest lifetime, 0.1 Gyr, of the MS+WD systems. In other words, the metallicity
effect is more strongly required in the presence of the young population of SNe Ia.
Asymptotic Giant Branch (AGB) stars Stars with initial masses between
about 0.8 8M (depending on metallicity) produce light elements such as C and N,
while the contribution of heavier elements are negligibly small in the galactic chemical
evolution (solid lines). The nucleosynthesis yields of AGB stars involves uncertainties due
to the computational difficulty of the thermal pulse. We introduce the new calculation by
Karakas (2009), which is based on Karakas & Lattanzio (2007). The Na overproduction
problem has been solved with the updated reaction rates (Karakas 2009).
Needless to say, the envelope mass and pre-existing heavy elements are returned by
stellar winds from all stars. Surprisingly, in many hydrodynamical simulations, this contribution is not included, which should affect the metallicity and star formation history.
3. Chemodynamical Model
The details of our chemodynamical models are described in Kobayashi (2004), and can
be summarized as follows. i) The Smoothed Particle Hydrodynamics (SPH) method is
adopted with individual smoothing lengths and timesteps. For the simulations of individual galaxies, the gravity is calculated in direct summation using the special purpose
computer GRAPE (GRAvity PipE). For cosmological simulations, we basically use an
SPH code GADGET-2 by Springel (2005), being modified with the physical processes
in Kobayashi (2004). ii) Radiative cooling is computed using a metallicity-dependent
cooling function (Sutherland & Dopita 1993). iii) Our star formation criteria are the
same as in Katz (1992); (1) converging flow; ( v)i < 0, (2) rapid cooling; tcool < tdyn ,
and (3) Jeans unstable gas; tdyn < tsound . The star formation timescale is proportional
to the dynamical timescale (tsf 1c tdyn ), where the star formation timescale parameter
c = 0.1 is adopted (Kobayashi 2005). If a gas particle satisfies the above star formation
criteria, a fractional part of the mass of the gas particle turns into a star particle. Since
an individual star particle has a mass of 1057 M , it dose not represent a single star, but
339
Figure 2. Age-metallicity relations (upper panels) and [O/Fe]-[Fe/H] relations (lower panels)
in the solar neighborhood (a and b), thick disk (c and d), and bulge (e and f). The contours
show the mass density for the simulation. The dots show the observations of stars in the solar
neighborhood.
an association of many stars. The mass of the stars associated with each star particle is
distributed according to an initial mass function (IMF). We adopt a Salpter IMF with a
slope x = 1.35, which give good agreement with many observations except for the [/Fe]
problem in elliptical galaxies. iv) For the feedback of energy and heavy elements, we
do not adopt the instantaneous recycling approximation. Via stellar winds, SNe II, and
SNe Ia, thermal energy and heavy elements are ejected from an evolved star particle as
functions of time and metallicity, and are distributed to all surrounding gas particles out
to a constant radius of 1 kpc (4), or with the number of feedback neighbors NFB = 405
(5). v) The photometric evolution of a star particle is identical to the evolution of a
simple stellar population (SSP). SSP Spectra are taken from Kodama & Arimoto (1997)
as a function of age and metallicity.
340
Chiaki Kobayashi
5. Cosmological Simulations
The story is not closed yet. Since the resolution is not enough to describe the small-scale
physics, the modelling of feedback involves a parameter. The model has to be checked
with a different scale of observational constraints. The most stringent one is the massmetallicity relation of galaxies. We simulate the evolution of dark matter, gas, and stellar
systems from the cosmological initial condition with H0 = 70 km s1 Mpc1 , m = 0.3,
= 0.7, b = 0.04, n = 1, and 8 = 0.9. The initial condition is calculated in a 10h1
Mpc cubic box with periodic boundary conditions with NDM = Ngas = 963 .
With the larger energy ejection by HNe, the SFR starts to be suppressed from z 6
onwards, and is overall reduced by a factor of 3 at 0
<z
< 3. We then succeed in reproducing both the observed cosmic SFRs and stellar density evolution (K07). The present
stellar fraction is less than 10% being consistent with the recent observational estimate
(Fukugita & Peebles 2004), while 25% of baryons turn into stars without feedback.
The metal enrichment timescale depends on the environment. In large galaxies, enrichment takes place so quickly that [O/H] reaches 1 at z 7, which is consistent
341
Figure 3. (a) Mean metallicities of cold gas (T < 104 K) within 10 kpc, plotted against the
total stellar mass. The lines are observational data: Tremonti et al. (2004, z = 0), Savaglio et
al. (2005, z = 0.7), Erb et al. (2006, z = 2.2), and Maiolino et al. (2008, z = 3.5). (b) Mean
stellar metallicities within 10 kpc, V-band luminosity-weighted. Kobayashi & Arimoto (1999,
solid line) from metallicity gradients, Gallazzi et al. (2005, dashed line), and Quider et al. (2009,
z = 2 3) for lensed galaxy.
with the sub-solar metallicities of the Lyman break galaxies (Pettini et al. 2001). The
low metallicities of DLA systems (Prochaska et al. 2003) are also consistent with our
galaxies, provided these systems are dwarf galaxies or the outskirts of massive galaxies.
The low [C/H] of the IGM (Schaye et al. 2003) can be explained if the IGM is enriched
only by SNe II and HNe. The average metallicity of the universe reaches [O/H] 2
and [Fe/H] 2.5 at z 4 , but reaches the same values at z 3 in the IGM.
How are heavy elements ejected from galaxies to the IGM? In the simulation, we can
trace the orbit of gas particles over time. Exploiting this, we define as wind particles
those that are not in galaxies now, but have been in galaxies before (Fig. 15 in K07).
In this simulation, 10% of baryons turn into stars, 10% of the gas stays in galaxies
( 8% is hot), and 20% is ejected as galactic winds. The rest, half of the baryons,
never accretes onto galaxies. Tracing the orbits of gas particles, we can also examine from
which galaxies the wind gas particles are ejected, and measured the ejected wind mass
from each galaxy (Fig. 16 in K07). Winds are efficiently ejected from small galaxies, with
80% of accreted baryons being ejected from Mtot 1011 M galaxies. A similar relation
is also found for the ejected metal fraction, i.e. the ratio between the wind metal mass
to the total metal mass. It is interesting that the wind fraction and the ejected metal
fraction correlate well with the stellar metallicity. Based on this finding, we conclude that
the origin of the mass-metallicity relation is the mass-dependent galactic winds.
Figure 3 shows of the mass-metallicity relations of simulated galaxies (dots) and observations (lines). For the gas-phase metallicity, there are uncertainties such as inhomogeneity and aperture effect, and a large scatter is seen. The simulation is comparable
to observations upto z 2, but there is an offset at z = 3.5. This may be a problem,
but the normalization of observations needs to be checked. For the stellar metallicity, a
tight relation is present since z = 5 in the simulation. At high redshifts, observations are
available for only 2 lensed galaxies, which is remarkably consistent with the simulation.
References
Kobayashi,
Kobayashi,
Kobayashi,
Kobayashi,
Kobayashi,
Abstract.
The Galactic structure and composition remain as one of the greatest open problems in
modern astrophysics. We show here that there are chemical similarities between the Galactic
bulge and local thick disk red giant stars. This finding puts strong constraints on the IMF, SFR
and chemical enrichment timescale of the bulge and thick disk. Our results are based upon a
detailed elemental abundance analysis of 80 high S/N and high resolution optical spectra of
giant stars, in the range 1.5 < [Fe/H] < +0.5.
Keywords. stars: abundances, Galaxy: bulge, Galaxy: disk
1. Introduction
Our own Galaxy offers a remarkable opportunity for testing different scenarios of galaxy
formation and assembly. In this sense, the bulge and disk play a crucial role because
these two Galactic subcomponents mostly contribute to the Galaxy total stellar mass
(e.g. Wyse 2009). Furthermore, the bulk of bulge and thick disk stars appear to be old
(Ortolani et al. 1995; Zoccali et al. 2003; Bensby et al. 2007).
On one hand, the bulges old ages and enhancement of -elements as revealed by
the detailed chemical analysis of giant stars in the field (e.g. McWilliam & Rich 1994;
Cunha & Smith 2006; Zoccali et al. 2006; Lecureur et al. 2007; Fulbright et al. 2007;
Melendez et al. 2008; Ryde et al. 2009a,b) and in globular clusters (e.g. Melendez et al.
2003; Alves-Brito et al. 2006; Barbuy et al. 2008) , suggest that the Galactic bulge
must have formed rapidly during intensive star formation (the classical scenario). On
the other hand, the Galactic bulges boxy shape is also morphologically consistent with
a pseudo-bulge scenario, which indicates the bulge must have formed as a consequence
of secular evolution through dynamical instability of an already established inner disk
(see Kormendy & Kennicutt 2004, for a review).
Melendez et al. (2008) analysed high resolution infrared spectra of both bulge and thick
disk giants with similar stellar parameters and found that the bulge is in fact chemically
very similar to the thick disk in [(C,N,O)/Fe]. Here we present the results based on
oxygen and other -elements (Mg, Si, Ti and Ca) as derived from high resolution optical
data.
343
Figure 1. Mean [/Fe] vs. [Fe/H] for the sample. Refer to the text for more details.
metal-rich ([Fe/H] > 0.5) bulge (top), thick disk (middle) and both bulge and thick
disk (bottom) stars is also displayed. As one can see, similar relations can fit both stellar
populations, with a scatter as low as = 0.03 dex. This finding suggests that the bulge
and local thick disk giants have essentially identical chemical abundance ratios. Thus, the
bulge and local thick disk stars experienced similar formation timescales, star formation
rates and initial mass functions. Our results and conclusions are discussed in detail in
Alves-Brito et al. (2009, A&A, submitted).
Interestingly, Bensby et al. (2009) have found that microlensed dwarfs in the Galactic
bulge present [/Fe] abundance ratios similar to those of dwarfs in the Galactic thick
disk, which agrees with our results for giant stars.
Acknowledgments
AAB acknowledges financial support from CAPES (4685-06-7) and FAPESP (04/002879).
References
Alves-Brito, A., Barbuy, B., Zoccali, M., et al. 2006, A&A, 460, 269
Barbuy, B., Alves-Brito, A., Ortolani, S., et al. 2008, PhST, 133a, 4032
Bensby, T., Zenn, A. R., Oey, M. S., Feltzing, S. 2007, ApJ, 663, 13
Bensby, T., Feltzing, S., Johnson, J. A., et al. 2009, in press, arXiv0908.2779
Cunha, K. & Smith, V. V. 2006, ApJ, 651, 491
Fulbright, J. P., McWilliam, A., Rich, R. M. 2007, ApJ, 661, 1152
Kormendy, J. & Kennicutt, Jr., R. C. 2004, ARA&A, 42, 603
Lecureur, A., Zoccali, M., Hill, V., et al. 2007, A&A, 465, 799L
McWilliam, A. & Rich, R. M. 1994, ApJS, 91, 749
Melendez, J., Barbuy, B., Bica, E., et al. 2003, A&A, 411, 417
Melendez, J., Asplund, M., Alves-Brito, A., et al. 2008, A&A, 484, L21
Ortolani, S., Renzini, A., Gilmozzi, R., et al. 1995, Nature, 377, 701
Ryde, N., Edvardsson, B., Gustafsson, B., et al. 2009a, A&A, 496, 701
Ryde, N., Edvardsson, B., Gustafsson, B., et al. 2009b, A&A, in press, arXiv:0910.0448
Wyse, R. 2009, IAUS, 258, 11
Zoccali, M., Renzini, A., Ortolani, S., et al. 2003, A&A, 399, 931
Zoccali, M., Lecureur, A., Barbuy, B., et al. 2006, A&A, 457, 1
Universidade de S
ao Paulo, Brazil
2
Universit`
a di Padova, Italy
3
Pontificia Universidad Catolica de Chile, Chile
4
Observatoire de la C
ote dAzur, France
5
Universidade Federal do Rio Grande do Sul, Brazil
6
Osservatorio Astronomico di Padova, Italy
7
Observatoire de Paris-Meudon, France
Abstract. Very few abundance analyses of individual stars in metal-poor globular clusters in the
galactic bulge are available. The main purpose of this study is to derive abundances in individual
stars of such clusters, in order to establish their abundance pattern, trying to characterize the
oldest bulge stellar populations.
Keywords. globular clusters, galactic bulge, metal-poor stars, abundances
1. Introduction
We have identified a sample of globular clusters in the Galactic bulge that show a
moderate metallicity of [Fe/H]-1.0, combined with an extended blue horizontal branch
(EHB), given in Table 1. Lee et al. (2007) suggested that NGC 6522 is among relics of
the first building blocks that first assembled to form the Galactic nucleus, and that are
now observed as relatively metal-poor EHB globular clusters.
We carried out high resolution spectroscopy of HP 1 using UVES data, and NGC 6558
and NGC 6522 using FLAMES in GIRAFFE mode, all at the VLT UT2.
3. Conclusions
Lee (1992) pointed out that RR Lyrae in the Galactic bulge have a peak metallicity of
[Fe/]-1.0. This population should be older than the halo, because being more metal-rich
these stars should be more massive, and expected to populate the red HB, whereas given
that they populate the RR Lyrae gap, then a lower mass and a lower age are implied.
344
345
Terzan 4
1.8
8.3
1.6
50
Terzan 9
1.95
4.9
2.0
HP 1
1.21
6.4
1.0
46
NGC 6522
0.45
7.4
0.86
25
NGC 6558
0.38
7.7
0.97
197
AL 3
0.36
1.3
Terzan 10
2.4
4.8
1.0
NGC 6540
0.60
3.5
1.0
Abundance ratios in HP 1, NGC 6558 and NGC 6522 show enhancements of the elements O, Mg and Si, whereas Ti and Ca enhancements are shallower. This might be
a signature of nucleosynthesis having occurred at very early times in the Galactic bulge.
References
Barbuy, B., Zoccali, M., Ortolani, S. et al. 2006, A&A, 449, 349
Barbuy, B, Zoccali, M., Ortolani, S. et al. 2007, AJ, 134, 1613
Barbuy, B, Zoccali, M., Ortolani, S. et al. 2009, A&A, in press
Lee, Y.-W., Gim, H.S., DInescu, D.I. 2007, ApJ, 661, L49
Lee, Y.-W. 1992, AJ, 104, 1780
Origlia, L., Rich, M. 2004, AJ, 127, 3422
Zoccali, M., Lecureur, A., Hill, V. et al. 2008 A&A, 486, 177
Figure 1. Abundance pattern of the bulge globular clusters HP 1, NGC 6558, NGC 6522,
UKS 1 and Terzan 4.
The formation and evolution of bulges are an integral and central aspect of galaxy
formation and evolution. Much of what we know about the formation and evolution of
the Milky Way bulge comes from giant stars. However, the underlying assumption that
the giants accurately represent all the stars has not yet been rigorously tested (e.g.,
Santos et al. 2009). A true picture of the star formation history in the Galactic bulge
requires the study of dwarf stars. Under normal circumstances, at the distance of the
Galactic bulge, dwarf stars are too faint to acquire the high-resolution spectra that are
crucial for abundance analysis. However, microlensing events offer, during a short time,
the opportunity to obtain spectra of dwarf stars in the Bulge.
In 2009 we have obtained high-resolution spectra for several microlensed dwarf and
subgiant stars using UVES on the VLT. Combined with previous events (Johnson et al.
2007, 2008; Cohen et al. 2008, 2009; Bensby et al. 2009a, 2009b) the total number of
dwarf and sub-giant star in the Bulge that we have analysed is now 13.
These 13 microlensed dwarf and subgiant stars have an average metallicity of h[Fe/H]i =
0.03 0.4. A two-sided KS-test (see top panel in Fig. 1) does not allow us to reject
the null-hypothesis that the MDF from the microlensed dwarfs and the MDF for the 500
Bulge giants from Zoccali et al. (2008) are identical, even adopting a loose significance
level of 0.1. More microlensed events would help refining the comparison. It is, however,
evident that the super-metal-rich MDF proposed by Cohen et al. (2009), is starting to
shift toward lower metallicities.
Bottom panels of Fig. 1 show the abundance trends for the microlensed Bulge dwarf
346
347
Figure 1. Top panel shows the two-sample KS-test between the Zoccali et al. (2008) giant stars
(full line) and the 13 microlensed dwarf and subgiant stars (dashed line). Bottom panels show
[X/Fe] versus [Fe/H]. Thin and thick disc stars from Bensby et al. (2003, 2005, and 2010 in
prep) are marked by black dots, and the microlensed Bulge dwarfs by red (larger) circles.
and sub-dwarf stars compared to nearby thin and thick disc dwarf stars. Two things can
be taken from this figure: 1) The bulge stars do have a wide spread in metallicity; 2) The
microlensed Bulge dwarfs fall together on the plot with dwarf stars that are known to
belong to the Galactic thick disc, i.e., high alpha-element abundances relative to iron.
Regarding the Bulge membership for these microlensed dwarf stars, theoretical calculations for the distance to microlensed sources, assuming a constant disk density and
an exponential bulge, show that the distance to the sources is strongly peaked in the
Bulge, with the probability of having D < 7 kpc very small (Kane & Sahu 2006). Also,
with regard to the sources being in the disk on the other side of the Bulge, the position
of the stars in the OGLE/MOA color-magnitude diagram strongly argues against that.
Basically, the stars are not faint enough, especially when considering the fact that there
will be additional dust as we enter the disc on the other side.
The data set, the analysis, and results for other -elements will be presented in Bensby
et al. (in prep.), and results for r- and s-process elements in Johnson et al. (in prep.).
References
Bensby, T., Feltzing, S., Johnson, J. A., et al. 2009b, ApJ, 699, L174
Bensby, T., Feltzing, S., & Lundstr
om, I. 2003, A&A, 410, 527
Bensby, T., Feltzing, S., Lundstr
om, I., & Ilyin, I. 2005, A&A, 433, 185
Bensby, T., Johnson, J. A., Cohen, J., et al. 2009a, A&A, 499, 737
Cohen, J. G., Huang, W., Udalski, A., Gould, A., & Johnson, J. A. 2008, ApJ, 682, 1029
Cohen, J. G., Thompson, I. B., Sumi, T., et al. 2009, ApJ
Johnson, J. A., Gal-Yam, A., Leonard, D. C., et al. 2007, ApJ, 655, L33
Johnson, J. A., Gaudi, B. S., Sumi, T., Bond, I. A., & Gould, A. 2008, ApJ, 685, 508
Kane, S. R., & Sahu, K. C., 2006, ApJ, 637, 752
Santos, N. C., Lovis, C., Pace, G., Melendez, J., & Naef, D., 2009, A&A, 493, 309
Zoccali, M., Hill, V., Lecureur, A., et al. 2008, A&A, 486, 177
Abstract. At present none of galactic chemical evolution (GCE) models provides a self-consistent
description of observed trends for all iron-peak elements with metallicity simultaneously. The
question is whether the discrepancy is due to deficiencies of GCE models, such as stellar yields,
or due to erroneous spectroscopically-determined abundances of these elements in metal-poor
stars. The present work aims at a critical reevaluation of the abundance trends for several odd
and even-Z Fe-peak elements, which are important for understanding explosive nucleosynthesis
in supernovae.
Keywords. Line: formation, Line: profiles, Stars: abundances, Nucleosynthesis
As a rule, abundances of Fe-peak elements in the atmospheres of the Sun and metalpoor stars are calculated using a local thermodynamic equilibrium (LTE) assumption
for the element line formation. However, recent studies indicate that LTE breaks down
for majority of metals in stellar atmospheres, in particular, for neutral atoms with low
ionization potential. Our earlier investigations of Mn and Co in cool stars (Bergemann
& Gehren 2007, 2008, Bergemann et al. 2009) confirm large departures from LTE in
the excitation-ionization balance of these elements, which are mainly stipulated by overionization of neutral atoms. As a result, the LTE-based abundances of Mn and Co in
metal-poor stars are severely underestimated. Under NLTE, the trend of [Mn/Fe] with
[Fe/H] is only slightly subsolar, whereas [Co/Fe] ratios steadily increase with decreasing
Fe abundances in the disk and halo stars.
In this work, we have expanded our sample of disk and halo stars from Bergemann et al.
(2009) with 9 members of the thin and thick disks. Contrary to the prevalent belief that
NLTE effects on differential abundances are minor for less metal-poor stars, we find that
NLTE abundance corrections for Co are significant even at [Fe/H] 0.8, log NLTE
log LTE +0.25 dex. Our new data reveal a well-defined increase of [Co/Fe] ratios at
[Fe/H] 0.5 (Fig. 1). An ostensive dichotomy of [Co/Fe] values seen for the stars with
1 < [Fe/H] < 0.5 is not significant. Apparently, there is a large continuous spread of
Co abundances in stars with mildly sub-solar metallicities, which can be explained by
GCE models with radial migration (Sch
onrich & Binney 2009).
We find similar NLTE effects on abundances of Cr in metal-poor stars. The details
of statistical equilibrium calculations will be published elsewhere. We only note that
quantum-mechanical photoionization cross-sections for Cr I are now available Nahar
(2009). This allowed us to put constraints on poorely-known cross-sections for inelastic collisions with H I, which are computed according to Drawin (1969). Under NLTE,
ionization equilibrium in Cr I/Cr II is shifted to lower number densities of Cr I, thus
348
349
1.2
0.8
0.8
0.4
0.4
[Cr/Fe]
[Co/Fe]
requiring larger Cr abundances to fit the observed spectral lines of Cr I. The main stellar parameter that determines the sign and magnitude of NLTE abundance corrections
is metallicity: the effect of overionization on the line opacity monotonously increases
with decreasing [Fe/H]. In addition to the metallicity effect, high temperatures control
overionization at higher [Fe/H], whereas at low [Fe/H] the effect of low gravity is more
important. Hence, NLTE line formation and abundance corrections for giants and dwarfs
at a given [Fe/H] are different for transitions involving different levels of Cr I.
0.0
0.4
0.8
0.4
NLTE
LTE
[Fe/H]
0.0
0.8
NLTE
LTE
[Fe/H]
Figure 1. [Co/Fe] and [Cr/Fe] ratios in metal-poor stars as a function of metallicity. The
abundances of Co and Cr are determined from the lines of neutral atoms.
NLTE abundances of Cr were calculated for a sample of dwarfs and subgiants (Fig. 1)
with atmospheric parameters taken from Bergemann et al. (2009). In short, the effective
temperatures and surface gravities are determined from Balmer line profiles and hipparchos parallaxes, respectively. The iron abundances and microturbulent velocities were
obtained from Fe II line profile fitting under LTE requiring that the derived Fe abundances are independent of the line strength. Ionization equilibrium of Cr in metal-poor
stars is satisfied when we apply a very small scaling factor to cross-sections for inelastic
collisions with H I, SH 6 0.05. With this choice of SH , our study yields [Cr/Fe] 0
throughout the range of metallicities analyzed here, 3 6 [Fe/H] 6 0. The metallicityindependent [Cr/Fe] ratios in metal-poor stars are well reproduced by most of the GCE
models (e.g. Samland 1998, Kobayashi et al. 2006). On the other side, these same models
tend to underestimate the NLTE [Mn/Fe] and [Co/Fe] ratios in metal-poor stars. It is
tempting to relate the discrepancy to stellar yields used in the GCE models. According
to Kobayashi et al. (2006), it is possible to find a combination of SN II explosion parameters, like mass cut and amount of mixing, to reproduce [Cr/Fe], [Mn/Fe], and [Co/Fe]
simultaneously. This problem will be investigated further.
References
Bergemann, M., & Gehren, T. 2007, A&A 473, 291
Bergemann, M., & Gehren, T. 2008, A&A 492, 823
Bergemann, M., Pickering, J. C., & Gehren, T. 2009, MNRAS (accepted), eprint arXiv:0909.2178
Sch
onrich, R., & Binney, J. 2009, MNRAS 396, 203
Nahar, S. 2009, JQSRT 110, 2148
Samland, M. 1998, ApJ 496, 155
Kobayashi, C., Umeda, H., Nomoto, K., Tominaga, N., & Ohkubo, T. 2006, ApJ 653, 1145
2. Abundance Distributions
Figure 1 presents individual generalized histograms for each element studied (except
for Fe), for our sample (solid lines) and that of Shi et al.(2004)(dash dot lines), AllendePrieto et al.(2004)(dot lines) and Takeda(2007)(dash lines). The generalized histograms
350
351
Figure 1: Probability density functions for our sample in comparison with that from
other authors. The pdfs constructed using our data is shown by solid lines and that by
Shi et al.(2004), Allende-Prieto et al.(2004) and Takeda(2007) are shown by dash dot
lines, dash lines and dot lines, respectively.
are constructed using a Gaussian kernel estimator so the errors on the abundances could
be explicitly included in the analysis. It is also shown on Figure 1 the mean value and
standard deviation for each of the pdfs. The sample of Shi et al.(2004) and Takeda(2007)
are composed, in the average, of stars preferentially brighter than those of our sample,
explaining the apparent regular shift in the distribution peak of each element. This could
indicate that their stars are younger and richer than ours. The sample of Allende-Prieto
et al.(2004) is more similar with ours because it is volume-complete.
3. Discussion
The mean values and standard deviations of chemical abundances, showed in Figure
1, indicate that elements having larger average abundances present smaller dispersion.
The results are inconsistent with simple models for the chemical element in the Galaxy
with delayed productions (Pagel, 1989) form whih it can be shown that elements having
productions more delayed than a typical SN II byproduct should present larger average
abundance and smaller dispersions at the present time. The contradiction comes from the
fact that the elements showing this behavior in our data are those mainly produced by
SN II (Ca and Si). This discrepancy indicates that other complex processes, such as infall,
time varying star formation rate or initial mass function, have affected the abundance
distribution of solar neighborhood stars. For more details, see Braganca et al. 2010 (in
preparation).
References
Allende-Prieto, C., Barklem, P.S., Lambert, D.L., Cunha, K., 2004, A&A, 420, 183
Edvardsson, B., Anderson, J., Gustafsson, B., Lambert, D.L., Nissen, P.E., & Tomkin, J. 1993,
A&A, 275, 101
Pagel, B.E.J., 1989, RMxAA, 18, 161
Shi, J.R., Gehren, T., Zhao, G., 2004, A&A, 423 ,683
Takeda, Y., 2007, P.A.S.J., 59, 335
Laborat
orio Nacional de Astrofsica/MCT
Rua Estados Unidos, 154 - 37504-364 - Itajub
a, MG, Brazil
email: bruno@lna.br marilia@lna.br
2
Observat
orio Nacional/MCT
Rua Gal. Jose Cristino, 77 - 20921-400 - Rio de Janeiro, RJ, Brazil
email: daflon@on.br
3
Dr. Remeis-Sterwarte Bamberg, Universit
at Erlangen-N
urnberg
Sternwartstr. 7, D-96049 Bamberg, Germany
email: przybilla@sternwarte.uni-erlangen.de
Abstract. In this work, we present temperature, surface gravity, metallicity, microturbulence
and element abundances, determined from a detailed spectroscopic analysis for a sample of 9
Herbig Ae stars, based on high resolution, high S/N spectra.
Keywords. stars: abundances, stars: premain-sequence
1. Introduction
The study of the chemical composition is an important way to understand the evolution of stars. In the case of pre-main-sequence (PMS) stars these studies are even more
necessary since there is still few data available in the literature. Furthermore, the study
of the chemical composition of intermediate mass PMS stars (Herbig Ae/Be stars) may
contribute to understand the relationship between the frequency of planets and stellar
metallicity. In this work we present the stellar parameters and abundance analysis of a
sample of 9 Herbig Ae stars based on spectra obtained with the FEROS spectrograph
(ESO).
2. Analysis
Effective temperatures could not be determined using photometric calibrations for
these stars, because the circumstellar reddening is unknown. The preliminary determination of the temperature and gravity are obtained through the spectral classification, based
on types determined by Torres (1999) and the calibration by Gray & Corbally (1994).
These values are then refined by fitting theoretical H profiles from Kurucz (1993) to
the observed spectra.
For stars with enough measurable FeI lines in their spectra the excitation equilibrium
was used to refine the temperature. The gravity was estimated by the FeI/FeII ionization
equilibrium. Metallicity was determined through the curves of growth of Fe I and II.
The curves were calculated with the code Renoir (by M. Spite) and using the same
atmospheric models as the H fitting. The Teff values derived from Fe lines agree very well
with the Teff obtained from the fitting of H profiles (< T (H) T (Fe) >= 49 111K).
Abundance analysis was based on the fully-blanketed and plane-parallel LTE model
Based on observations collected at the European Southern Observatory, Chile.
352
353
Teff
[K]
101412 10000
142666 7100
144432 7250
145718 7500
100453 7150
139614 7700
141569 10000
163296 8500
169142 7400
-0.05
0.00
0.10
0.20
0.00
-0.20
0.00
0.00
-0.30
9.04
8.5
8.3
8.3
8.25
8.4
8.5
8.2
8.55
8.66
8.72
8.68
8.59
8.7
8.67
8.71
8.75
8.62
8.84
9.57
9.86
8.85
8.23
8.87
7.81
7.18
7.25
7.32
7.25
6.94
6.5
6.09
6.29
6.3
5.95
6.02
6.11
5.69
4.31
5.09
4.79
4.89
4.67
5.06
4.9
5.02
4.49
2.52
3.08
2.81
atmospheres calculated with the ATLAS9 code (Kurucz 1993) for a constant microturbulent velocity of 2km/s and solar composition. LTE abundances were determined by
fitting synthetic spectra calculated with program LINFOR (originally developed by H.
Holweger, M. Steffen, and W. Steenbock) to the selected spectral regions. The theoretical
profiles were broadened to account for effects of rotation, limb darkening and instrumental
profile. CI lines yield carbon solar abundance (log(C) = 8.410.03, Asplund 2003). The
mean oxygen abundance derived is 0.2 dex higher than in the Sun (log(O) = 8.660.03,
Asplund 2003) and should be analyzed with caution. In an independent determination
log(O) departures from LTE were considered in the line formation calculations with the
newest version of the program DETAIL (Giddings 1981). The adopted oxygen model
atom is described in Przybilla et al. (2000). The synthetic line profiles were calculated
with the SURFACE code (Butler & Giddings 1985), assuming Voigt profile functions.
These profiles were then broadened by means of convolution with the rotational profile,
including vsini, limb darkening, and instrumental profile.
IAG, University of S
ao Paulo, 05508-900, S
ao Paulo-SP, Brazil.
2
email: cavichia@astro.iag.usp.br
Abstract.
New abundances of planetary nebulae located towards the bulge of the Galaxy are derived
based on observations made at LNA (Brazil). We present accurate abundances of the elements
He, N, S, O, Ar, and Ne for 56 PNe located towards the galactic bulge. The data shows a
good agreement with other results in the literature, in the sense that the distribution of the
abundances is similar to those works. From the statistical analysis performed, we can suggest
a bulge-disk interface at 2.2 kpc for the intermediate mass population, marking therefore the
outer border of the bulge and inner border of the disk.
Keywords. Planetary nebulae, chemical abundances, chemical evolution, Milky Way
1. Introduction
Bulge and disk may have formed in different ways such as the disk inside-out formation
model (Chiappini et al. 2001), and the model of multiple infalls onto the bulge (Costa et
al. 2008), so that we would expect that these differences should appear in the abundance
distributions of these structures. Many authors have compared the abundance distributions of the bulge and the disk and find no clear differences (Chiappini et al. 2009,
Escudero et al. 2004, Exter et al. 2004, Cuisinier et al. 2000). Nevertheless, in these works
abundances of the solar neighborhood or the whole disk were used in order to compare
the abundance distributions. Until now, few studies made an effort to investigate whether
or not the radial abundance gradient of the disk extends toward the galactic center, as
for example those from Smartt et al. (2001) or Gutenkunst et al. (2008). In this context,
the present study intends to shed light in this field by comparing the PNe abundance
distributions of the inner disk and the bulge using PNe statistical distance scales.
2. Method
Spectrophotometry observations in the optical domain were made at LNA observatory
(Brazil) for a sample of 56 planetary nebulae located in the direction of the the galactic
bulge. The data were reduced following standard reduction procedures with the IRAF
software (see Escudero et al. 2004 for details).
The Stanghellini et al. (2008) (SSV08) statistical distance scale was used to study the
distribution of chemical abundances across the disk-bulge interface. Additionally, new
distances were derived for 46 objects whose distances were not available in the same
paper. The method consists in establishing a galactocentric distance that divides the
sample into two groups: group I, composed by those PNe with distances lower than the
limit, and group II with objects whose distances are higher than the limit settled. Then
the galactocentric distance that divides the groups is varied from 0.1 to 3.6 kpc, in 0.7
kpc steps. A Kolmogorov-Smirnov test was then applied to each step in order to find the
distance in which the chemical properties of these regions better separates.
354
355
Figure 1. Abundance distributions for each chemical element for groups I and II using the
SSV08 distance scale and a galactocentric distance for the separation set at 2.2 kpc. Unfilled
histograms represent group I objects and filled histograms are for group II. The number of
objects in each distribution is shown at the top.
The comparison between the two populations shows that, on the average, group I
(bulge) objects have slightly lower abundances than those from the group II (inner-disk),
although this difference is not larger than the errors in individual abundances. Taking
into account the results derived in this work as well as other evidences from the literature,
and using the SSV08 distance scale, we propose a galactocentric distance of 2.2 kpc to
mark the transition between the bulge and inner-disk of the Galaxy.
References
Chiappini, C., Matteucci, F., & Romano, D. 2001, ApJ, 554, 1044
Chiappini, C., Gorny, S., Stasi
nska, G., & Barbuy, B. 2009, A&A, 494, 591
Costa, R. D. D., Maciel, W. J., & Escudero, A. V. 2008, Baltic Astronomy, 17, 321
Cuisinier, F., Maciel, W. J., K
oppen, J., Acker, A., & Stenholm, B. 2000, A&A, 353, 543
Escudero, A. V., Costa, R. D. D., & Maciel, W. J. 2004, A&A, 414, 211
Exter, K. M., Barlow, M. J., & Walton, N. A. 2004, MNRAS, 349, 1291
Gutenkunst, S., Bernard-Salas, J., Pottasch, S. R., et al. 2008, ApJ, 680, 1206
Smartt, S. J., Venn, K. A., Dufton, P. L., et al. 2001, A&A, 367, 86
Stanghellini, L., Shaw, R. A., & Villaver, E. 2008, ApJ, 689, 194 (SSV08)
1. Introduction
Post-AGB objects present a dust envelope formed by intense mass loss during the
previous AGB phase. The spectra of PAGB stars show many absorption lines formed
in their atmospheres and also emission lines formed in the envelope. Most of PAGB
present low to intermediate temperatures with spectral types F-G. However, some PAGB
present B-type spectra and are called hot-PAGB stars (Parthasarathy & Pottasch 1986).
On the other hand, some OB supergiants with high galactic latitudes and presenting
dust envelopes and IRAS colors similar to those of PNe have been identified (Conlon
et al., 1992). A possible evolutionary connection between these objects can be tested
by comparing chemical abundances for samples of hot-PAGB and halo OB supergiant
stars. In this work we present preliminary results for four PAGB stars without any
abundance result previously published. These objects are IRAS14331-6435, IRAS170741845, IRAS18023-3409 and IRAS17203-1534.
2. Observations
High resolution spectra (R=48000) of the objects have been obtained with FEROS
spectrograph coupled to the ESO 2.2m telescope at La Silla (Chile). The spectra cover
the wavelenght range 3700 to 9200
A and have S/N 120 at 4500
A.
3. Analysis
The absorption spectra of the sample stars were analyzed in non-LTE using the program SYNSPEC. The synthetic profiles were produced interpolating in the grid of atmospheric models BSTAR2006 (Lanz & Hubeny 2007) that was previously generated with
the code TLUSTY (Hubeny & Lanz 1995).
The atmospheric parameters effective temperature Tef f , surface gravity logg and the
microturbulent velocity were determined simultaneously from the analysis of HeI, OII
and NII line profiles. The solution for a given pair Tef f , logg and for was chosen
356
357
as to have similar abundance values for weak and intermediate lines and to have lower
abundance dispersion.
The abundances are determined from the fit of non-LTE theoretical line profiles to the
observed spectra. Lines of HeI, NII and OII were fitted independently and the average
abundance with the corresponding line-to-line scatter was adopted as the star abundance.
Figure 1 shows the abundances for the sample stars relative to Sun.
Figure 1. Helium, nitrogen and oxygen abundances relative to Sun (Asplund, Grevesse &
Sauval 2006) for the hot Post-AGB stars.
Observat
orio Nacional, Rua General Jose Cristino 77, CEP 20921-400, Rio de Janeiro Brazil
email: daflon@on.br
2
National Optical Astronomy Observatories, 950 N. Cherry Ave., Tucson, AZ 85719
3
New Mexico State University, 1320 Frenger St., Las Cruces, NM 88003
Observatoire de Gen`eve, Universite de Gen`eve, 51 Chemin des Maillettes, CH-1290 Sauverny,
Switzerland
Abstract. Sulfur abundances are derived for a sample of ten B MS star members of the Orion
association. The analysis is based on LTE model atmospheres and non-LTE line formation
theory by means of spectrum synthesis analysis of Sii and Siii lines. The abundance distribution
obtained for the Orion targets is homogeneous within the errors in the analysis: A(S)=7.150.05.
This abundance result is in agreement with the solar value and with results for the Orion nebula.
The sulfur abundances for Orion combined with previous results for other OB-type stars produce
a relatively shallow sulfur abundance gradient with a slope of 0.0370.012 dex Kpc1 .
Keywords. stars: early-type, stars: abundances, (Galaxy:) open clusters and associations: individual (Ori OB1)
1. Introduction
The -element sulfur is of great astrophysical interest: it is among the ten most abundant elements in the universe and its production occurs both during hydrostatic and
explosive oxygen-burning phases in massive star evolution and SNii. In addition, sulfur
is not expected to locked-up significantly into grains. Thus, a comparison between nebular
and stellar sulfur abundances may involve relatively small corrections for grain depletion;
if true, sulfur then would provide a direct connection between the stellar and gas-phase
abundances in the ISM, as well as nebular abundances in galactic and extra-galactic
environments.
This work aims to define the present-day stellar sulfur abundance in a sample of young
B-star members of the Orion association. These abundances can be used to help define
the present-day sulfur abundance in the solar neighborhood, as well as constrain sulfur
depletions onto grains.
2. Observations
The targets are 10 MS early B-type star members of the Ori OB1 association, selected
from the sample of Cunha & Lambert (1994). The data are high resolution (R 35 000)
spectra obtained with the ARCES on the 3.5m telescope at the APO. The spectra cover
the wavelength range between 3480-10260
A and have SNR > 100.
3. Analysis
The stellar parameters (Teff , log g, and ) adopted in this analysis are from Cunha
& Lambert (1994). Sulfur abundances were derived from non-LTE synthetic profiles
358
359
Figure 1. Radial gradient of sulfur abundances: the red triangle represents the average sulfur
abundance for Orion B stars and the black open circles, the stellar abundances for the sample
studied by Daflon & Cunha (2004).
computed for 16 Sii and 3 Siii lines. The non-LTE calculations were done with programs
DETAIL/SURFACE, using LTE model atmospheres (Kurucz 1993) plus a sulfur model
atom from Vrancken et al. (1996). The abundances and V sin i of the synthetic profiles
were varied until they match the observed Sii and Siii lines. For most stars in our sample,
sulfur abundances derived from Sii and Siii lines were found to agree within 0.10 dex. The
total errors in the derived Sii and Siii abundances are 0.11 dex and 0.10 dex, respectively.
The sulfur abundances show no trend with effective temperature, suggesting that these
results are probably free of major systematics within this Teff range.
4. Results
The average sulfur abundance for our sample is A(S)= 7.150.05 and it is found to
be in perfect agreement with the Solar System value A(S)=7.14 0.05 (Asplund et al.
2006), suggesting that little, if any, chemical evolution of sulfur has taken place in the
Solar vicinity in the last 4.5 Gyrs. The abundances of main-sequence B stars overlap
with the nebular abundance derived by Esteban et al. (2004) for Orion. The similarity
between these abundances indicates that sulfur is undepleted in the Orion Nebula.
The sulfur abundance results for Orion are added to a database of abundances previously published for OB main sequence stars along the Galactic Disk (Daflon & Cunha
2004), producing a gradient of sulfur abundance of 0.0370.012 dex Kpc 1 (Fig. 1).
References
Asplund, M., Grevesse, N., & Sauval, A. J. 2006, Nuclear Physics A, 777, 1
Cunha, K. & Lambert, D. 1994, ApJ, 426, 170
Daflon, S. & Cunha, K. 2004, ApJ, 617, 1115
Esteban, C. et al. 2004, MNRAS, 355, 229
Kurucz, R. 1993, CD-ROM No. 13. Cambridge, Mass.: SAO
Vrancken, M., Butler, K & Becker, S. R. 1996, A&AS, 311, 661
Abstract. We report the spectroscopic analysis of six kinematical members of the Zeta Reticuli
Moving Group, one of them for the first time. We confirm the existence of the Group by establishing a common abundance pattern for four kinematical members. High resolution spectra
yielded abundances of Si, Ca, Fe, Ni and Ba, and others. Effective temperatures were derived
from the excitation & ionization equilibria of Fe lines of four stars. For these, and the remaining
two members, temperatures were derived from colors and the fitting of theoretical spectra to
the H line, and ages and masses were estimated from theoretical HR diagrams. We suggest
that the Group is physical being metal-poor and 6 Gyr old.
Keywords. techniques: spectroscopic, stars: abundances, stars: kinematics, Galaxy: evolution.
1. Introduction
Stellar Kinematic Groups (SKGs) are assemblages of stars which share approximately
the same vectors of Galactic space velocity. Presumably they constitute a link between
gravitationally bound systems, such as the open clusters, and the field stars, and they
are supposed to share the same features of these systems, as coeval age and chemical
composition . The rarity of old kinematic groups bears witness to the processes that tear
them apart, probably encounters with massive objects, such spiral arms and molecular
clouds, in time scales of a billion years or less. Therefore, the majority of these groups
must be young but some relatively old groups have already been considered as is the case
of the Zeta Reticuli group (del Peloso et al. 2000).
2. Methodology
We report a detailed spectroscopic analysis of four objects of this group (with FEROS
data - 2001), beyond two new candidates selected kinematically (with OPD/LNA data 2005). The spectroscopic parameters were determined from the model atmosphere analysis (NMARCS) of nearly 100 lines of Fe I and 10 lines of Fe II, through excitation/ionization
equilibria. Effective temperature was determined by three different constraints: excitation & ionization equilibria of Fe lines, photometry and the modeling of the H profile,
whenever possible. The chemical abundance of each element (Ca, Ba, Sc, Y, Ti, V, Ce,
Cr, Mn, Fe, Co, Ni e Si) was determined by a program that uses the atmosphere model
derived for star, in set with the measure of the equivalent widths gotten for each element.
360
361
3. Results
For HD158614, we obtained Teff from the H modelling. This temperature was derived
by fitting the observed wings of H, using the automated procedure described in Lyra &
Porto de Mello (2005), and we found a Teff =5573 50K. The other objects have their H
temperatures derived from previous works. We constructed the theoretical HR diagram
for the group according to Yi et al. (2003) with the new atmospheric parameters found for
the group members, using HIPPARCOS parallaxes (fig. 1). Evolutionarily, we attach to
the group a 6 Gyr , considering the associated errors [(Teffs )30K, ()0.04kms1 ,
(log g)0.13dex, (Fe/H)0.05dex].
Figure 1. H-R diagram of the probable members of the Ret moving group. The solid lines,
numbered 4 to 7, are the Yale isochrones for the metacility -0.20 dex.
1. Introduction
Radial gradients of metallicity are important constraints to the chemical evolution
models. Observations of different classes of young objects show that the elemental abundances decrease from the Galactic Center towards the edge of the Galactic disk. The
radial distribution of the chemical properties of the Galaxy can be studied from B stars
and photoionized nebula. Concerning the stellar studies, Daflon & Cunha (2004) analyzed 69 Galactic OB stars from RG =4.7 - 13 kpc, and found a radial metallicity gradient
with a slope of -0.042 dex/kpc. However some objects of the sample at R G 10 kpc show
abundance values lower than expected for that region of the Galactic Plane. Some results
from open clusters (Twarog et al. 1997) and in cepheids (Caputo et al. 2001; Andrievsky
et al. 2004) also suggest a discontinuity in the abundance distribution of the Galactic disk
in that region. The sample of Daflon & Cunha (2004) is not enough (especially at R G
10 kpc, where only five stars were analysed) to conclude if such discontinuity really exists
or if this minimun in the abundance distribution is an artifact. Concerning the studied
elements, nebular analyses are focalized in O, S, Ar and Ne abundances and rarely in He
and N. In the case of B stars, despite of showing intense He lines in their spectra, most
of the works are focused in metals. In both cases the distribution of He abundance, a key
element for models of Galactic chemical evolution, is not well established. The objective
of this work is to analyze the discontinuity of the abundance gradient at RG 9-11 kpc,
beginning with helium.
2. The Sample
Our sample consists of 13 main sequence B stars members of the clusters NGC 2264
(RG =10.3 kpc), NGC 2362 (RG =9.3 kpc), NGC 2367 (RG =9.8 kpc) and NGC 2384
(RG =9.8 kpc). The spectra were obtained with spectrograph FEROS + ESO 2.2m telescope (under the agreement ESO-ON).
362
363
3. Analysis
The effective temperature (Tef f ) was calculated from a photometric calibration for the
reddening free parameter Q (Daflon et al. 1999). The surface gravity (log g) was derived
from the fits of theoretical wings of H calculated in non-LTE to the observed profiles.
We determined non-LTE Helium abundance (log(He)) for stars with vsini<150 km/s,
using 9 lines of He i. The He abundance were determined from spectral synthesis with
SYNPLOT, interpolating in the grids of non-LTE model atmospheres OSTAR2002 and
BSTAR2006, calculated with the program TLUSTY. We calculated He abundances for
three values of microturbulence (=0, 5 and 10 km/s) and chose the value that produce
an He abundance independent of the line strenght (Fig. 1 (left)). In Fig. 1 (right) are
shown the preliminary results of the He abundances as a function of Tef f for the stars of
the sample. From this figure, it seems to exist a trend of the He abundances with Tef f .
This result still needs further investigation.
Figure 1. Variation of He abundance with for the star CD-24 5180 (left). Log(He) vs. T ef f
for the sample stars (right) with their respective line-to-line scatter. The dotted line represents
the Solar He abundance (10.93 0.01) from Asplund et al. 2006.
Abstract.
We present high-resolution spectroscopic measurements of the abundances of the -like element titanium (Ti) and s-process elements yttrium (Y) and lanthanum (La) for M giant candidates of (a) the Sagittarius (Sgr) dwarf spheroidal + tidal tail system, (b) the TriangulumAndromeda (TriAnd) Star Cloud, and (c) the Galactic Anticenter Stellar Structure (GASS, or
Monoceros Stream). All three systems show abundance patterns unlike the Milky Way but typical of dwarf galaxies. The Sgr system abundance patterns resemble those of the Large Magellanic
Cloud. GASS/Mon chemically resembles Sgr but is distinct from TriAnd, a result that does not
support previous suggestions that TriAnd is a piece of the Monoceros Stream.
Keywords. galaxies: interactions, Galaxy: halo, Galaxy: structure, stars: abundances
1. Results
We obtained echelle spectra of M giant stars spatially and kinematically associated with
the Sagittarius (Sgr), Triangulum-Andromeda (TriAnd) and Galactic Anticenter Stellar
Structure (GASS) systems using the KPNO 4-m, ARC 3.5-m and TNG 3.6-m telescopes.
The majority of Sgr stars show peculiar abundance patterns compared to those of nominal
Milky Way (MW) stars (left panel, Fig. 1) and, as a group, form a coherent picture of
chemical enrichment of the Sgr dwarf spheroidal (dSph) from [Fe/H] = 1.4 to solar
abundance (Chou et al. 2007, 2009). The overall [Ti/Fe], [Y/Fe], [La/Fe] and [La/Y]
trends with [Fe/H] for the Sgr stream + core stars resemble those seen in the Large
Magellanic Cloud (LMC) and other dSphs, only shifted from them by [Fe/H]+0.4
(LMC) and +1 dex (dSphs), respectively (middle panel, Fig. 1); these relative shifts
reflect the faster and/or more efficient chemical evolution of Sgr compared to the other
satellites. By tracking the evolution of abundance patterns along the Sgr stream we can
follow the time variation of the chemical make-up of dSph stars donated to the Galactic
halo. This evolution demonstrates that while the bulk of the stars currently in the Sgr
dSph are quite unlike those of the Galactic halo, an increasing number of stars farther
along the Sgr stream have abundance patterns like metal poor MW halo stars, a trend
that shows clearly how the Galactic halo could have been contributed by present day
satellite galaxies even if the present chemistry of those satellites is now different from
typical halo field stars. We also have analyzed the chemical abundances of a group of
M giants found among the Sgr leading arm stars at the North Galactic Cap (NGC)
but having radial velocities unlike the infalling Sgr leading arm debris there. Through
chemical fingerprinting we conclude that these mostly receding northern hemisphere
M giants are also Sgr stars likely trailing arm debris overlapping the Sgr leading arm.
The origin of the recently identified GASS ring, also known as the Monoceros Stream,
364
365
0.5
0.5
[Ti/Fe]
[Ti/Fe]
[Y/Fe]
0.5
MW trend
TriAnd
0.5
0.5
GASS
[Y/Fe]
Sgr
MW trend
Sgr
[La/Fe]
0.5
dSphs
LMC
0.5
0
[La/Y]
[La/Fe]
0.5
1
MW trend
M05+S07
0.5
Sgr core
Sgr LN
0
Sgr LS
NGC
1.5
0.5
[Fe/H]
1.5
0.5
[Fe/H]
1.5
0.5
[Fe/H]
0.5
Figure 1. Left: [Ti/Fe], [Y/Fe], [La/Fe] and [La/Y] as a function of [Fe/H]. The open circles
are giants from the Sgr core (including data from Monaco et al. 2005 [M05] and Sbordone et al.
2007 [S07], open squares), filled circles are from the Sgr leading arm in the northern hemisphere
(LN), diamonds are from the leading arm in the southern hemisphere (LS), and asterisks are
the NGC group. The solid line is a fit to the Galactic trend for halo and disk field stars (see
discussion in Chou et al. 2009). Middle: Same as left, but comparing Sgr (circles) with the
LMC (crosses) and other dSphs (triangles), after applying a shift of +0.4 dex in [Fe/H] for LMC
and +1 dex in [Fe/H] for other dSphs. All LMC and dSph data are taken from the literature
see Chou et al. (2009). Right: The distribution as a function of [Fe/H] of the abundance ratios
of [Ti/Fe], [Y/Fe] and [La/Fe] for TriAnd stars (triangles), GASS stars (open squares) and Sgr
stars (open circles).
remains controversial e.g., whether it is truly a tidal stream or a part of the MW disk.
We find that the Ti and s-process abundances of GASS stars are similar to those of Sgr
stars (right panel, Fig. 1) consistent with the notion that GASS stars formed in the
chemical environment of a dwarf galaxy that subsequently disrupted. However, differences
in [La/Fe] at high [Fe/H] suggest that the GASS progenitor may have enriched faster
than Sgr, which evolved at a rate slow enough to allow the yields of low metallicity AGB
stars to enrich its younger populations.
Contrary to the proposal of Pe
narrubia et al. (2005) that TriAnd is a distant part
of the Monoceros Stream, we find that, although the chemical patterns of TriAnd are
characteristic of other dwarf galaxies that contain metal-rich M giant stars, the TriAnd
patterns are different enough from those of GASS/Mon specifically (especially [Y/Fe];
right panel, Fig. 1) to suggest that the TriAnd Star Cloud is likely an independent halo
substructure unrelated to the GASS/Monoceros Stream.
References
Chou, M.-Y., et al. 2007, ApJ, 670, 346
Chou, M.-Y., et al. 2009, ApJ, submitted
Monaco, L., et al. 2005, A&A, 441, 141
Pe
narrubia, J., et al. 2005, ApJ, 626, 128
Sbordone, L., et al. 2007, A&A, 465, 815
Observat
orio Nacional MCT, Brazil (email: demello@on.br),
2
Observat
orio do Valongo UFRJ, Brazil
Abstract. We present chemical abundances obtained by fitting synthetic spectra to the FEROS
data of 12 C-J stars, normal and silicate carbon one. The Li and 13 C abundance, as well as the sprocess elements abundance indicates no evidence of a diferent mechanism of formation between
the two kinds of C-J stars. We also studied the elements, and that suggests a scenario that put
the carbon C-R stars as a possible progenitor of the beither C-J stars. We made also available
a study of the radial velocity of some stars of the sample calculated at different epochs. That
allowed us to make some discussion about the probable binary system of some silicon C-J carbon
stars. Based on these abundances and radial velocities, we discuss possible evolutive sequences
for the C-J.
Keywords. techniques: radial velocities, stars: abundances, stars: AGB and post-AGB, stars:
carbon
1. Introduction
There are still many open questions about C-J type carbon stars. Estimates of sprocess elements abundances for these low 12 C/13 C stars suggest that they achieved
their high carbon abundances through a mechanism other than the dredge-up processes
that cause the carbon enrichment in the more common C-N carbon stars. C-J stars are
also notable, among their chemical peculiarities, for abnormally high abundances of 7 Li.
Additionally, the circumstellar envelope of most C-J stars is carbon-rich, in accordance
with their photospheric composition. However, there is a peculiar C-J stars referred as
silicate carbon stars, due to an emission feature at 9,8m, typical of oxygen-rich stars.
Because of these unique traits of the C-J class, little is known about how these objects
evolve.
2. Radial Velocity
The radial velocity measures were applied to both correct the FEROS spectra for
the chemical abundance analysis as well as to investigate the possibility of a binary
nature to the C-J silicate carbon stars. Until now, the most plausible scenery for C-J
silicate carbon stars seems to be a binary system, composed by an unseen, low-luminosity
companion. The oxygen- rich material that was shed by mass loss when the primary star
was an M giant can be stored in a circumbinary disk (Morris (1987), Evans (1990)), or
in a circumstellar disk around the low-luminosity companion (circum-companion disk,
Yamamura et al. (2000)) even after the primary star becomes a carbon star.
V433 Pup seems to have a certain variation in the radial velocity, however, small
in comparison with the errors. MT Hya, MC79 2-11 and C* 1003 have a variation of
3.4kms1 in 7 years, 3kms1 in 3 years and 9,2kms1 in 35 months, respectively,
366
367
which points to a binary system that may enclose an oxigen-rich disk. BM Gem is one
of the most studied star of our sample. All others either do not present any relevant
variation at the radial velocity or do not have more the one measure.
3. Chemical Abundances
From all the element abundances calculated, the most relevant ones were the 7 Li and
C, which are produced in the stellar interior and it only reach the outer part of the
atmosphere where they are detected, if besides the dredge-up process also happens an
extra mechanism of mixture. It is supposed that both abundances vary together (Abia
et al. (1993)).
The 12 C/13 C distribution of of the sample fluctuate around 5.5; and an average value
for the C-J is 5.9; which is a close result to 4.7 2.8 obtained by Ohnaka & Tsuji
(1999). Most of our targets has an uncommon abundance of 7 Li together with low ratios
of 12 C/13 C. Figure 1 presents the correlation of those parameters for our targets with
the ones published by Abia & Isern (2000). It is noticeable that our results do not flee
from the behavior found by previous authors. As more 13 C the star have, higher is its
7
Li abundance. It seems to confirm, the existence of a single extra mechanism of mixture
capable to take both elements from the inner parts to the outer atmosphere.
13
Figure 1. Left panel: Correlation between 7 Li and 13 C abundance values. Right panel: a close
up of the region where our sample is located.
The C-J and their supposed progenitors, C-R, have different areas in the graphic.
However, any separation among normal and silicate C-J do not seems to exist, reinforcing
the idea that both, as well as the C-R, experiment the same process of production. On the
other hand, 7 Li and 13 C can only be detected if an extra process of mixture takes place
(HBB or CBP). Therefore, it is also possible to infer an evolutionary sequence C-R C-J,
if we consider that the process capable to enrich with carbon the photosphere also carries
7
Li and 13 C.
References
Abia et al. 1993, A&A, 275, 96
Abia, C. & Isern, J. 2000, MNRAS, 177, 89
Evans, T. L. 1990, MNRAS, 243, 336-348
Morris, M. 1987, PASP, 99, 1115-1122
Yamamura et al. 2000, A&A, 363, 629-639
Nearly a decade ago, Helmi et al. (1999) reported the detection of a group of low
metallicity stars whose angular momentum components clumped together more than
would be expected for a random distribution of halo stars. Members of this stellar stream
are scattered throughout all parts of the sky, indistinguishable from ordinary halo
stars, yet they are on very similar, eccentric orbits, indicating that the stars remember
where they originated. We have obtained high resolution (R 30,00045,000) and high
S/N spectra of 9 members of this stream using the 2.7m Smith Telescope at McDonald
Observatory and the 6.5m Magellan-Clay Telescope at Las Campanas Observatory. We
have derived abundances for 2030 elements in each star.
The [C/Fe] ratios in these stars are subsolar ( 0.3). The [/Fe] ratios are enhanced
and typical of metal-poor halo stars, as are the Fe-peak species to Fe ratios. In all cases,
there is no star-to-star dispersion within the uncertainties. The neutron-capture elements
show clear signatures of enrichment by the main component of the rapid (r) neutroncapture process for Z > 62 (0.2 < [Eu/Fe] < +0.5), while Ba to Nd show very mild
overabundances relative to the scaled-solar r-process. Sr, Y, Zr, and Mo show somewhat
higher overabundances. This pattern is found in all stars in the sample. The [Pb/Fe] ratio
is not enhanced as would be expected for enrichment from the slow (s) neutron-capture
process at low metallicity; this, together with the subsolar [C/Fe] ratios, suggests that the
excess of BaNd is not due to the s-process. The increasing excess relative to the main rprocess with decreasing Z for the neutron-capture elements, shown in Figure 1, suggests
that the weak component of the r-process (e.g., HD 122563 as an empirical example of
this enrichment pattern; Honda et al. 2007) may contribute to these elements. There is
no star-to-star dispersion for the metal-rich stars in the sample (2.2 < [Fe/H] < 1.5),
but the two most metal-poor stars (2.5 < [Fe/H] < 2.3) have somewhat lower levels
of neutron-capture enrichment.
The overall chemical homogeneitymuch more than the rest of the stellar halo
suggests that these stars do share a common origin. Also, these enrichment patterns are
368
369
not characteristic of any of the known Milky Way dSph or uFd systems. The globular
cluster M15 ([Fe/H] = 2.6; Sobeck et al. 2009) has similar kinematic properties to the
stars in the stream. This cluster has a wide dispersion of enrichment by the main rprocess (+0.2 < [Eu/Fe] < +1.0) at a single metallicity, which is a unique characteristic
among globular clusters. The level of r-process enrichment in M15 is also very similar
to that found in the stream, with a total dispersion of order 1 dex occurring around a
metallicity of [Fe/H] 2.5 in both groups of stars. Figure 1 shows the neutron-capture
abundances in the stream compared with those from three stars in M15; the abundances
are identical for every element in common. The lighter elements in M15 stars are also
very similar to the stars in the stream. The combination of kinematic and abundance
similarities is evidence that these groups of stars may in fact share a common origin.
References
Helmi, A., White, S. D. M., de Zeeuw, P. T., & Zhao, H. 1999, Nature, 204, 53
Honda, S., Aoki, W., Ishimaru, Y, & Wanajo, S. 2007, ApJ, 666, 1189
Sneden, C., Lawler, J. E., Cowan, J. J., Ivans, I. I., & Den Hartog, E. A. 2009, ApJS, 182, 80
Sobeck, J. S. et al. 2009, in prep.
Figure 1. Comparison of the mean abundances in the stream stars (squares) and the globular
cluster M15 (circles). CS 22892052, which is strongly enriched by the main r-process (Sneden
et al. 2009), is illustated by the line. All abundances are normalized at Eu.
371
Figure 1. No correlation was found between N (Li) and [F e/H], [F e/H] and Tef f , N (Li)
and log(g) showing that few or no systematic trend was induced by the calculations.
2.9]dex. Our results confirm that the stars identified by Torres, C. A. O. et al. (2006a)
and da Silva, L. et al. (2009), as members of the association, are indeed young stars,
having parameters and abundances compatible with the membership in AB Doradus
and therefore the association stars can be studied like a whole. The results also seem
to confirm some of the hypothesis raised by Travaglio, C.; et al. (2001) concerning the
behavior of lithium abundance in the pre-main sequence and main sequence evolution.
References
Alonso, A.; Arribas, S. & Martinez-Roger, C. 1996, A&A 313, 873
da Silva, L; Torres, C. A. O.; de la Reza, R.; Quast, G. R; Melo, C. H; & Sterzik, 2009, A&A
accepted
Gray, R. O. & Corbally, C. J. 1994, AJ 107, 742
Houdashelt, M. L.; Bell, R. A. & Sweigart, A. V. 2000, AJ 119, 1448
Padgett, D. L. 1996, ApJ 451, 1053
Torres, C. A. O.; Quast, G. R.; da Silva, L.; de La Reza, R.; Melo, C. H. F.; & Sterzik, M. 2006,
A&A 460, 695
Torres, C. A. O.; Quast, G.R; Melo, C.H; & Sterzik, 2006, Handbook of Star Forming Regions,
Volume II: The Southern Sky ASP Monograph Publications v5, p757 Edited by Bo Reipurth
Travaglio, C.; Randich, S.; Galli, D.; Lattanzio, J.; Elliott, L. M.; Forestini, M.; & Ferrini, F.
2001 2001, ApJ 559, 909
Zuckerman, B.; Song, I.; & Bessell, M. S. The AB Doradus Moving Group 2004, ApJ 613
Abstract. We present the results of a light element abundance analysis of three solar-type
main sequence (MS) dwarfs and three red giant branch (RGB) clump stars in the Hyades open
cluster using high-resolution and high signal-to-noise spectroscopy. The CNO abundances of
each group (MS or RGB) are in excellent star-to-star agreement and confirm that the giants
have undergone first dredge-up mixing. The observed abundances are compared to predictions
of a standard stellar model based on the Clemson-American University of Beirut (CAUB) stellar
evolution code. The model reproduces the observed evolution of the N and O abundances, as
well as the previously derived 12 C/13 C ratio, but it fails to predict the observed level of 12 C
depletion in the giants. More tellingly, the sum of the observed giant CNO abundances does not
equal that of the dwarfs.
Keywords. nuclear reactions, nucleosynthesis, abundances open clusters and associations:
individual (Hyades), stars: abundances, stars: atmospheres, stars: evolution, stars: interiors
1. Introduction
The CN cycle is the dominant energy source powering the cores of stars more massive
than the Sun. 12 C and 14 N act as catalysts in the conversion of four 1 H nuclei into
a single 4 He nucleus. In the presence of 16 O, the ON cycle can also contribute to the
energy production, and the two cycles working together make up the well-known CNO
bi-cycle. There is no net loss of CNO nuclei in the CNO bi-cycle; however, the relative
number of each nuclei changes due to different lifetimes to proton capture. The slowest
reaction in the cycle, 14 N(p, )15 O, creates a bottleneck and results in, relative to the
initial abundances, a depletion of 12 C, an enhancement of 14 N, and if the ON cycle is
active, a depletion of 16 O.
At the end of core H burning, stars experience the first dredge-up (Iben 1964), the
expansion of the convective envelope from the surface layers down into the interior of
the star. The first dredge-up mixes material processed by the CN and possibly the ON
cycles up to the surface layers where the products of these core reactions can be observed.
Stellar evolution models quantitatively predict changes in the surface abundances of red
giant branch (RGB) stars that have experienced the first dredge-up. Here, we present the
results of a CNO abundance analysis of three main sequence (MS) and three RGB clump
giants in the Hyades open cluster. Using the dwarf abundances as a proxy for the initial
composition of the giants, we compare the observed changes in the surface abundance of
the giants to a standard stellar evolution model (Schuler et al. 2009).
372
373
2. Analysis
We have used the Clemson-American University of Beirut (CAUB) stellar evolution
code (The et al. 2007) to model the evolution of a 2.5 M star, the approximate mass
of the Hyades giants. The model is characterized by a metallicity Z = 0.025 or [Fe/H]
= +0.10, the approximate metallicity of the Hyades cluster. The model was run through
the core He burning phase, ending at about 785 Myr. The Hyades open cluster has an
age estimated to be 600 Myr (Perryman et al. 1998), and the giants currently reside on
the cluster RGB clump.
High-resolution, high signal-to-noise (S/N) spectra of the Hyades stars were obtained
with the Harlan J. Smith 2.7-m telescope and 2dcoude cross-dispersed echelle spectrometer at The McDonald Observatory. The spectra have a nominal resolution of R =
/ = 60, 000 and s/n ratios of 150-200 for the dwarfs and 400-600 for the giants.
The CNO abundances were derived assuming local thermodynamic equilibrium (LTE)
using the LTE spectral line analysis and spectrum synthesis software package MOOG
(Sneden 1973) along with model atmospheres interpolated from the grids of R.L. Kurucz.
Stellar parameters are taken from Schuler et al. (2006). Carbon abundances were derived
from C2 Swan lines at 5086.3 and 5135.6
A, the [C I] forbidden line at 8727
A, and for one of
the giants, 10 CH lines near 4325
A. For N, three CN lines at 6706.7, 6707.5, and 6707.8
A were
used, adopting the C abundances derived from the lines mentioned above. Oxygen abundances
have been derived previously from the [O I] forbidden line at 6300
A by Schuler et al. (2006) using
the same McDonald spectra discussed above. For each group (dwarfs or giants), the star-to-star
abundances are in excellent agreement.
Observat
orio Nacional, Brazil,1
email: licio@on.br
2
Laborat
orio Nacional de Astrofsica, Brazil
3
European Southern Observatory, Germany
4
European Southern Observatory, Chile
Keywords. stars: premain-sequence, stars: evolution, stars: abundances
1. Introduction
In a recent paper, da Silva et al (2009), we report results of Li abundance analysis for nine
young stellar associations, defined in Torres et al (2008), from a high-resolution optical spectroscopic survey searching for associations containing young stars (SACY), among optical counterparts of ROSAT All-Sky X-ray sources in the Southern Hemisphere. They have applied a
convergence method in the (UVW) velocity space and have determined nine nearby young associations in the sample. As they are young and with different ages, those associations form an
interesting laboratory to test the Li depletion theory, as a function of the star age.
2. Analysis
For this research, most of the spectroscopic observations were performed with the highresolution (RP 50000) FEROS spectrograph Kaufer et al (1999) at the 1.5m ESO telescope at
La Silla (Chile). A smaller set of data ( 30%) was collected at the coude spectrograph attached
to the 1.60m telescope at the Observat
orio do Pico dos Dias, LNA, Brazil, with a RP 9000.
To obtain the Li abundance we used a LTE code and the atmospheric models of Kurucz and
Castelli (www.user.oat.ts.astro.it/castelli). A larger part of the stars effective temperatures Teff
were determined from the (V-Ic ) color index. The other atmosphere model parameters, not determined by us, were fixed a priori: the metalicity as 0.1 and log g as 4.5, for dwarfs, and 4.0, for
sub-giants. The used spectral classification is that of Torres et al (2008). The microturbulence
velocity also was fixed as 2 km/s. We presented in da Silva et al (2009) our determination of
Li abundance for nine associations, discovered or better determined in the SACY analysis, the
associations of Chamaleontis ( ChA)(6, 6), TW Hydrae (TWA)(8, 9), Pictoris ( PA)(10,
1), Octans (OctA)(20?,8), Tucana-Horologium (THA)(45, 7), Columba (ColA)(30, 4), Carina
(CarA)(30, 5), Argus (ArgA)(40, 3), AB Doradus (ABDA)(70,2). The numbers between parenthesis are the association age in My and its number on Fig. 1, respectively. The main data
of those associations (i.e., number of members, mean distance and age) were determinated by
Torres et al (2008). The results for those associations presented in da Silva et al (2009) are final
but for ABDA and ColA. These two associations have few new members discovered from recent
observations at La Silla (ESO). Their final results will be presented very soon in a next paper.
However, the introduction of these new members do not change significantly their mean curves
(see below).
374
Li abundances in SACY
375
3. Implications
Fig. 1 shows the mean curves of the diagrams (Teff vs. Li abundance) of those associations. In
da Silva et al (2009) is shown that the errors of the Li abundance determination can not explain
the dispersion of the points for a given Teff. Then, this dispersion is real. Stars with vsin(i)
smaller than 20 km/s are, in average, under those of vsin(i) larger, shown a larger Li depletion.
The diagrams of the stellar Li abundances vs Teff corresponding to the program associations
above show, for Teff < 5000 K, a clear separation among the stars associations. As we can see on
Fig. 1, the curve of ChA, what has 6 My, is upper the other curves; that of ColA, with 30
My, is in the middle and the curve corresponding to AB Dor, with 70 My, is below the other
ones. In da Silva et al (2009) we showed that the AB Dor stars are in the band corresponding to
the Pleiades stars ( 100 My), given one more evidence that those associations have the same
age, what is in agreement with a recent 3D Galactic dynamical analysis indicating a common
origin of these two groups (Ortega et al (2007)). We showed also the importance of the rotational
velocity in the Li depletion: stars with larger rotational velocity have a Li depletion smaller than
stars with smaller rotational velocity.
References
da Silva, L., Torres,C. A. O., Quast, G. R., de la Reza, R., Melo, C. H. F. and Sterzik, M. 2009,
A&A, in press
Kaufer, A., Stahl, O., Tubbesing, S., Norregaard, P., Avila, G., Francois, P., Pasquini, L.,
Pizzella, A., 1999, The Messenger 95, 8
Ortega, V. G.; Jilinski, E.; de la Reza, R.; Bazzanella, B. Monthly Notices of the Royal Astronomical Society, 2007, Volume 377, 441
Torres,C. A. O., Quast, G. R., da Silva, L., de la Reza, R., Melo, C. H. F. and Sterzik, M. 2006,
A&A, 460, 695
Torres, C. A. O., Quast, G. R.,, Melo, C. H. F. and Sterzik, M. 2008, Handbook of Star forming
Regions, Volume II: The Southern Sky ASP Monograph Publications, Vol. 5. Edited by Bo
Reipurth, p 757
Abstract. From high resolution spectra taken with the spectrograph FIES on the Nordic Optical
Telescope, La Palma, we measure abundances of oxygen and -elements in order to characterize
stars which from their dynamical properties are suspected to have originated in disrupted satellites. We find that the chemical composition of investigated stars is homogeneous and distinct
from Galactic disk dwarfs, which is providing further evidence of their extragalactic origin.
Keywords. Stars: abundances, Galaxy: formation
1. Introduction
The formation and evolution of the Milky Way is quite complex and still not fully
understood. Helmi et al. (2006) have used a homogeneous data set of about 14.500
F- and G-type stars from the Nordstr
om et al. (2004) catalogue which has complete
kinematic, metallicity and age parameters to search for signatures of past accretion in
the Milky Way. From correlations between orbital parameters: apocentre (A), pericentre
(P) and z-angular momentum (Lz ) so called APL space, Helmi et al. identified three
new coherent groups of stars and suggested that those might correspond to remains of
disrupted satellites. Stars in those groups cluster around regions of roughly constant
eccentricity, they have distinct metallicity [Fe/H] and age distribution.
Our aim is to measure the detailed elemental abundances of stars belonging to those
groups and make a comparison with Galactic disk stars. In this contribution the preliminary chemical composition results are presented for the third stellar group which is of a
typical metallicity arround 0.8 dex and a single isochrone age of 14 Gyr.
377
Figure 1. Chemical element abundance ratios in the investigated stars suspected to belong to
a disrupted satellite (black dots). The data for the Milky Way disk dwarfs by Edvardsson et al.
(1993, plus signs). The Galactic chemical evolution models by Pagel & Tautvaisiene (1995) are
shown by solid lines.
. The atomic
The oxygen abundances were determined from the [O i] line at 6300 A
data of Ni i were taken from Johansson et al. (2003). The magnesium abundances were
determined from the Mg i lines 6318.7 and 6319.2
A and the non-LTE corrections up to
0.03 dex were calculated and applied using the programme package described by Gratton
et al. (1999). More details about the method of analysis can be found in our previous
papers (e.g. Tautvaisiene et al. 2005).
379
Te = Te [Ag (ag , bg )]
log Lpuls = f1 [(P, Te , M ), (Y, Z), ] = f1 [(P, Te , M ), (Y, aZ , bZ ), ]
log Lev = f2 [(M, t9 ), (Y, Z), (lMLT , , EOS)]
where (Y, Z) represents the abundances in He and heavy metals ([F e/H] [Z/X]),
t9 age of the star; (lMLT , , EOS) - the ingredients of the stellar evolutive calculations;
(M, L, Te ) - the physical parameters of the star, and where
The first relation represents the asteroseismic relation between the observed and theoretical period values (radial pulsating modes), where the theoretical period was inferred
from a linear non-adiabatic model (Suran 1985, 1986, 1990).
The second relation is a temperature-amplitude relation in a g pass-band: log Te =
ag Ag + bg
The third relation must be a variant of the general (P LM Te Y Z) relation:
log L = a0 log M + [a + b log P + c log Te ] + dY + e[F e/H] + f
The fourth relation represents the accommodation of our results with the points in
the evolutive HR diagram. The input parameters are: {P, Ag } and the output parameters
are: {(M, L, Te ), [Z/X]}.
References
Cox, A. 1980, ARA&A, 18, 15
Suran, M. D. 1985, Top. Astrophys. Astron. Space Sci., 1, 1
Suran, M. D. 1986, Top. Astrophys. Astron. Space Sci., 2, 28
Suran, M. D. 1990, Contrib. Astron. Obs. Skalnat Pleso, 20, 77
1,2
1
V. Hill 3 S. Andrievsky
and S. Korotin 4
R.
1. Introduction
In the framework of he ESO Large Program First Stars, high resolution (R=45000) high
S/N ratio spectra have been obtained for a sample of Extremely Metal Poor Stars (EMP stars),
35 giants and 18 turnoff stars. Among them 37 have a very low metallicity: [Fe/H] < 2.9.
The trends in abundance ratios have been presented in Cayrel et al. (2004) and Bonifacio et
al. (2009) from LTE computations of the abundances of the different elements.
However the abundance of light metals such as Na, Al and K had been determined from
resonance lines, particularly sensitive to NLTE effects. Moreover Bonifacio et al. (2009) found
for some other elements such as magnesium, chromium etc. a significant disagreement between
the mean value of [X/Fe] in the giants and the turnoff stars, and it has been even shown that the
abundances deduced from Cr I and Cr II were different. These anomalies suggest the importance
of NLTE effects in EMP stars.
380
381
Figure 1. Variation of [O/Mg], [Na/Mg], [Al/Mg] and [K/Mg] in the Galaxy. The open circles
represent our sample of giants,the black filled circles our sample of turnoff stars. The thin dots
are the measurements of Gehren et al. (2004, 2006) Mashonkina et al. (2008) and Zhang et al.
(2006). For Na only the unmixed stars have been plotted.
It can be seen (Fig. 1) that:
O/Mg at low metallicity is close to the solar value ([O/Mg] 0.0)
the ratios [Na/Mg], [Al/Mg], and [K/Mg] have a negative slope in the interval 3.5 <
[Mg/H] < 2.5. The error on [O/Mg] at low metallicity is too large to firmly establish any
trend in this region.
Between [Mg/H] = 2.5 and 1.5, [Na/Mg], [Al/Mg], and [K/Mg] are stable and then increase toward the solar value.
References
Andrievsky S.M., Spite M., Korotin S.A., Spite F., Bonifacio P., Cayrel R., et al. (2007), A&A,
464, 1081
Andrievsky S.M., Spite M., Korotin S.A., Spite F., Bonifacio P., Cayrel R., et al. . (2008), A&A,
481, 481
Bonifacio P., Molaro P., Sivarani T., Cayrel R., Spite M., Spite F. et al. (2007), A&A, 462, 851
Bonifacio P., Spite M., Cayrel R., Hill V., Spite F., Francois P., Plez B., Ludwig H., Caffau E.,
Molaro P., et al. (2009), A&A, in press.
Cayrel R., Depagne E., Spite M., Hill V.,Spite F., Francois P., Plez B., et al. (2004), A&A, 416,
117
Gehren T., Liang Y.C., Shi J.R. et al. (2004), A&A, 413, 1045
Gehren T., Shi J.R., Zhang H.W. et al. (2006), A&A, 451, 1065
Korotin S.A., Andrievsky S.M., Luck R.E. (1999), A&A, 351, 168
Mashonkina L., Zhao G., Gehren T. et al. (2008), A&A, 478, 529
Zhang H.W., H.W.,Gehren T., Butler K. et al. (2006), A&A, 457, 645
Universidade de S
ao Paulo, IAG, Rua do Mat
ao 1226, S
ao Paulo 05508-900, Brazil
Observatoire de Gen`eve; Chemin des Maillettes 51, CH-1290 Sauverny, Switzerland
3
Dept. of Astronomy & Space Physics, Uppsala University, SE-75120 Uppsala, Sweden
4
Universidade do Vale do Paraba, S
ao Jose dos Campos 12244-000, Brazil
2
1. Introduction
Grenon (1989, 1999) selected a sample of about 6000 dwarf stars from the NLTT
catalogue (see Grenon 1999), and studied them by means of Geneva photometry, radial
velocity and Hipparcos astrometry. A sub-sample of SMR stars was revealed. In this work
we report preliminary oxygen abundances for 21 of these SMR stars. Our aim is to shed
light on the origin of this population and to place it in the global scenario of the Galaxy.
382
383
described in Cayrel et al. (1991) and Barbuy et al. (2003). Photospheric 1D models for
the sample were extracted from the MARCS grid (Gustafsson et al. 2008).
3. Conclusion
Stellar parameters of 21 SMR stars were derived from high resolution spectra obtained
with FEROS. Oxygen abundance were obtained for 16 of these stars. Values in the range
+0.10 < [O/Fe] < +0.35 were found, except for HD 77338, which has a solar O/Fe ratio.
In Figure 1 we compare our results with those for the galactic bulge and thick disk.
The oxygen overabundance (<[O/Fe]> 0.2) suggests that these stars are bulge-like,
probably formed in the inner disk at early times.
References
Barbuy, B., Perrin, M.-N., Katz, D., Coelho, P., Cayrel, R., Spite, M., & vant Veer-Menneret,
C. 2003, A&A, 404, 661
Bensby, T., Feltzing, S., Lundstr
om, I. 2004, A&A, 415, 155.
Cayrel, R., Perrin, M.-N., Barbuy, B., & Buser, R. 1991, A&A, 247, 108
Grenon, M. 1989, Ap&SS, 156, 29
Grenon, M. 1999, Ap&SS, 265, 331
Grevesse, N. & Sauval, J.N. 1998, SSRev, 85, 161
Gustafsson, B., Edvardsson, B., Eriksson, K., Jrgensen, U. G., Nordlund,
A., Plez, B. 2008,
A&A, 486, 951
Fulbright, J. P., McWilliam, A., Rich, R. Michael. 2007, ApJ, 661, 1152
Kaufer, A., Stahl, O., Tubbesing, S., Norregaard, P., Avila, G., Francois, P., Pasquini, L., &
Pizzella, A. 2000, Proc. SPIE, 4008, 459
Lecureur, A., Hill, V., Zoccali, M., Barbuy, B., G
omez, A., Minniti, D., Ortolani, S., Renzini,
A. 2007, A&A, 465, 799
Mel`endez, J., Asplund, M., Alves-Brito, A., Cunha, K., Barbuy, B., Bessell, M. S., Chiappini,
C., Freeman, K. C., Ramrez, I., Smith, V. V., Yong, D. 2008, A&A, 484, 21
Pompeia et al. 2009, in preparation.
Stetson, P. B. & Pancino, E. 2008, PASP, 120, 1332S
0.2
0.2
0.0
[O/Fe]
0.4
0.6
SMR
Bulge (Lecureur et al. 2007)
Bulge (Fulbright et al. 2007)
Bulge (Melndez et al. 2008)
Thick disk (Bensby et al. 2004)
Thick disk (Melndez et al. 2008)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
[Fe/H]
Figure 1. Derived [O/Fe] values compared with data for thick disk (Bensby et al. 2004) and
bulge (Lecureur et al. 2007, Fulbright et al. 2007, Mel`endez et al. 2008).
385
regions, with a corresponding increase in the outer region. These features must come
from the phenomenon that early wind enrichment sets up the initially steep gradient.
2. Hydrodynamical simulations
1
0.8
0.6
0.4
0.2
0
0
10
15
We numerically investigate the dynamical evolution of gas ejected from bulges in disk
galaxies using our own GRAPE-SPH codes. The left panel of figure shows how gas ejected
from a bulge evolves with time during hydrodynamical interaction between the gas and
the halo gas in a disk galaxy. Although a significant fraction of the gas once escapes
from the bulge (T = 0.11 Gyr), most of the gas can be finally returned to the disk plane
owing to gaseous pressure from the halo and the disk (T = 0.45 Gyr). The bulge gas
particles initially in the central region of the bulge can interact so strongly with the
disk gas particles immediately after the ejection that a thin disk can be formed from the
bulge gas at R < 3 kpc for a short time scale (T = 0.06 Gyr). The bulge gas particles
ejected from the bulge regions with larger |z| can be later accreted onto the outer part of
the disk and consequently can rotate around the center of the galaxy. The mean orbital
eccentricity of bulge gas particles in the solar neighborhood is 0.08 in this model, which
means that the orbits become almost circular, because the particles have acquired orbital
angular momentum during hydrodynamical interaction with the disk gas.
The right panel shows that almost 80% of the bulge gas can be accreted onto the inner
disk region with R 6 3 kpc in the model with hg = 105 cm3 where hg denotes the
gas density of halo. This preferential accretion can be clearly seen also in the model with
hg = 104 cm3 .
References
Bensby, T., Feltzing, S., Lundstr
om, I., & Ilyiin, I. 2005, A&A, 433, 185
Chen, L., Hou, J.L., & Wang, J.J. 2003, AJ, 125, 1397
Chiappini, C., Matteucci, F., & Romano, D. 2001, ApJ, 554, 1044
Feltzing, S., & Gustafsson, B. 1998, A&AS, 129, 237
Hou, J. L., Prantzos, N., & Boissier, S. 2000, A&A, 362, 921
Maciel, W.J., Lago, L.G., & Costa, R.D.D. 2006, A&A, 453, 587
Roskar, R., Debattista, V.P., Quinn, T.R., Stinson, G.S., & Wadsley, J. 2008, ApJ, 684, L79
Tsujimoto, T. 2007, ApJ, 665, L115
Superbubble H ii regions:
how self-enriched should they be?
Aida Wofford1
1
1. Introduction
A superbubble is the shell of gas produced when the winds and supernovae (SNe) of
massive stars in a cluster sweep up the ISM. Some superbubbles contain H ii regions. I
took advantage of the relatively simple geometry of superbubble H ii regions in order to
estimate how self-enriched they should be. Here, self-enrichment is the pollution of H ii
regions with ejecta from embedded massive stars, on spatial scales of a few 10 100 pc,
and on time scales of a few Myr.
If self-enrichment is significant, then 1) it could explain the lack of galaxies more metal
poor than I Zw 18; 2) it could explain why some H ii galaxies with Wolf-Rayet (WR)
star signatures show N and He enhancements; 3) if star formation is triggered by the
evolution of a previous generation of stars, then it could occur in a chemically enriched
environment; and 4) H ii regions do not reflect the composition of the gas out of which
their ionizing stars form, nor the composition of galaxies at large. Most giant H ii regions
do not show clear signs of self-enrichment but some do (e.g., Kobulnicky & Skillman
1996; L
opez-S
anchez et al. 2007).
2. Model
The central star cluster is point-like. The H ii region is a spherical shell of constant
density and inner radius r1 (L t3 /n0 )1/5 , as given by the standard superbubble
model of Mac Low, M.-M. & McCray, R. (1988), where L is the mechanical luminosity
of the stellar winds and supernovae, t is the age, and n0 is the density of the ambient
undisturbed ISM. At 5 Myr, I uniformly polluted the volume of radius r2 , where r2 is
the distance from the cluster to the outer edge of the H ii region, with the cumulative
mass ejected up to age t = 5 Myr, by single stellar populations of masses 104 , 105 , or
106 M . Note that chemical yields from supernovae were not included in the calculation
because they are highly uncertain for the most massive stars. The stellar mass loss rate
peaks at 4 Myr, during the stellar WR phase. At 5 Myr, the enriched WR star ejecta
has had enough time to reach the H ii region. I studied self-enrichment up to 5 Myr
386
387
because at t > 5 Myr three problems arise: 1) the ionizing luminosity from the stars is
insufficient to produce a fully ionized H ii region, 2) the radius of the H ii region becomes
comparable with the thickness of the Galactic disk, and the equilibrium time scale of
the H ii becomes greater than 1 Myr. The stellar radiative, mechanical, and chemical
feedbacks used as input to the models are from STARBURST99 (Leitherer et al. 1999;
V
azquez & Leitherer 2005) and are based on the 1994 Geneva non-rotating high-mass
loss stellar evolutionary tracks.
3. Results
Columns 1 5 of Table 1 give: cluster mass, adopted ISM density, H ii region inner
radius, and predicted H ii region C and O enhancements due to self-enrichment (with
respect to the initial ISM values 12+log(C/H) = 7.22 and 12+log(O/H) = 7.53), respectively. The maximum enhancements occur for M? = 106 M and n0 = 100 cm3 . They
would be unobservable considering the typical uncertainties in the chemical abundance
determinations of giant H ii regions. The predicted He and N enhancements are even
smaller and are not shown.
Table 1. Predictions of the extent of self-enrichment in C and O at 5 Myr.
M?
n0
M cm3
r1
pc
104
104
105
105
106
106
153
61
243
97
385
153
1
100
1
100
1
100
log(C/H) log(O/H)
0.014
0.017
0.016
0.017
0.017
0.017
0.017
0.021
0.019
0.021
0.020
0.021
4. Conclusion
At 5 Myr, my self-enrichment scenario produces insignificant (6 0.02 dex) enhancements in He, C, N, and O, in superbubble H ii regions with an initial metallicity of
Z = 0.001. The few enhancements detected so far are associated with younger H ii regions
at higher subsolar metallicities. In Wofford (2009), a detailed study of self-enrichment
in oxygen based on the model presented here can be found.
Acknowledgments
This work was supported by NSF grant AST 03-07118 to the University of Oklahoma
and partially supported by NASA grant N1317 to the Space Telescope Science Institute.
References
Kobulnicky, H. A. & Skillman, E. D. 1996, ApJ, 471, 211
Leitherer, C. & Schaerer, D. & Goldader, J. D. & Gonz
alez Delgado, R. M. & Robert, C. &
Kune, D. F. & de Mello, D. F. & Devost, D. & Heckman, T. M. 1999, ApJS, 123, 3
R. & Esteban, C. & Garca-Rojas, J. and Peimbert, M. & Rodrguez, M.
L
opez-S
anchez, A.
2007, ApJ, 656, 168
Mac Low, M.-M. and McCray, R. 1988, ApJ, 324, 776
V
azquez, G. A. & Leitherer, C. 2005, ApJ, 621, 695
Wofford, A. , MNRAS, 395, 1043
Session V
Extrasolar Planets:
The Chemical Abundance Connection
1. Introduction
The discovery to date of over 350 extrasolar planets has revolutionized not only the
field of exoplanet hunting, but also theoretical efforts to explain the mechanisms of planet
formation and orbital evolution. Exoplanet detections have revealed an incredible variety
of planet characteristics, including a wide range of masses, semi-major axes, and orbital
eccentricities (e.g., Butler et al. 2006). One of the first correlations to emerge from a
handful of exoplanets was the high metallicities of their host stars (Gonzalez 1997).
Figure 1 shows that the evidence for this metallicity correlation has only continued to
increase (e.g., Fischer & Valenti 2005; see chapter by Valenti in this volume). Given the
fact that planets are generally more metal-rich than the most metal-rich star, explaining
this correlation promises to place strong constraints upon planet formation and evolution
processes. The goal of this chapter is thus to briefly summarize current theoretical work
on planet formation models, with an emphasis on the theoretical implications for forming
planets around host stars with varied metallicities, which are presumed to be identical
to the metallicities of the protoplanetary disks in which their planets form.
391
392
Alan P. Boss
2. Core Accretion
All planet formation theorists agree that terrestrial planets form by the collisional accumulation of solids composed of heavy elements in the inner regions of protoplanetary
disks (e.g., Wetherill 1996). This process begins with the sedimentation of dust grains
to the disk midplane and their growth by coagulation, continuing through increasingly
energetic collisions between progressively larger bodies: m-sized boulders, km-sized planetesimals, 1000-km-sized planetary embryos, and ending with 10,000-km-sized terrestrial
planets.
Given this agreement, it is only natural that the generally accepted mechanism for
giant planet formation is core accretion, where roughly 10-Earth-mass, solid cores form
by collisional accumulation in the outer disk, growing larger than in the inner disk because
of the greater amount of solids available in the outer disk. Disk gas is accreted slowly at
first by the cores, but rapidly once the gaseous atmospheres become dynamically unstable
and collapse, leading to gas giant planet formation on a time scale of 1 Myr. In the
core accretion mechanism, ice giant planets are rock/ice cores that failed to accrete much
disk gas, presumably because most of the disk gas had been depleted by the time that
the cores grew large enough to accrete significant gas.
Wetherill (1996) showed that the masses of the terrestrial planets depend directly on
the total mass of the solids available for planet formation in the inner disk. I.e., when the
surface densities of solids is tripled, the maximum planet masses rise by about a factor of
three. One would expect the same dependence of core mass in the outer disk on surface
density of solids as in the inner disk, so that metal-rich disks should be increasingly able
to form core masses large enough to undergo dynamical gas accretion and complete their
evolution into gas giant planets.
The correlation of Doppler-detected planets with stellar metallicity (e.g., Fischer &
Valenti 2005) is cited as proof that core accretion is the only formation mechanism. Ida
& Lin (2004) have shown that core accretion is able to produce a metallicity correlation
consistent with the observations (Figure 2), once the model parameters are suitably
chosen (see the range of theoretical outcomes evident in Figure 2).
The evidence for large solid cores in a number of Doppler exoplanets that have been
Figure 1. Metallicity of planet host stars, from the Encyclopedia of Extrasolar Planets web
site: http://exoplanet.eu/.
393
3. Disk Instability
The competing scenario for giant planet formation is disk instability, where a gravitationally unstable disk forms spiral arms and self-gravitating clumps of gas and dust on
time scales of 0.001 Myr (Boss 1997, 1998; Mayer et al. 2002). Dust grains sediment
to the center of the clumps and form solid cores on time scales of 0.1 Myr, resulting
in gas giant protoplanets. Ice giant planets could be formed by UV photoevaporation of
the outermost gas giant protoplanets (Boss, Wetherill, & Haghighipour 2002; Boss 2003),
or by collisional accumulation, which must also be invoked to form terrestrial planets in
this necessarily hybrid scenario.
Disk instability requires a protoplanetary disk that is massive enough to be at least
marginally gravitationally unstable. For a solar-mass star, this typically requires a disk
with a mass of 0.1M . Estimated disk masses for T Tauri stars with ages of a few Myr
range from 0.01 to 0.1 M (Kitamura et al. 2002), suggesting that at least some protoplanetary disks are massive enough to form gas giant planets by disk instability. Current
estimates are that roughly 10% of nearby solar-type stars harbor gas giant planets with
masses greater than that of Saturn inside 3 AU (Cumming et al. 2008). Provided then
that at least 10% of protoplanetary disks are massive enough to undergo disk instability,
this mechanism could be a major contributer to the giant planet population.
Disk instability appears to be a strong candidate for forming planets in metal-poor
systems, such as the M4 pulsar planet, where the metallicity [Fe/H] = -1.5 (Sigurdsson
Figure 2. Predicted fraction (Ida & Lin 2004) of stars with gas giants formed by core accretion
(filled symbols) that are detectable by Doppler surveys (open circles), as a function of stellar
metallicity.
394
Alan P. Boss
et al. 2003), the giant planets orbiting HD 155358 and HD 47536, both of which have
[Fe/H] = -0.68 (Cochran et al. 2007), and the giant planet around HD 171028, with
[Fe/H] = -0.49 (Santos et al. 2007).
Boss (2002) found that disk instability proceeded in much the same way in models
where the dust grain opacities were varied by factors of 10 or 0.1 (Figure 4). Cai et al.
(2006) found that lowering the disk opacity by factors of 2 or 4 led to enhanced instability.
However, models by Mayer et al. (2007) found that lowering the opacity by a factor of 50
led to almost no difference in the outcome, confirming the results of Boss (2002). The
Mayer et al. (2007) models also suggested that disks with enhanced metallicity would
have higher mean molecular weights, leading to reduced gas pressure, which would aid
in gas giant planet formation and lead to a metallicity correlation for disk instability (cf.
Matsuo et al. 2007). Clearly this possibility is deserving of further investigation.
Helled et al. (2006) considered the question of the capture of planetesimals by a clump
formed by disk instability. They modeled the spherically symmetric contraction of a
Jupiter-mass clump for 3 105 yr, using a modified stellar evolution code, finding
that the protoplanets envelope would capture a large fraction of the planetesimals in its
feeding zone, leading to a heavy element enrichment over solar abundances (Figure 5).
Helled et al. (2008) considered the coagulation and sedimentation of dust grains in the
same Jupiter-mass, contracting clump, finding that a small heavy element core would
form, in basic agreement with the simplified analysis by Boss (1998). More recently,
Helled & Schubert (2009) showed that the heavy element enrichment of a giant gaseous
protoplanet depends strongly on the planets initial orbital distance and the assumed
surface density of solids. It is clear from this work that giant planets formed by disk
Figure 3. Mass accreted by a Saturn-like planet intended to represent HD 149026b in the core
accretion scenario (Dodson-Robinson & Bodenheimer 2009). Shortly after 1.5 Myr, the total
mass is 120 Earth-masses, and the solid core mass is 70 Earth-masses.
395
instability are likely to have highly non-solar compositions, just as is the case for planets
formed by core accretion.
4. Interior Models
Models of the interiors of Jupiter and Saturn by Guillot, Gautier, & Hubbard (1997)
and Guillot (1999) require considerably smaller core masses than were previously thought
to be the case, or even no core at all. Based on recent shock wave experiments, the bestfitting models of the Jovian interior restrict its core mass to be either 1M or to lie
in the range 0 to 4M (Saumon & Guillot 2004), depending on the equation of state
(EOS) that is assumed. In these interior models, Jupiters present core mass does not
appear to be large enough to trigger hydrodynamic accretion of disk gas, at least not
before the disk gas disappears. The suggestion has thus been made (Saumon & Guillot
2004) that Jupiters core (but not Saturns, for uncertain reasons) largely dissolved into
the overlying envelope after the planet formed, in which case the present Jupiter core
mass may not constrain the primordial core mass.
More recent work on the H-He EOS leads to a core mass of 14-18 M for Jupiter
(Militzer et al. 2008), while a different group finds a Jupiter core mass between 1 and 6
M (Nettelmann et al. 2008). Given that both core accretion and disk instability predict
the formation of sizable cores in gas giant planets, core erosion might need to be invoked
to explain the absence of a Jovian core for either formation mechanism.
There does seems to be agreement, however, between the different EOS calculations
that Saturn has a core mass of about 15 M . Helled & Schubert (2008) found that a
Saturn-mass protoplanet formed by disk instability would contract slower than a Jupitermass protoplanet, leading to the capture of more planetesimals and hence a larger core
Figure 4. Density contours in the equatorial plane in the high opacity 3D disk instability model
of Boss (2002), showing multiple clump formation in a disk with a radius of 20 AU. The inner 4
AU in radius is excluded from the calculation for time step reasons. A solar-mass protostar lies
at the center of the disk. Cross-hatched regions denote midplane densities above 1010 g cm3 .
396
Alan P. Boss
mass. In the core accretion scenario, Saturns formation is explained by shutting off gas
accretion at the right moment, as in the case for HD 149026s planet (Dodson-Robinson
& Bodenheimer 2009).
5. Metallicity Correlations
Support for the original suggestion by Gonzalez (1997) that a metallicity correlation
exists and might be explainable by self-pollution (i.e., stellar ingestion of inwardly migrating planets) has waxed and waned in the last decade, but seems to be resurging.
Pasquini et al. (2007) showed that giant stars with gas giant planets have the same
metallicities as giant stars without known planets (Figure 6), suggesting that the metallicity correlation seen in dwarf stars is a self-pollution effect that is erased by much more
extensive convective mixing in the giant star phase. Recently, Gonzalez (2008) has shown
that Li abundances in stars with planets correlate with the stellar surface temperature,
and concludes that this finding also supports the self-pollution hypothesis. There should
also be a correlation between the extent of inward orbital migration and metallicity, as
disks with higher dust grain opacities should have hotter midplanes, leading to higher
sound speeds and disk scale heights, and hence enhanced rates of migration through
gravitational interactions with a disk that evolves in the manner of a viscous accretion
disk. Clearly much more remains to be learned about the metallicity correlation and its
implications for planet formation, orbital migration, and pollution of the outer layers of
host stars.
6. Conclusions
The recent claims for the discovery by direct imaging of massive gas giant planets
orbiting at distances from their host stars beyond 20 AU (Figure 7) have been taken
by many as evidence that at least some gas giant planets form by disk instability, as
397
disk instability is able to operate at such distances (e.g., Boss 2003, but see Boss 2006),
whereas core accretion is unable to do so (e.g., Ida & Lin 2004) due to the low surface
density of solids at such distances. The low metallicities of the gas giants on wide orbits
discovered to date also argues in favor of formation by disk instability (Figure 7). We thus
appear to be in a situation where both core accretion and disk instability are required to
explain the full range of exoplanets discovered to date. In that case, the main theoretical
challenge is to try to determine which, if either, mechanism dominates. The question of
host star metallicities is certain to continue to play a leading role in the ongoing debate
over the formation mechanisms of giant planets.
I thank the American Astronomical Society for a Travel Grant that partially supported
my attendance at this Symposium, as well as travel support by the NASA Astrobiology
Institute through grant NNA09DA81A.
References
Boss, A. P. 1997, Science, 276, 1836
. 1998, ApJ, 503, 923
. 2002, ApJ, 567, L149
. 2003, ApJ, 599, 577
. 2006, ApJ, 637, L137
Boss, A. P., Wetherill, G. W., & Haghighipour, N. 2002, Icarus, 156, 291
Butler, R. P., et al. 2006, ApJ, 646, 505
Cai, K., et al. 2006, ApJL, 636, L149 (erratum 642, L173)
Figure 6. Metallicity distributions (Pasquini et al. 2007) of giant stars hosting detected planets (solid blue or dotted lines) compared to dwarf stars with detected planets (dashed red or
dash-dotted lines) and dwarf stars with planets with orbital periods greater than 180 days (grey
dotted or dash-dash-dotted lines). The giant star planet hosts do not show the strong metallicity
enhancement found for the dwarf star planet hosts (e.g., Fischer & Valenti 2005).
398
Alan P. Boss
Figure 7. Metallicity of planet host stars versus semi-major axes of the planets orbits, from
the Encyclopedia of Extrasolar Planets web site: http://exoplanet.eu/.
Cochran, W. D., Endl, M., Wittenmyer, R. A., & Bean, J. L. 2007, ApJ, 665, 1407
Cumming, A., et al. 2008, PASP, 120, 531
Dodson-Robinson, S. E., & Bodenheimer, P. 2009, ApJ, 695, L159
Fischer, D. A., & Valenti, J. 2005, ApJ, 622, 1102
Gonzalez, G. 1997, MNRAS, 285, 403
. 2008, MNRAS, 386, 928
Guillot, T. 1999, Planet. Space Sci., 47, 1183
Guillot, T., Gautier, D., & Hubbard, W. B. 1997, Icarus, 130, 534
Helled, R., Podolak, M., & Kovetz, A. 2006, Icarus, 185, 64
. 2008, Icarus, 195, 863
Helled, R., & Schubert, G. 2008, Icarus, 198, 156
. 2009, ApJ, 697, 1256
Ida, S., & Lin, D. N. C. 2004, ApJ, 616, 567
Kitamura, Y., et al. 2002, ApJ, 581, 357
Matsuo, T., et al. 2007, ApJ, 662, 1282
Mayer, L., Lufkin, G., Quinn, T., & Wadsley, J. 2007, ApJL, 661, L77
Mayer, L., Quinn, T., Wadsley, J., & Stadel, J. 2002, Science, 298, 1756
Militzer, B., et al. 2008, ApJ, 688, L45
Nettlemann, N., et al. 2008, ApJ, 683, 1217
Pasquini, L., et al. 2007, A&A, 473, 979
Santos, N. C., et al. 2007, A&A, 474, 647
Sato, B., et al. 2005, ApJ, 633, 465
Saumon, D., & Guillot, T. 2004, ApJ, 609, 1170
Sigurdsson, S., et al. 2003, Science, 301, 193
Wetherill, G. W. 1996, Icarus, 119, 219
1. Introduction
To date, more than 400 extrasolar giant planets are known to exist, the vast majority
being Neptune-mass or larger. The recent discovery of super-Earth-sized extrasolar terrestrial planets suggests that Earth-mass terrestrial planets may also be present within
many of these systems. Given that the number of planets in the galaxy is expected to
correlate inversely with planetary mass, it is expected that Earth-sized terrestrial planets
are much more common than giant planets (Marcy & Butler (2000)). The possibility of
terrestrial extrasolar planets has been examined by several authors, with many focussing
on the long-term dynamical stability of terrestrial planet forming regions (e.g. Raymond
& Barnes (2005)). Others have considered terrestrial planet formation in hypothetical
systems (Raymond et al. (2005), Mandell et al. (2007)) while only one study has examined terrestrial planet formation within a specific planetary system (Raymond et al.
(2006)). While such studies are undoubtably of great benefit to future planet-finding
missions, they do not provide any insight into the potential nature of a given terrestrial
planet.
At the same time, extrasolar planetary host stars are chemically distinct from the general stellar population, displaying significant variations in key planet-building elements
such as Fe, C, O, Mg and Si (Bond et al. (2008) and references therein). These enrichments are primordial in origin, established in the giant molecular cloud from which these
systems formed (e.g. Bond et al. (2008)). As a result, the chemistry of planet-building
materials in many of these systems is distinctly different from that in our Solar System. Consequently, it is likely that terrestrial extrasolar planets may have compositions
reflecting the enrichments observed in the host stars, possibly resulting in terrestrial
399
400
planets with compositions and mineralogies unlike any body yet observed within our
Solar System.
2.0
C/O
1.5
1.0
SiC
SiO
0.5
MgSiO3 &
SiO2
0.0
0.5
Mg2SiO4 &
MgO
1.0
1.5
2.0
2.5
Mg/Si
Figure 1. Mg/Si vs. C/O for known planetary host stars with reliable stellar abundances.
Stellar photospheric values were taken from Gilli et al. (2006) (Si, Mg), Beir
ao et al. (2005)
(Mg), Ecuvillon et al. (2004) (C) and Ecuvillon et al. (2006) (O). Solar values are shown by the
green circle and were taken from Asplund et al. (2006). The dashed line indicates a C/O value of
0.8 and marks the transitions between a silicate-dominated composition and a carbide-dominated
composition at 104 . Average 2- error bars shown in upper right.
401
402
however, with 2 confidence that at least 6 known planetary systems have C/O values
above 0.8 (10% of the sample) and would thus contain C and its associated phases as a
major planet forming material. These data clearly demonstrate that there are a significant number of systems in which terrestrial planets have compositions vastly different
to any planetary body observed in our Solar System. These bodies may represent a new
class of terrestrial planet.
The distribution of water vapor is also altered within the C-rich systems. Instead of
being present throughout the entire disk (as for systems with lower C/O values), water
vapor is only present in significant quantities at temperatures below 600 - 800K (generally
outside of 0.8 AU). Currently, one theory regarding the delivery of water to the early
Earth is that it occurred locally via adsorption onto solid grains that were later accreted
onto the Earth (Drake & Campins (2006)). Thus the restricted distribution of water
vapor may prohibit the formation of ocean-bearing planets in these systems.
Finally, the mass distribution of these high-C/O systems is intriguing. The combination
of a broad zone of refractory carbon-bearing solids and the relatively small amount of
water ice that condenses in these systems suggests that these systems may have more
solid mass located in the inner regions (inside 1 AU) of the disk than for Solar-like
disks. This mass distribution is the opposite of that expected in O-rich systems and
suggests that protoplanets may form more easily in the inner regions of these systems.
The full implications of this need to be examined by using alternative mass distributions
for extrasolar planetary formation simulations for both gas giant planets and smaller
terrestrial planets.
1. Introduction
Chemical abundances in the photospheres of cool main-sequence stars provide a fossil
record of the composition of the parent molecular core and the initial composition of the
circumstellar disk. Exceptions may occur, for example when light elements fuse at the
base of the convection zone or when a star preferentially accretes specific elements from a
an evolved stellar companion or perhaps an evolved circumstellar disk. Nonetheless, the
dependence of planet formation on disk abundances can be observationally constrained
by studying the frequency of observed planets as a function of photospheric abundance.
404
Jeff A. Valenti
Figure 1. Comparison of effective temperatures (left panel) and iron abundances (right panel)
from Fuhrmann (1998), Allende Prieto et al. (2004), Santos, Israelian, & Mayor (2004), relative
to values from the SPOCS catalog (Valenti & Fischer 2005). Systematic difference are evident.
Figure 2. Left: Distribution of iron abundances in a uniform planet search sample (Fischer &
Valenti 2005), for all stars (left bins) and for stars known to host giant planets (right bins).
Right: The fraction of stars known to host giant planets increases dramatically with increasing
iron abundance. Fischer & Valenti (2005) and Santos, Israelian, & Mayor (2004) obtain similar
results, despite different samples, spectrographs, and analysis techniques. A model by Ida & Lin
(2004) shows similar behavior.
0.15
0.3
15 %
0.15
[ Fe / H ] of Stellar Host
0.6
405
0.0
0.3
0.6
243 planets
M sin i > 20 M
1.07
0.1
1
Planet SemiMajor Axis (AU)
Snow
Line
10
Figure 3. Iron abundances of stars known to host planets more massive than 20 M . At least
85% of these planets have migrated inwards from beyond the snow line. The median semi-major
axis of known planets is 1.07 AU. The median iron abundance on both sides of this boundary
is 0.15, so there is no evidence that abundance affects the stopping point of inward migration.
median separation between known planets and their host star is 1.07 AU. Fig. 3 shows
that subsamples of known planets closer to and further from the star have similar iron
abundances, so the stopping point of migration does not depend on metallicity.
406
Jeff A. Valenti
0.6
MNep
[ Fe / H ] of Stellar Host
[ Fe / H ] of Stellar Host
0.6
0.3
0.0
Solar
0.3
0.6
10
100
1000
Mininum M sin i (M)
104
MNep
0.3
0.0
Solar
0.3
0.6
10
100
1000
Maximum M sin i (M)
104
Figure 4. Iron abundance of known planet hosts as a function of the least massive (left panel)
and most massive (right panel) known planet in the system. The maximum M sin i plot addresses
more directly whether stars with lower iron abundance than the Sun are able to form planets
less massive than Neptune, despite an observed dearth of more massive planets.
variations, which might imply low activity, old age, and hence low metallicity. These
detectability issues are best evaluated by the planet search team.
5. M Dwarf Abundances
M dwarfs are useful targets in the search for low mass planets because observed velocities are inversely proportional to stellar mass. However, M dwarf abundances are difficult
to measure spectroscopically due to inaccurate or missing molecular line data. Improved
line data are desperately needed, especially for TiO.
An alternate approach is to use broad-band photometric indexes that are calibrated
using binaries, where the primary is warm enough to obtain a reliable spectroscopic
abundance. Using binaries, Johnson & Apps (2009) have redetermined the photometric
[ Fe / H ] of Stellar Host
0.6
MNep
0.3
0.0
Solar
0.3
0.6
10
100
1000
Maximum M sin i (M)
104
Figure 5. Similar to Fig. 4, but with iron abundances for M dwarfs in Johnson & Apps (2009)
adjusted upwards ( symbols). Even after this correction, 5 of 6 planet hosts that do not have
a known planet more massive than Neptune tend to have lower iron abundance than the Sun.
More work is required to characterize sample and detection biases in searches for systems that
only have planets less massive than Neptune.
407
Abstract. Rapid rotation in red giant stars may be one signature of the past engulfment of a
planetary companion. Models of the future tidal interaction of known exoplanet host stars with
their planets show that many of these stars will accrete one or more of their planets, and the
orbital angular momentum of these accreted planets is sometimes sufficient to spin up the host
stars to a level commonly accepted as rapid rotation for giant stars. Planets accreted during
the red giant phase should leave behind a chemical signature in the form of unusual abundance
patterns in the host red giants atmosphere. Proposed signatures of planet accretion include
the enhancement of Li and 12 C; both species are generally depleted in giant star atmospheres
by convection but could be replenished by planet accretion. Moreover, accreted planets may
preferentially enhance the stellar abundance of refractory elements assuming that the refractory
nature of these elements leads to their relative enhancements in the planets themselves. Here
we present preliminary results of a search for these predicted chemical signatures through high
resolution spectroscopic abundance analysis of both rapidly rotating giant stars (i.e., stars with
a higher probability of having experienced planet accretion) and normally rotating giant stars.
We find that the rapid rotators are enhanced in Li relative to the slow rotators a result
consistent with Li replenishment through planet absorption.
Keywords. stars: abundances, stars: chemically peculiar,stars: rotation
1. Background
Generally, red giants are slow rotators with v sin i 2 km s1 . A small fraction of
red giants, however, have v sin i in excess of 8 km s1 ; these are the rapid rotators.
Planet accretion is a possible mechanism for spinning up rapidly rotating giants. Modeling
the tidal interaction of exoplanet host stars and their planets, Carlberg et al. (2009),
hereafter C09, found that the lower red giant branch (RGB) was one of the best late
stellar evolutionary stages to find evidence for planet induced rapid rotation. Giants
are good candidates for searching for chemical evidence of this accretion because they
are chemically distinct from planets in ways that main sequence stars are not. Giant
stars process light elements, reducing their surface abundances up to several orders of
magnitude. Consequently, a planet with unprocessed abundances of these light elements
can significantly enhance a giant stars atmospheric abundance. For this work, we focus
on lithium, an element that is easily measured in the optical spectra of giant stars.
408
409
410
Carlberg et al.
Figure 1. Lithium abundances versus effective temperature for the rapid and slow rotators.
Some of the slow rotators are giant stars with planets (SWP). The line indicates the general
trend of A(Li) with temperature as established from the data of B89. The region below the
sensitivity of B89 is hatched.
Figure 2. Histogram of the offset of A(Li) from the temperature trend (A(Li)/A(Li) ) of the
rapid and slow rotator samples. Histograms on the left are from this work, while histograms on
the right add stars from de Medeiros et al. (2000) and Drake et al. (2002).
stage. In the right panel of Figure 3, we show a prediction for the enhanced rotation expected from planet ingestion for the modeled future evolution of 99 known main sequence
stars with planets (SWPs) by C09. The density plot shows the fraction of SWPs passing
through each bin as rapid rotators compared to the number of stars passing through
the bin regardless of rotation. Only SWPs expected to be rapid rotators at some point
in their future RGB lifetimes are included. From this plot, we see that the Li-enhanced
rapid rotators do tend to coincide with regions that have higher fractions of SWPs passing through as rapid rotators. However, our giants do not completely overlap with the
future evolution of the less massive, more metal rich SWPs in this parameter space.
The mean chemical abundances of lithium in rapid and slow rotating red giants are
consistent with, albeit not conclusive evidence for, the hypothesis that planet accretion is
411
Table 1. Stars with unusually high A(Li) for their effective temperature, ordered by v sin i.
Star
G0928+73.2600
G0804+39.4755
Tyc3340-01195-1
Tyc0647-00254-1
Tyc5904-00513-1
3.71
1.23
1.24
1.99
0.80
-0.25
-0.39
-0.18
0.01
-0.47
Figure 3. Temperature and log g of giants. Triangles and squares are slow and rapid rotators,
respectively. (L) Solid lines give [Fe/H]=0.20 isochrones for five ages. (R) Color scale indicates
the fraction of exoplanet host stars passing through each bin as rapid rotators from C09.
responsible for spinning up rapid rotators and that the lithium enhancement is a chemical
signature of such accretion. This evidence includes (1) differences in A(Li) between the
slow and rapid rotators, as illustrated in Figure 2, (2) no clustering of the high A(Li)
rapid rotators on the TGP as would be expected if lithium was internally regenerated at
a specific evolutionary stage, and (3) high A(Li) rapid rotators on the TGP coincident
with where one expects to see planet accretion induced spin-up. However, much work
remains. In addition to completing the lithium analysis on the remainder of our sample,
we intend to look for ancillary evidence of chemical enrichment from planet accretion
including enhancements of 12 C and refractory elements in the rapid rotators.
References
Bizyaev, D., Smith, V. V., Arenas, J., Geisler, D., Majewski, S. R., Patterson, R. J., Cunha, K.,
Del Pardo, C., Suntzef, N. B., and Gieren, W. 2006, AJ, 131, 1784
Brown, J. A., Sneden, C., Lambert, D. L., & Dutchover, E. J. 1989, ApJS, 71, 293
Carlberg, J. K., Majewski, S. R., & Arras, P. 2009, ApJ, 700, 832
de Medeiros, J. R., & Mayor, M. 1999, A&AS, 139, 433
de Medeiros, J. R., do Nascimento, J. D., Jr., Sankarankutty, S., Costa, J. M., & Maia, M. R. G.
2000, A&A, 363, 239
Drake, N. A., de la Reza, R., da Silva, L., & Lambert, D. L. 2002, AJ, 123, 2703
Girardi, L., Bressan, A., Bertelli, G., & Chiosi, C. 2000, A&AS, 141,
Kurucz, R. L. 1994, http://kurucz.harvard.edu
Sneden, C. A. 1973, PhD thesis, The University of Texas at Austin
Centro de Astrofsica da Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal
email: jorge@astro.up.pt
2
Max-Planck-Institut f
ur Astrophysik, Germany
3
Institutionen f
or fysik och astronomi, Uppsala universitet, Sweden
4
Research School of Astronomy & Astrophysics, Australian National University, Australia
Abstract. For more than 140 years the chemical composition of our Sun has been considered
typical of solar-type stars. Our highly differential elemental abundance analysis of unprecedented
accuracy ( 0.01 dex) of the Sun relative to solar twins, shows that the Sun has a peculiar
chemical composition with a 20% depletion of refractory elements relative to the volatile
elements in comparison with solar twins. The abundance differences correlate strongly with the
condensation temperatures of the elements. A similar study of solar analogs from planet surveys
shows that this peculiarity also holds in comparisons with solar analogs known to have close-in
giant planets while the majority of solar analogs without detected giant planets show the solar
abundance pattern. The peculiarities in the solar chemical composition can be explained as
signatures of the formation of terrestrial planets like our own Earth.
Keywords. Sun: abundances, Sun: fundamental parameters, solar system: formation, stars:
abundances, stars: fundamental parameters, planetary systems: formation, Galaxy: abundances
1. Introduction
For many years people have wondered about how our Sun compares to other stars
and to whether our existence is related to special properties of our solar system. Angelo
Secchi compared the Sun to many bright stars (Secchi 1868). He classified the stars in
three types: type I which is the modern class A and early F, type II which are M stars,
and type III comprising modern class G, K and early F, and called by Secchi tipo solare
(solar type), due to their spectroscopic similarity to the Sun. Furthermore, he concluded
that le stelle di questo terzo tipo mostrano di avere una composizione identica a quella
del nostro Sole (Secchi 1868), meaning that solar type stars have identical composition
to our Sun. Further works (e.g. Payne 1925; Edvardsson et al. 1993; Reddy et al. 2003)
have not conclusively shown whether the Sun has a normal composition or not, due to the
relatively large (& 0.05) remaining systematic errors (Gustafsson 2008). Thus, for more
than 140 years father Secchis conclusion on the universal solar composition of the
Sun has remained valid. In order to make further progress, it is important to eliminate
many of the systematic errors ( 0.05-0.1 dex) that plague stellar chemical composition
analyses (Asplund 2005). Solar analogs, which are G0-G5 dwarfs, and solar twins, stars
almost identical to the Sun (Cayrel de Strobel 1996), are important in this context, in
particular solar twins, because due to their similarity to the Sun a highly differential
analysis will cancel most systematic errors.
Thus, the first step in accurate comparisons of the Sun to other stars is to find solar
twins. After many years of intensive search (see review by Cayrel de Strobel 1996), Porto
412
413
de Mello & da Silva (1997) found the first solar twin, 18 Sco [HD 146233], which is much
closer to the Sun than previous candidates like 16 Cyg B [HD 186427] (see e.g. Fig. 2
of Melendez et al. 2006). More recent surveys have largely increased the number of solar
twins in the field (Melendez et al. 2006; Melendez & Ramrez 2007; Takeda et al. 2007;
Melendez et al. 2009; Ramrez et al. 2009) as well as in the open cluster M67 (Pasquini et
al. 2008). The most productive survey at high resolution (R 60,000 - 110,000) is being
undertaken by our group. Most Northern solar twin candidates were observed with the
2.7m telescope at the McDonald observatory, and complemented with data obtained at
the Keck observatory. The Southern targets were observed using the Magellan Clay 6.5 m
telescope at Las Campanas, complemented with recent (August 2009) VLT observations.
Our solar twin project started in 2002, when only one solar twin was known. In order
to improve our chances of finding solar twins, we first expanded the color-temperature
relations of Alonso et al. (1996) to other photometric systems (Melendez & Ramrez
2003), allowing us to use existing photometry in other systems (e.g. Geneva) to select
the best targets. We later improved the calibrations adding more stars and including new
homogeneous systems like Tycho (BT VT ) and 2MASS (Ramrez & Melendez 2005).
Although our first solar twin proposal in 2004 did not fly, the same year a more exciting
proposal on Li in halo stars was granted a few nights with Keck. A small amount of that
observing time was devoted to twin candidates, resulting in the discovery of the second
solar twin (HD 98618; Melendez et al. 2006) and revealing that our temperature scale
(and that of Alonso et al. 1996), although precise, have probably a zero-point issue. Our
new selection of candidates from the Hipparcos catalog took into account a preliminary
zero-point offset, which has been recently confirmed by analyses of solar twins and used
for improved temperature calibrations (Casagrande et al. 2009).
New solar twin proposals at McDonald and Magellan at Las Campanas (through Australian access) during 2006 were successful. The first observing run at McDonald in April
2007 allowed us to identify the best solar twin known to date, HIP 56948 (Melendez &
Ramrez 2007), which is not only very similar to the Sun physically, but has also a low Li
abundance similar to solar. Its status of best solar twin has been recently confirmed by
Takeda & Tajitsu (2009). On the other hand, the Magellan observations at Las Campanas
are opening new windows for astrophysics of the 0.01 dex level in chemical abundance
accuracy. New observations taken recently at the VLT with UVES and CRIRES, promise
to achieve even better precision (0.005 dex, 1%), and to use solar twins to set tight
constraints on Li and Be depletion in the Sun (e.g. do Nascimento et al. 2009).
414
J. Melendez et al.
Figure 1. Differences between [X/Fe] of the Sun and the mean values in the solar twins as a
function of Tcond . The abundance pattern shows a break at Tcond 1200 1250 K. The solid
lines are fits to the abundance pattern, while the dashed lines represent the standard deviation
from the fits. The low element-to-element scatter from the fits for the refractory ( = 0.007 dex)
and volatile ( = 0.011 dex) elements confirms the high precision of our work. Observational
errors (including errors in both the Sun and solar twins) are shown with dotted error bars, while
the errors in the mean abundance of the solar twins are shown with solid error bars.
the solar twins. Also, a radial gradient exists, with greater enhancement of refractories
at smaller heliocentric distances (Palme 2000). The break in the chemical abundance
pattern at Tcond 1200 K (Fig . 1), suggests that the volatiles retained their original
abundances both in the Sun and the solar twins. Such temperatures are only encountered
in the inner parts (< 3 AU) of proto-planetary disks (Ciesla 2008), which also suggests
that the abundance pattern is related to the presence of terrestrial planets. Furthermore,
iron meteorites (like those we see in museums around the world) were probably formed
in the terrestrial planet region and may be a remnant of Earth forming material (Bottke
et al. 2006), hence being perhaps formed from the iron missing in the Sun.
The amount of dust-depleted gas required to explain the solar abundances is similar
to the mass of refractories locked up in Mercury, Venus, Earth and Mars (Melendez et
al. 2009). Thus, it is tempting to speculate that the formation of the terrestrial planets
might have given the Sun its special surface composition. The disk masses during the
T Tauri phase are 0.02 M but larger values are possible, providing thus enough
material to change the solar photospheric composition. However it may be a problem
with time-scales since proto-planetary disks are observed to have typical life-times < 10
Myr (Sicilia-Aguilar et al. 2006), but the deep solar convection zone at that time would
have erased the planetary signatures, unless the proto-solar nebula was abnormally long
lived so that the dust-depleted gas was accreted 20 Myr later when the solar convection
zone was thin. Another possibility is that the early Sun was never fully convective, as
shown by the dynamical star-formation calculations with a time-dependent convection
treatment of Wuchterl & Tscharnuter (2003) and Wuchterl & Klessen (2001), making it
easier to imprint the signatures of other Earths (Nordlund 2009).
415
We have also analyzed a sample of 10 solar analogs from planet surveys, four having
inner giant planets, while for the other six no planets have been detected yet. Our analysis
shows that all solar analogs with giant planets differ from the Sun but closely resemble
most solar twins. Thus, the odd solar composition is not due to giant planets as such,
but probably related to terrestrial planets. Indeed, only two of the six solar analogs
without close-in giant planets have abundances that differ significantly from the solar
pattern. The fraction of stars with the solar pattern seems thus tentatively related to the
presence of giant planets on close orbits: 0% when having such planets and 10 20%
for solar-type stars in general, and 50%-70% without close-in giant planets. Although
the statistics has to be improved considerably, the numbers are clearly tantalizing.
1. Introduction
Fourteen years after the discovery of 51 Pegb (Mayor & Queloz 1995), the aims of
Doppler surveys for planets are now evolving fast. On the one hand, existing surveys
are extending their time baseline and/or are achieving higher velocity precision (6 1 m
s1 , see for example Lovis et al. 2006), to continue searching for planets at increasingly
larger orbital distances (e.g., Fischer et al. 2007) and with increasingly smaller masses
(e.g., Mayor et al. 2008). On the other hand, early evidence for a strong relationship
between the physical properties of stars and the likelihood that they harbor planets has
prompted both theoretical analyses attempting to reconcile the observed trends within
the framework of planet formation models as well as renewed experimental efforts to
put such trends on firmer statistical grounds and thus thoroughly test the theoretical
explanations put forth to explain their existence.
For example, dedicated radial-velocity (RV) surveys of M dwarfs (e.g., Endl et al. 2006),
of Hertzsprung gap sub-giants (Johnson et al. 2006), heavily evolved stellar samples
belonging to the red-giant branch and clump regions of the H-R diagram (Sato et al.
2003; Setiawan et al. 2005; Lovis & Mayor 2007; Niedzielski et al. 2007) and early-type
dwarfs (Galland et al. 2005) are investigating the predictions for a positive correlation
416
417
between the mass of the host and the occurrence rate and mass of planets (Ida & Lin
2005).
Another important relationship uncovered so far between planet characteristics and
frequencies and host properties is quantified by the strong positive correlation between
planet frequency fp and stellar metallicity [Fe/H]. On the one hand, the observational
evidence (Gonzalez 1997; Santos et al. 2004; Fischer & Valenti 2005) has found theoretical
support within the context of the core accretion model of giant planet formation (Ida
& Lin 2004). Doppler surveys biased toward high-metallicity stellar samples (Fischer et
al. 2005; Bouchy et al. 2005) have begun in recent years, prompted by the enhanced
chances of finding large numbers of planets. On the other hand, the possible bimodality
of the fp -[Fe/H], with a flat tail for [Fe/H]6 0.0 (Santos et al. 2004) might indicate
that more than one formation mechanism is at work (e.g., Boss 2002). In order to better
characterize the dependence of giant planet frequency in the metal-poor regime, two
surveys have started monitoring stellar samples with [Fe/H]6 0.5. A southern sample
of 100 metal-deficient stars is being monitored with HARPS (Santos et al. 2007), while
our group (Sozzetti et al. 2006) has focused on a northern sample ( & 25 o ) of 200
objects. We provide here a summary of the results obtained during our three-year long
observing campaign with Keck/HIRES (see Sozzetti et al. (2009a) for details).
2. Radial-velocity Results
As described in Sozzetti et al. (2006), all stars are drawn from the Carney-Latham
and Ryan samples of metal-poor, high-velocity field stars (e.g., Ryan & Norris 1991;
Carney et al. 1994). Based on a decade-long radial-velocity monitoring with the CfA
Digital Speedometers (Latham 1992) none of our program stars showed signs of velocity
variation at the 0.5 to 1.0 km s1 level. We selected targets with V 6 12 mag, Teff 6 6000
K, and 2.0 .[Fe/H]6 0.6.
All observations were collected with HIRES and its I2 cell (except for one Iodine-free
template exposure per target) on the Keck 1 telescope (Vogt et al. 1994). To extract the
RV information from each star+iodine spectrum, we perform a full spectral modeling
which includes the reconstruction of the asymmetries, spatial and temporal variations
in the HIRES instrumental profile at the time of observation. Our algorithm follows a
procedure based on the methodology developed for the AFOE spectrograph (Korzennik
et al. 2000), and adapted for the processing of HIRES spectra, as described in Sozzetti
et al. (2009b). The median RV uncertainty for the stars in our sample is RV ' 9 m s1 .
All stars were probed for excess variability, based on the close relative agreement of
three statistical tests (F -test, 2 -test, and Kuipers test). Nine stars were identified as
variables based on the above criteria. These nine objects (HD 7424, G197-45, G 237-84,
G 63-5, G 135-46, HD 192718, HD 210295, G 27-44, and G 28-43) all exhibit an RV
scatter in the measurements at least four times larger than the nominal average internal
error.
For six of the above mentioned objects, the RV data were better described by a linear
slope, indicating the presence of a massive companion orbiting with a period greatly
exceeding the duration of the observations (>> 3 yr). As for the other three variables
(HD 210295, G 27-44, and G 28-43), the RV residuals were not significantly improved
after fitting for a linear trend. The three stars clearly exhibit significantly non-linear RV
variations (see Sozzetti et al. (2009a) for details). Both the long-term, low-amplitude
radial-velocity variables as well as those showing non-linear trends are the objective of
dedicated follow-up work with direct imaging techniques at infrared wavelengths, which
will be presented elsewhere.
418
Figure 1. Left: Survey completeness for companions of given mass, orbital period, and eccentricity. The limits shown are for 50% and 95% completeness and for three realizations with
different values of e. Right: Percentage of planet hosting stars as a function of metallicity (0.25
dex bins) for the sample constructed combining our survey stars with those of Fischer & Valenti
(2005). The increasing trend in the fraction of stars with planets as a function of metallicity
is well fitted with a power law, but the data are compatible with a constant occurrence rate
fp ' 1% for [Fe/H]6 0.0.
3. Survey Completeness
The quantitative determination of the sensitivity of an RV survey such as ours to
planetary companions of given mass and period relies upon detailed numerical simulations
of synthetic datasets of RV observations of a population of planetary systems of varying
orbital properties and masses. The ability to recover, with a given level of statistical
significance, the presence of a planetary signal or not translates into lower limits on the
detectable (minimum) companion mass as a function of e.g. period and eccentricity (for
an assumed mass of the central star). For this purpose, we utilized a statistical approach
based on 2 - and F -tests to detect excess residuals above an assumed level of Gaussian
noise. As a result of this exercise, our survey would have detected, with a 99.5% confidence
level, over 95% of all companions on low-eccentricity orbits with velocity semi-amplitude
K & 100 m s1 , or Mp sin i & 4.1 MJ (P/yr)(1/3) , for orbital periods P 6 3 yr (left panel
of Figure 1).
None of the stars in our sample exhibits radial-velocity variations compatible with the
presence of Jovian planets with periods shorter than the survey duration. The resulting
average frequency of gas giants orbiting metal-poor dwarfs with 2.0 6[Fe/H]6 0.6 is
fp < 0.67% (see Sozzetti et al. (2009a) for details).
4. Discussion
We do not detect any massive planets (Mp sin i & 1 4 MJ ) within 2 AU of metalpoor stars with 2.0 6[Fe/H]6 0.6, and constrain their occurrence rate to be no larger
than fp ' 1%. All long-period trends identified in our survey are compatible with being
induced by brown dwarf or stellar companions. We examine the implications of this null
result in the context of the observed correlation between the rate of occurrence of giant
planets and the metallicity of their main-sequence solar-type stellar hosts. By combin-
419
ing our dataset with the Fischer & Valenti (2005) uniform sample, we confirm that
the likelihood of a star to harbor a planet more massive than Jupiter within 2 AU can
be expressed as a quadratic function of the hosts metallicity (right panel of Figure 1).
These findings appear to be circumstantial evidence in favor of the core accretion mechanism of giant planet formation (Pollack et al. 1996). However, the data for stars with
1.0 6[Fe/H]6 0.0 are compatible, in a statistical sense, with a constant occurrence rate
fp ' 1%. This evidence could be read as supportive of the alternative disk instability
model (Boss, 2000). However, though very useful, our results are not resolutive, and a
number of observational avenues should be pursued (e.g., lowering the mass sensitivity
threshold, increasing the sample sizes, and extending the time baseline of planet surveys)
to expand and improve the statistics and thus further constrain proposed models.
Acknowledgements
A.S. gratefully acknowledges for partial support the Kepler mission (NASA Cooperative Agreement NCC 2-1390) and the Italian Space Agency (Contract ASI-Gaia
I/037/08/0).
References
Boss, A. P. 2000 ApJ, 536, L101
Boss, A. P. 2002, ApJ, 567, L149
Bouchy, F., et al. 2005, A&A, 444, L15
Carney, B. W., et al. 1994, AJ, 107, 2240
Endl, M., et al. 2006, ApJ, 649, 436
Fischer, D. A., & Valenti, J. 2005, ApJ, 622, 1102
Fischer, D. A., et al. 2005, ApJ, 620, 481
Fischer, D. A., et al. 2007, ApJ, 675, 790
Galland, F., et al. 2005, A&A, 444, L21
Gonzalez, G. 1997, MNRAS, 285, 403
Korzennik, S. G., et al. 2000, ApJ, 533, L147
Ida, S., & Lin, D. N. C. 2004, ApJ, 616, 567
Ida, S., & Lin, D. N. C. 2005, ApJ, 626, 1045
Johnson, J. A., et al. 2006, ApJ, 652, 1724
Latham, D. W. 1992, ASP-CS, 32, 110
Lovis, C., et al. 2006, Nature, 441, 305
Lovis, C., & Mayor, M. 2007, A&A, 472, 657
Mayor, M., & Queloz, D. 1995, Nature, 378, 355
Mayor, M., et al. 2008, A&A, 493, 639
Niedzielski, A., et al. 2007, ApJ, 669, 1354
Pollack, J. B., et al. 1996, Icarus, 124, 62
Ryan, S. G., & Norris, J. E. 1991, AJ, 101, 1835
Santos, N. C., Israelian, G., & Mayor, M. 2004, A&A, 415, 1153
Santos, N. C., et al. 2007, A&A, 474, 647
Sato, B., et al. 2003, ApJ, 597, L157
Setiawan, J., et al. 2005, A&A, 437, L31
Sozzetti, A., et al. 2006, ApJ, 649, 428
Sozzetti, A., et al. 2009a, ApJ, 697, 544
Sozzetti, A., et al. 2009b, ApJ, 691, 1145
Vogt, S. S., et al. 1994, Proc. SPIE, 2198, 362
1. Introduction
Presently, exoplanet sample sizes are sufficiently so large that what were previously seen
as marginal trends can be checked against more detailed models of planetary formation.
The first and apparently main trend discovered in these exoplanetary samples were the
planetmetallicity connection (Gonz
alez 1997, Fischer & Valenti 2005). The orbital period
distribution also points to the existence of planetary populations which may be linked to
migration (Udry, Mayor & Santos 2003). We show here that a new planetary classification
emerge from the orbital angular momentum distibution.
421
Figure 1. Panel A shows the massangular momentum diagram for planets known by mid-2007.
Two apparent linear relations between mass and angular momentum can be seen (marked by
solid lines). A nonparametric regression model was used to define these two relations. Solid
squares and open circles planets represents planets having semimajor axis larger and smaller
than 0.67 AU, respectively. Solar System planets are identified by their traditional symbols.
Panel B shows the angular momentum p.d.f.s for the selected mass bin in Panel A as an example,
showing the bimodality in L.
For each mass bin we obtained the modes of the resulting L p.d.f.s. Then, we were able
to define a relation between the orbital angular momentum and the planetary mass using
a linear fit to these modes for each mass bin. From this, we can classify each planet in
populations according to which L M relation it probably follows. We will call these
populations and . See an example of this separation in Fig 1b.
The distribution of planetary mass for both populations indicates that population
has more low mass planets than population , suggesting that Population planets
had a runaway migration process (for instance, see Trilling et al. 2002), explaning why
Population presents less angular momentum than Population which is mainly composed by more massive planets and would have had a different migration process. The
Kolmogorov-Smirnov test on the cumulative distribution of planetary host stars mass
indicate that we can reject the hyphotesis that the distributions come from the same
population with a 95% confidence level. This can also indicate that both populations
may have been through different kinds of migration (e.g., Kennedy & Kenyon 2009).
References
Butler R. P., et al. 2006, ApJ, 646, 505
Gonz
alez, G. 1997, MNRAS, 285, 403
Fisher, D. A., & Valenti, J. 2005, ApJ, 622, 1102
Kennedy, G. M., Kenyon, S. J., ApJ, 695, 1210
Tarter, M. E. & Lock, M. D., 1993, Model-Free Curve Estimation, Chapman & Hall: New York,
p. 55
Trilling, D., Lunine, J., & Benz, W. 2002,
Udry, S., Mayor, M. & Santos, N. C. 2003
Abstract. We explore the non-standard mixing history of five solar twins to determine as
precisely as possible their mass and age. For this, we computed a grid of evolutionary models
with non-standard mixing at given metallicities with the Toulouse-Geneva code for a range
of stellar masses. We choose the evolutionary model that best fit the low lithium abundances
observed in the solar twins. Our best model for each solar twin provides a mass and age solution
constrained by their Li content and Teff determination. Li depletion due to the additional
mixing in solar-twins is strongly mass dependent. An accurate lithium abundance measurement
connected with non-standard models provides a more precise information about the age and
mass better than that determined only by classical methods.
Keywords. stars: abundances, (stars:) Hertzsprung-Russell diagram, turbulence
2. Evolutionary Models
Stellar evolution calculations were computed with the Toulouse-Geneva stellar Evolution Code TGEC. Details of the physics can be found in Richard et al. (1996, 2004),
Hui-Bon-Hoa (2008), and do Nascimento et al. (2009). Rotation-induced mixing is computed as described in Theado & Vauclair (2003). The calibration method of the models
is based on the Richard et al. (1996) prescription, with the solar values of Richard et al.
(2004). We obtained for our best-fit solar model L = 3.8499 1033 erg.s1 , R = 6.95938
1010 cm, and log N(Li) = 1.13 at an age = 4.576 Gyrs. The sound velocity profile
of our best-fit model is consistent with that deduced from helioseismology inversions by
Basu et al. (1997). The input parameters for the other masses are the same as for the
1.00 M model.
422
423
!7" * % $ )
!7. 6 ) + "
-
!4+ 5 5 % * 6
!#. % * . /
& + )10 2 3 ,
Figure 1. Comparison between masses and ages determined by TGEC models (filled symbols)
and masses and ages estimated by the observations (open symbols). Squares correspond to the
solar twins observed by Melendez & Ramrez (2007) and circles to the solar twins observed by
Takeda et al. (2007). The errors bars are as described in the text.
3. First Issues
For each star, we found the mass of the model passing through the observed point in a
HR diagram. The masses inferred are within the range expected for the mass of a solar
twin ( 5% of the solar mass). The precision of the mass determination is directly linked
to the precision of the Teff estimations from the observations. In a log N(Li)-Teff diagram,
the track passing through the observed point provides the most probable modeling of the
observed star, and the values of our mass and age estimations. The mass determined with
this method are consistent with precedent results with a better precision (see Figure 1).
An accurate Li abundance measurement and non-standard models provide more precise
informations about the age and mass of solar twins, more precisely than determined by
classical methods, and then a more realistic characterization of solar twin stars.
References
Basu, S., Christensen-Dalsgaard, J., et al. 1997, MNRAS, 292, 243
do Nascimento, J.-D., Jr., Castro, M., et al. 2009, A&A, 501, 687
Hui-Bon-Hoa, A. 2008, Ap&SS, 316, 55
Melendez, J., & Ramrez, I. 2007, ApJ, 669, L89
Richard, O., Vauclair, S., Charbonnel, C., Dziembowski, W.A. 1996, A&A 312, 1000
Richard, O., Thado, S., Vauclair, S. 2004, SoPh, 220, 243
Takeda, Y., Kawanomoto, S., Honda, S., Ando, H., Sakurai, T. 2007, A&A, 468, 663
Theado, S., Vauclair, S. 2003, ApJ 587, 784
van Leeuwen, F. 2007, ASSL, 350
Observat
orio Nacional, Rua Jose Cristino 77, S
ao Cristov
ao, 20921-400, Rio de Janeiro,
Brazil
email: carolina@on.br, delareza@on.br, claudio@on.br
2
S
ao Paulo State University - UNESP, Grupo de Din
amica Orbital & Planetologia,
Guaratinguet
a, CEP 12.516-410, Brazil
email: ocwinter@pq.cnpq.br, rcassia@feg.unesp.br
3
Sobolev Astronomical Institute, St. Petersburg State University, Universitetski pr. 28,
St. Petersburg 198504, Russia
email: drake@on.br
1. Introduction
Gonzalez (1997) and Santos et al. (2001) first showed the presence of metallicity excess
in stars with giant planets (SWP). The relative distribution of the stellar abundances
of refractory elements, with a high condensation temperature (TC ), with respect to the
volatile elements (low TC ), has been widely used in the literature as a tool to investigate
the nature of the metal enrichment in SWP. In this work we present other interpretation
of these gradients by studying the relationship between the metallicity of these CoRoT
systems and the migration of the planets resulting from interaction with planetesimals.
Table 1. Stellar parameters of the CoRoT exoplanets parent stars
Star Name
CoRoT-2
CoRoT-4
Teff [K]
log g
[km/s]
[Fe/H]
v sin i [km/s]
8.5 1.0
5.5 1.0
425
Figure 1. Relative abundances distribution of elements for CoRoT-2 (red squares) and
CoRoT-4 (blue circles) stars in function of their condensation temperatures
3. Numerical simulation
We numerically integrated the CoRoT-2 and CoRoT-4 systems, within the circular
restricted three-body problem: star-planet-planetesimals. We assumed that large planets
observed close to their parent stars actually have been formed at larger distances but
migrated inward due to lose energy and excess energy used to disperse the planetesimals.
A detailed description of the planet migration used in this work is presented in Winter et
al. (2007). For simplicity, we assume that in a certain epoch of the disk evolution, more
properly between 20-30 Myr, the disk is formed by a sea of planetesimals defining three
representative zones: refractory (R), intermediate (I) and volatile (V).
Figure 2. Histogram of the percentage of planetesimals from R, I and V zones that fall on
Corot-4 for the case of aPi = 5 AU. Zones: Refractory 0.0.30.1 AU, 17801360 K; Intermediate,
0.11.56 AU, 1360200 K and Volatiles, 1.564.5 AU, < 200 K. The time of integrations are
from left to rigth: 100 000 yr, 10 000 yr and 1000 yrs.
4. Results
We present the abundance results in function of TC for CoRoT-2 and CoRoT-4 stars.
We find that a flat distribution is the rule for them. In the case Corot-4 system the
accretion shows large and similar contributions of I and V particles and a very small contribution of pure refractory elements as presented in Figure 2. In other words, accretion
is mainly cool and warm and not hot as largely mentioned in the literature. In the
case of the system 2, an unrealistic large disk mass is necessary to bring the planet to
their observed final distance with respect to the star. In adittion, CoRoT-2 star deserves
a special attention due to its youth indicators represented respectively by the Ca II H
and K lines and its strong Li resonance line. This age can be obtained by means of the
equivalent width (EW) and abundance of Li. Considering all these matches, we infer an
age of 120 Myr for CoRoT-2.
References
Gonzalez, G. 1997, MNRAS, 285, 403
Kurucz, R. L. 1993, VizieR Online Data Catalog, 6039, 0
Santos, N. C., Israelian, G., and Mayor, M. 2001, A&A, 373, 1019
Sneden, C. A. 1973, PhD thesis, University of Texas at Austin.)
Winter, O. C., de la Reza, R., Domingos, R. C., Boldrin, L. A. G., & Chavero, C. 2007, MNRAS,
378, 1418
Divis
ao de Astrofsica, Instituto Nacional de Pesquisas Espaciais, Brazil
email: dasilvar@das.inpe.br
Abstract. We have been analyzing a large sample of solar-like stars with and without planets
in order to homogeneously measure their photospheric parameters and Carbon abundances.
Our sample contains around 200 stars in the solar neighborhood observed with the ELODIE
spectrograph, for which the observational data are publicly available. We performed spectral
synthesis of prominent bands of C2 and C I lines, aiming to accurately obtain the C abundances.
We intend to contribute homogeneous results to studies that compare elemental abundances in
stars with and without known planets. New arguments will be brought forward to the discussion
of possible chemical anomalies that have been suggested in the literature, leading us to a better
understanding of the planetary formation process. In this work we focus on the C abundances
in both stellar groups of our sample.
Keywords. stars: fundamental parameters, stars: abundances, planetary systems
1. Introduction
The Sun is widely thought to be formed from material representative of local physical
conditions in the Galaxy at the time of its formation and to represent a standard chemical
composition. This homogeneity hypothesis has been often put in question because of many
improvements in the observations techniques and data analysis. With the discovery of
new planetary systems, the already known heterogeneity sources (stellar nucleosynthesis,
stellar formation process) have gained a new perspective and brought new questions.
It is now a fact that stars with giant planets are, on average, richer in metals than
those for which no planet was detected (Gonzalez 2006 and references therein). Some
authors suggested that this kind of anomaly may not only involve the content of heavy
elements but also some light elements like Li, C, N, and O (Ecuvillon et al. 2006 and
references therein). Other authors found no difference in the abundance of light elements
when comparing stars with and without planets (see Ecuvillon et al. 2004 and references
therein). The studies above are not conclusive yet. We need new tests, using more accurate
and homogeneous data, with a larger number of stars. Abundances of these elements in
solar-like stars will bring new information that will surely help to distinguish the different
stellar and planetary formation processes. This is the purpose of our work, in which we
analyzed a sample of 200 stars to homogeneously measure the photospheric parameters
and C abundances.
427
which data are publicly available (Moultaka et al. 2004). We obtained Teff , [Fe/H], log(g),
and (through the excitation equilibrium of neutron iron and the ionization equilibrium
between Fe I and Fe II). We also performed spectral synthesis of some regions containing
bands of the C2 Swan System (e.g. at 5165), using the MOOG LTE code (Sneden 2000,
http://verdi.as.utexas.edu/moog.html).
4. Conclusions
The preliminary results on C abundances are presented here for a large number of
nearby solar-like stars, based on homogeneous photospheric parameters obtained from
spectra with high signal-to-noise ratio and high resolution. Our analysis used public spectra from the ELODIE database, which represent about 90% of all data. The remaining
10% include many stars with detected planets and having many observations that shall
be analyzed in the same way as soon as they become available to the scientific community,
since they will contribute to more reliable conclusions.
Figure 1. Left: spectral synthesis applied to the solar spectrum for several C abundances (in
steps of 0.1 dex). The best fit gives [C/Fe] = 0.01 dex. Right: [C/Fe] versus [Fe/H] diagram
comparing stars with and without planets.
References
Baranne, A., Queloz, D., Mayor, M., et al. 1996, A&AS, 119, 373
Ecuvillon, A., Israelian, G., Santos, N.C., et al. 2004, A&A, 426, 619
Ecuvillon, A., Israelian, G., Santos, N.C., et al. 2006, A&A, 449, 809
Gonzalez, G. 2006, PASP, 118, 1494
Moultaka, J., Ilovaisky, S.A., Prugniel, P., & Soubiran, C. 2004, PASP, 116, 693
1. Introduction
Ices such as H2 O, CO, CO2 , NH3 can be found in a variety of astrophysical environments (e.g. icy grain mantles, the satellites and rings of giant planets, comets, dense
clouds, protoplanetary disks). These ices are exposed to irradiation by energetic particles:
proton and UV radiation are the dominant component in most cases, but also heavy ions
are present in galactic cosmic rays and also in solar (and stellar) wind. Irradiation effects
induced by electrons, UV photons, protons, and low energy heavy ions have been studied
since some decades. Nevertheless, few experimental studies using swift heavy ions in order
to simulate the heavy ion component of cosmic rays do exist. Therefore, we performed
irradiations of ices with Ni ions in the electronic energy loss regime, where energy deposition in inelastic collisions with target electrons dominates, and ion penetration depths
are rather high.
428
429
3. Conclusion
We have irradiated CO and CO2 ices with swift nickel ions in the electronic energy
loss regime. The destruction of the ice molecules and formation of new species were
observed. An interesting result is the important role of heavy ion induced sputtering,
because of the quadratic dependence on the electronic energy loss of the projectile. We
estimated the specific sputtering flux (SSF) by multiplying the ion flux in cosmic rays
by the corresponding sputtering yield, which was estimated as a function of velocity by
calculating the energy loss with the widely used SRIM code (www.srim.org) and assuming
Y S2e . In the case of cosmic rays in dense clouds, the heavy ion SSF is dominant over
that of protons, even if the abundance of heavy ions is at least three orders of magnitude
lower than the proton one [Seperuelo Duarte et al. (2009)].
References
Seperuelo Duarte, E., Boduch, P., Rothard, H., Been, T., Dartois, E., Farenzena, L.S. & da
Silveira, E.F. 2009, Astron. Astrophys., 502, 599
Palumbo, M.E., Leto, P., Siringo, C. & Trigilo, C. 2008 Astrophys. J., 685, 1033
Cottin, H., Moore, M.H. & Benilan, Y., 2003 Astrophys. J., 590, 874
Jamieson,C.S., Mebel, A.M.& Kaiser, R. I. 2006, Astrophys. J. Suppl. Series, 163, 184
Brown, W.L., Augustyniak, W.M., Simmons, E., Marcantonio, K.J., Lanzerotti, L.J., Johnson,
R.E., Borring, J.W., Reimann, C.T., Foti, G. & Pirronello, V., 1982 Nucl. Instr. Meth.,
198, 1
In Fig. 1 we present the derived Be abundances as a function of effective temperature. Be shows a trend similar to previous works if we exclude the subgiant stars with
temperatures below 4900 K. Overall, no clear difference seems to exist between planet
hosts and comparison stars. Be abundances decrease from a maximum near Teff = 6100
K towards both higher and lower temperatures although there is an object that does
not follow this trend. Very little, if any, depletion has occurred for stars of this effective temperature. A similar maximum for the Li abundances is also found at about the
same temperature. For higher temperatures the values of log N(Be) decrease forming the
well-known Be gap for F stars, a feature that has a counterpart for Li. The maximum at
6100 K may be attributed to Galactic chemical evolution effects, since most of the stars
in the temperature interval between 6000 and 6200 K are particularly metal-rich, and
Galactic Be abundances are known to increase with the metallicity (Rebolo et al. 1988,
Boesgaard et al. 1999). However, high Be abundances are not observed in low temperature metal-rich stars. The fact that Be abundances decrease at low temperatures can be
seen as evidence that Be is burned during the main sequence evolution of these stars but
this low abundances are not predicted by models (Pinsonneault et al. 1990).
Be and Li abundances are shown in the right panel of Fig. 1. This figure tells us that
overall, and taking into account only dwarfs, there seems to be a correlation between the
depletions of Li and Be, in the sense that stars that have depleted their Be have also
430
431
Figure 1. Left panel: Beryllium abundances as a function of effective temperature for stars
with (open stars) and without (filled circles) known planets from Santos et al. (2002, 2004) and
G
alvez Ortiz et al. (2009, submitted). Abundances below log N(Be) 0.8 are not predicted by
models (Pinsonneault et al. 1990) at Tef f < 5500 K. Right panel: Beryllium abundances as a
function of lithium bundances for stars with (red symbols) and without (black symbols) known
planets from G
alvez Ortiz et al. (2009, submitted).
strongly depleted their Li. We can see that objects with temperatures above 6000 K are
situated in the upper right corner where there is a maximum of Li and Be values. In the
temperature regime between 5600 and 6000 K a large dispersion in the Li abundances is
observed while Be remains close to the primordial value. Stars with temperatures below
5600 K are mostly situated in the left upper-middle panel where both Li and Be are
clearly depleted. This temperature seems to mark the onset of strong Be depletion while
severe Li depletion begins at 5900 K.
The results presented for Be can be explained in terms of the Galactic chemical evolution. A comparison of the Be abundances of planet hosts and single stars has revealed
that, perhaps with a few exceptions, the two samples follow the same behaviour in the
log N(Be) vs. Tef f plot. Nevertheless, we have found a small group of stars that present
particularly high Be abundances that could signal pollution effects, although other explanations are possible. Measurements of 6 Li/7 Li ratio will allow to find signatures of
planetary accretion.
References
Boesgaard, A. M., Deliyannis, C. P., King, J. R. et al. 1999, AJ, 117, 1549
Fischer, D. A. & Valenti, J. 2005, ApJ, 622, 1102
G
alvez-Ortiz, M.C., Delgado Mena, E., Santos, N.C., Israelian, G. et al. 2009, A&A, submitted
Israelian, G., Santos, N. C., Mayor, M. & Rebolo, R. 2003, A&A, 405, 753
Israelian, G., Santos, N. C., Mayor, M. & Rebolo, R. 2004, A&A, 414, 601
Kurucz, R. L. 1993, CD-ROMs, ATLAS9 Stellar Atmospheres Programs (Cambridge: Smithsonian Astrophys. Obs.)
Pinsonneault, M. H., Kawaler, S. D. & Demarque, P. 1990, ApJS, 74, 501
Rebolo, R., Abia, C., Beckman, J. E. & Molaro, P. 1988, A&A, 193, 193
Santos, N. C., Garca Lpez, R. J., Israelian, G., Mayor, M., Rebolo, R., Garca-Gil, A., Prez de
Taoro, M. R. & Randich, S. 2002, A&A, 386, 1028
Santos, N. C., Israelian, G., Randich, S., Garca Lpez, R. J. & Rebolo, R. 2004, A&A, 425, 1013
Santos, N. C., Israelian, G., Garca Lpez, R. J., Mayor, M., Rebolo, R., Randich, S., Ecuvillon,
A. & Domnguez Cerdea, C. 2004, A&A, 427, 1085
Sousa, S. G., Santos, N. C., Mayor, M., Udry, S., Casagrande, L., Israelian, G., Pepe, F., Queloz,
D., Monteiro, M. J. P. F. G. 2008, A&A, 487, 373
Observat
orio Nacional, Rua General Jose Cristino, 77, 20921-400,
S
ao Crist
ov
ao, Rio de Janeiro, RJ, Brazil; email: luan@on.br
2
NOAO, 950 N Cherry Ave., Tucson, AZ 85719, USA
Abstract.
The study of chemical abundances in stars with planets is an important ingredient for the
models of formation and evolution of planetary systems. In order to determine accurate abundances, it is crucial to have a reliable set of atmospheric parameters. In this work, we describe
the homogeneous determination of effective temperatures, surface gravities and iron abundances
for a large sample of stars with planets as well as a control sample of stars without giant planets.
Our results indicate that the metallicity distribution of the stars with planets is more metal rich
by 0.13 dex than the control sample stars.
Keywords. stars: abundances, stars: atmospheres, (stars:) planetary systems: formation
1. Introduction
More than 400 stars with planets have been detected up to this date. One of the
interesting properties of these stars concerns their metallicity distribution. Several studies
(e.g., Fischer & Valenti 2005) have confirmed the result first shown by Gonzalez (1997):
stars with giant planets are systematically metal-rich (by 0.2 dex) relative to field FGK
dwarfs not known to harbor planets. Two hypotheses have been proposed to explain
this excess: primordial enrichment or pollution. Current results (e.g., Fischer & Valenti
2005) show that the frequency of planets increases significantly for higher metallicities,
thus giving strong support for the primordial hypotheses. Evidence for the occurence of
pollution is still ambiguous (for a comprehensive review, see Gonzalez 2006).
433
40
Stars with Planets
Stars without Planets
35
30
25
20
15
10
0
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.1
0.2
0.3
0.4
0.5
[Fe/H]
Figure 1. The metallicity distributions for a sample of 125 planet host stars (solid line) and a
control sample with 149 stars without giant planets.
set to eliminate the strong lines that are sensitive to the microturbulent velocity value,
decoupling the Tef f determination from the microturbulence determination. The microturbulent velocities were varied until the slopes of A(Fe I) versus log(EW/) were zero.
Finally, surface gravities were derived from ionization equilibrium between Fe I and Fe II
species. The abundances were derived in LTE using an updated version of the spectrum
synthesis code MOOG (Sneden 1973). The model atmospheres adopted in the analysis
were interpolated from the ODFNEW grid of ATLAS9 models (Castelli & Kurucz 2004).
Figure 1. (a) The (apocenter,pericenter) distribution of SWGP. The line is for (Ra+Rp)/2=8
kpc. (b) The histograms are for SWGP with (Ra+Rp)/2<8kpc (thin line) and (Ra+Rp)/2>8
kpc (thick line)
434
435
2. Two caveats
The planet-metallicity correlation fails at low metallicities (-0.7<[Fe/H]<-0.3 dex) and
when the host stars are giants. Giant stars hosts are known to have a metallicity not
skewed towards high metallicity. This last point is easy to interpret: since most giants are
young objects, the radial migration effect cannot be significant in their case. Migration
is a process which time scale is the gigayear. Hence, we dont expect young populations
to contain objects more metal-rich than the solar vicinity.
The second case where the correlation breaks down is at -0.7<[Fe/H]<-0.3 dex, where
stellar populations in the solar vicinity can be divided into two groups: the thin and the
thick disks, which differentiate both by their -elements content and their asymmetric
drift. It has been shown in Haywood (2008, A&A 482, 673) that in this metallicity
interval, giant planets are found preferentially on thick disk stars, in a proportion of
10 objects being either thick disk or transition objects (between the thin and the thick
disks) and 2 objects (one dwarf, HD 171028, and one giant, HD 170693) compatible with
belonging to the metal-poor thin disk. There is a hint here that distance to the galactic
center may play a role by noting that metal-poor thin disk stars have a probable origin
in the outer disk.
Figure 2. The planet-metallicity correlation is well reproduced assuming that stars originating
in the inner galactic disk have a higher percentage of giant planets, with no need to assume a
relation between the metallicity of stellar disks and the rate of giant planet formation.
1. Introduction
Chemical abundances in solar-type stars still represent an open issue, in many respects:
the photospheric abundance of elements with high first ionization potentials, like Argon
and Neon, is poorly known; coronal and photospheric abundances appear to be different
in many cases, with a possible dependence on unconventional parameters like the stellar
activity level; photospheres of planet-hosting stars are known to be metal-rich, but does
this peculiarity apply also to the chemical composition of the outer stellar atmosphere?
Figure 1. Left panel: Photospheric and coronal abundances of Boo for individual elements,
sorted by increasing FIP, relative to the solar photospheric composition (Anders & Grevesse
1989). Right panel: Comparison of coronal to photospheric abundance ratios of FG dwarfs with
different activity levels.
436
437
3. Conclusions
Tau Bootis is a coronal X-ray source with a level of magnetic activity intermediate
between that of the quiet Sun (Peres et al. 2000) and those of younger FG stars. In
Fig. 1 (right panel) we compare the coronal to photospheric abundance ratios of Boo
(method 2) with those of the Sun (Feldman & Laming 2000) and two other intermediateactivity G-type stars: Eri (Sanz-Forcada et al. 2004), another planet-hosting star, and
1 Cet (Telleschi et al. 2005) without any known planet.
The pattern of coronal to photospheric abundance ratios vs. FIP for Boo is almost flat
(no FIP bias, with possible exception for Ni), and suggests that its coronal characteristics
are similar to those of other intermediate-activity stars with planets, like Eri.
References
Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
Butler, R.P., Marcy, G.W., Williams, E., Hauser, H., Shirts, P. 1997, ApJ, 474, L115
Feldman, U. , Laming, J.M. 2000, Physica Scripta, 61, 222
Gonzalez, G., Laws, C. 2007, MNRAS, 378, 1141
Peres G., Orlando S., Reale F., Rosner R., Hudson H. 2000, ApJ, 528, 537
Sanz-Forcada, J., Brickhouse, N.S., Dupree, A.K. 2003, ApJS, 145, 147
Sanz-Forcada, J., Favata, F., Micela, G. 2004, A&A, 416, 281
Scelsi, L., Maggio, A., Peres, G., Gondoin, P. 2004, A&A, 413, 643
Telleschi, A., G
udel, M., Briggs, K., Audard, M., Ness, J.-U., Skinner, S. 2005, A&A, 622, 653
1. Introduction
Before describing the evolution of molecules at interstellar medium (ISM), we have to
know which reactions occur in such environment. Several molecules have been identified
in the ISM, leading to laboratory studies of possible reactions that end up with the
production of these molecules. Such molecules are continuously created and destroyed,
depending on the reactions that they are involved and their rates. The abundance of
some molecules can be connected with the abundance of many others. This study is
based on the previous work by Herbst & Klemperer (1973). Here, instead of dealing only
with the molecular abundances, we make an attempt to connect the evolution of the
abundance of many molecules with the chemical evolution of the Galaxy itself, by using
a self-consisted model. This way, we can see how the evolution of these molecules are
afected by the Galactic evolution.
2. Data
We are working with the OSU 2009 data base, that contains 6046 reactions and 468
species. To solve the differential equations that describe the evolution of abundances
of such molecules, we used the NAHOON code. Each one of these molecules has a
differential equation that describes the time evolution of its abundance. By varying the
initial chemical composition of the interstellar clouds, we can study the behaviour of these
http://www.physics.ohio-state.edu/ eric/
http://www.obs.u-bordeaux1.fr/amor/VWakelam/Valentine%20Wakelam/Downloads.html
438
Evolution of molecules
439
molecules with the change of some parameters, like metallicity and/or temperature, for
example.
3. Development
Once we have the database, it is necessary to choose the astrophysical parameters. We
use typical values for these parameters, such as 10 K for the temperature of the molecular
cloud, density of 2 105 H cm3 and ionization rate by cosmic rays of 1017 s1 . We do
not consider reactions occuring at the solid-phase, that is, we are working with a database
of gas-phase reactions only. The system of differential equations is built by the Nahoon
program, using the DLSODE code to solve it. Astrophysical equations are solved to find
the elemental abundances of many elements, all reffered to the abundance of H. The solar
abundances of these elements are taken from Asplund et al. (2005). The molecular cloud
chemical composition is chosen for six metallicities: [Fe/H]= 2.5, 2.0, 1.5, 1.0, 0.5
and 0.0. This way, we work with six different initial chemical compositions determined by
these metallicities. Initially, we consider the cloud is neutral, with all the carbons at CO
molecules and all the hidrogens at H2 molecules. The initial compositions were chosen
following the predictions of a model for the Galactic chemical evolution (Timmes et al.
1995).
4. Results
Important molecules for life are abundantly produced at molecular clouds at the gasphase, being H2 O, CO2 , H2 CO, NH3 , CH4 and HCN the most abundants. Complex
molecules, such as cyanopolyynes and many carbon molecules, are likely to be destroyed
more efficiently than produced, in the last 90% of the cloud lifetime, as the metallicity
increases. This is so because richer clouds are more oxidant, and the oxygen reacts efficiently with carbons to form other molecules, as CO2 , trapping C molecules that are
needed to form more complex molecules. Most of the molecules tend to be destroyed
after 10% of the molecular cloud lifetime nearly 105 years , regardless of the metallicity, and the rate of destruction is, on average, smaller for lower metalicities. Simple
oxygen molecules tend to be more abundant at molecular clouds having solar metallicity, contrary to what can be seen for complex oxygen molecules. Nearly 80% of the
studied molecules reach the peak of their abundances before 105 years, suggesting that
the molecular cloud collapses before the chemical balance is reached. Comparison with
other works (e.g., Shalabiea 2001) show that our model produces similar results for the
molecular abundances at the gas phase when we consider a molecular cloud having solar
metallicity. However, these estimates are not yet consistent with observations for many
molecules studied, showing that only gas-phase reactions underestimate the abundance
of several molecules.
References
Asplund, M., Grevesse, N., & Sauval, A. 2005 ASP Conference Series, v. 336
Herbst, E., & Klemperer, W. 1973, ApJ, 185, 505
Shalabiea, O. M. 2001, Astronomy and Astrophysics, 370, 1044
Timmes, F. X., Woosley, S. E., & Weaver, T. Q. 1995, ApJ Supplement Series, 98, 617
Abstract. We present experimental studies on the interaction of soft X-rays on gas-phase and
solid-phase amino acids and nucleobases in an attempt to verify if these molecules (supposed to
be formed in molecular clouds/protostellar clouds) can survive long enough to be observed or
even to be found in meteorites. Measurements have been undertaken employing 150 eV photons
under high vacuum conditions at the Brazilian Synchrotron Light Laboratory (LNLS). The produced ions from the gas-phase experiments (glycine, adenine and uracil) have been mass/charge
analyzed by time-of-flight spectrometer. The analysis of solid phase samples (glycine, DL-proline,
DL-valine, adenine and uracil) were performed by a Fourier transform infrared spectrometer
coupled to the experimental chamber. Photodissociation cross sections and halflives were determined and extrapolated to astrophysical environments. The nucleobases photostability was up
to two orders of magnitude higher than for the amino acids.
Keywords. astrochemistry, astrobiology, molecular processes, molecular data, methods: laboratory, ISM: clouds, X-rays: ISM
1. Introduction
The search for amino acids, nucleobases and related compounds in the interstellar
medium/comets has been performed at least in the last 30 years. Recently, some traces
(upper limits) of these molecules (e.g. glycine and pyrimidine) have been detected in
molecular clouds (MC), protoplanetary disks (PPD) and in comets (Kuan et al. (2003);
Elsila et al. (2009)). The search for these biomolecules in meteorites, on the contrary,
has been revealed an amazing number, up to several parts per million (Kissel & Krueger
(1987)). This chemical dichotomy between the molecular inventories found in meteorites
and interstellar medium and comets remains a big puzzle in astrochemistry field and in
the investigation about the origin of life.
441
used the Brazilian Synchrotron Light Laboratory (LNLS) facility, located at Campinas,
Brazil. The measurements have been undertaken employing 150 eV photons ( 4 10 11
photons cm2 s1 ) from the toroidal grating monocromator (TGM) beamline under high
vacuum conditions. For the gas-phase experiments (glycine, adenine and uracil) the soft
X-ray beam had intercepted perpendicularly the molecular beam. The produced ions were
mass/charge analyzed by Time-of-Flight mass spectrometry (TOF-MS) employed in a
photoelectron-photoion coincidence mode (Figure 1a). In the case of solid phase samples
(glycine, DL-proline, DL-valine, adenine and uracil), in-situ analysis were performed by
a Fourier transform Infrared spectrometry (FTIR) coupled to the experimental chamber
(Figure 1b). Further experimental information is given elsewhere (Pilling et al. (2009))
The results show that the gaseous amino acids are virtually destroyed by soft Xrays while the nucleobases have a survivability of a few percents. The dissociation cross
section of glycine in the gas phase is about 10 times lower than in the solid-phase. In the
case of nucleobases both solid and gas-phase present similar dissociation cross section.
Condensed amino acids can survive at least 7 105 and 7 108 years into dense
MC and PPD, respectively.
Figure 1. a) Mass spectra of the ionic species produced by the interaction of 150 eV photons on
two gaseous samples. b) Infrared spectra of two solid-phase samples before and after irradiation
by 150 eV photons. Details can be found elsewhere (Pilling et al. (2009))
3. Conclusions
Experimental study on the interaction of 150 eV photons on gaseous and solid-phase
biomolecules expected to be found interstellar dense clouds have been performed. The
nucleobases photostability is up to two orders of magnitude higher than for the amino
acids. The determined lifetime for these molecules is of the order or even grater than
MC and PPD lifetime. This result corroborates the scenario in which during planetary
formation (and after) these molecules, trapped into and onto dust grains, meteoroids and
comets, could be delivered into the planets/moons possibly allowing pre-biotic chemistry
in such environments where water was also found in liquid state.
References
Elsila, J., et al. 2009 Met. & Planetary Science, In press
Kissel, J. & Krueger, F. R. 1987 Nature, 326, 755
Kuan, Y.-J., Charnley, S. B., Huang, H.-C., Tseng, W.-L., & Kisiel, Z. 2003, ApJ, 593, 848
Pilling, S., Andrade, D.P.P., Marinho R.T., et al. 2009 A&A, to be submited
1. Introduction
The observation of molecules in the gas phase deeply inside these cold and dense
regions, where the gas sticking efficiency on grains is close to unity, suggests that they
are indeed energetically active regions (e.g. (Ehrenfreund & Charnley (2000)). In these
highly obscured regions, cosmic rays are the main energy source for chemical reactions.
Although the flux of heavy ions (e.g. Fe, Ni, Si, Mg, ...) is about 3-4 orders of magnitude
lower than that of protons (Roberts et al. (2007)), their effects play an important role
on interstellar grains since they can deposit about 100 times more energy than the light
ions (protons and alpha particles) inside the grains. The amount of species released per
impact to gas phase due to heavy ions could be 4-5 orders of magnitude higher than
those for protons (Seperuelo Duarte et al. (2009)).
443
Accelerateur National dIons Lourds) in Caen, France. In-situ Fourier transformed infrared (FTIR) spectra were recorded at different fluences, up to 21013 Ni ions cm2 . Further experimental information is given elsewhere(Seperuelo Duarte et al. (2009); Pilling
et al. (2009)).
The dissociation cross sections: H2O 2 1013 , CO = 1.4 1013 and N H3 =
1.9 1013 cm2 were determined from the evolution of the molecular column density
as a function of ion fluence. Figure 1a presents a comparison between the IR spectra
of interstellar ices around the embedded protostar W33A obtained by Infrared Space
Observatory (ISO) with the irradiated water-ammonia-CO ice at a fluence 2 10 13 Ni
ions cm2 . The bands observed from the radiolysis reproduces very well the OCN (2165
cm1 ) and CO (2139 cm1 ) IR features observed in the W33A spectrum. Other spectral
features like the broad IR peaks observed at 1650 and 1450 cm1 are also similar to the
astronomical source. The IR analysis of organic residue at 300 K suggests the presence
of containing CN bearing molecules, amides, esters and possibly zwitterionic glycine
(NH+
3 CH2 COO ) (Figure 1b). The complete description of the results can be found
elsewhere Pilling et al. (2009).
Figure 1. a) Comparison between the IR spectra of protostar W33A obtained by Infrared Space
Observatory (ISO) with the irradiated H2 O:NH3 :CO (1:0.6:0.4) 13 K ice at a fluence 2 1013 Ni
ions cm2 . (b) Comparison between the irradiated ice and the 300 K residue. Vertical dashed
lines indicate the frequencies of some vibration modes of zwitterionic glycine (NH+
3 CH2 COO ).
Details can be found elsewhere (Pilling et al. (2009))
3. Conclusions
Experimental study on the interaction of energetic Ni ions on ammonia-containing
interstellar ice analog have been performed to simulate the physical chemistry induced
by cosmic rays inside dense regions of interstellar medium. Dissociation cross section
have been determined allowing an estimative for half life of the studied species of about
2-3 106 years in real astrophysical real situation. This value is in a good agreement
with the half lives of typical dense molecular environments in the interstellar medium.
Although they represent only a small fraction ( 1%) of the cosmic rays flux, some effects
(e.g. molecular sputtering and ice compaction) promoted by heavy ions on interstellar
ice grains are more intense than those promoted by protons.
References
Ehrenfreund, P. & Charnley, S.B. 2000, ARAA, 38, 427
Roberts, J.F., Rawlings, J.M.C., Viti, S. & Williams, D. A. 2007, MNRAS, 382, 733
Seperuelo Duarte E., Boduch P., Rothard H., Been T., Dartois E., et al. 2009 A&A, 502, 599
Pilling, S., Seperuelo Duarte E., da Silveira, E.F., Balanzat E., et al. 2009 A&A, Aceppted
Session VI
Abstract. I review the instrumentation plans of the future ELT projects, with particular emphasys on one category of instruments which will be particularly relevant for this Symposium
community: high resolution spectrographs.
Keywords. Instrumentation, high-resolution spectroscopy
1. Introduction
After the great excitement produced by the wealth of high quality data from the 8m
class telescopes, an even more exciting perspective is opening for ground based astronomy:
the era of the Extremely Large Telescopes (ELTs). There are currently three projects
which a funded study phase: the 42m diameter European ELT (E-ELT, Gilmozzi 2008) ,
the 30m diameter Thirty Meter Telescope (TMT, Nelson 2008), and the 21-25m diameter
Giant Magellan Telescope (GMT, Johns 2008). The step between the current class of
telescope and the ELTs is of more than a factor 3 in diameter, larger than what we have
experimented in the last 15 yrs when passing from the 4m telescopes to the 8-10 meters
ones. This contribution is biased in two ways: first it concentrates mostly on the E-ELT
instrumentation plan, which is the one the author is more familiar with; second it is
biased towards high resolution spectroscopy, because this is the most used technique for
deriving detailed chemical abundances. I will show, however, that the three ELTs have
quite similar instrumentation plans, and therefore most conclusions and cases valid for
the E-ELT can be applied the other ELT telescopes as well.
2. Why ELTs
The first advantage of having an ELT is obvious: obtain an enormous photon collecting
area. What is perhaps less obvious is that this enormous power can be used not only to
reach fainter objects, but also to increase the precision in our spectroscopy.
The second advantage is that if an ELT can work (close to) at the diffraction limit, than
we have a tremendous increase in angular resolution; this is capital for several reasons.
This implies that we can obtain the highest gain for faint magnitudes because for these
objects sky is the main source of noise, and less contribution from sky is recorded. In
addition it will be possible to beat confusion, to obtain very high contrast and, for those
like me struggling with instrumentation, making the IR spectrographs feasible. In fact,
given that R tg d ins/D tel), (where is the aperture in the sky, R is the
resolving power, and dins and Dtel the instrument and telescope pupils respectively), to
obtain for a 1 arcsec seeing Resolving power R=100000 at a 40 m telescope (and a very
steep tg = 4 echelle grating) requires an instrument pupil of 4m!
It is therefore clear that Adaptive Optics is absolutely essential for ELTs, and indeed a
full zoo of AO types is being developed, also in prevision of its use at the ELTs: GLAO,
LTAO , MOAO , MCAO, SCAO, XAO are all acronyms to which we will soon be used.
447
448
Luca Pasquini
Figure 1. From Evans et al. 2008: EAGLE simulation of the MC cluster NGC346 posed at a
distance of 1.3 Mpc as observed by an ELT. The image is 1.6x1.6 arcsec. The first image the
correction is made with Ground Layer AO (GLAO), Multi Object AO (MOAO) , and Laser
Tomography AO (LTAO)
Figure 1 ( Evans et al. 2008), shows a simulated image of the SMC cluster NGC346
projected at 1.3 Mpc, in a 1.6x1.6 arcsec field, as seen with GLAO, MOAO and LTAO
corrections. This figure is very instructive, showing clearly as in the study of stellar
population in external galaxies the power of AO is needed, to beat the confusion at faint
magnitudes.
3. Which Instruments
The first question in analyzing the instrumentation plans for ELTs, it is to first establish
how many will be typically hosted. TMT and GMT have Nasmyth platforms which
plan to host several instruments/each; the E-ELT will in addition be equipped with two
gravity invariant platforms below the Nasmyth ones, and a coude focus. Each ELT
plans at present to host up to 8-9 instruments simultaneously. Gedanken Experiments
with realistic dimensions have been made, and showing that it shall be feasible to host
3-4 instruments on the Nasmyth platforms simultaneously.
Figure 2 shows the suite of instruments presently studied for the E-ELT, together with
a very brief description and the names of the PIs. As for the VLT most instruments are
studied within the ESO community by consortia of institutes.
Figure 3 shows the parameter space covered by the studied instrumentation, in the
Field of View vs. wavelength and in the Resolving power vs. wavelength diagrams. The
parameter spaces are very well covered; from the diffraction limited instruments (with
a typically limited FoV) to the multi-object instruments using the whole 10 arcmin
FoV; in resolving power, from the imagers to the two optical and IR high resolutions
spectrographs, all the space is covered.
As anticipated above, both TMT and GMT have very similar instrumentation plans,
with 9 and 7 instruments respectively. Not all the instrument studied will necessarily be
built, and surely not all will be selected for first light. TMT has already chosen their
initial complement of 3 instruments (IRIS, IRMS and WFOS), while for the E-ELT and
GMT the selection process is expected to be finished within 2010.
ELT Instrumentation
449
Figure 2. List of instruments (with their main characteristics) under phase A study for the
E-ELT project. Similar complements are under study for GMT and TMT
Figure 3. Space covered by the ELT studied instruments: Field of View (left) and Resolving
Power (right) vs wavelength. (Courtesy of M. Kissler-Patig)
The E-ELT spectrographs able to provide some substantial high resolution capability
are three: CODEX (Pasquini et al. 2005), SIMPLE(Origlia & Oliva 2009) and OPTIMOS
(OPTIMOS web page ).
CODEX is the optical (range 380-710 nm), high resolution (R=120000), high stabil-
450
Luca Pasquini
ity spectrograph. It has an aperture of 0.8 arsceconds in the sky and aims at the highest
Doppler precision, 2 cms1 over a period of many years.
SIMPLE is the AO assisted (aperture on the sky of 0.04 arcseconds) IR high resolution (R up to 100000) spectrograph. One of its peculiar characteristics is the very large
spectral range (0.9-2.5 m ) covered simultaneously in one exposure.
OPTIMOS/EVE is a Multi-Object fibre optical and IR spectrograph. It has a resolving power from 5000 up to 50000 with a coverage of about /32 and a central wavelength
tunable between 370 and 1700 nm. Thanks to image slicers an aperture of 0.9 arcseconds
on the sky is used.
4.1. IGM and Stars: The faint end
One of the most interesting aspects of this Symposium has been to link chemical evolution
from high redshift to planets. Several speakers have shown how it will be fundamental to
push the magnitude limits of objects observed to determine the abundance of intervening
IGM. How does the metallicity evolve with redshift? How homogeneous the IGM is ? We
need to observed many more high redshift objects to determine the IGM Zn abundance
and OPTIMOS will allow to reach QSOs and galaxies as faint as V=25. Another quest
is to expand the range of redshifts covered: our present knowledge stops at Z4 because
the Zn lines fall in the IR for higher redshift. With SIMPLE it will be possible to extend
these studies at much higher redshifts. With a limiting magnitude around J,H and K
19, many QSOs will become available for these studies.
Figure 4. Simulations of Li observations in a bulge main sequence solar star (V=21.5) with
10 hour of OPTIMOS/EVE at the E-ELT. The spectrum has a quality lower, but comparable
with what obtained with VLT+UVES on 17 magnitude stars in the Globular Cluster 47 Tuc
(Bonifacio et al. 2007)
For stars as well the E-ELT will represent a tremendous breakthrough, allowing, for
instance, to obtain high resolution spectroscopy of low mass stars in the whole Galaxy
and in nearby satellites. The understanding of the origin and primordial value of Li,
requires, for instance, to study this element in other galaxies and in other populations of
our own Galaxy to test whether its abundance in metal poor stars is the same of the Spite
plateau( Spite & Spite 1982) irrespective of the birth place. Figure 4 shows a very exciting
ELT Instrumentation
451
simulation of a Li spectrum of a late type main sequence star (V=21.5) in the bulge as
observed by OPTIMOS-EVE at the E-ELT in 10 hours. The spectrum is compared with
an observed UVES spectrum of a star in the Globular Cluster 47 Tuc (Bonifacio et
al. 2007) and the quality is quite comparable. Observing tens of stars in the bulge it
will be possible therefore to derive a detailed history of the evolution of Li, search for
correlations with other elements are present in this population. Of course similar studies
will be possible for a number of chemical elements. Similarly, it will be possible to obtain
R=100000 resolution (and S/N ratio 50) spectra for stars up to magnitudes J,H,K=17
with SIMPLE.
4.2. IGM and Stars: The high S/N ratio
The huge ELT collecting area allows un-precedent high S/N ratio and high precision
spectroscopy. By acquiring for a sample of objects S/N ratio observations as high s
S/N=2000 with CODEX, it will be possible to study IGM at very low column densities,
as low as the median IGM density (NHI13.5). Figure 4 presents the CODEX simulations
for such a case, showing as tiny features will be measurable; an impossible task even for
excellent instruments such as UVES at the VLT. By measuring these lines , for instance
in QSO pairs or groups it will be possible to distinguish between competing mechanism
of IGM enrichment so far proposed.
Figure 5. Simulations of IGM CIV lines at different HI column densities. Only with the CODEX
resolving power and S/N ratio it will be possible to measure metallicity at low IGM density.
The last figure below shows the UVES spectrum of CIV corresponding to an NHI density more
than 10 times higher than what will be measurable with CODEX. (Courtesy of V. DOdorico)
For stars one interesting proposal is to use CODEX to determine the age of different
452
Luca Pasquini
5. Concluding Remarks
In the ELT era smaller telescopes will still be needed. ELTs will have some limitations
and will not be able to perform all kind of observations. FoV is, for example, limited, and
in addition the implementation of instruments able to use the full FoV is usually quite
complex. Another example might be UV science: it might not be possible to procure
mirror coatings which are extremely efficient both in the IR and at the atmospheric cut
off. A second consideration concerns accessibility: there are about a dozen of 6-10 m class
telescopes available at present for the community, and they are very often heavily over
subscribed. The ELTs will be at most three. We may expect that it will be extremely
difficult to obtain large amount of observing time for single programs.
Provided funds are secured, all the three ELT projects are foreseen to start operations
around year 2018. We have a very exciting era in front of us!
I would like to thank S. DOdorico (E-ELT Instrument Project Office), P. Bonifacio
(Optimos-EVE), J-G. Cuby (Eagle), P. Hickson (TMT), E. Oliva (Simple), S. Schectman
(GMT) for sharing information and figures about their projects.
References
Bonifacio, P., Pasquini, L., et al. 2007, A&A , 470, 153
Evans, C.J., Lehnert, M.D. 2008, Proc. SPIE, 7014, 201
Gilmozzi, R. 2008, Proc. SPIE 6986. 698604
Johns, N. 2008, Proc. SPIE 6986. 698603
Kasper, M. et al. 2008, Proc. SPIE 7015. 63
Nelson, J. 2008, Proc. SPIE 6986. 698602
OPTIMOS web page , http://gepi.obspm.fr/les-projets/instrumentation-e-elt/article/optimoseve?lang=en
Origlia, L., Oliva, E. 2005, Earth, Moon and Planets, 105, 123
Pasquini, L. et al. 2005, The Messenger, 122, 10
Spite, F., Spite M. 1982, A&A 115, 357
Abstract.
Although originally conceived as primarily an extragalactic survey, the Sloan Digital Sky
Survey (SDSS-I), and its extensions SDSS-II and SDSS-III, continue to have a major impact
on our understanding of the formation and evolution of our host galaxy, the Milky Way. The
sub-survey SEGUE: Sloan Extension for Galactic Exploration and Understanding, excuted as
part of SDSS-II, obtained some 3500 square degrees of additional ugriz imaging, mostly at lower
Galactic latitudes, in order to better sample the disk systems of the Galaxy. Most importantly,
it obtained over 240,000 medium-resolution spectra for stars selected to sample Galactocentric
distances from 0.5 to 100 kpc. In combination with stellar targets from SDSS-I, and the recently completed SEGUE-2 program, executed as part of SDSS-III, the total sample of SDSS
spectroscopy for Galactic stars comprises some 500,000 objects.
The development of the SEGUE Stellar Parameter Pipeline has enabled the determination
of accurate atmospheric parameter estimates for a large fraction of these stars. Many of the
stars in this data set within 5 kpc of the Sun have sufficiently well-measured proper motions to
determine their full space motions, permitting examination of the nature of much more distant
populations represented by members that are presently passing through the solar neighborhood.
Ongoing analyses of these data are being used to draw a much clearer picture of the nature of
our galaxy, and to supply targets for detailed high-resolution spectrscopic follow-up with the
worlds largest telescopes. Here we discuss a few highlights of recently completed and ongoing
investigations with these data.
Keywords. astronomical data bases: surveys, Galaxy: halo, structure, methods: data analysis,
stars: abundances
1. Introduction
The era of the massive surveys of Galactic stars is now very much underway. Our
understanding of the history of the formation and evolution of the stellar populations
in the Galaxy is presently being revolutionized, based on results from two primary surveys, the Sloan Digital Sky Survey (SDSS; York et al. 2000), in particular its dedicated
sub-surveys Sloan Extension for Galactic Exploration and Understanding (SEGUE-1 and
SEGUE-2; Yanny et al. 2009) and the RAdial Velocity Experiment (RAVE; Zwitter et al.
2008; Wyse, this volume). Collectively, these two surveys are in the process of providing
detailed spectroscopic information for a million or more stars, sampling all of the known
stellar populations in the Galaxy. Although refinements in the procedures for extracting
estimates of the stellar atmospheric parameters (Teff , log g, [Fe/H]) and individual element abundances (in the case of SDSS/SEGUE, [/Fe] and [C/Fe] ratios; for RAVE,
many more) are still underway, the wealth of information already available provides the
453
454
Timothy C. Beers
Figure 1. Simple visualization of the SDSS/SEGUE stellar database, for roughly 400,000 stars
with available atmospheric parameter estimates from the SSPP. The sphere has a radius of 10
kpc, centered on the Sun. The colors scale with metallicity, from green (metal-poor stars) to red
(metal-rich stars).
basis for numerous investigations and follow-up observations. Here we summarize a few
of the ongoing projects that are being enabled by the SDSS/SEGUE database.
455
Based on the derived atmospheric parameters from the SSPP, and distances estimated
from photometric parallaxes, Figure 1 shows a simple visualization of the SDSS/SEGUE
database. The colors are coded to represent metallicity, and clearly indicate the presence of stars from the disk populations close to the Galactic plane, and an extended
population of stars comprising representatives of the inner- and outer-halo populations.
More sophisticated visualizations, encoding other available observables such as full space
motions, are presently being developed.
1.0
2.0
3.0
4.0
5.0
6.0
15,000
3,000
400
5
2
0
150,000+
30,000+
1,000+
5
2
0
456
Timothy C. Beers
Figure 2. Observed metallicity distribution functions (MDFs) for the full sample of
SDSS/SEGUE DR-7 calibration stars as a function of vertical distance from the Galactic plane.
The black histograms represent the MDFs obtained at different cuts of |Z|, while the red arrows
denote the locations of the metallicity peaks of the MDF for the thick disk (0.6), the MWTD
( 1.3), the inner halo (1.6), and the outer halo (2.2), respectively.
Galaxy. The metallicity distribution function (MDF) of the inner halo peaks at [Fe/H]
= 1.6, with tails exending to higher and lower metallicities. By comparison, the outer
halo comprises stars that exhibit a more spherical spatial density distribution with an
axial ratio 0.9. Outer-halo stars possess a wide range of eccentricities, exhibit a clear
retrograde net rotation, and are drawn from an MDF that peaks at [Fe/H] = 2.2, a
factor of four lower than that of the inner-halo population.
This work has now been extended by Carollo et al. (2009) (see also Carollos article,
this volume), making use of the calibration star sample through DR-7. The full sample
of 32360 stars represents an increase of 60% relative to the numbers used in the previous
analysis. The larger sample, which includes many more stars from the disk populations
due to SEGUE observations, has been used to derive, among many other results, estimates
of the velocity ellipsoids for the inner- and outer-halo populations, as well as for the
metal-weak thick disk (MWTD) and canonical thick-disk populations.
One can obtain an impression of the contribution of the various populations to the
observed MDF from inspection of Figure 2, which shows its variation with height above
the Galactic plane. Examination of the left-hand column of panels in this figure shows
how the MDF changes from the upper panel, in which there are obvious contributions
from the thick-disk, MWTD, and inner-halo components in the cuts close to plane, to
the lower panel, with an MDF dominated by inner-halo stars. In the right-hand column
of panels, with distances from the plane greater than 5 kpc, the transition from innerhalo dominance to a much greater contribution from outer-halo stars is obvious. This
demonstration is, by design, independent of any errors that might arise from derivation
of the kinematic parameters, and provides confirmation of the difference in the chemical
457
properties of the inner- and outer-halo populations originally suggested by Carollo et al.
(2007).
Although it might appear possible to attempt mixture-model analysis to obtain MDFs
for each of the individual components, we recall that biases in the selection of the stars
in the calibration-star sample would confound such an attempt. Other samples of SDSS
stars are being studied for this purpose, and will be reported on in due course.
458
Timothy C. Beers
log10 N
Rmean (kpc)
0.5
1.0
1.5
2.0
2.5
10
[/Fe]
0.4
0.3
0.2
0.1
0.0
V (km s-1)
0.5
150
200
Eccentricity
250
[/Fe]
0.4
0.3
0.2
0.1
0.0
-1.0
-0.5
[Fe/H]
0.0
-1.0
-0.5
[Fe/H]
0.0
Figure 3. Distribution of number densities, mean orbital radii, rotation velocities, and orbital
eccentricities for F- and G-type dwarfs within 2 kpc of the Galactic plane, located between
8 < R < 9 kpc, in bin sizes of 0.05 dex in [Fe/H] by 0.025 dex in [/Fe]. Rmean is an average of
Rmax and Rmin , the maximum and minimum projected distance on the plane reached by a star
over the course of its orbit. Each pixel has more than 20 stars.
459
1.5
1.5
1.0
1.0
0.5
0.5
0.0
3800
4000
4200
4400
4600
0.0
3800
4000
[C/H]= -0.75
4200
4400
4600
4310
4320
4400
4600
[C/H]= -0.59
1.4
2.0
1.2
1.5
1.0
1.0
0.8
0.5
0.6
0.4
4280
4290
4300
4310
4320
0.0
4280
4290
1.0
1.0
0.5
0.5
0.0
3800
4000
4200
4300
1.5
4400
4600
0.0
3800
4000
4200
Figure 4. Two examples of the fitting procedure for determination of [C/Fe] in metal-poor stars.
The upper panels show a portion of the original spectrum (black) overlayed with a synthetic
spectrum (red) with the listed parameters and [C/Fe] = 0. The middle panels show the region
around the CH G-band, with a red line showing the best fit. The green line is a division of the
original spectrum by the fit spectrum, which should be close to 1.0 for a successful fit. The lower
panels show the result of the best fit with the listed [C/H] levelLeft: A moderately metal-poor
star. Right: A very metal-poor star.
that they can be usefully separated into metallicity sub-classes, which they refer to as
subdwarfs (sdM), extreme subdwarfs (esdM), and ultra subdwarfs (usdM), in order of decreasing metal content. Such stars were specifically targeted in SEGUE-2. As
a result of this follow-up, there are now more than 6000 such stars known, an order of
magnitude increase in the numbers that were known just a few years ago.
Because sdM, esdM, and usdM stars are selected in a completely different way than
other probes in SDSS/SEGUE, they provide a useful comparison to kinematic analyses
of, e.g., the calibration stars studied by Carollo et al. (2007) and Carollo et al. (2009).
It is thus of great interest to note, as shown in Figure 5, that the kinematics of this very
local sample reflects the dichotomy of the halo that has been previously claimed. As can
be appreciated from inspection of this figure, the sdM stars (the centroid indicated by the
smaller circle) have local kinematics that can be associated with the thick disk, the esdM
stars (the centroid indicated by the large circle) with the inner-halo population, and of
greatest interest, the tendency of the usdM stars to exhibit a net retrograde rotation,
which can be associated with the outer-halo population.
Funding for SDSS-III has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, and the U.S. Department of
Energy. The SDSS-III web site is http://www.sdss3.org/.
SDSS-III is managed by the Astrophysical Research Consortium for the Participating
Institutions of the SDSS-III Collaboration including the University of Arizona, the Brazilian Participation Group, University of Cambridge, University of Florida, the French Participation Group, the German Participation Group, the Michigan State/Notre Dame/JINA
Participation Group, Johns Hopkins University, Lawrence Berkeley National Laboratory,
Max Planck Institute for Astrophysics, New Mexico State University, New York University, the Ohio State University, University of Portsmouth, Princeton University, University of Tokyo, the University of Utah, Vanderbilt University, University of Virginia,
University of Washington and Yale University.
460
Timothy C. Beers
Figure 5. Left: Distribution of W vs. V velocities for the sample of sdM stars (small dots),
esdM stars (filled triangles), and usdM stars (open squares) from SDSS/SEGUE. The smaller
circle indicates the centroid for the sdM stars, while the larger circle is that for the esdM stars.
Stars with a velocity component V > 220 km s1 are on prograde orbits, while those with
V > 220 km s1 are on retrograde orbits. The tendency for the usdM stars to occupy orbits
with large retrograde rotations is clear. Right: Distribution of W vs. U velocities for esdM stars
(filled triangles) and usdM stars (open circles) stars. Note the higher energy orbits associated
with the usdM stars, suggesting membership in the outer-halo population.
References
Abadi, M. G., et al. 2003, ApJ, 597, 21
Abazajian, K.N., et al. 2009, ApJS, 182, 543
Adelman-McCarthy, J.K. et al. 2007, ApJS, 172, 634
Allende Prieto, C. et al. 2008, AJ, 136, 2070
An, D., et al. 2009a, ApJ, 700, 523
An, D., et al. 2009b, ApJL, submitted (ArXiv:0907.1082)
Beers, T.C. et al. 1985, AJ, 90, 2089
Beers, T.C. et al. 1992, AJ, 103, 1987
Brook, C. et al. 2007, ApJ, 658, 60
Carollo, D., et al. 2007, Nature, 450, 1020
Carollo, D., et al. 2009, ApJ, submitted (ArXiv:0903.3019)
Christlieb, N., et al. 2008, A&A, 484, 721
et al. 2008a, ApJ, 684, 287
Ivezic, Z.,
et al. 2008b, ArXiv:0805.2366
Ivezic, Z.,
Johnston, K.V., et al. 2008, ApJ, 689, 936
Kazantzidis, S. et al. 2008, ApJ, 688, 254
Lee, Y. S., et al. 2008a, AJ, 136, 2022
Lee, Y. S., et al. 2008b, AJ, 136, 2050
Lepine, S. 2009, AIP Conf. Proc., 1094, p. 545
Newberg, H.J., et al. 2007, ApJ, 668, 221
Schlaufman, K.C., et al. 2009, ApJ, 703, 2177
Roskar, R., et al. 2008, ApJ, 684, L79
Sch
onrich, R., & Binney, J. 2009a, MNRAS, 396, 203
Sch
onrich, R., & Binney, J. 2009b, MNRAS, in press (ArXiv:0907.1899)
Tumlinson, J. 2007, ApJ, 664, L63
York, D. G., et al. 2000, AJ, 120, 1579
Zwitter, T., et al. 2008,AJ, 136, 421
1. Near-Field Cosmology
Analysis of the properties of old stars in our own Milky Way galaxy, and in nearby
galaxies, can be used to infer physical conditions in the early Universe, when those
stars formed. Many of the properties of stars are approximately conserved over long
timescales, even in the hierarchical merging scenario for galaxy formation. We thus can
trace the evolution of any one individual galaxy through time, from the fossil record
imprinted in the motions and chemical abundances of its stars for example, the Milky
Way, with current capabilities, and even beyond the Local Group with planned ELTs.
This complements beautifully the one-age snapshots of many galaxies that are available
from direct studies of high-redshift objects. Further, when one observes high-redshift
systems, one sees the total sum of all their stars together, and the interpretation of
this integrated light is subject to several degeneracies (e.g. between stellar age and
metallicity, or between star-formation rate and stellar Initial Mass Function) that limit
the robustness of conclusions that may be drawn. In contrast, when individual stars
are observed, these degeneracies can be broken. Modeling of stellar motions further can
map dark matter distributions in three-dimensions, critical information for discriminating
different dark matter candidates.
A major prediction of CDM cosmology is that much of the stellar populations of
present-day large galaxies, like the Milky Way, formed in systems of much lower mass. The
bulge and stellar halo are expected to be built-up during mergers, with predominantly
dissipationless mergers dominating for the halo, and both dissipational and dissipationless
mergers contributing to the bulge. Stellar disks can be thickened by mergers, and satellites
on the right orbits can even contribute stars (and dark matter!) directly to both the thick
and thin disks.
Spectroscopy provides much of the necessary astrophysics to test these, and other,
461
462
models, and, ideally, together with astrometric imaging surveys, allows the acquisition
and analysis of full three-dimensional space motions, elemental abundances and spatial
distributions of large samples of stars. Spectra provide the line-of-sight velocities, and
the stellar parameters effective temperature, gravity and chemical abundances with
an accuracy that depends on wavelength region, signal-to-noise and spectral resolution.
The panoramic imaging surveys being planned and implemented including Gaia need
matching spectroscopic surveys to fulfil their full potential in deciphering the nearby
Universe. The physics of galaxy formation and evolution that can then be addressed
includes deciphering the spatially resolved star-formation histories of galaxies, the massassembly histories which may be quite different from the star-formation histories, and
their links to growth of a central black hole. Further, the role and form of feedback
that is necessary in CDM theories to regulate star formation must be consistent with the
fossil record in the stars that form. The nature of dark matter is constrained not only
by the halo potential well shape, but also by the merger histories of galaxies, the effects
of which again are written in the properties of the stars both within and outwith
galaxies.
463
3. Applications
3.1. The smooth stellar halo
Photometric techniques can obviously be applied to fainter targets than can spectroscopic techniques, and prior to multi-object spectrographs, had a significant multiplex
advantage also. In terms of chemical abundances, the line-blanketing in the U-band offers the most powerful approach, utilising a comparison of (U B) in a given star to
that of a star of the same (B V ) in a cluster of known metallicity (e.g. the Hyades).
Sandage (1969) developed the calibration of (U B)0.6 , using a normalisation to a
fixed value of (B V ) = 0.6 to take out the temperature sensitivity. This is valid only
for a limited range of effective temperature, essentially F/G main sequence stars, and
it saturates at low metallicities, 1.5 dex (e.g. Carney 1979, his Fig. 3), understood
in terms of the opacity from metals becoming comparable to that from helium, so that
reducing the metallicity results in little further reduction in overall opacity. Within the
range of validity, it provides metallicities accurate to 0.2 dex, for 1% photometry. The
advent of wide-field, deep photometry from the Sloan Digital Sky Survey has opened the
opportunity to develop and apply such techniques using the photometry in the SDSS
filters. The result of such an analysis, based on ugr photometry, calibrated using the
SDSS stellar parameters from the spectroscopic pipeline, is vividly illustrated in Ivezic et
al. (2008): the halo, out to 10 kpc from the Sun, is remarkably uniform in metallicity,
with no gradients or scatter about the mean value of [Fe/H] = 1.46 dex their quoted
deviation about the median of 1.5 dex is less than 0.05 dex. Even more distant largescale overdensities, such as the Virgo overdensity at 1020 kpc, have the same mean
iron abundance as the rest of the field halo, based on photometric metallicities from gri
photometry (An et al. 09). This is somewhat unexpected, implying very well-mixed gas
across the star-forming regions that created the halo. The simulations of Johnston et
al. (2008) suggest that inhomogeneities should have been observed. Spectroscopic confirmation of the spatially invariant Gaussian metallicity distribution of the (inner?) halo
would provide rather stringent constraints on theories.
Spectroscopy also provides line-of-sight velocities, allowing the identification of kinematically defined subsystems in addition to metallicity distributions. The spectroscopic
component of the initial SDSS galaxy survey utilised Galactic stars for spectrophotometric calibration and removal of telluric features. These stars can be used to investigate
the Galactic stellar halo and thick disk, although their non-standard selection functions
preclude the detailed analysis called for above (dedicated stellar surveys should provide
that). Carollo et al. (2008, 2009) demonstrated that this calibration-star sample is best-fit
by a two-component smooth halo, with the components distinguished by both kinematics
and metallicity, and also radial distribution the outer halo is more metal-poor and in
464
retrograde rotation, compared with the non-rotating inner halo. This is reminiscent of
the seminal results of Searle & Zinn (1978), based on globular clusters, with much later
interpretation in terms of hierarchical clustering CDM models of Galaxy formation.
3.2. Elemental abundances and kinematics
The combination of detailed elemental abundances with kinematic and/or spatial phase
space information is immensely powerful. This was beautifully illustrated by the compilation of elemental abundance data for stars in the Milky Way and its satellite galaxies
that was shown at this Symposium by Andreas Koch (see his contribution for details, also
Koch 2009; Tolstoy et al. 2009; Geisler et al. 2007). With no coding to indicate parent
galaxy, the plot of [Fe/H] against [Ca/Fe] is largely a scatter plot. However, once stars
are coded by parent galaxy, it is evident that each galaxy has its own locus in this plane.
This can be understood in terms of the wide diversity of star-formation histories from
satellite to satellite (e.g. Hernandez, Gilmore & Valls-Gabaud, 2000; Orban et al. 2008),
and the relatively long duration of star formation in the vast majority of satellites, so
that Type Ia supernovae are important contributors to the chemical enrichment. A consequence is that at a given [Fe/H], [/Fe] is typically lower in stars in dwarf galaxies than
in the field halo, reflecting the longer duration of star formation in the former compared
to the latter. As noted above, an invariant IMF leads to invariant elemental abundance
ratios (with some scatter if mixing and/or sampling of the mass function is incomplete) if
only core-collapse supernovae contribute, with nothing to distinguish the parent system.
The accretion and merging of independent star-forming satellites, with typical extended
star-formation histories, may be expected to produce structure in the element abundance
patterns of stars in the host larger galaxy. Elemental abundances can thus be used to
define and map substructure, and place constraints on the minor-merger history of the
larger galaxy. It may be noted that at the lowest iron abundances in each satellite, there
should be stars with enhanced [/Fe], consistent with the field halo, being the first stars
to form within that system, and this is now being observed (e.g. Cohen & Huang 2009;
Frebel et al. 2009; Norris et al. 2010a).
For the Milky Way at least (at present) one can go further than simply coding stars by
their parent galaxy, by (probabilistically) assigning stars to individual stellar components
of the parent galaxy. This is illustrated in Figure 1, based on Ruchti et al. (2010; part
of Greg Ruchtis PhD thesis at JHU). The main scientific thrust of this work is to
identify the low-metallicity stars of the thick and thin disks, as probes for the early
evolution of these components, and compare the derived elemental abundances with the
predictions of theories. For example, build-up of the disks by late accretion, directly into
the disks, of stars from merging satellite galaxies (e.g. Abadi et al. 2003) would predict
metal-poor stars with low values of [/Fe]. The RAVE spectroscopic survey (Steinmetz
et al. 2006) provides moderate-resolution spectra of bright stars, allowing estimates of
overall metallicity and distances. The stars are sufficiently bright that estimates of proper
motion are available, giving full space motions. From this, one can select a sample of
candidate metal-poor stars with disk kinematics, for follow-up echelle spectroscopy (the
RAVE stars are ideal targets for 4-8m class telescopes). Our present sample is shown in
Fig. 1, where elemental abundances for some 170 stars are shown, coded by the stellar
component to which the star most probably belongs.
There are several conclusions from this figure. The disks extend to low metallicity,
[Fe/H] 1 dex for the thin disk and 1.75 dex for the thick disk. There is remarkably little scatter in the patterns of the elemental abundance ratios, and at low
iron abundances, all components show the same enhanced [/Fe] (unlike stars in satellite
galaxies). This implies pre-enrichment by the same underlying massive-star IMF, with
465
Figure 1. Elemental abundance ratios for metal-poor disk and halo stars, initially selected
from the medium-resolution RAVE spectroscopic survey. The stars have been assigned to a
given population based on a combination of kinematic and positional criteria. Open circles
represent thin-disk stars, open squares thick-disk stars, open triangles are halo stars, asterisks
are thin/thick disk, and crosses are thick disk/halo. The thick and thin disks clearly extend to
low iron abundances, and at these low abundances all populations have the same (enhanced)
value of [/Fe], unlike the bulk of stars in satellite galaxies. The most robust conclusion is that
the same massive-star IMF pre-enriched halo, thick disk plus thin disk, and that for all three
this IMF was well-sampled, and there was also good mixing.
rapid (short-duration) star formation taking place in regions that are well-mixed, and
of high enough mass that the IMF is well-sampled. There is no evidence supporting a
variation of the massive-star IMF. Neither is there any evidence to support late accretion
of stars from satellite galaxies directly into the disks, as this would be expected to create
metal-poor disk stars with low [/Fe] which is not observed.
It is also apparent that the stellar halo could form by the merging of stars formed
in any system(s) in which star-formation is short-lived, and chemical evolution is truncated/inefficient so that the mean metallicity is kept low; these systems could be star
clusters, galaxies, or transient structures in which stars form during accretion of gas into
the Galaxy; identifying stars with enhanced [/Fe] is not sufficient to identify halo progenitors. The addition of complementary, independent age information that the bulk of
field halo stars are old allows us to constrain further possible progenitors, and rules out
late accretion of typical luminous dwarf galaxies (e.g. Unavane, Wyse & Gilmore 1996).
The sample of Fig. 1 was defined using a rather complex selection function and thus
understanding the relative numbers of thick and thin disk stars requires modelling (underway). As discussed in the contributions of Bensby and by Reddy to these proceedings
(see also Bensby et al. 2007a), there is, in some samples of local disk stars,kinematic
confusion, in that stars selected on the basis of their kinematics to be thick-disk members can follow the elemental abundance trend of the thin disk, as opposed to that of
the thick disk. Using age information as part of the criteria used to assign a given star
to a given component appears to minimise the confusion (Bensby & Feltzing 2009)
the older stars follow the thick disk elemental abundance trends. However, we know that
the velocity distribution function in the local thin disk (probed by the Hipparcos satel-
466
lite) contains moving groups and is non-Gaussian, in at least the radial and azimuthal
motions (e.g. Dehnen 1998; Famaey et al. 2005). Resonances in the disk plane due to transient spiral arms and the Galactic bar provide plausible explanations for this kinematic
structure (e.g. Dehnen 2000; de Simone et al. 2004; Bensby et al. 2007b). The detected
dynamical substructures have cold vertical velocity dispersions, lower than typical for the
thin disk (Famaey et al. 2005). The adoption of Gaussian velocity distributions for the
underlying Galactic components when making assignments of stars to a given component
is clearly an over-simplification, one that should and could be abandoned. Large, unbiased surveys such as RAVE can be used to determine the actual velocity distributions,
for use instead.
Dynamical interactions between stars and transient spiral arms can also lead to radial
migration of resonant stars (e.g. Sellwood & Binney, 2002; Roskar et al. 2008a,b), which
in turn can be expected to imprint a signature in the elemental abundance pattern
of local stars, reflecting the variation in star-formation history and chemical evolution
across the region of migration (e.g. Sch
onrich & Binney 2009), which may even have been
detected (Haywood 2008). However, as noted in Sch
onrich & Binney (2009), the efficiency
of migration should be a function of several parameters such as velocity dispersion, and
theoretical understanding of this is incomplete.
3.3. Ultra-faint satellite galaxies
Present samples of stars with elemental abundances from high-spectral resolution observations are restricted to a thousand bright field stars (e.g. compliation of Roederer
2009), and hundreds of stars in satellite galaxies (e.g. Koch, this volume); more data are
clearly needed! The ultra-faint satellite dwarf galaxies are clearly very interesting for
study, with intriguing results even from very small samples. Our results for radial-velocity
members of Bo
otes I (discovered by Belokurov et al. 2006; luminosity LV 3 104 L ,
distance 60 kpc; Martin, de Jong & Rix 2008) and of Segue 1 (discovered by Belokurov et al. 2007; LV 335L , distance 25 kpc; Martin et al. 2008) are shown in
Fig. 2 (see also Norris et al. 2008; Norris et al. 2010b). Here I show derived iron and
carbon abundances, based on intermediate-resolution spectra (from the 2-degree field of
view, multi-object, fibre-fed spectrograph AAOmega on the Anglo-Australian Telescope)
covering the blue Ca II K-line (from which the iron abundance is derived) and the CH
G-band (synthesis of which is the basis for the carbon abundance). Thus the iron abundances are on the same scale as many of the surveys of metal-poor field halo stars (see
Beers & Christlieb 2005).
The left panel of Fig. 2 shows the iron abundances as a function of the projected radial
distance of a given star from the centre of its parent galaxy, with the filled (blue) squares
denoting stars that are radial-velocity members of Segue 1, and the open (red) star
symbols denoting radial-velocity members of Boo I. The vertical arrows along the x-axis
mark the half-light radii of these two systems (S for Segue 1 and B for Boo I). The wide
field-of-view of AAOmega allows efficient mapping of these very sparse, nearby systems
(and is critical given the apparent contamination in these lines-of-sight by stars from
the Sagittarius dSph; see Niederste-Ostholt et al. 2009). Three points are immediately
obvious: the mean level of enrichment is low, with a fraction of extremely metal-poor
stars (those with [Fe/H] < 3.0 dex; see Table 1 of Beers & Christlieb 2005); there
are members well beyond the nominal half-light radius; and there is a large internal
metallicity dispersion in each system, suggestive of self-enrichment. The right panel of
Fig. 2 shows the carbon-to-iron ratio in these stars; here the striking result is the very
high values of [C/Fe] for the two most metal-poor stars in Segue 1 (indeed, the values
derived for [Fe/H] for these two C-rich stars are rather uncertain given the weakness,
467
Figure 2. Iron and carbon abundances for stars that are radial-velocity members of the Milky
Way satellite systems Bo
otes I (open red star symbols) and Segue 1 (filled blue square symbols).
The mean level of chemical enrichment is low, and there is a large dispersion within each system.
Similarly to the field halo, the most iron-poor stars are carbon-rich.
468
each of the disk and halo. The ubiquitous intermediate-age tracers, such as AGB stars
(e.g. van Loon et al. 2003; Uttenthaler et al. 2008) and OH/IR stars (e.g. van der Veen &
Habing 1990) need to be placed in context. The planned surveys APOGEE and HERMES
should certainly help.
Evolved stars are clearly the only feasible tracers for which large samples are available.
Elemental abundances of unevolved bulge stars are possible to obtain if one can exploit
fortuitous magnification by a microlensing event. This is clearly going to be a rare occurrence, leading to small samples. However, the first results are intriguing and show a very
high mean metallicity, apparently not consistent with the previous results from giant
stars (Johnson et al. 2007, 2008; Cohen et al. 2008, 2009; Bensby et al. 2009), leading to
the suggestion that the giant phase of stellar evolution does not yield unbiased tracers,
but lacks the highest metallicity stars. The statistics of microlensing favour a location at
the distance of the bulge for the source stars, but there remains the possibility that they
are in the inner disk, rather than the bulge. Again, we need good mapping of the inner
disk bulge transition.
4. Concluding remarks
Large scale spectroscopic surveys of Galactic stars - and of stars in Local Group galaxies
- are clearly required for both kinematic/dynamic analyses and chemical abundance determinations. While several moderate-size surveys are underway or being planned, such
as SEGUE-II, RAVE, HERMES, APOGEE, these are not sufficient for global understanding, nor will they match the planned deep imaging surveys. Spectroscopic surveys
at both moderate- and high-spectral resolution are needed, and the large samples for
statistical analyses call for wide-field multi-object spectrographs.
Acknowledgements
I am very grateful to the organizers, to the Brazilian Astronomical Society (SAB) and
to the IAU for their financial support, enabling me to participate.
References
Abadi, M., Navarro, J., Steinmetz, M. & Eke, V. 2003, ApJ, 597, 21
An, D. et al. 2009, ApJ submitted (arXiv 0907.1082)
Beers, T. & Christlieb, N. 2005, ARAA, 43, 531
Belokurov, V. et al. 2006, ApJ, 647, L111
Belokurov, V. et al. 2007, ApJ, 654, 897
Bensby, T. & Feltzing, S. 2009, these proceedings (arXiv:0908.3907)
Bensby, T., Zenn, A., Oey, S. & Feltzing, S. 2007a, ApJ, 663, L13
Bensby, T., Oey, S., Feltzing, S. & Gustaffson, B. 2007b, ApJ, 655, L89
Bensby, T. et al. 2009, ApJ, 699, L174
Carney, B. 1979, ApJ, 233, 211
Carollo, D. et al. 2009, ApJ, submitted (arXiv:0909.3019)
Carollo, D. et al. 2008, Nature, 451, 216
Cohen, J.G. & Huang, W. 2009, ApJ, 701, 1053
Cohen, J.G. et al. 2009, ApJ, 699, 66
Cohen, J.G. et al. 2008, ApJ, 682, 1029
Dehnen, W. 1998, AJ, 115, 2384
Dehnen, W. 2000, AJ, 119, 800
Famaey, B. et al. 2005, A&A, 430, 165
Frebel, A., Simon, J.D., Geha, M., & Willman, B. 2009, ApJ in press (arXiv:0902.2395)
469
Abstract.
Galaxies are complex non-linear systems, evolving on all time-scales. Isolating whatever set of
physical processes was important at each major phase of their evolution requires artefacts which
resolve the timescales of dominant physical processes. These are the chemical elements, and
stellar kinematics. I consider what surveys are required to make progress in Galaxy evolution
mapping, in the era of Gaia.
Keywords. Surveys, stellar spectroscopy, Gaia, galaxy evolution
The Future
471
One type of feedback, iterative feedback between models and observations, is driving
positive progress in identifying what we cannot predict, and hence where improved observations are essential to guide progress. This form of feedback clearly has a substantial
effect on astrophysical progress. The relevant precision observations are focussed on our
only available local records of star formation and baryon assembly over time, which are
the chemical element abundances and the kinematics of stars.
Considering which astrophysical questions are least-well posed, and where the disagreement between simple prediction and observation is greatest, defines both the scale
of required next-generation observations, and which subset of observation space is likely
to provide the greatest progress. Combining that with known future observational advances - especially the astrophysical products of the Gaia mission, can help us answer
the planners request to say what we want next. Actually, we know what we want next
- massive amounts of high-quality data. What we need to define is how much data on
which targets. For that one needs also to ask: why?
A starting point to illustrate which specific surveys are likely to be most productive
is to look at a deliberately over-simplified version of galaxy evolution model predictions,
to then compare with observation, and notice the most severe inconsistencies.
Generic predictions (all of which have multiple approaches to their modification) for
disk galaxies include the following:
Extended disks form late, after a redshift of unity, or a lookback time of 8 Gyr,
in order to avoid losing too much angular momentum during active merging at earlier
times; observation shows thin disks, even in their outer parts, to be generally very old,
implying very early formation.
A large disk galaxy should have many hundreds of surviving satellite dark haloes
at the present day, which will disrupt the inner disk; observationally, satellites are less
common and more massive than anticipated.
The stellar halo is formed from disrupted satellites; the Sgr dwarf and many other
lumps and bumps support this in teh outer halo, but the dominant halo field star element
ratios are strongly inconsistent.
Minor mergers (a mass ratio of 10 20% between the satelite and the disk) into a
disk heat it, forming a thick disk out of a pre-existing thin disk, and create torques that
drive gas into the central (bulge?) regions, observationally, thick disks seem common on
disk galaxies, but are not thin-disk extensions - a single merger seems implicated.
More significant mergers transform a disk galaxy into an S0 or even an elliptical;
observationally, late-type disk galaxies are extremely common.
Subsequent accretion of gas can reform a thin disk; young disks are hard to find.
Stars can be accreted into the thin disk from suitably massive satellites (dynamical
friction must be efficient) and if to masquerade as stars formed in the thin disk, must be
on suitable high angular momentum, prograde orbits; no useful observational limits are
yet available.
Given that list, a simple comparison with local observations, and some straightforward
astrophysics, it is feasible to identify the scale of observational requirements. The detailed
comparison with observations is provided in the many contributions to this Symposium.
Here I just note the elementary astrophysics, and draw conclusions.
1.1. The evolution of Stellar Populations
Quantifying the evolution of the Galaxy requires that we deduce the past history of star
formation, chemical enrichment, and baryon material assembly. This can be achieved
by exploitation of the chemical element clock, which recognises that different chemical
elements are created on different timescales. Fortunately, element creation timescales
472
Gerry Gilmore
Figure 1. A summary overview of the dominant processes affecting chemical element ratio
distributions in stellar populations. While only alpha elements are presented here for clarity,
similar analyses for each family of elements can distinguish the dominant physics associated with
the specific element creation and distribution timescale. This figure is adapted from Figure 1 of
Gilmore & Wyse (1991).
are comparable to galaxy assembly timescales, which define star formation timescales,
and so measurement of chemical element ratios provides sufficient temporal resolution
to quantify galaxy evolution. Quantifying the assembly history of dark matter is even
more fundamental, but remains indirect, largely from general deductions about assembly
histories of stellar systems, and from kinematic and phase-space substructure analyses.
What information do we actually need? Leaving aside spatial distributions and kinematics as self-evident, the most fundamental is to define the generic distribution function
of metallicity, for example [Fe/H], for each identifiable structural component of the
Galaxy. Some progress is possible here from photometric studies, rather approximately
using the SDSS-derivation of the well-known ultra-violet excess UBV technique 0.6 , more
accurately using the Stromgren narrow-band technique (uvby) or the various similar systems. These techniques work well for quick-look results on very large samples, have a
spectacular multiplex advantage, and with suitable care can define mean and sometimes
even a dispersion in abundances. It is a mistake to overstate their value: the broader
the band the poorer the information content, and all these systems saturate at low and
high abundances. To do better one needs higher resolution than the R 10-50 available
photometrically.
In Figure 1 is a top-level summary of the evolutionary patterns of a system with no
substantial gas inflow, and with efficient mixing. This indicates what types of information
is available from spectroscopy, and implies what we need to quantify observationally. We
now consider this fuigure, starting at the lowest abundances.
Perhaps the least-appreciated aspect of this figure is the sensitivity of the amplitude of
the alpha elements to the stellar initial mass function. This means that we have available
The Future
473
direct and robust information on the high-mass stellar IMF at every metallicity, down
to very low values appropriate to the first stars in the Universe, and whatever stellar
contribution was involved in reionization of the Universe at redshits of order 10-20. This
fundamental information requires determination of the distribution function of elements
in low metallicity stars. Astonishing progress is being made here, in part from detailed
follow-up of huge low-dispersion surveys used to identify good candidates, and in part
from detailed studies of specific dwarf galaxies. Similar studies of the Galactic bulge
should find even more extreme objects, if the model predictions are correct. Discussion
of the current studies is elsewhere in this volume, so I just note one thing here: the
distribution function of alpha elements is strongly asymmetric, with a clear upper bound,
and a tail down to solar-like values. While some few high values are found, these are
consistent with the failure of assumptions about well-mixed gas, and well-sampled IMFs.
Looking at Figure 1, one sees the evidence for a universal high-mass IMF with plausible
yields is strong, and getting stronger. Interestingly, this result is frequently interpreted
to deduce limits on the assembly history ot the Galactic field halo. In fact it provides no
information at all on that. The only information is on the stellar IMF, and a minimum
scale of mixing of the ISM. The surface density of these extreme stars is so low their
detailed analysis will remain a topic for specific studies - the role of surveys will be to
identify candidates.
1.1.1. The Galactic halo and satellites
To study the Galactic halo one needs to look at metallicities where most halo stars
are found. that is, to study the halo one needs to study the true distribution function of
elements in the range 2 6[Fe/H]6 1. This is a much neglected region of parameter
space, where substantial progress could be made. A high quality survey (R > 20000) is
needed. One can quantify numbers. The halo metallicity distribution function (DF) is
roughly gaussian, with mean -1.5dex, dispersion about 0.5dex. Gaussians can be physically generated - by processes involving the central limit theorem - or can be generated
by observational error. It is the deviation away from a gaussian which contains the information. To quantify this at the few percent level where one could usefully test assembly
models requires consistent data on 10000 stars. Given that, one would then be in a position to ask more specific follow-up questions, and in particular to quantify the metal-rich
end of the halo DF. A sample of this size is eminently achievable in feasible observing
campaign, even with current low-multiplex echelle spectrographs. Data reduction will
need some automation!
1.1.2. The thin disk
High quality spectroscopy of nearby GFK stars has become a productive industry
(e.g., Edvardsson et al. 1993; Feltzing & Gustafsson 1998; Chen et al. 2000; Nissen et
al. 2000; Fulbright 2002; Reddy et al. 2003, 2006; Takeda 2007; Ecuvillon et al. 2004;
Gilli et al. 2006; Bensby et al. 2003; Ramrez et al. 2007; Fuhrmann 1998, 2004, 2008).
The remarkable outcomes are the narrowness of the thin disk abundance distribution,
and the essentially scatter-free distribution of the abundance ratios for thin disk stars
at given [Fe/H]. For example, Reddy et al. (2003) failed to detect cosmic scatter in the
abundance ratios. This has strong consequences for Galactic disk evolution, invalidating
most obvious models. It also invalidates most obvious observational strategies: apparently
ones ambition to apply chemical tagging to identify stars places of origin is doomed
to fail. The stellar IMF is remarkable uniform, ISM mixing remarkably efficient, star
formation remarkably uniform, the local disk remarkably old.
There is much kinematic substructure in the Galactic disk, some large-scale gradients
474
Gerry Gilmore
and much still inconclusive discussion of radial migration and late satellite accretion.
there is a bar and a warp. It seems the next step in advancing knowledge of the thin disk
is to quantify the abundance distribution function in these subsystems, and especially
far from the solar neighbourhood - at least one scale length radially in and radially out.
This, while feasible, will require studies of giants through much extinction, and will not
be easy.
1.1.3. The thick disk
The thick disk was defined through star counts more than 25 years ago (Gilmore &
Reid 1983) and is now well-established as a distinct component, not the tail of the stellar
halo or of the thin disk. Its origins remain the source of considerable debate. Locally,
some 5 10% of stars are in the thick disk; the vertical scale-height is 1 kpc, and
radial scale-length 3 kpc. Thick disks with roughly similar properties are seen in the
resolved stellar populations of many nearby spirals (e.g. Yoachim & Dalcanton 2006,
2008). Thus quantitative analysis of the Galactic thick disk is of wide implication. Given
its convenience - high latitude, nearby - large survey samples could be readily achieved.
Defining the abundance and kinematic (sub-)structure would be a direct test of galaxy
models, and the early accretion history of the Galaxy.
Stars in the solar neighborhood can apparently be reliably labelled, at least probabilistically, into thin and thick disks stars. Although the distributions of the velocity
components and metallicities overlap, considering kinematics (U V W ) as well as [Fe/H],
makes their definition clear. The age distributions have probably very little overlap (see,
e.g., Fig. 24 in Reddy et al. 2006), while the alpha element distributions also seem disjoint. Is it really true that the Galaxy has two discreet disks? Where is the evidence
for whatever process amplified the distinction? Again here, large samples are needed to
define the wings of the distribution functions, and any substructure.
One can imagine a survey of a galactic slice, covering mid-latitudes from longitudes 90
to 270, and so sensitive to the crucial angular momentum orbital vector. This might be
500-1500pc above the Galactic plane, covering both the thick disk and halo, and whatever
else is in there we have yet to notice (see eg Gilmore, Wyse & Norris (2002)). Again,
a prelimary survy of order 10,000 stars would be suffient to define what we know, to
quantify the apparently remarkable consequences for early galaxy mergers, and define
the next stages for further progress.
1.1.4. The inner disk and Bulge
Remarkably little remains known of the inner disk, the Galactic nucleus, and the bulge.
Studies are finally becoming underway. These suggest an old alpha-enhanced population
in the bulge, with a solar-like mean abundance, and a broad metallicity distribution
function. Given the diversity of types of galactic bulge, and the range of formation mechanisms (Kormendy etal (2009)), all of which may apply, one should presume a very
complex system. Very substantial efforts will be needed to make progress here, since we
know so little as yet.
2. Summary
Gaia is coming. Substantial and fundamental advances in our understanding of galaxy
(and Galaxy) formation, assembly and evolution, and much more, will become feasible.
In order to get full-value from the Gaia opportunity complementary surveys are critical. These include near-IR photometry, to help define extinction, high spatial resolution
The Future
475
1. Summary
To present a summary of IAU Symposium 265 it is perhaps best to first step back
and ask the question Why did we meet in Rio to discuss chemical abundances?. Part
of the scientific rationale was to host a meeting that brought together researchers who
probe chemical abundances and chemical evolution in all of the different astrophysical
environments. All meetings should be planned such that they have an outcome and, with
such a diverse set of abundance specialists brought together in one place to talk about
their favorite topics, the stated outcome for IAUS265 was to provide, within our current
understanding of the universe, a unified picture of the production of chemical elements
over cosmic time; such a view is what I think of as cosmochemistry.
This broad goal of IAUS265 was nicely framed near the beginning of the meeting week
by two review lectures: one was an Invited Discourse by Simon White entitled Evolution
of Structure in the Universe, while the other was the Plenary Review for IAUS 265 by
Stan Woosley entitled Origins and Evolution of the Elements. These lectures could be
paraphrased, albeit too simply, as the formation and evolution of spatial mass structure
and the formation and evolution of chemical structure. Taken together, the excellent
reviews by White and Woosley illustrate how intimately related are the increasingly
complex gravitational structure and chemical evolution of our universe.
As envisioned by the meeting organizers, the following 4 days of presentations spanned
the entire spectral range of chemical studies from the very early universe down to the
present day. These proceedings are the printed record of the presentations and indeed
satisfy the stated goal of IAUS265 in providing an up-to-date picture of the origins and
evolution of the chemical elements across cosmic time. In thinking about the connections
between the early universe and the present-day universe, I categorize the meeting topics
into four broad areas and briefly note some thoughts about each.
1.1. The Early Universe and the First Stars
As befits its important role as one of the pillars upon which rests our picture of a Big
Bang cosmology, Big Bang nucleosynthesis (BBN) was the first topic and we were reminded that it remains an important probe of the early universe, as well as non-standard
physics. Predictions from BBN with 3 families of neutrinos are consistent with the results
from studies of the Cosmic Microwave Background and with the observationally measured abundances of 2 H, 3 He, and 4 He. The currently measured abundances of 7 Li in the
most metal-poor stars, however, do not agree with what is predicted from the CMB. This
difference could be due to some process that is removing 7 Li from the atmospheres of
the halo stars (such as unknown types of mixing or diffusion) or to perhaps new physics.
As if to enter oncue in response to the 7 Li CMB discrepency, the possible detections
of 6 Li in metal-poor turn-off stars may provide the clues needed to solve the 7 Li mysteryif
the observers are really measuring lithium-6! Since the detections of 6 Li rest on accurately
476
477
Summary
7
mapping a rather small, red asymmetry in the much stronger Li doublet, the discussion
of lithium-6 led quite naturally into the topic of 3d model atmospheres and the limits of
using 1d models. Since the 3d models with convection produce spectral lines with small
red asymmetries, the question as to how much the 6 Li detections might be affected by
this remains under active discussion.
Moving along in time, the meeting turned to discussions of the first stars. Due to a lack
of cooling processes at zero metallicity, the expectation is that the very first stars should
define a mass function that is strongly biased towards higher masses, perhaps with some
extremely massive objects inhabiting the universe at this time and driving reionization.
At such low metallicities, approaching the limit of no heavy elements, it was noted how
stellar rotation may have a major affect on early nucleosynthesis and thus the initial
phases of chemical evolution.
The increasing fraction of carbon enhanced metal-poor (CEMP) stars with decreasing
[Fe/H] is a relatively new result and may be providing us with clues about some of the
early massive stars that dominated chemical yields at this time.
1.2. Chemical Evolution in the High Redshift Universe
Thanks to the current generation of large-aperture telescopes (6.5m to 10m), distant faint
objects with redshifts having zseveral can have their chemical abundance patterns
scrutinized. The types of objects that are being probed at high-z include QSOs, Lyman
break galaxies (LBG), damped Lyman- systems (DLA), Lyman- forest gas, or the
ISM within the host galaxies of gamma-ray bursts (GRB). These objects are providing a
wealth of chemical information about the universe when it was much younger and nicely
complements the results from the nearby, but very old, low-mass stars. The gas size
scales that are sampled by these different objects also vary over a wide range, with the
OSOs being quite small spatially ( tens of pc), while an ever-increasing spatial scale is
provided by the LBGs, then the DLAs, and finally the Lyman- forest systems, which
span the largest scales (105 pc).
We now know that in some of these objects, particularly the OSOs, chemical enrichment
occurs rapidly, with many quasars containing gas with super-solar metallicities. There is
evidence that in many of these systems chemical enrichment took place very early in the
history of the universe, at equivalent redshifts of perhaps z9-10.
The range of conditions that can be probed using the variety of high-z objects is
providing information on how SN feedback affects chemical evolution, as well as the
importance of the AGN phase of a galaxy in shaping its heavy-element enrichment. These
processes as they occur in young galaxies may influence the relation between central black
hole mass and the inner galaxy environment.
1.3. Mass Assembly and Galaxy Formation
Chemical abundance distributions are providing new insights into how galaxies form and
evolve. We are now able to probe chemistry in a growing variety of galaxy types, such
as our own barred spiral, along with a number of smaller irregular and dwarf spheroidal
galaxies. The abundance patterns in these smaller galaxies differ significantly from what
is observed in the Milky Way disk and halo, with these differences telling us how star
formation and the return of metal-enriched gas back into the local ISM varies in different
types of galaxies. The distinctive abundance patterns being found in small galaxies can
be used to constrain what types of objects may have been incorporated into the Milky
Way halo. Of considerable interest are the newly discovered ultra-faint dwarf galaxies
and whether there is any relationship between these tiny systems and the Galactic halo.
Continuing and increasingly sophisticated studies of the different components of the
478
Verne V. Smith
Milky Way, such as the thin disk, thick disk, inner and outer halos, or bulge and galactic
center are painting an ever more detailed picture of our home galaxy. For example, there
may be evidence now that the thick disk and bulge are chemically indistinguishable. It
is also clear that there is substructure in the halo, with there being an inner and outer
halo having different abundance signatures. Within the halo stellar families have been
identified, via so-called chemical tagging, which can be associated with parent galaxies,
such as the captured system Cen or the Sagittarius dwarf galaxy.
To incorporate all of these new abundances from several galaxies into a coherent view
of chemical evolution requires more complex models, which include both chemistry and
dynamics. Processes such as infall or outflow must be modelled, as well as the dynamic
interactions between different parts of a galaxy. These chemo-dynamic models are the
tools with which we can improve our interpretations of the observed abundance patterns.
1.4. Extrasolar Planets and the Chemical Connection
The topic of exoplanets is a relatively new addition to cosmochemistry, but will be an
increasingly important one in the coming years due to the clear correlations between the
liklihood of the presence of large planets with host-star metallicity. Such chemical correlations in certain types of planetary systems may shed light on the relative importance
of planet formation via core accretion or disk instability.
The possibility of interactions between parent stars and their planetary families (such
as injestion of the planet by its host star, or the process of planetary formation) which may
influence stellar chemistry is of interest. If it turns out that the presence of an underlying
terrestrial-like group of planets can be gleaned from an accurate measurement of the
stellar abundance distribution, the impact on our field will be enormous.
2. Concluding Remarks
After listening to all of the talks and the discussion, I came away from the meeting
with a few impressions that are summarized below.
As a group of researchers, we are literally awash in observations compared to not
that many years ago. There are also impressive theoretical and modeling tools to help
interpret our observations. The picture we have put together of cosmochemistry continues
to grow in detail, with more connections between what were once thought of as disparate
fields. I believe that we know much better now how the universe has chemically evolved
from early simplicity to ever-growing complexity. But we are on the verge of painting an
even bigger and much richer picture over the coming decade.
Current large-scale surveys, and planned even larger surveys, will require management skills in order to efficiently exploit this new knowledge: good archives with uniform
data sets, visualization tools, and improved communication. Such management skills are
not necessarily required in small projects with just a few people, but the large surveys
are really complex missions that demand more planning and more teamwork. Our needed
set of skills is evolving.
To maximize the science that can be extracted from the chemical abundances that
are derived from the spectra that we gather and will gather, we must continue to push
on improvements in the analysis. Many talks during the meeting emphasized this, with
demonstrations of improved stellar abundance analysis techniques that may take us well
below the 0.10 dex limits that many of us seem content to work within, down to less than
0.05 dex or even approaching 0.01 dex! 3d modeling, NLTE, inclusion of magnetic fields,
rotational effects a lot more physics involved in the analysis of the beautiful data sets
that continue to be collected.
Summary
479
Finally, on the horizon are the ELT projects, which currently number 3. These new
telescopes will allow us to push back on all fronts and as a community we need more than
one, ideally with sites in both northern and southern hemispheres and with first-light
instruments that include high-resolution spectrographs. In whatever ways we can as a
research community, we should be united in our support of these efforts.
I would like to take this opportunity to thank the organizers of this meeting, both
for putting together such a varied and interesting set of sessions and for giving me the
chance to present closing comments.
Abstract.
The Apache Point Observatory Galactic Evolution Experiment (APOGEE) is a large-scale,
near-infrared (H-band), high-resolution (R 30, 000), high S/N (&100) spectroscopic survey of
Milky Way stellar populations. APOGEE will operate from 1.51-1.68m, a region that includes
useful absorption lines from at least fifteen chemical species including , odd-Z, and iron peak
elements. The APOGEE instrument has a novel design featuring 300 science fibers feeding light
to a mosaiced VPH grating and a six-element camera encased in a liquid nitrogen-cooled cryostat.
A three year bright-time observing campaign will enable APOGEE to observe approximately
100,000 red giants across the Galactic bulge, disk and halo.
Keywords. galaxies: interactions, Galaxy: halo, Galaxy: structure, stars: abundances
481
APOGEE
Refractive Camera
(6) Elements: Silicon & Fused Silica
Mosaic Volume
Phase Holographic
(VPH) Grating
Detector Mosaic
(3) H2RG Arrays
Collimating Mirror
1.4 m
Fold Mirror
Pseudo-Slit for
300 fibers
Fold Mirror
LN2-cooled Cryostat
2.3 m
Figure 1. Solid model of the 2.31.3-m APOGEE cryostat showing the primary features of the
spectrograph optical train. The cryostat, housed remotely from the telescope, is linked to the
telescope focal plane by an 40-m fiber optic train.
manner and with statistically large samples of stars the metallicity distributions and
abundance patterns for numerous chemical species across the different Galactic stellar
populations; (2) deriving spatial variations of these chemical distributions within these
populations; (3) studying the processes of star formation, feedback, chemical mixing
and chemical evolution in the Milky Way with sensitivity to numerous nucleosynthetic
pathways; (4) constraining the evolution of the initial mass function, by inference from the
differences in chemical yields produced by stars of different masses; and (5) using these
extensive chemical data combined with dynamical information provided by APOGEE
radial velocities, proper motions from other surveys, as well as Galactic chemical evolution
models to unravel the overall formation and evolution of the Milky Way.
Author Index
Aden, D. 227, 346
Adelman, S.J. 213
Allen, D.M. 118
Allende Prieto, C. 304
Alves-Brito, A. 342
Amarante, J. A. S. 420
Andersen, J. 120
Andrade, D. P. P. 440
Andrievsky, S. 380
Aoki, W. 61, 111 , 124
Arnett, W.D. 106
Arras, P. 408
Asplund, M. 71, 90, 342, 346, 412
Aykutalp, A. 63
Chiba, M. 267
Chieffi, A. 54
Chou, M.Y. 364
Christlieb, N. 75, 132
Colgan, S.W.J. 249
Costa-Mello, D.R. 356
Costa, R.D.D. 247, 317, 354
Coutinho, L. H. 440
Cowan, J.J. 46
Craven, T.W. 249
Cunha, K. 358, 364, 408, 432
Daz, A. I. 243
da Costa, J. S. 422
da Silva, L. 360, 374
Da Silva, R. 426
da Silveira, E. F. 428, 442
Daflon, S. 352, 356, 358, 362
Dartois, E. 428
de Ara
ujo, F. X. 432
de Brito, A. N. 440
de Castilho, R. B. 440
de la Reza, R. 358, 374, 424, 432
De Medeiros, J. R. 422
de Mello, A.B. 366
de Souza, G. G. B. 440
Decressin, T. 94
Dekovic, M.S. 211
Delgado Mena, E. 430
do Nascimento Jr., J.-D. 422
do Nascimento, E. M. 440
Dobrovolskas, V. 209
Docenko, D. 167
Domaracka, A. 428, 442
Domingos, R. C. 424
Dominguez Cerde
na, C. 430
Drake, N.A. 130, 241 , 424
Dufour, R.J. 249
Eggenberger, P. 98
Ekstrom, S. 98
Erb, D. 147
Farenzena, L. S. 428
Feltzing, S. 227, 300, 346
Ferreira, L.D. 360
Firpo, V. 243
Flores M. 181
Francois, P. 75, 120, 380
Frebel, A. 237
Freeman, K.C. 263, 267
Freytag, B. 75, 201
482
Author Index
Fujimoto, M.Y. 90, 128
Fulbright, J. 279
Galvez-Ortiz, M. C. 430
Gal-Yam, A. 346
Galli, D. 134
Gallino, R. 46
Geha, M. 237
Gehren, T. 197, 348
Georgy, C. 98
Ghezzi, L. 432
Gilmore, G. 470
Gitterman, E.D. 249
Glebbeek, E. 117
Glover, S.C.O. 65
Gomez, A. 344
Goncalves, D.R. 159
Gonzalez Hern
andez, J.I. 75, 122,
430
Gould, A. 346
Gratton, R. 134
Grebel, E.K. 227
Grenon, M. 382
Grupp, F. 197
Guillermo Bosch, G. 243
Guillermo H
agele, G. 243
Gustafsson, B. 187, 382, 412
Habe, A. 128
Hamann, F. 171, 183
Hammer, F. 181
Hansen, C.J. 67
Haywood, M. 434
Hearty, F. 479
Heger, A. 3
Hensler, G. 325
Hill, V. 75, 120, 219, 344, 380
Hirschi, R. 98
Hoffman, R.D. 3
Holtzman, J. 358
Honda, S. 61, 124
Hubeny, I. 213
Huja, D. 73
Hyung, S. 77
Ishizuka, C. 90
Israelian, G. 430
Ito, H. 124
Ivanauskas, A. 209
Izumiura, H. 77
Izzard, R.G. 117
James, G. 346
Jappsen, A. 65
Joggerst, C.C. 42
Johnson, J. 126
Johnson, J.A. 346
483
Jurkic, T. 211
Karakas, A.I. 54, 57
Katime-Santrich, O.J. 370
Katsuta, Y. 90
Kennedy, C.R. 126, 132
Kholtygin, A. 313
King, J.R. 372
Kirby, E. 237
Klessen, R.S. 65
Klevas, J. 209
Kobayashi, C. 336
Koch, A. 227
Komiya, Y. 90, 128
Korn, A. 197
Korotin, S. 380
Korzennik, S. 416
Kotnik-Karuza, D. 211
Krticka, J. 69
Kucinskas, A. 209
Kubas, D. 346
Lagarde, L. 205
Lago, A. F. 440
Laird, J. B. 416
Lambert, D.L. 54
Lanfranchi, G.A. 245
Lanz, T. 213
Latham, D. W. 416
Lauretta, D. S. 399
Lee, Y.S. 267
Lepine, J.R.D. 251
Limongi M. 54
Lo, C.C. 249
Lorenz-Martins, S. 366
Low, M.M. 65
Lucatello, S. 346
Ludwig, H.-G. 23, 75, 122, 201 , 209
Lugaro, M. 57
Maciel, W.J. 317 350, 354
Maeder, A. 98
Maggio, A. 436
Magrini, L. 159
Maiolino, R. 179
Majewski, S.R. 364, 408, 480
Marcon-Uchida, M.M. 247
Marconi, A. 179
Marinho, R. T. 440
Martnez-Delgado, D. 364
Martin, C.L. 163
Mashonkina, L. 197
Masseron, T. 126
Matsuoka, E. 179
Matteucci, F. 245, 247
McNabb, I.A. 249
McWilliam, A. 279
484
Author Index
Meakin, C. 106
Melendez, J. 71, 342, 346, 412, 422
Melo, C. 374
Mesz
aros, A. 73
Meynet, G. 69, 98
Milanova, Y. 313
Milone, A. 426
Minniti, D. 344
Molaro, P. 75, 120
Morisset, C. 94
Moriya, T. 34
Morrell, N. 243
pa, J. 73
R
Roberts, L. 3
Rocha-Pinto, H. J. 255, 350, 420,
438
Rodrigues M. 181
Roederer, I.U. 368
Rood, R. T. 408
Rossi, S. 118, 126, 132
Rothard, H. 428, 442
Rubin, R.H. 249
Ryan, S.G. 118
Ryde, N.A. 285
Nagao, T. 179
Nidia Morrell, N. 243
Nishimura, T. 90
Nomoto, K. 34
Nordstr
om, B. 120, 376
Norris, J.E. 267
Sana, H. 346
Santos, N. C. 430
Sanz-Forcada, J. 436
Sartori, M. 352
Savaglio, S. 139
Sbordone, L. 75, 122
Scannapieco, E. 163
Scarano Jr., S. 251
Scelsi, L. 436
Schiavon, R. P. 479
Schuler, S.C. 372, 432
Seperuelo Duarte E. 428, 442
Shectman, S.A. 368
Shi, J. 197
Simmerer, J. 346
Simon, J.D. 237
Simon, L.E. 171, 183
Simpson, J.P. 249
Siqueira-Mello, C. 120
Sivarani, T. 75, 126, 132
Skrutskie, M. F. 479
Smiljanic, R. 134
Smith, V. V. 364, 408, 432, 476
Sneden, C. 46, 120, 368
Sozzetti, A. 416
Spaans, M. 63
Spite, F. 75, 120, 380
Spite, M. 75, 120, 380
Stancliffe, R.J. 117
Stasi
nska, G. 94
Stefanik, R. P. 416
Steffen, M. 23, 75, 201
Steigman, G. 15
Sterzik, M. 374
Stonkute, E. 376
Suda, T. 90, 128
Sunyaev, R.A 167
Suran, M.D. 378
OBrien, D. P. 399
Ortolani, S. 344
Otsuka, M. 77
Owocki, S.P. 69
Pasquini, L. 134, 447
Patterson, R.J. 364, 408
Pauldrach, A.W.A. 249
Pe
na, M. 155
Penteado, E. M. 438
Pereira, C.B. 130, 241, 356, 424
Peters, G.J. 213
Pilling, S. 428, 440, 442
Placco, V.M. 126, 132
Plez, B. 75, 120
Pols, O. 117
Pompeia, L. 382
Porto de Mello, G.F. 350, 360, 422
Prakapavicius, D. 209
Prester, D.D. 211
Preston, G.W. 368
Primas, F. 67, 120
Przybilla, N. 352
Puech M. 181
Quast, G. R. 370, 374
Ramrez, I. 71, 412
Randich, S. 134
Rangel, R.H.O. 350
Rauch, T. 94
Rebolo, R. 430
Recchi, S. 325
Reddy, B. 289
Renzini, A. 344
Rich, R.M. 279
Tajitsu, A. 77
Taniguchi, Y. 179
Tautvaisiene, G. 376
Theado, S. 422
The, L. 372
Author Index
Thompson, I.B. 368
Tominaga, N. 34, 124
Torres, C.A.O. 370, 374
Torres, G. 416
Tovmassian, G. 94
Trevisan, M. 382
Tsangarides, S.A. 118
Tsujimoto, T. 384
Udalski, A. 346
Valenti, J. 403
Vant Veer, C. 75
Whalen, D.J. 42
Williams, M. 263
Willman, B. 237
Wilson, J. C. 479
Winter, O. C. 424
Wofford, A. 386
Woosley, S. E. 3
Wylie-de Boer, E. 263
Wyse, R. 461
Yamada, S. 90
Yong, D. 54, 412
Zevallos Herencia, M. I. 362
Zoccali, M. 271, 344
485